commit 39d0c8f0dba554f90d4a8c19587d029a485ddaf9 Author: krish@sentientgeeks.com Date: Thu Sep 30 23:13:39 2021 +0530 refactored code update diff --git a/.editorconfig b/.editorconfig new file mode 100644 index 0000000..0009889 --- /dev/null +++ b/.editorconfig @@ -0,0 +1,5 @@ +[*] +indent_style = space +indent_size = 2 +trim_trailing_whitespace = true +insert_final_newline = true diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000..1b6a9a2 --- /dev/null +++ b/.gitignore @@ -0,0 +1,15 @@ +build/ +node_modules/ +deps/librdkafka +npm-debug.log + +docs + +deps/* +!deps/*.gyp +!deps/windows-install.* + +.DS_Store + +package-lock.json +.vscode \ No newline at end of file diff --git a/.gitmodules b/.gitmodules new file mode 100644 index 0000000..4d4e7fb --- /dev/null +++ b/.gitmodules @@ -0,0 +1,3 @@ +[submodule "deps/librdkafka"] + path = deps/librdkafka + url = https://github.com/edenhill/librdkafka.git diff --git a/.jshintignore b/.jshintignore new file mode 100644 index 0000000..b43bf86 --- /dev/null +++ b/.jshintignore @@ -0,0 +1 @@ +README.md diff --git a/.jshintrc b/.jshintrc new file mode 100644 index 0000000..09968b8 --- /dev/null +++ b/.jshintrc @@ -0,0 +1,23 @@ +{ + "node": true, + "mocha": true, + "browser": false, + "boss": true, + "curly": true, + "debug": false, + "devel": false, + "eqeqeq": true, + "evil": true, + "forin": false, + "latedef": false, + "noarg": true, + "nonew": true, + "nomen": false, + "onevar": false, + "plusplus": false, + "regexp": false, + "undef": true, + "strict": false, + "white": false, + "eqnull": true +} diff --git a/.npmignore b/.npmignore new file mode 100644 index 0000000..c6793a2 --- /dev/null +++ b/.npmignore @@ -0,0 +1,10 @@ +deps/* +!deps/librdkafka.gyp +!deps/librdkafka +!deps/windows-install.* +.gitmodules +Dockerfile +deps/librdkafka/config.h +build +.github +.vscode diff --git a/.travis.yml b/.travis.yml new file mode 100644 index 0000000..6124d08 --- /dev/null +++ b/.travis.yml @@ -0,0 +1,67 @@ +language: node_js + +cache: + directories: + - node_modules +node_js: + - "4" + - "6" + - "8" + - "10" + - "12" + - "13" + - "14" + - "15" + - "16" +sudo: required +services: docker +before_install: + - if [[ "$TRAVIS_OS_NAME" == "linux" && "$TRAVIS_BUILD_STAGE_NAME" =~ Test.* ]]; then ./run_docker.sh; fi + # - if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then brew install openssl; fi + - if [[ "$TRAVIS_OS_NAME" == "windows" ]]; then ./win_install.bat; fi +install: + - if [[ "$TRAVIS_BUILD_STAGE_NAME" =~ Test.* ]]; then npm install; fi + +script: + - if [[ "$TRAVIS_BUILD_STAGE_NAME" =~ Test.* ]]; then make lint && make test && make check; else echo $TRAVIS_BUILD_STAGE_NAME; fi + +jobs: + include: + - stage: test_on_mac + os: osx + osx_image: xcode10 + env: CPPFLAGS=-I/usr/local/opt/openssl/include LDFLAGS=-L/usr/local/opt/openssl/lib + node_js: + - "10" + - stage: test_on_win + os: windows + node_js: + - "10" + # - stage: create_doc + # provider: script + # before_deploy: + # - openssl aes-256-cbc -K $encrypted_a2e08d5c220e_key -iv $encrypted_a2e08d5c220e_iv -in deploy.enc -out /tmp/deploy -d + # - eval "$(ssh-agent -s)" + # - chmod 600 /tmp/deploy + # - ssh-add /tmp/deploy + # script: + # - "./make_docs.sh" + # if: (tag =~ ^v) + - stage: deploy + os: linux + node_js: + - "10" + script: skip + if: (tag =~ ^v) + before_deploy: + - npm install --no-save semver + deploy: + provider: npm + skip_cleanup: true + email: webmakersteve@gmail.com + api_key: + secure: "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" + on: + tags: true + after_deploy: + - cat package.json diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md new file mode 100644 index 0000000..563d54b --- /dev/null +++ b/CONTRIBUTING.md @@ -0,0 +1,192 @@ +# Contributing to `node-rdkafka` + +:+1::tada: First off, thanks for taking the time to contribute! :tada::+1: + +The following is a set of guidelines for contributing to `node-rdkafka` +which is hosted in the [Blizzard Organization](https://github.com/blizzard) +on GitHub. This document lists rules, guidelines, and help getting started, +so if you feel something is missing feel free to send a pull request. + +#### Table Of Contents + +[What should I know before I get started?](#what-should-i-know-before-i-get-started) + * [Contributor Agreement](#contributor-agreement) + +[How Can I Contribute?](#how-can-i-contribute) + * [Reporting Bugs](#reporting-bugs) + * [Suggesting Enhancements](#suggesting-enhancements) + * [Pull Requests](#pull-requests) + +[Styleguides](#styleguides) + * [Git Commit Messages](#git-commit-messages) + * [JavaScript Styleguide](#javascript-styleguide) + * [C++ Styleguide](#c++-styleguide) + * [Specs Styleguide](#specs-styleguide) + * [Documentation Styleguide](#documentation-styleguide) + +[Debugging](#debugging) + * [Debugging C++](#debugging-c) + +## What should I know before I get started? + +### Contributor Agreement + +Not currently required. + +## How can I contribute? + +### Reporting Bugs + +Please use __Github Issues__ to report bugs. When filling out an issue report, +make sure to copy any related code and stack traces so we can properly debug. +We need to be able to reproduce a failing test to be able to fix your issue +most of the time, so a custom written failing test is very helpful. + +Please also note the Kafka broker version that you are using and how many +replicas, partitions, and brokers you are connecting to, because some issues +might be related to Kafka. A list of `librdkafka` configuration key-value pairs +also helps. + +### Suggesting Enhancements + +Please use __Github Issues__ to suggest enhancements. We are happy to consider +any extra functionality or features to the library, as long as they add real +and related value to users. Describing your use case and why such an addition +helps the user base can help guide the decision to implement it into the +library's core. + +### Pull Requests + +* Include new test cases (either end-to-end or unit tests) with your change. +* Follow our style guides. +* Make sure all tests are still passing and the `linter` does not report any issues. +* End files with a new line. +* Document the new code in the comments (if it is JavaScript) so the + documentation generator can update the reference documentation. +* Avoid platform-dependent code. +
**Note:** If making modifications to the underlying C++, please use built-in + precompiler directives to detect such platform specificities. Use `Nan` + whenever possible to abstract node/v8 version incompatibility. +* Make sure your branch is up to date and rebased. +* Squash extraneous commits unless their history truly adds value to the library. + +## Styleguides + +### General style guidelines + +Download the [EditorConfig](http://editorconfig.org) plugin for your preferred +text editor to automate the application of the following guidelines: + +* Use 2-space indent (no tabs). +* Do not leave trailing whitespace on lines. +* Files should end with a final newline. + +Also, adhere to the following not enforced by EditorConfig: + +* Limit lines to 80 characters in length. A few extra (<= 5) is fine if it helps + readability, use good judgement. +* Use `lf` line endings. (git's `core.autocrlf` setting can help) + +### Git Commit Messages + +Commit messages should adhere to the guidelines in tpope's +[A Note About Git Commit Messages](http://tbaggery.com/2008/04/19/a-note-about-git-commit-messages.html) + +In short: + +* Use the imperative mood. ("Fix bug", not "Fixed bug" or "Fixes bug") +* Limit the first line to 50 characters or less, followed by a blank line + and detail paragraphs (limit detail lines to about 72 characters). +* Reference issue numbers or pull requests whenever possible. + +### JavaScript Styleguide + +* Place `module.exports` at or near the top of the file. + * Defined functions are hoisted, so it is appropriate to define the + function after you export it. + * When exporting an object, define it first, then export it, and then add + methods or properties. +* Do not use ES2015 specific features (for example, do not use `let`, `const`, + or `class`). +* All callbacks should follow the standard Node.js callback signature. +* Your JavaScript should properly pass the linter (`make jslint`). + +### C++ Styleguide + +* Class member variables should be prefixed with `m_`. +* Use a comment when pointer ownership has changed hands. +* Your C++ should properly pass the `cpplint.py` in the `make lint` test. + +### Specs Styleguide + +* Write all JavaScript tests by using the `mocha` testing framework. +* All `mocha` tests should use exports syntax. +* All `mocha` test files should be suffixed with `.spec.js` instead of `.js`. +* Unit tests should mirror the JavaScript files they test (for example, + `lib/client.js` is tested in `test/client.spec.js`). +* Unit tests should have no outside service dependencies. Any time a dependency, + like Kafka, exists, you should create an end-to-end test. +* You may mock a connection in a unit test if it is reliably similar to its real + variant. + +### Documentation Styleguide + +* Write all JavaScript documentation in jsdoc-compatible inline comments. +* Each docblock should have references to return types and parameters. If an + object is a parameter, you should also document any required subproperties. +* Use `@see` to reference similar pieces of code. +* Use comments to document your code when its intent may be difficult to understand. +* All documentation outside of the code should be in Github-compatible markdown. +* Make good use of font variations like __bold__ and *italics*. +* Use headers and tables of contents when they make sense. + +## Editor + +I began using Visual Studio code to develop on `node-rdkafka`. If you use it you can configure the C++ plugin to resolve the paths needed to inform your intellisense. This is the config file I am using on a mac to resolve the required paths: + +`c_cpp_properties.json` +``` +{ + "configurations": [ + { + "name": "Mac", + "includePath": [ + "${workspaceFolder}/**", + "${workspaceFolder}", + "${workspaceFolder}/src", + "${workspaceFolder}/node_modules/nan", + "${workspaceFolder}/deps/librdkafka/src", + "${workspaceFolder}/deps/librdkafka/src-cpp", + "/usr/local/include/node", + "/usr/local/include/node/uv" + ], + "defines": [], + "macFrameworkPath": [ + "/Library/Developer/CommandLineTools/SDKs/MacOSX10.14.sdk/System/Library/Frameworks" + ], + "compilerPath": "/usr/bin/clang", + "cStandard": "c11", + "cppStandard": "c++17", + "intelliSenseMode": "clang-x64" + } + ], + "version": 4 +} +``` + +## Debugging + +### Debugging C++ + +Use `gdb` for debugging (as shown in the following example). + +``` +node-gyp rebuild --debug + +gdb node +(gdb) set args "path/to/file.js" +(gdb) run +[output here] +``` + +You can add breakpoints and so on after that. diff --git a/LICENSE.txt b/LICENSE.txt new file mode 100644 index 0000000..be3db86 --- /dev/null +++ b/LICENSE.txt @@ -0,0 +1,20 @@ +The MIT License (MIT) +Copyright (c) 2016 Blizzard Entertainment + +Permission is hereby granted, free of charge, to any person obtaining a copy of +this software and associated documentation files (the "Software"), to deal in +the Software without restriction, including without limitation the rights to +use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies +of the Software, and to permit persons to whom the Software is furnished to do +so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS +IN THE SOFTWARE. diff --git a/Makefile b/Makefile new file mode 100644 index 0000000..ef64d46 --- /dev/null +++ b/Makefile @@ -0,0 +1,94 @@ +NODE-GYP ?= node_modules/.bin/node-gyp + +# Sick of changing this. Do a check and try to use python 2 if it doesn't work +PYTHON_VERSION_FULL := $(wordlist 2,4,$(subst ., ,$(shell python --version 2>&1))) +PYTHON_VERSION_MAJOR := $(word 1,${PYTHON_VERSION_FULL}) + +ifeq ($(PYTHON_VERSION_MAJOR), 2) +PYTHON = python +else +PYTHON = python2 +endif + +NODE ?= node +CPPLINT ?= cpplint.py +BUILDTYPE ?= Release +TESTS = "test/**/*.js" +E2E_TESTS = $(wildcard e2e/*.spec.js) +TEST_REPORTER = +TEST_OUTPUT = +CONFIG_OUTPUTS = \ + build/bindings.target.mk \ + build/Makefile \ + build/binding.Makefile build/config.gypi + +CPPLINT_FILES = $(wildcard src/*.cc src/*.h) +CPPLINT_FILTER = -legal/copyright +JSLINT_FILES = lib/*.js test/*.js e2e/*.js + +PACKAGE = $(shell node -pe 'require("./package.json").name.split("/")[1]') +VERSION = $(shell node -pe 'require("./package.json").version') + +GYPBUILDARGS= +ifeq ($(BUILDTYPE),Debug) +GYPBUILDARGS=--debug +endif + +.PHONY: all clean lint test lib docs e2e ghpages check + +all: lint lib test e2e + +lint: cpplint jslint + +cpplint: + @$(PYTHON) $(CPPLINT) --filter=$(CPPLINT_FILTER) $(CPPLINT_FILES) + +jslint: node_modules/.dirstamp + @./node_modules/.bin/jshint --verbose $(JSLINT_FILES) + +lib: node_modules/.dirstamp $(CONFIG_OUTPUTS) + @PYTHONHTTPSVERIFY=0 $(NODE-GYP) build $(GYPBUILDARGS) + +node_modules/.dirstamp: package.json + @npm update --loglevel warn + @touch $@ + +$(CONFIG_OUTPUTS): node_modules/.dirstamp binding.gyp + @$(NODE-GYP) configure + +test: node_modules/.dirstamp + @./node_modules/.bin/mocha $(TEST_REPORTER) $(TESTS) $(TEST_OUTPUT) + +check: node_modules/.dirstamp + @$(NODE) util/test-compile.js + +e2e: $(E2E_TESTS) + @./node_modules/.bin/mocha --exit $(TEST_REPORTER) $(E2E_TESTS) $(TEST_OUTPUT) + +define release + NEXT_VERSION=$(shell node -pe 'require("semver").inc("$(VERSION)", "$(1)")') + node -e "\ + var j = require('./package.json');\ + j.version = \"$$NEXT_VERSION\";\ + var s = JSON.stringify(j, null, 2);\ + require('fs').writeFileSync('./package.json', s);" && \ + git commit -m "release $$NEXT_VERSION" -- package.json && \ + git tag "$$NEXT_VERSION" -m "release $$NEXT_VERSION" +endef + +docs: node_modules/.dirstamp + @rm -rf docs + @./node_modules/jsdoc/jsdoc.js --destination docs \ + --recurse -R ./README.md \ + -t "./node_modules/toolkit-jsdoc/" \ + --tutorials examples ./lib + +gh-pages: node_modules/.dirstamp + @./make_docs.sh + +release-patch: + @$(call release,patch) + +clean: node_modules/.dirstamp + @rm -f deps/librdkafka/config.h + @$(NODE-GYP) clean diff --git a/bench/consumer-raw-rdkafka.js b/bench/consumer-raw-rdkafka.js new file mode 100644 index 0000000..82b9343 --- /dev/null +++ b/bench/consumer-raw-rdkafka.js @@ -0,0 +1,76 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Kafka = require('../'); +var count = 0; +var total = 0; +var store = []; +var host = process.argv[2] || 'localhost:9092'; +var topic = process.argv[3] || 'test'; + +var consumer = new Kafka.KafkaConsumer({ + 'metadata.broker.list': host, + 'group.id': 'node-rdkafka-bench-s', + 'fetch.wait.max.ms': 100, + 'fetch.message.max.bytes': 1024 * 1024, + 'enable.auto.commit': false + // paused: true, +}, { + 'auto.offset.reset': 'earliest' +}); + +var interval; + +consumer.connect() + .once('ready', function() { + consumer.subscribe([topic]); + consumer.consume(); + }) + .on('rebalance', function() { + console.log('rebalance'); + }) + .once('data', function() { + interval = setInterval(function() { + console.log('%d messages per second', count); + if (count > 0) { + store.push(count); + } + count = 0; + }, 1000); + }) + .on('data', function(message) { + count += 1; + total += 1; + }); + +function shutdown() { + clearInterval(interval); + + if (store.length > 0) { + var calc = 0; + for (var x in store) { + calc += store[x]; + } + + var mps = parseFloat(calc * 1.0/store.length); + + console.log('%d messages per second on average', mps); + + } + + var killTimer = setTimeout(function() { + process.exit(); + }, 5000); + + consumer.disconnect(function() { + clearTimeout(killTimer); + process.exit(); + }); + +} diff --git a/bench/consumer-subscribe.js b/bench/consumer-subscribe.js new file mode 100644 index 0000000..c124cd6 --- /dev/null +++ b/bench/consumer-subscribe.js @@ -0,0 +1,77 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Kafka = require('../'); +var count = 0; +var total = 0; +var store = []; +var host = process.argv[2] || 'localhost:9092'; +var topic = process.argv[3] || 'test'; + +var consumer = new Kafka.KafkaConsumer({ + 'metadata.broker.list': host, + 'group.id': 'node-rdkafka-bench', + 'fetch.wait.max.ms': 100, + 'fetch.message.max.bytes': 1024 * 1024, + 'enable.auto.commit': false + // paused: true, +}, { + 'auto.offset.reset': 'earliest' +}); + +var interval; + +consumer.connect() + .once('ready', function() { + consumer.subscribe([topic]); + consumer.consume(); + }) + .once('data', function() { + interval = setInterval(function() { + console.log('%d messages per second', count); + if (count > 0) { + store.push(count); + } + count = 0; + }, 1000); + }) + .on('data', function(message) { + count += 1; + total += 1; + }); + +process.once('SIGTERM', shutdown); +process.once('SIGINT', shutdown); +process.once('SIGHUP', shutdown); + +function shutdown() { + clearInterval(interval); + + if (store.length > 0) { + var calc = 0; + for (var x in store) { + calc += store[x]; + } + + var mps = parseFloat(calc * 1.0/store.length); + + console.log('%d messages per second on average', mps); + + } + + var killTimer = setTimeout(function() { + process.exit(); + }, 5000); + + consumer.disconnect(function() { + clearTimeout(killTimer); + process.exit(); + }); + +} diff --git a/bench/kafka-consumer-stream.js b/bench/kafka-consumer-stream.js new file mode 100644 index 0000000..6e240ff --- /dev/null +++ b/bench/kafka-consumer-stream.js @@ -0,0 +1,100 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Writable = require('stream').Writable; + +var Kafka = require('../'); +var count = 0; +var total = 0; +var store = []; +var host = process.argv[2] || 'localhost:9092'; +var topic = process.argv[3] || 'test'; + +var stream = Kafka.createReadStream({ + 'metadata.broker.list': host, + 'group.id': 'node-rdkafka-benchs', + 'fetch.wait.max.ms': 100, + 'fetch.message.max.bytes': 1024 * 1024, + 'enable.auto.commit': false + // paused: true, +}, { + 'auto.offset.reset': 'earliest' +}, { + fetchSize: 16, + topics: [topic] +}); + +// Track how many messages we see per second +var interval; + +var isShuttingDown = false; + +stream + .on('error', function(err) { + console.log('Shutting down due to error'); + console.log(err.stack); + shutdown(); + }) + .once('data', function(d) { + interval = setInterval(function() { + if (isShuttingDown) { + clearInterval(interval); + } + console.log('%d messages per second', count); + if (count > 0) { + // Don't store ones when we didn't get data i guess? + store.push(count); + // setTimeout(shutdown, 500); + } + count = 0; + }, 1000).unref(); + }) + .on('end', function() { + // Can be called more than once without issue because of guard var + console.log('Shutting down due to stream end'); + shutdown(); + }) + .pipe(new Writable({ + objectMode: true, + write: function(message, encoding, cb) { + count += 1; + total += 1; + setImmediate(cb); + } + })); + +process.once('SIGTERM', shutdown); +process.once('SIGINT', shutdown); +process.once('SIGHUP', shutdown); + +function shutdown() { + if (isShuttingDown) { + return; + } + clearInterval(interval); + isShuttingDown = true; + if (store.length > 0) { + var calc = 0; + for (var x in store) { + calc += store[x]; + } + + var mps = parseFloat(calc * 1.0/store.length); + + console.log('%d messages per second on average', mps); + } + + // Destroy the stream + stream.destroy(); + + stream.once('end', function() { + console.log('total: %d', total); + }); + +} diff --git a/bench/producer-raw-rdkafka.js b/bench/producer-raw-rdkafka.js new file mode 100644 index 0000000..d8fabcc --- /dev/null +++ b/bench/producer-raw-rdkafka.js @@ -0,0 +1,125 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Kafka = require('../'); +var crypto = require('crypto'); +var count = 0; +var total = 0; +var totalComplete = 0; +var verifiedComplete = 0; +var errors = 0; +var store = []; +var started; +var done = false; +var host = process.argv[2] || '127.0.0.1:9092'; +var topicName = process.argv[3] || 'test'; +var compression = process.argv[4] || 'gzip'; +var MAX = process.argv[5] || 10000000; + +var producer = new Kafka.Producer({ + 'metadata.broker.list': host, + 'group.id': 'node-rdkafka-bench', + 'compression.codec': compression, + 'retry.backoff.ms': 200, + 'message.send.max.retries': 10, + 'socket.keepalive.enable': true, + 'queue.buffering.max.messages': 100000, + 'queue.buffering.max.ms': 1000, + 'batch.num.messages': 1000 +}); + +// Track how many messages we see per second +var interval; +var ok = true; + +function getTimer() { + if (!interval) { + interval = setTimeout(function() { + interval = false; + if (!done) { + console.log('%d messages per sent second', count); + store.push(count); + count = 0; + getTimer(); + + } else { + console.log('%d messages remaining sent in last batch <1000ms', count); + } + }, 1000); + } + + return interval; +} + +var t; + +crypto.randomBytes(4096, function(ex, buffer) { + + producer.connect() + .on('ready', function() { + getTimer(); + + started = new Date().getTime(); + + var sendMessage = function() { + try { + var errorCode = producer.produce(topicName, null, buffer, null); + verifiedComplete += 1; + } catch (e) { + console.error(e); + errors++; + } + + count += 1; + totalComplete += 1; + if (totalComplete === MAX) { + shutdown(); + } + if (total < MAX) { + total += 1; + + // This is 100% sync so we need to setImmediate to give it time + // to breathe. + setImmediate(sendMessage); + } + }; + + sendMessage(); + + }) + .on('event.error', function(err) { + console.error(err); + process.exit(1); + }) + .on('disconnected', shutdown); + +}); + +function shutdown(e) { + done = true; + + clearInterval(interval); + + var killTimer = setTimeout(function() { + process.exit(); + }, 5000); + + producer.disconnect(function() { + clearTimeout(killTimer); + var ended = new Date().getTime(); + var elapsed = ended - started; + + // console.log('Ended %s', ended); + console.log('total: %d messages over %d ms', total, elapsed); + + console.log('%d messages / second', parseInt(total / (elapsed / 1000))); + process.exit(); + }); + +} diff --git a/bench/producer-rdkafka.js b/bench/producer-rdkafka.js new file mode 100644 index 0000000..3f1aba5 --- /dev/null +++ b/bench/producer-rdkafka.js @@ -0,0 +1,117 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Kafka = require('../'); +var crypto = require('crypto'); +var count = 0; +var total = 0; +var totalComplete = 0; +var store = []; +var host = process.argv[2] || '127.0.0.1:9092'; +var topicName = process.argv[3] || 'test'; +var compression = process.argv[4] || 'gzip'; +var MAX = process.argv[5] || 1000000; + +var stream = Kafka.Producer.createWriteStream({ + 'metadata.broker.list': host, + 'group.id': 'node-rdkafka-bench', + 'compression.codec': compression, + 'retry.backoff.ms': 200, + 'message.send.max.retries': 10, + 'socket.keepalive.enable': true, + 'queue.buffering.max.messages': 100000, + 'queue.buffering.max.ms': 1000, + 'batch.num.messages': 1000, +}, {}, { + topic: topicName, + pollInterval: 20 +}); + +stream.on('error', function(e) { + console.log(e); + process.exit(1); +}); + +// Track how many messages we see per second +var interval; +var done = false; + +function log() { + console.log('%d messages per sent second', count); + store.push(count); + count = 0; +} + +crypto.randomBytes(4096, function(ex, buffer) { + + var x = function(e) { + if (e) { + console.error(e); + } + count += 1; + totalComplete += 1; + if (totalComplete >= MAX && !done) { + done = true; + clearInterval(interval); + setTimeout(shutdown, 5000); + } + }; + + function write() { + if (!stream.write(buffer, 'base64', x)) { + return stream.once('drain', write); + } else { + total++; + } + + if (total < MAX) { + // we are not done + setImmediate(write); + } + + } + + write(); + interval = setInterval(log, 1000); + stream.on('error', function(err) { + console.log(err); + }); + // stream.on('end', shutdown); + +}); + + +process.once('SIGTERM', shutdown); +process.once('SIGINT', shutdown); +process.once('SIGHUP', shutdown); + +function shutdown() { + + if (store.length > 0) { + var calc = 0; + for (var x in store) { + calc += store[x]; + } + + var mps = parseFloat(calc * 1.0/store.length); + + console.log('%d messages per second on average', mps); + console.log('%d messages total', total); + + } + + clearInterval(interval); + + stream.end(); + + stream.on('close', function() { + console.log('total: %d', total); + }); + +} diff --git a/bench/seed.sh b/bench/seed.sh new file mode 100755 index 0000000..5395aaa --- /dev/null +++ b/bench/seed.sh @@ -0,0 +1,38 @@ +#!/bin/bash +kafka_root=${KAFKA_ROOT:-/opt/kafka} +# Generate and insert some messages + +OS=$(uname -s) + +function initializeTopic { + topic=$1 + host=$2 + msg_size=$3 + batch_size=$4 + batch_count=$5 + + if [ $host == "localhost:9092" ]; then + ${kafka_root}/bin/kafka-topics.sh --create --zookeeper localhost:2181 \ + --replication-factor 1 --partitions 1 --topic ${topic} + fi + + echo "Generating messages (size: ${msg_size})" + : > /tmp/msgs # Truncate /tmp/msgs + for i in $(seq 1 ${batch_size}); do + if [ $OS == 'Darwin' ]; then + printf %s\\n "$(head -c${msg_size} /dev/urandom | base64)" >> /tmp/msgs + else + printf %s\\n "$(head --bytes=${msg_size} /dev/urandom | base64 --wrap=0)" >> /tmp/msgs + fi + done + + echo "Done generating messages" + + for i in $(seq 1 ${batch_count}); do + echo "Adding $(wc -l /tmp/msgs) messages to topic ${topic}" + "${kafka_root}/bin/kafka-console-producer.sh" \ + --broker-list ${host} --topic ${topic} < /tmp/msgs + done +} + +initializeTopic "librdtesting-01" "localhost:9092" "4096" "5000" "2000" diff --git a/binding.gyp b/binding.gyp new file mode 100644 index 0000000..529d5d9 --- /dev/null +++ b/binding.gyp @@ -0,0 +1,154 @@ +{ + "variables": { + # may be redefined in command line on configuration stage + # "BUILD_LIBRDKAFKA%": " line.startsWith('#def')); +const definedLines = precompilerDefinitions.map(definedLine => { + const content = definedLine.split(' ').filter(v => v != ''); + + return { + command: content[0], + key: content[1], + value: content[2] + }; +}); + +const defines = {}; + +for (let item of definedLines) { + if (item.command == '#define') { + defines[item.key] = item.value; + } +} + +function parseLibrdkafkaVersion(version) { + const intRepresentation = parseInt(version); + + const major = (intRepresentation & majorMask) >> (8 * 3); + const minor = (intRepresentation & minorMask) >> (8 * 2); + const patch = (intRepresentation & patchMask) >> (8 * 1); + const rev = (intRepresentation & revMask) >> (8 * 0); + + return { + major, + minor, + patch, + rev + }; +} + +function versionAsString(version) { + return [ + version.major, + version.minor, + version.patch, + version.rev === 255 ? null : version.rev, + ].filter(v => v != null).join('.'); +} + +const librdkafkaVersion = parseLibrdkafkaVersion(defines.RD_KAFKA_VERSION); +const versionString = versionAsString(librdkafkaVersion); + +if (pjs.librdkafka !== versionString) { + console.error(`Librdkafka version of ${versionString} does not match package json: ${pjs.librdkafka}`); + process.exit(1); +} diff --git a/ci/checks/librdkafka-exists.js b/ci/checks/librdkafka-exists.js new file mode 100644 index 0000000..c156476 --- /dev/null +++ b/ci/checks/librdkafka-exists.js @@ -0,0 +1,13 @@ +const path = require('path'); +const fs = require('fs'); + +const root = path.resolve(__dirname, '..', '..'); +const librdkafkaPath = path.resolve(root, 'deps', 'librdkafka'); + +// Ensure librdkafka is in the deps directory - this makes sure we don't accidentally +// publish on a non recursive clone :) + +if (!fs.existsSync(librdkafkaPath)) { + console.error(`Could not find librdkafka at path ${librdkafkaPath}`); + process.exit(1); +} diff --git a/ci/librdkafka-defs-generator.js b/ci/librdkafka-defs-generator.js new file mode 100644 index 0000000..60f95f7 --- /dev/null +++ b/ci/librdkafka-defs-generator.js @@ -0,0 +1,193 @@ +const fs = require('fs'); +const path = require('path'); + +const LIBRDKAFKA_VERSION = require('../package.json').librdkafka; +const LIBRDKAFKA_DIR = path.resolve(__dirname, '../deps/librdkafka/'); + +function getHeader(file) { + return `// ====== Generated from librdkafka ${LIBRDKAFKA_VERSION} file ${file} ======`; +} + +function readLibRDKafkaFile(file) { + return fs.readFileSync(path.resolve(LIBRDKAFKA_DIR, file)).toString(); +} + +function extractConfigItems(configStr) { + const [_header, config] = configStr.split(/-{5,}\|.*/); + + const re = /(.*?)\|(.*?)\|(.*?)\|(.*?)\|(.*?)\|(.*)/g; + + const configItems = []; + + let m; + do { + m = re.exec(config); + if (m) { + const [ + _fullString, + property, + consumerOrProducer, + range, + defaultValue, + importance, + descriptionWithType, + ] = m.map(el => (typeof el === 'string' ? el.trim() : el)); + + const splitDescriptionRe = /(.*?)\s*?
.*?:\s.*?(.*?)\*/; + const [_, description, rawType] = splitDescriptionRe.exec(descriptionWithType); + + configItems.push({ + property, + consumerOrProducer, + range, + defaultValue, + importance, + description, + rawType, + }); + } + } while (m); + + return configItems.map(processItem); +} + +function processItem(configItem) { + // These items are overwritten by node-rdkafka + switch (configItem.property) { + case 'dr_msg_cb': + return { ...configItem, type: 'boolean' }; + case 'dr_cb': + return { ...configItem, type: 'boolean | Function' }; + case 'rebalance_cb': + return { ...configItem, type: 'boolean | Function' }; + case 'offset_commit_cb': + return { ...configItem, type: 'boolean | Function' }; + } + + switch (configItem.rawType) { + case 'integer': + return { ...configItem, type: 'number' }; + case 'boolean': + return { ...configItem, type: 'boolean' }; + case 'string': + case 'CSV flags': + return { ...configItem, type: 'string' }; + case 'enum value': + return { + ...configItem, + type: configItem.range + .split(',') + .map(str => `'${str.trim()}'`) + .join(' | '), + }; + default: + return { ...configItem, type: 'any' }; + } +} + +function generateInterface(interfaceDef, configItems) { + const fields = configItems + .map(item => + [ + `/**`, + ` * ${item.description}`, + ...(item.defaultValue ? [` *`, ` * @default ${item.defaultValue}`] : []), + ` */`, + `"${item.property}"?: ${item.type};`, + ] + .map(row => ` ${row}`) + .join('\n') + ) + .join('\n\n'); + + return `export interface ` + interfaceDef + ' {\n' + fields + '\n}'; +} + +function addSpecialGlobalProps(globalProps) { + globalProps.push({ + "property": "event_cb", + "consumerOrProducer": "*", + "range": "", + "defaultValue": "true", + "importance": "low", + "description": "Enables or disables `event.*` emitting.", + "rawType": "boolean", + "type": "boolean" + }); +} + +function generateConfigDTS(file) { + const configuration = readLibRDKafkaFile(file); + const [globalStr, topicStr] = configuration.split('Topic configuration properties'); + + const [globalProps, topicProps] = [extractConfigItems(globalStr), extractConfigItems(topicStr)]; + + addSpecialGlobalProps(globalProps); + + const [globalSharedProps, producerGlobalProps, consumerGlobalProps] = [ + globalProps.filter(i => i.consumerOrProducer === '*'), + globalProps.filter(i => i.consumerOrProducer === 'P'), + globalProps.filter(i => i.consumerOrProducer === 'C'), + ]; + + const [topicSharedProps, producerTopicProps, consumerTopicProps] = [ + topicProps.filter(i => i.consumerOrProducer === '*'), + topicProps.filter(i => i.consumerOrProducer === 'P'), + topicProps.filter(i => i.consumerOrProducer === 'C'), + ]; + + let output = `${getHeader(file)} +// Code that generated this is a derivative work of the code from Nam Nguyen +// https://gist.github.com/ntgn81/066c2c8ec5b4238f85d1e9168a04e3fb + +`; + + output += [ + generateInterface('GlobalConfig', globalSharedProps), + generateInterface('ProducerGlobalConfig extends GlobalConfig', producerGlobalProps), + generateInterface('ConsumerGlobalConfig extends GlobalConfig', consumerGlobalProps), + generateInterface('TopicConfig', topicSharedProps), + generateInterface('ProducerTopicConfig extends TopicConfig', producerTopicProps), + generateInterface('ConsumerTopicConfig extends TopicConfig', consumerTopicProps), + ].join('\n\n'); + + fs.writeFileSync(path.resolve(__dirname, '../config.d.ts'), output); +} + +function updateErrorDefinitions(file) { + const rdkafkacpp_h = readLibRDKafkaFile(file); + const m = /enum ErrorCode {([^}]+)}/g.exec(rdkafkacpp_h); + if (!m) { + throw new Error(`Can't read rdkafkacpp.h file`) + } + const body = m[1] + .replace(/(\t)|( +)/g, ' ') + .replace(/\n\n/g, '\n') + .replace(/\s+=\s+/g, ': ') + .replace(/[\t ]*#define +(\w+) +(\w+)/g, (_, define, original) => { + const value = new RegExp(`${original}\\s+=\\s+(\\d+)`).exec(m[1])[1]; + return ` ${define}: ${value},`; + }) + + // validate body + const emptyCheck = body + .replace(/(( \/\*)|( ?\*)).*/g, '') + .replace(/ ERR_\w+: -?\d+,?\n/g, '') + .trim() + if (emptyCheck !== '') { + throw new Error(`Fail to parse ${file}. It contains these extra details:\n${emptyCheck}`); + } + + const error_js_file = path.resolve(__dirname, '../lib/error.js'); + const error_js = fs.readFileSync(error_js_file) + .toString() + .replace(/(\/\/.*\n)?LibrdKafkaError.codes = {[^}]+/g, `${getHeader(file)}\nLibrdKafkaError.codes = {\n${body}`) + + fs.writeFileSync(error_js_file, error_js); + fs.writeFileSync(path.resolve(__dirname, '../errors.d.ts'), `${getHeader(file)}\nexport const CODES: { ERRORS: {${body.replace(/[ \.]*(\*\/\n \w+: )(-?\d+),?/g, ' (**$2**) $1number,')}}}`) +} + +(async function updateTypeDefs() { + generateConfigDTS('CONFIGURATION.md'); + updateErrorDefinitions('src-cpp/rdkafkacpp.h'); +})() diff --git a/ci/prepublish.js b/ci/prepublish.js new file mode 100644 index 0000000..3526be8 --- /dev/null +++ b/ci/prepublish.js @@ -0,0 +1,4 @@ +require('./checks/librdkafka-exists'); +require('./checks/librdkafka-correct-version'); +require('./librdkafka-defs-generator.js'); +require('./update-version'); diff --git a/ci/update-version.js b/ci/update-version.js new file mode 100644 index 0000000..d14cdfc --- /dev/null +++ b/ci/update-version.js @@ -0,0 +1,119 @@ +const path = require('path'); +const semver = require('semver'); +const { spawn } = require('child_process'); +const fs = require('fs'); + +const root = path.resolve(__dirname, '..'); +const pjsPath = path.resolve(root, 'package.json'); +const pjs = require(pjsPath); + +function parseVersion(tag) { + const { major, minor, prerelease, patch } = semver.parse(tag); + + // Describe will give is commits since last tag + const [ commitsSinceTag, hash ] = prerelease[0] ? prerelease[0].split('-') : [ + 1, + process.env.TRAVIS_COMMIT || '' + ]; + + return { + major, + minor, + prerelease, + patch, + commit: commitsSinceTag - 1, + hash + }; +} + +function getCommandOutput(command, args, cb) { + let output = ''; + + const cmd = spawn(command, args); + + cmd.stdout.on('data', (data) => { + output += data; + }); + + cmd.on('close', (code) => { + if (code != 0) { + cb(new Error(`Command returned unsuccessful code: ${code}`)); + return; + } + + cb(null, output.trim()); + }); +} + +function getVersion(cb) { + // https://docs.travis-ci.com/user/environment-variables/ + if (process.env.TRAVIS_TAG) { + setImmediate(() => cb(null, parseVersion(process.env.TRAVIS_TAG.trim()))); + return; + } + + getCommandOutput('git', ['describe', '--tags'], (err, result) => { + if (err) { + cb(err); + return; + } + + cb(null, parseVersion(result.trim())); + }); +} + +function getBranch(cb) { + if (process.env.TRAVIS_TAG) { + // TRAVIS_BRANCH matches TRAVIS_TAG when TRAVIS_TAG is set + // "git branch --contains tags/TRAVIS_TAG" doesn't work on travis so we have to assume 'master' + setImmediate(() => cb(null, 'master')); + return; + } else if (process.env.TRAVIS_BRANCH) { + setImmediate(() => cb(null, process.env.TRAVIS_BRANCH.trim())); + return; + } + + getCommandOutput('git', ['rev-parse', '--abbrev-ref', 'HEAD'], (err, result) => { + if (err) { + cb(err); + return; + } + + cb(null, result.trim()); + }); +} + +function getPackageVersion(tag, branch) { + let baseVersion = `v${tag.major}.${tag.minor}.${tag.patch}`; + + if (tag.commit === 0 && branch === 'master') { + return baseVersion; + } + + baseVersion += '-'; + + if (branch !== 'master') { + baseVersion += (tag.commit + 1 + '.' + branch); + } else { + baseVersion += (tag.commit + 1); + } + + return baseVersion; +} + +getVersion((err, tag) => { + if (err) { + throw err; + } + + getBranch((err, branch) => { + if (err) { + throw err; + } + + pjs.version = getPackageVersion(tag, branch); + + fs.writeFileSync(pjsPath, JSON.stringify(pjs, null, 2)); + }) + +}); diff --git a/config.d.ts b/config.d.ts new file mode 100644 index 0000000..5666b89 --- /dev/null +++ b/config.d.ts @@ -0,0 +1,1023 @@ +// ====== Generated from librdkafka 1.6.1 file CONFIGURATION.md ====== +// Code that generated this is a derivative work of the code from Nam Nguyen +// https://gist.github.com/ntgn81/066c2c8ec5b4238f85d1e9168a04e3fb + +export interface GlobalConfig { + /** + * Indicates the builtin features for this build of librdkafka. An application can either query this value or attempt to set it with its list of required features to check for library support. + * + * @default gzip, snappy, ssl, sasl, regex, lz4, sasl_gssapi, sasl_plain, sasl_scram, plugins, zstd, sasl_oauthbearer + */ + "builtin.features"?: string; + + /** + * Client identifier. + * + * @default rdkafka + */ + "client.id"?: string; + + /** + * Initial list of brokers as a CSV list of broker host or host:port. The application may also use `rd_kafka_brokers_add()` to add brokers during runtime. + */ + "metadata.broker.list"?: string; + + /** + * Alias for `metadata.broker.list`: Initial list of brokers as a CSV list of broker host or host:port. The application may also use `rd_kafka_brokers_add()` to add brokers during runtime. + */ + "bootstrap.servers"?: string; + + /** + * Maximum Kafka protocol request message size. Due to differing framing overhead between protocol versions the producer is unable to reliably enforce a strict max message limit at produce time and may exceed the maximum size by one message in protocol ProduceRequests, the broker will enforce the the topic's `max.message.bytes` limit (see Apache Kafka documentation). + * + * @default 1000000 + */ + "message.max.bytes"?: number; + + /** + * Maximum size for message to be copied to buffer. Messages larger than this will be passed by reference (zero-copy) at the expense of larger iovecs. + * + * @default 65535 + */ + "message.copy.max.bytes"?: number; + + /** + * Maximum Kafka protocol response message size. This serves as a safety precaution to avoid memory exhaustion in case of protocol hickups. This value must be at least `fetch.max.bytes` + 512 to allow for protocol overhead; the value is adjusted automatically unless the configuration property is explicitly set. + * + * @default 100000000 + */ + "receive.message.max.bytes"?: number; + + /** + * Maximum number of in-flight requests per broker connection. This is a generic property applied to all broker communication, however it is primarily relevant to produce requests. In particular, note that other mechanisms limit the number of outstanding consumer fetch request per broker to one. + * + * @default 1000000 + */ + "max.in.flight.requests.per.connection"?: number; + + /** + * Alias for `max.in.flight.requests.per.connection`: Maximum number of in-flight requests per broker connection. This is a generic property applied to all broker communication, however it is primarily relevant to produce requests. In particular, note that other mechanisms limit the number of outstanding consumer fetch request per broker to one. + * + * @default 1000000 + */ + "max.in.flight"?: number; + + /** + * Non-topic request timeout in milliseconds. This is for metadata requests, etc. + * + * @default 60000 + */ + "metadata.request.timeout.ms"?: number; + + /** + * Period of time in milliseconds at which topic and broker metadata is refreshed in order to proactively discover any new brokers, topics, partitions or partition leader changes. Use -1 to disable the intervalled refresh (not recommended). If there are no locally referenced topics (no topic objects created, no messages produced, no subscription or no assignment) then only the broker list will be refreshed every interval but no more often than every 10s. + * + * @default 300000 + */ + "topic.metadata.refresh.interval.ms"?: number; + + /** + * Metadata cache max age. Defaults to topic.metadata.refresh.interval.ms * 3 + * + * @default 900000 + */ + "metadata.max.age.ms"?: number; + + /** + * When a topic loses its leader a new metadata request will be enqueued with this initial interval, exponentially increasing until the topic metadata has been refreshed. This is used to recover quickly from transitioning leader brokers. + * + * @default 250 + */ + "topic.metadata.refresh.fast.interval.ms"?: number; + + /** + * **DEPRECATED** No longer used. + * + * @default 10 + */ + "topic.metadata.refresh.fast.cnt"?: number; + + /** + * Sparse metadata requests (consumes less network bandwidth) + * + * @default true + */ + "topic.metadata.refresh.sparse"?: boolean; + + /** + * Apache Kafka topic creation is asynchronous and it takes some time for a new topic to propagate throughout the cluster to all brokers. If a client requests topic metadata after manual topic creation but before the topic has been fully propagated to the broker the client is requesting metadata from, the topic will seem to be non-existent and the client will mark the topic as such, failing queued produced messages with `ERR__UNKNOWN_TOPIC`. This setting delays marking a topic as non-existent until the configured propagation max time has passed. The maximum propagation time is calculated from the time the topic is first referenced in the client, e.g., on produce(). + * + * @default 30000 + */ + "topic.metadata.propagation.max.ms"?: number; + + /** + * Topic blacklist, a comma-separated list of regular expressions for matching topic names that should be ignored in broker metadata information as if the topics did not exist. + */ + "topic.blacklist"?: any; + + /** + * A comma-separated list of debug contexts to enable. Detailed Producer debugging: broker,topic,msg. Consumer: consumer,cgrp,topic,fetch + */ + "debug"?: string; + + /** + * Default timeout for network requests. Producer: ProduceRequests will use the lesser value of `socket.timeout.ms` and remaining `message.timeout.ms` for the first message in the batch. Consumer: FetchRequests will use `fetch.wait.max.ms` + `socket.timeout.ms`. Admin: Admin requests will use `socket.timeout.ms` or explicitly set `rd_kafka_AdminOptions_set_operation_timeout()` value. + * + * @default 60000 + */ + "socket.timeout.ms"?: number; + + /** + * **DEPRECATED** No longer used. + * + * @default 1000 + */ + "socket.blocking.max.ms"?: number; + + /** + * Broker socket send buffer size. System default is used if 0. + * + * @default 0 + */ + "socket.send.buffer.bytes"?: number; + + /** + * Broker socket receive buffer size. System default is used if 0. + * + * @default 0 + */ + "socket.receive.buffer.bytes"?: number; + + /** + * Enable TCP keep-alives (SO_KEEPALIVE) on broker sockets + * + * @default false + */ + "socket.keepalive.enable"?: boolean; + + /** + * Disable the Nagle algorithm (TCP_NODELAY) on broker sockets. + * + * @default false + */ + "socket.nagle.disable"?: boolean; + + /** + * Disconnect from broker when this number of send failures (e.g., timed out requests) is reached. Disable with 0. WARNING: It is highly recommended to leave this setting at its default value of 1 to avoid the client and broker to become desynchronized in case of request timeouts. NOTE: The connection is automatically re-established. + * + * @default 1 + */ + "socket.max.fails"?: number; + + /** + * How long to cache the broker address resolving results (milliseconds). + * + * @default 1000 + */ + "broker.address.ttl"?: number; + + /** + * Allowed broker IP address families: any, v4, v6 + * + * @default any + */ + "broker.address.family"?: 'any' | 'v4' | 'v6'; + + /** + * **DEPRECATED** No longer used. See `reconnect.backoff.ms` and `reconnect.backoff.max.ms`. + * + * @default 0 + */ + "reconnect.backoff.jitter.ms"?: number; + + /** + * The initial time to wait before reconnecting to a broker after the connection has been closed. The time is increased exponentially until `reconnect.backoff.max.ms` is reached. -25% to +50% jitter is applied to each reconnect backoff. A value of 0 disables the backoff and reconnects immediately. + * + * @default 100 + */ + "reconnect.backoff.ms"?: number; + + /** + * The maximum time to wait before reconnecting to a broker after the connection has been closed. + * + * @default 10000 + */ + "reconnect.backoff.max.ms"?: number; + + /** + * librdkafka statistics emit interval. The application also needs to register a stats callback using `rd_kafka_conf_set_stats_cb()`. The granularity is 1000ms. A value of 0 disables statistics. + * + * @default 0 + */ + "statistics.interval.ms"?: number; + + /** + * See `rd_kafka_conf_set_events()` + * + * @default 0 + */ + "enabled_events"?: number; + + /** + * Error callback (set with rd_kafka_conf_set_error_cb()) + */ + "error_cb"?: any; + + /** + * Throttle callback (set with rd_kafka_conf_set_throttle_cb()) + */ + "throttle_cb"?: any; + + /** + * Statistics callback (set with rd_kafka_conf_set_stats_cb()) + */ + "stats_cb"?: any; + + /** + * Log callback (set with rd_kafka_conf_set_log_cb()) + */ + "log_cb"?: any; + + /** + * Logging level (syslog(3) levels) + * + * @default 6 + */ + "log_level"?: number; + + /** + * Disable spontaneous log_cb from internal librdkafka threads, instead enqueue log messages on queue set with `rd_kafka_set_log_queue()` and serve log callbacks or events through the standard poll APIs. **NOTE**: Log messages will linger in a temporary queue until the log queue has been set. + * + * @default false + */ + "log.queue"?: boolean; + + /** + * Print internal thread name in log messages (useful for debugging librdkafka internals) + * + * @default true + */ + "log.thread.name"?: boolean; + + /** + * If enabled librdkafka will initialize the PRNG with srand(current_time.milliseconds) on the first invocation of rd_kafka_new() (required only if rand_r() is not available on your platform). If disabled the application must call srand() prior to calling rd_kafka_new(). + * + * @default true + */ + "enable.random.seed"?: boolean; + + /** + * Log broker disconnects. It might be useful to turn this off when interacting with 0.9 brokers with an aggressive `connection.max.idle.ms` value. + * + * @default true + */ + "log.connection.close"?: boolean; + + /** + * Background queue event callback (set with rd_kafka_conf_set_background_event_cb()) + */ + "background_event_cb"?: any; + + /** + * Socket creation callback to provide race-free CLOEXEC + */ + "socket_cb"?: any; + + /** + * Socket connect callback + */ + "connect_cb"?: any; + + /** + * Socket close callback + */ + "closesocket_cb"?: any; + + /** + * File open callback to provide race-free CLOEXEC + */ + "open_cb"?: any; + + /** + * Application opaque (set with rd_kafka_conf_set_opaque()) + */ + "opaque"?: any; + + /** + * Default topic configuration for automatically subscribed topics + */ + "default_topic_conf"?: any; + + /** + * Signal that librdkafka will use to quickly terminate on rd_kafka_destroy(). If this signal is not set then there will be a delay before rd_kafka_wait_destroyed() returns true as internal threads are timing out their system calls. If this signal is set however the delay will be minimal. The application should mask this signal as an internal signal handler is installed. + * + * @default 0 + */ + "internal.termination.signal"?: number; + + /** + * Request broker's supported API versions to adjust functionality to available protocol features. If set to false, or the ApiVersionRequest fails, the fallback version `broker.version.fallback` will be used. **NOTE**: Depends on broker version >=0.10.0. If the request is not supported by (an older) broker the `broker.version.fallback` fallback is used. + * + * @default true + */ + "api.version.request"?: boolean; + + /** + * Timeout for broker API version requests. + * + * @default 10000 + */ + "api.version.request.timeout.ms"?: number; + + /** + * Dictates how long the `broker.version.fallback` fallback is used in the case the ApiVersionRequest fails. **NOTE**: The ApiVersionRequest is only issued when a new connection to the broker is made (such as after an upgrade). + * + * @default 0 + */ + "api.version.fallback.ms"?: number; + + /** + * Older broker versions (before 0.10.0) provide no way for a client to query for supported protocol features (ApiVersionRequest, see `api.version.request`) making it impossible for the client to know what features it may use. As a workaround a user may set this property to the expected broker version and the client will automatically adjust its feature set accordingly if the ApiVersionRequest fails (or is disabled). The fallback broker version will be used for `api.version.fallback.ms`. Valid values are: 0.9.0, 0.8.2, 0.8.1, 0.8.0. Any other value >= 0.10, such as 0.10.2.1, enables ApiVersionRequests. + * + * @default 0.10.0 + */ + "broker.version.fallback"?: string; + + /** + * Protocol used to communicate with brokers. + * + * @default plaintext + */ + "security.protocol"?: 'plaintext' | 'ssl' | 'sasl_plaintext' | 'sasl_ssl'; + + /** + * A cipher suite is a named combination of authentication, encryption, MAC and key exchange algorithm used to negotiate the security settings for a network connection using TLS or SSL network protocol. See manual page for `ciphers(1)` and `SSL_CTX_set_cipher_list(3). + */ + "ssl.cipher.suites"?: string; + + /** + * The supported-curves extension in the TLS ClientHello message specifies the curves (standard/named, or 'explicit' GF(2^k) or GF(p)) the client is willing to have the server use. See manual page for `SSL_CTX_set1_curves_list(3)`. OpenSSL >= 1.0.2 required. + */ + "ssl.curves.list"?: string; + + /** + * The client uses the TLS ClientHello signature_algorithms extension to indicate to the server which signature/hash algorithm pairs may be used in digital signatures. See manual page for `SSL_CTX_set1_sigalgs_list(3)`. OpenSSL >= 1.0.2 required. + */ + "ssl.sigalgs.list"?: string; + + /** + * Path to client's private key (PEM) used for authentication. + */ + "ssl.key.location"?: string; + + /** + * Private key passphrase (for use with `ssl.key.location` and `set_ssl_cert()`) + */ + "ssl.key.password"?: string; + + /** + * Client's private key string (PEM format) used for authentication. + */ + "ssl.key.pem"?: string; + + /** + * Client's private key as set by rd_kafka_conf_set_ssl_cert() + */ + "ssl_key"?: any; + + /** + * Path to client's public key (PEM) used for authentication. + */ + "ssl.certificate.location"?: string; + + /** + * Client's public key string (PEM format) used for authentication. + */ + "ssl.certificate.pem"?: string; + + /** + * Client's public key as set by rd_kafka_conf_set_ssl_cert() + */ + "ssl_certificate"?: any; + + /** + * File or directory path to CA certificate(s) for verifying the broker's key. Defaults: On Windows the system's CA certificates are automatically looked up in the Windows Root certificate store. On Mac OSX this configuration defaults to `probe`. It is recommended to install openssl using Homebrew, to provide CA certificates. On Linux install the distribution's ca-certificates package. If OpenSSL is statically linked or `ssl.ca.location` is set to `probe` a list of standard paths will be probed and the first one found will be used as the default CA certificate location path. If OpenSSL is dynamically linked the OpenSSL library's default path will be used (see `OPENSSLDIR` in `openssl version -a`). + * + * @default probe + */ + "ssl.ca.location"?: string; + + /** + * CA certificate as set by rd_kafka_conf_set_ssl_cert() + */ + "ssl_ca"?: any; + + /** + * Comma-separated list of Windows Certificate stores to load CA certificates from. Certificates will be loaded in the same order as stores are specified. If no certificates can be loaded from any of the specified stores an error is logged and the OpenSSL library's default CA location is used instead. Store names are typically one or more of: MY, Root, Trust, CA. + * + * @default Root + */ + "ssl.ca.certificate.stores"?: string; + + /** + * Path to CRL for verifying broker's certificate validity. + */ + "ssl.crl.location"?: string; + + /** + * Path to client's keystore (PKCS#12) used for authentication. + */ + "ssl.keystore.location"?: string; + + /** + * Client's keystore (PKCS#12) password. + */ + "ssl.keystore.password"?: string; + + /** + * Enable OpenSSL's builtin broker (server) certificate verification. This verification can be extended by the application by implementing a certificate_verify_cb. + * + * @default true + */ + "enable.ssl.certificate.verification"?: boolean; + + /** + * Endpoint identification algorithm to validate broker hostname using broker certificate. https - Server (broker) hostname verification as specified in RFC2818. none - No endpoint verification. OpenSSL >= 1.0.2 required. + * + * @default none + */ + "ssl.endpoint.identification.algorithm"?: 'none' | 'https'; + + /** + * Callback to verify the broker certificate chain. + */ + "ssl.certificate.verify_cb"?: any; + + /** + * SASL mechanism to use for authentication. Supported: GSSAPI, PLAIN, SCRAM-SHA-256, SCRAM-SHA-512, OAUTHBEARER. **NOTE**: Despite the name only one mechanism must be configured. + * + * @default GSSAPI + */ + "sasl.mechanisms"?: string; + + /** + * Alias for `sasl.mechanisms`: SASL mechanism to use for authentication. Supported: GSSAPI, PLAIN, SCRAM-SHA-256, SCRAM-SHA-512, OAUTHBEARER. **NOTE**: Despite the name only one mechanism must be configured. + * + * @default GSSAPI + */ + "sasl.mechanism"?: string; + + /** + * Kerberos principal name that Kafka runs as, not including /hostname@REALM + * + * @default kafka + */ + "sasl.kerberos.service.name"?: string; + + /** + * This client's Kerberos principal name. (Not supported on Windows, will use the logon user's principal). + * + * @default kafkaclient + */ + "sasl.kerberos.principal"?: string; + + /** + * kinit -t "%{sasl.kerberos.keytab}" -k %{sasl.kerberos.principal} | low | Shell command to refresh or acquire the client's Kerberos ticket. This command is executed on client creation and every sasl.kerberos.min.time.before.relogin (0=disable). %{config.prop.name} is replaced by corresponding config object value. + * + * @default kinit -R -t "%{sasl.kerberos.keytab}" -k %{sasl.kerberos.principal} \ + */ + "sasl.kerberos.kinit.cmd"?: string; + + /** + * Path to Kerberos keytab file. This configuration property is only used as a variable in `sasl.kerberos.kinit.cmd` as ` ... -t "%{sasl.kerberos.keytab}"`. + */ + "sasl.kerberos.keytab"?: string; + + /** + * Minimum time in milliseconds between key refresh attempts. Disable automatic key refresh by setting this property to 0. + * + * @default 60000 + */ + "sasl.kerberos.min.time.before.relogin"?: number; + + /** + * SASL username for use with the PLAIN and SASL-SCRAM-.. mechanisms + */ + "sasl.username"?: string; + + /** + * SASL password for use with the PLAIN and SASL-SCRAM-.. mechanism + */ + "sasl.password"?: string; + + /** + * SASL/OAUTHBEARER configuration. The format is implementation-dependent and must be parsed accordingly. The default unsecured token implementation (see https://tools.ietf.org/html/rfc7515#appendix-A.5) recognizes space-separated name=value pairs with valid names including principalClaimName, principal, scopeClaimName, scope, and lifeSeconds. The default value for principalClaimName is "sub", the default value for scopeClaimName is "scope", and the default value for lifeSeconds is 3600. The scope value is CSV format with the default value being no/empty scope. For example: `principalClaimName=azp principal=admin scopeClaimName=roles scope=role1,role2 lifeSeconds=600`. In addition, SASL extensions can be communicated to the broker via `extension_NAME=value`. For example: `principal=admin extension_traceId=123` + */ + "sasl.oauthbearer.config"?: string; + + /** + * Enable the builtin unsecure JWT OAUTHBEARER token handler if no oauthbearer_refresh_cb has been set. This builtin handler should only be used for development or testing, and not in production. + * + * @default false + */ + "enable.sasl.oauthbearer.unsecure.jwt"?: boolean; + + /** + * SASL/OAUTHBEARER token refresh callback (set with rd_kafka_conf_set_oauthbearer_token_refresh_cb(), triggered by rd_kafka_poll(), et.al. This callback will be triggered when it is time to refresh the client's OAUTHBEARER token. + */ + "oauthbearer_token_refresh_cb"?: any; + + /** + * List of plugin libraries to load (; separated). The library search path is platform dependent (see dlopen(3) for Unix and LoadLibrary() for Windows). If no filename extension is specified the platform-specific extension (such as .dll or .so) will be appended automatically. + */ + "plugin.library.paths"?: string; + + /** + * Interceptors added through rd_kafka_conf_interceptor_add_..() and any configuration handled by interceptors. + */ + "interceptors"?: any; + + /** + * A rack identifier for this client. This can be any string value which indicates where this client is physically located. It corresponds with the broker config `broker.rack`. + */ + "client.rack"?: string; + + /** + * Enables or disables `event.*` emitting. + * + * @default true + */ + "event_cb"?: boolean; +} + +export interface ProducerGlobalConfig extends GlobalConfig { + /** + * Enables the transactional producer. The transactional.id is used to identify the same transactional producer instance across process restarts. It allows the producer to guarantee that transactions corresponding to earlier instances of the same producer have been finalized prior to starting any new transactions, and that any zombie instances are fenced off. If no transactional.id is provided, then the producer is limited to idempotent delivery (if enable.idempotence is set). Requires broker version >= 0.11.0. + */ + "transactional.id"?: string; + + /** + * The maximum amount of time in milliseconds that the transaction coordinator will wait for a transaction status update from the producer before proactively aborting the ongoing transaction. If this value is larger than the `transaction.max.timeout.ms` setting in the broker, the init_transactions() call will fail with ERR_INVALID_TRANSACTION_TIMEOUT. The transaction timeout automatically adjusts `message.timeout.ms` and `socket.timeout.ms`, unless explicitly configured in which case they must not exceed the transaction timeout (`socket.timeout.ms` must be at least 100ms lower than `transaction.timeout.ms`). This is also the default timeout value if no timeout (-1) is supplied to the transactional API methods. + * + * @default 60000 + */ + "transaction.timeout.ms"?: number; + + /** + * When set to `true`, the producer will ensure that messages are successfully produced exactly once and in the original produce order. The following configuration properties are adjusted automatically (if not modified by the user) when idempotence is enabled: `max.in.flight.requests.per.connection=5` (must be less than or equal to 5), `retries=INT32_MAX` (must be greater than 0), `acks=all`, `queuing.strategy=fifo`. Producer instantation will fail if user-supplied configuration is incompatible. + * + * @default false + */ + "enable.idempotence"?: boolean; + + /** + * **EXPERIMENTAL**: subject to change or removal. When set to `true`, any error that could result in a gap in the produced message series when a batch of messages fails, will raise a fatal error (ERR__GAPLESS_GUARANTEE) and stop the producer. Messages failing due to `message.timeout.ms` are not covered by this guarantee. Requires `enable.idempotence=true`. + * + * @default false + */ + "enable.gapless.guarantee"?: boolean; + + /** + * Maximum number of messages allowed on the producer queue. This queue is shared by all topics and partitions. + * + * @default 100000 + */ + "queue.buffering.max.messages"?: number; + + /** + * Maximum total message size sum allowed on the producer queue. This queue is shared by all topics and partitions. This property has higher priority than queue.buffering.max.messages. + * + * @default 1048576 + */ + "queue.buffering.max.kbytes"?: number; + + /** + * Delay in milliseconds to wait for messages in the producer queue to accumulate before constructing message batches (MessageSets) to transmit to brokers. A higher value allows larger and more effective (less overhead, improved compression) batches of messages to accumulate at the expense of increased message delivery latency. + * + * @default 5 + */ + "queue.buffering.max.ms"?: any; + + /** + * Alias for `queue.buffering.max.ms`: Delay in milliseconds to wait for messages in the producer queue to accumulate before constructing message batches (MessageSets) to transmit to brokers. A higher value allows larger and more effective (less overhead, improved compression) batches of messages to accumulate at the expense of increased message delivery latency. + * + * @default 5 + */ + "linger.ms"?: any; + + /** + * How many times to retry sending a failing Message. **Note:** retrying may cause reordering unless `enable.idempotence` is set to true. + * + * @default 2147483647 + */ + "message.send.max.retries"?: number; + + /** + * Alias for `message.send.max.retries`: How many times to retry sending a failing Message. **Note:** retrying may cause reordering unless `enable.idempotence` is set to true. + * + * @default 2147483647 + */ + "retries"?: number; + + /** + * The backoff time in milliseconds before retrying a protocol request. + * + * @default 100 + */ + "retry.backoff.ms"?: number; + + /** + * The threshold of outstanding not yet transmitted broker requests needed to backpressure the producer's message accumulator. If the number of not yet transmitted requests equals or exceeds this number, produce request creation that would have otherwise been triggered (for example, in accordance with linger.ms) will be delayed. A lower number yields larger and more effective batches. A higher value can improve latency when using compression on slow machines. + * + * @default 1 + */ + "queue.buffering.backpressure.threshold"?: number; + + /** + * compression codec to use for compressing message sets. This is the default value for all topics, may be overridden by the topic configuration property `compression.codec`. + * + * @default none + */ + "compression.codec"?: 'none' | 'gzip' | 'snappy' | 'lz4' | 'zstd'; + + /** + * Alias for `compression.codec`: compression codec to use for compressing message sets. This is the default value for all topics, may be overridden by the topic configuration property `compression.codec`. + * + * @default none + */ + "compression.type"?: 'none' | 'gzip' | 'snappy' | 'lz4' | 'zstd'; + + /** + * Maximum number of messages batched in one MessageSet. The total MessageSet size is also limited by batch.size and message.max.bytes. + * + * @default 10000 + */ + "batch.num.messages"?: number; + + /** + * Maximum size (in bytes) of all messages batched in one MessageSet, including protocol framing overhead. This limit is applied after the first message has been added to the batch, regardless of the first message's size, this is to ensure that messages that exceed batch.size are produced. The total MessageSet size is also limited by batch.num.messages and message.max.bytes. + * + * @default 1000000 + */ + "batch.size"?: number; + + /** + * Only provide delivery reports for failed messages. + * + * @default false + */ + "delivery.report.only.error"?: boolean; + + /** + * Delivery report callback (set with rd_kafka_conf_set_dr_cb()) + */ + "dr_cb"?: boolean | Function; + + /** + * Delivery report callback (set with rd_kafka_conf_set_dr_msg_cb()) + */ + "dr_msg_cb"?: boolean; + + /** + * Delay in milliseconds to wait to assign new sticky partitions for each topic. By default, set to double the time of linger.ms. To disable sticky behavior, set to 0. This behavior affects messages with the key NULL in all cases, and messages with key lengths of zero when the consistent_random partitioner is in use. These messages would otherwise be assigned randomly. A higher value allows for more effective batching of these messages. + * + * @default 10 + */ + "sticky.partitioning.linger.ms"?: number; +} + +export interface ConsumerGlobalConfig extends GlobalConfig { + /** + * Client group id string. All clients sharing the same group.id belong to the same group. + */ + "group.id"?: string; + + /** + * Enable static group membership. Static group members are able to leave and rejoin a group within the configured `session.timeout.ms` without prompting a group rebalance. This should be used in combination with a larger `session.timeout.ms` to avoid group rebalances caused by transient unavailability (e.g. process restarts). Requires broker version >= 2.3.0. + */ + "group.instance.id"?: string; + + /** + * The name of one or more partition assignment strategies. The elected group leader will use a strategy supported by all members of the group to assign partitions to group members. If there is more than one eligible strategy, preference is determined by the order of this list (strategies earlier in the list have higher priority). Cooperative and non-cooperative (eager) strategies must not be mixed. Available strategies: range, roundrobin, cooperative-sticky. + * + * @default range,roundrobin + */ + "partition.assignment.strategy"?: string; + + /** + * Client group session and failure detection timeout. The consumer sends periodic heartbeats (heartbeat.interval.ms) to indicate its liveness to the broker. If no hearts are received by the broker for a group member within the session timeout, the broker will remove the consumer from the group and trigger a rebalance. The allowed range is configured with the **broker** configuration properties `group.min.session.timeout.ms` and `group.max.session.timeout.ms`. Also see `max.poll.interval.ms`. + * + * @default 10000 + */ + "session.timeout.ms"?: number; + + /** + * Group session keepalive heartbeat interval. + * + * @default 3000 + */ + "heartbeat.interval.ms"?: number; + + /** + * Group protocol type. NOTE: Currently, the only supported group protocol type is `consumer`. + * + * @default consumer + */ + "group.protocol.type"?: string; + + /** + * How often to query for the current client group coordinator. If the currently assigned coordinator is down the configured query interval will be divided by ten to more quickly recover in case of coordinator reassignment. + * + * @default 600000 + */ + "coordinator.query.interval.ms"?: number; + + /** + * Maximum allowed time between calls to consume messages (e.g., rd_kafka_consumer_poll()) for high-level consumers. If this interval is exceeded the consumer is considered failed and the group will rebalance in order to reassign the partitions to another consumer group member. Warning: Offset commits may be not possible at this point. Note: It is recommended to set `enable.auto.offset.store=false` for long-time processing applications and then explicitly store offsets (using offsets_store()) *after* message processing, to make sure offsets are not auto-committed prior to processing has finished. The interval is checked two times per second. See KIP-62 for more information. + * + * @default 300000 + */ + "max.poll.interval.ms"?: number; + + /** + * Automatically and periodically commit offsets in the background. Note: setting this to false does not prevent the consumer from fetching previously committed start offsets. To circumvent this behaviour set specific start offsets per partition in the call to assign(). + * + * @default true + */ + "enable.auto.commit"?: boolean; + + /** + * The frequency in milliseconds that the consumer offsets are committed (written) to offset storage. (0 = disable). This setting is used by the high-level consumer. + * + * @default 5000 + */ + "auto.commit.interval.ms"?: number; + + /** + * Automatically store offset of last message provided to application. The offset store is an in-memory store of the next offset to (auto-)commit for each partition. + * + * @default true + */ + "enable.auto.offset.store"?: boolean; + + /** + * Minimum number of messages per topic+partition librdkafka tries to maintain in the local consumer queue. + * + * @default 100000 + */ + "queued.min.messages"?: number; + + /** + * Maximum number of kilobytes of queued pre-fetched messages in the local consumer queue. If using the high-level consumer this setting applies to the single consumer queue, regardless of the number of partitions. When using the legacy simple consumer or when separate partition queues are used this setting applies per partition. This value may be overshot by fetch.message.max.bytes. This property has higher priority than queued.min.messages. + * + * @default 65536 + */ + "queued.max.messages.kbytes"?: number; + + /** + * Maximum time the broker may wait to fill the Fetch response with fetch.min.bytes of messages. + * + * @default 500 + */ + "fetch.wait.max.ms"?: number; + + /** + * Initial maximum number of bytes per topic+partition to request when fetching messages from the broker. If the client encounters a message larger than this value it will gradually try to increase it until the entire message can be fetched. + * + * @default 1048576 + */ + "fetch.message.max.bytes"?: number; + + /** + * Alias for `fetch.message.max.bytes`: Initial maximum number of bytes per topic+partition to request when fetching messages from the broker. If the client encounters a message larger than this value it will gradually try to increase it until the entire message can be fetched. + * + * @default 1048576 + */ + "max.partition.fetch.bytes"?: number; + + /** + * Maximum amount of data the broker shall return for a Fetch request. Messages are fetched in batches by the consumer and if the first message batch in the first non-empty partition of the Fetch request is larger than this value, then the message batch will still be returned to ensure the consumer can make progress. The maximum message batch size accepted by the broker is defined via `message.max.bytes` (broker config) or `max.message.bytes` (broker topic config). `fetch.max.bytes` is automatically adjusted upwards to be at least `message.max.bytes` (consumer config). + * + * @default 52428800 + */ + "fetch.max.bytes"?: number; + + /** + * Minimum number of bytes the broker responds with. If fetch.wait.max.ms expires the accumulated data will be sent to the client regardless of this setting. + * + * @default 1 + */ + "fetch.min.bytes"?: number; + + /** + * How long to postpone the next fetch request for a topic+partition in case of a fetch error. + * + * @default 500 + */ + "fetch.error.backoff.ms"?: number; + + /** + * **DEPRECATED** Offset commit store method: 'file' - DEPRECATED: local file store (offset.store.path, et.al), 'broker' - broker commit store (requires Apache Kafka 0.8.2 or later on the broker). + * + * @default broker + */ + "offset.store.method"?: 'none' | 'file' | 'broker'; + + /** + * Controls how to read messages written transactionally: `read_committed` - only return transactional messages which have been committed. `read_uncommitted` - return all messages, even transactional messages which have been aborted. + * + * @default read_committed + */ + "isolation.level"?: 'read_uncommitted' | 'read_committed'; + + /** + * Message consume callback (set with rd_kafka_conf_set_consume_cb()) + */ + "consume_cb"?: any; + + /** + * Called after consumer group has been rebalanced (set with rd_kafka_conf_set_rebalance_cb()) + */ + "rebalance_cb"?: boolean | Function; + + /** + * Offset commit result propagation callback. (set with rd_kafka_conf_set_offset_commit_cb()) + */ + "offset_commit_cb"?: boolean | Function; + + /** + * Emit RD_KAFKA_RESP_ERR__PARTITION_EOF event whenever the consumer reaches the end of a partition. + * + * @default false + */ + "enable.partition.eof"?: boolean; + + /** + * Verify CRC32 of consumed messages, ensuring no on-the-wire or on-disk corruption to the messages occurred. This check comes at slightly increased CPU usage. + * + * @default false + */ + "check.crcs"?: boolean; + + /** + * Allow automatic topic creation on the broker when subscribing to or assigning non-existent topics. The broker must also be configured with `auto.create.topics.enable=true` for this configuraiton to take effect. Note: The default value (false) is different from the Java consumer (true). Requires broker version >= 0.11.0.0, for older broker versions only the broker configuration applies. + * + * @default false + */ + "allow.auto.create.topics"?: boolean; +} + +export interface TopicConfig { + /** + * Application opaque (set with rd_kafka_topic_conf_set_opaque()) + */ + "opaque"?: any; +} + +export interface ProducerTopicConfig extends TopicConfig { + /** + * This field indicates the number of acknowledgements the leader broker must receive from ISR brokers before responding to the request: *0*=Broker does not send any response/ack to client, *-1* or *all*=Broker will block until message is committed by all in sync replicas (ISRs). If there are less than `min.insync.replicas` (broker configuration) in the ISR set the produce request will fail. + * + * @default -1 + */ + "request.required.acks"?: number; + + /** + * Alias for `request.required.acks`: This field indicates the number of acknowledgements the leader broker must receive from ISR brokers before responding to the request: *0*=Broker does not send any response/ack to client, *-1* or *all*=Broker will block until message is committed by all in sync replicas (ISRs). If there are less than `min.insync.replicas` (broker configuration) in the ISR set the produce request will fail. + * + * @default -1 + */ + "acks"?: number; + + /** + * The ack timeout of the producer request in milliseconds. This value is only enforced by the broker and relies on `request.required.acks` being != 0. + * + * @default 30000 + */ + "request.timeout.ms"?: number; + + /** + * Local message timeout. This value is only enforced locally and limits the time a produced message waits for successful delivery. A time of 0 is infinite. This is the maximum time librdkafka may use to deliver a message (including retries). Delivery error occurs when either the retry count or the message timeout are exceeded. The message timeout is automatically adjusted to `transaction.timeout.ms` if `transactional.id` is configured. + * + * @default 300000 + */ + "message.timeout.ms"?: number; + + /** + * Alias for `message.timeout.ms`: Local message timeout. This value is only enforced locally and limits the time a produced message waits for successful delivery. A time of 0 is infinite. This is the maximum time librdkafka may use to deliver a message (including retries). Delivery error occurs when either the retry count or the message timeout are exceeded. The message timeout is automatically adjusted to `transaction.timeout.ms` if `transactional.id` is configured. + * + * @default 300000 + */ + "delivery.timeout.ms"?: number; + + /** + * **EXPERIMENTAL**: subject to change or removal. **DEPRECATED** Producer queuing strategy. FIFO preserves produce ordering, while LIFO prioritizes new messages. + * + * @default fifo + */ + "queuing.strategy"?: 'fifo' | 'lifo'; + + /** + * **DEPRECATED** No longer used. + * + * @default false + */ + "produce.offset.report"?: boolean; + + /** + * Partitioner: `random` - random distribution, `consistent` - CRC32 hash of key (Empty and NULL keys are mapped to single partition), `consistent_random` - CRC32 hash of key (Empty and NULL keys are randomly partitioned), `murmur2` - Java Producer compatible Murmur2 hash of key (NULL keys are mapped to single partition), `murmur2_random` - Java Producer compatible Murmur2 hash of key (NULL keys are randomly partitioned. This is functionally equivalent to the default partitioner in the Java Producer.), `fnv1a` - FNV-1a hash of key (NULL keys are mapped to single partition), `fnv1a_random` - FNV-1a hash of key (NULL keys are randomly partitioned). + * + * @default consistent_random + */ + "partitioner"?: string; + + /** + * Custom partitioner callback (set with rd_kafka_topic_conf_set_partitioner_cb()) + */ + "partitioner_cb"?: any; + + /** + * **EXPERIMENTAL**: subject to change or removal. **DEPRECATED** Message queue ordering comparator (set with rd_kafka_topic_conf_set_msg_order_cmp()). Also see `queuing.strategy`. + */ + "msg_order_cmp"?: any; + + /** + * Compression codec to use for compressing message sets. inherit = inherit global compression.codec configuration. + * + * @default inherit + */ + "compression.codec"?: 'none' | 'gzip' | 'snappy' | 'lz4' | 'zstd' | 'inherit'; + + /** + * Alias for `compression.codec`: compression codec to use for compressing message sets. This is the default value for all topics, may be overridden by the topic configuration property `compression.codec`. + * + * @default none + */ + "compression.type"?: 'none' | 'gzip' | 'snappy' | 'lz4' | 'zstd'; + + /** + * Compression level parameter for algorithm selected by configuration property `compression.codec`. Higher values will result in better compression at the cost of more CPU usage. Usable range is algorithm-dependent: [0-9] for gzip; [0-12] for lz4; only 0 for snappy; -1 = codec-dependent default compression level. + * + * @default -1 + */ + "compression.level"?: number; +} + +export interface ConsumerTopicConfig extends TopicConfig { + /** + * **DEPRECATED** [**LEGACY PROPERTY:** This property is used by the simple legacy consumer only. When using the high-level KafkaConsumer, the global `enable.auto.commit` property must be used instead]. If true, periodically commit offset of the last message handed to the application. This committed offset will be used when the process restarts to pick up where it left off. If false, the application will have to call `rd_kafka_offset_store()` to store an offset (optional). **NOTE:** There is currently no zookeeper integration, offsets will be written to broker or local file according to offset.store.method. + * + * @default true + */ + "auto.commit.enable"?: boolean; + + /** + * **DEPRECATED** Alias for `auto.commit.enable`: [**LEGACY PROPERTY:** This property is used by the simple legacy consumer only. When using the high-level KafkaConsumer, the global `enable.auto.commit` property must be used instead]. If true, periodically commit offset of the last message handed to the application. This committed offset will be used when the process restarts to pick up where it left off. If false, the application will have to call `rd_kafka_offset_store()` to store an offset (optional). **NOTE:** There is currently no zookeeper integration, offsets will be written to broker or local file according to offset.store.method. + * + * @default true + */ + "enable.auto.commit"?: boolean; + + /** + * [**LEGACY PROPERTY:** This setting is used by the simple legacy consumer only. When using the high-level KafkaConsumer, the global `auto.commit.interval.ms` property must be used instead]. The frequency in milliseconds that the consumer offsets are committed (written) to offset storage. + * + * @default 60000 + */ + "auto.commit.interval.ms"?: number; + + /** + * Action to take when there is no initial offset in offset store or the desired offset is out of range: 'smallest','earliest' - automatically reset the offset to the smallest offset, 'largest','latest' - automatically reset the offset to the largest offset, 'error' - trigger an error (ERR__AUTO_OFFSET_RESET) which is retrieved by consuming messages and checking 'message->err'. + * + * @default largest + */ + "auto.offset.reset"?: 'smallest' | 'earliest' | 'beginning' | 'largest' | 'latest' | 'end' | 'error'; + + /** + * **DEPRECATED** Path to local file for storing offsets. If the path is a directory a filename will be automatically generated in that directory based on the topic and partition. File-based offset storage will be removed in a future version. + * + * @default . + */ + "offset.store.path"?: string; + + /** + * **DEPRECATED** fsync() interval for the offset file, in milliseconds. Use -1 to disable syncing, and 0 for immediate sync after each write. File-based offset storage will be removed in a future version. + * + * @default -1 + */ + "offset.store.sync.interval.ms"?: number; + + /** + * **DEPRECATED** Offset commit store method: 'file' - DEPRECATED: local file store (offset.store.path, et.al), 'broker' - broker commit store (requires "group.id" to be configured and Apache Kafka 0.8.2 or later on the broker.). + * + * @default broker + */ + "offset.store.method"?: 'file' | 'broker'; + + /** + * Maximum number of messages to dispatch in one `rd_kafka_consume_callback*()` call (0 = unlimited) + * + * @default 0 + */ + "consume.callback.max.messages"?: number; +} \ No newline at end of file diff --git a/configure b/configure new file mode 100755 index 0000000..8d25db3 --- /dev/null +++ b/configure @@ -0,0 +1,15 @@ +#!/usr/bin/env bash + +# This file is intended to be run on unix systems to configure librdkafka +# inside the submodules + +# This does not get run on windows which uses the build in solutions file + +# Get script directory +scriptdir=$(cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd) + +pushd ./deps/librdkafka &> /dev/null + +./configure --prefix="${scriptdir}/build/deps" --libdir="${scriptdir}/build/deps" $* + +popd &> /dev/null diff --git a/cpplint.py b/cpplint.py new file mode 100644 index 0000000..4472624 --- /dev/null +++ b/cpplint.py @@ -0,0 +1,6325 @@ +#!/usr/bin/env python +# +# Copyright (c) 2009 Google Inc. All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are +# met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above +# copyright notice, this list of conditions and the following disclaimer +# in the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Google Inc. nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +"""Does google-lint on c++ files. + +The goal of this script is to identify places in the code that *may* +be in non-compliance with google style. It does not attempt to fix +up these problems -- the point is to educate. It does also not +attempt to find all problems, or to ensure that everything it does +find is legitimately a problem. + +In particular, we can get very confused by /* and // inside strings! +We do a small hack, which is to ignore //'s with "'s after them on the +same line, but it is far from perfect (in either direction). +""" + +import codecs +import copy +import getopt +import math # for log +import os +import re +import sre_compile +import string +import sys +import unicodedata + + +_USAGE = """ +Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...] + [--counting=total|toplevel|detailed] [--root=subdir] + [--linelength=digits] + [file] ... + + The style guidelines this tries to follow are those in + http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml + + Every problem is given a confidence score from 1-5, with 5 meaning we are + certain of the problem, and 1 meaning it could be a legitimate construct. + This will miss some errors, and is not a substitute for a code review. + + To suppress false-positive errors of a certain category, add a + 'NOLINT(category)' comment to the line. NOLINT or NOLINT(*) + suppresses errors of all categories on that line. + + The files passed in will be linted; at least one file must be provided. + Default linted extensions are .cc, .cpp, .cu, .cuh and .h. Change the + extensions with the --extensions flag. + + Flags: + + output=vs7 + By default, the output is formatted to ease emacs parsing. Visual Studio + compatible output (vs7) may also be used. Other formats are unsupported. + + verbose=# + Specify a number 0-5 to restrict errors to certain verbosity levels. + + filter=-x,+y,... + Specify a comma-separated list of category-filters to apply: only + error messages whose category names pass the filters will be printed. + (Category names are printed with the message and look like + "[whitespace/indent]".) Filters are evaluated left to right. + "-FOO" and "FOO" means "do not print categories that start with FOO". + "+FOO" means "do print categories that start with FOO". + + Examples: --filter=-whitespace,+whitespace/braces + --filter=whitespace,runtime/printf,+runtime/printf_format + --filter=-,+build/include_what_you_use + + To see a list of all the categories used in cpplint, pass no arg: + --filter= + + counting=total|toplevel|detailed + The total number of errors found is always printed. If + 'toplevel' is provided, then the count of errors in each of + the top-level categories like 'build' and 'whitespace' will + also be printed. If 'detailed' is provided, then a count + is provided for each category like 'build/class'. + + root=subdir + The root directory used for deriving header guard CPP variable. + By default, the header guard CPP variable is calculated as the relative + path to the directory that contains .git, .hg, or .svn. When this flag + is specified, the relative path is calculated from the specified + directory. If the specified directory does not exist, this flag is + ignored. + + Examples: + Assuming that src/.git exists, the header guard CPP variables for + src/chrome/browser/ui/browser.h are: + + No flag => CHROME_BROWSER_UI_BROWSER_H_ + --root=chrome => BROWSER_UI_BROWSER_H_ + --root=chrome/browser => UI_BROWSER_H_ + + linelength=digits + This is the allowed line length for the project. The default value is + 80 characters. + + Examples: + --linelength=120 + + extensions=extension,extension,... + The allowed file extensions that cpplint will check + + Examples: + --extensions=hpp,cpp + + cpplint.py supports per-directory configurations specified in CPPLINT.cfg + files. CPPLINT.cfg file can contain a number of key=value pairs. + Currently the following options are supported: + + set noparent + filter=+filter1,-filter2,... + exclude_files=regex + linelength=80 + + "set noparent" option prevents cpplint from traversing directory tree + upwards looking for more .cfg files in parent directories. This option + is usually placed in the top-level project directory. + + The "filter" option is similar in function to --filter flag. It specifies + message filters in addition to the |_DEFAULT_FILTERS| and those specified + through --filter command-line flag. + + "exclude_files" allows to specify a regular expression to be matched against + a file name. If the expression matches, the file is skipped and not run + through liner. + + "linelength" allows to specify the allowed line length for the project. + + CPPLINT.cfg has an effect on files in the same directory and all + sub-directories, unless overridden by a nested configuration file. + + Example file: + filter=-build/include_order,+build/include_alpha + exclude_files=.*\.cc + + The above example disables build/include_order warning and enables + build/include_alpha as well as excludes all .cc from being + processed by linter, in the current directory (where the .cfg + file is located) and all sub-directories. +""" + +# We categorize each error message we print. Here are the categories. +# We want an explicit list so we can list them all in cpplint --filter=. +# If you add a new error message with a new category, add it to the list +# here! cpplint_unittest.py should tell you if you forget to do this. +_ERROR_CATEGORIES = [ + 'build/class', + 'build/c++11', + 'build/deprecated', + 'build/endif_comment', + 'build/explicit_make_pair', + 'build/forward_decl', + 'build/header_guard', + 'build/include', + 'build/include_alpha', + 'build/include_order', + 'build/include_what_you_use', + 'build/namespaces', + 'build/printf_format', + 'build/storage_class', + 'legal/copyright', + 'readability/alt_tokens', + 'readability/braces', + 'readability/casting', + 'readability/check', + 'readability/constructors', + 'readability/fn_size', + 'readability/function', + 'readability/inheritance', + 'readability/multiline_comment', + 'readability/multiline_string', + 'readability/namespace', + 'readability/nolint', + 'readability/nul', + 'readability/strings', + 'readability/todo', + 'readability/utf8', + 'runtime/arrays', + 'runtime/casting', + 'runtime/explicit', + 'runtime/int', + 'runtime/init', + 'runtime/invalid_increment', + 'runtime/member_string_references', + 'runtime/memset', + 'runtime/indentation_namespace', + 'runtime/operator', + 'runtime/printf', + 'runtime/printf_format', + 'runtime/references', + 'runtime/string', + 'runtime/threadsafe_fn', + 'runtime/vlog', + 'whitespace/blank_line', + 'whitespace/braces', + 'whitespace/comma', + 'whitespace/comments', + 'whitespace/empty_conditional_body', + 'whitespace/empty_loop_body', + 'whitespace/end_of_line', + 'whitespace/ending_newline', + 'whitespace/forcolon', + 'whitespace/indent', + 'whitespace/line_length', + 'whitespace/newline', + 'whitespace/operators', + 'whitespace/parens', + 'whitespace/semicolon', + 'whitespace/tab', + 'whitespace/todo', + ] + +# These error categories are no longer enforced by cpplint, but for backwards- +# compatibility they may still appear in NOLINT comments. +_LEGACY_ERROR_CATEGORIES = [ + 'readability/streams', + ] + +# The default state of the category filter. This is overridden by the --filter= +# flag. By default all errors are on, so only add here categories that should be +# off by default (i.e., categories that must be enabled by the --filter= flags). +# All entries here should start with a '-' or '+', as in the --filter= flag. +_DEFAULT_FILTERS = ['-build/include_alpha'] + +# We used to check for high-bit characters, but after much discussion we +# decided those were OK, as long as they were in UTF-8 and didn't represent +# hard-coded international strings, which belong in a separate i18n file. + +# C++ headers +_CPP_HEADERS = frozenset([ + # Legacy + 'algobase.h', + 'algo.h', + 'alloc.h', + 'builtinbuf.h', + 'bvector.h', + 'complex.h', + 'defalloc.h', + 'deque.h', + 'editbuf.h', + 'fstream.h', + 'function.h', + 'hash_map', + 'hash_map.h', + 'hash_set', + 'hash_set.h', + 'hashtable.h', + 'heap.h', + 'indstream.h', + 'iomanip.h', + 'iostream.h', + 'istream.h', + 'iterator.h', + 'list.h', + 'map.h', + 'multimap.h', + 'multiset.h', + 'ostream.h', + 'pair.h', + 'parsestream.h', + 'pfstream.h', + 'procbuf.h', + 'pthread_alloc', + 'pthread_alloc.h', + 'rope', + 'rope.h', + 'ropeimpl.h', + 'set.h', + 'slist', + 'slist.h', + 'stack.h', + 'stdiostream.h', + 'stl_alloc.h', + 'stl_relops.h', + 'streambuf.h', + 'stream.h', + 'strfile.h', + 'strstream.h', + 'tempbuf.h', + 'tree.h', + 'type_traits.h', + 'vector.h', + # 17.6.1.2 C++ library headers + 'algorithm', + 'array', + 'atomic', + 'bitset', + 'chrono', + 'codecvt', + 'complex', + 'condition_variable', + 'deque', + 'exception', + 'forward_list', + 'fstream', + 'functional', + 'future', + 'initializer_list', + 'iomanip', + 'ios', + 'iosfwd', + 'iostream', + 'istream', + 'iterator', + 'limits', + 'list', + 'locale', + 'map', + 'memory', + 'mutex', + 'new', + 'numeric', + 'ostream', + 'queue', + 'random', + 'ratio', + 'regex', + 'set', + 'sstream', + 'stack', + 'stdexcept', + 'streambuf', + 'string', + 'strstream', + 'system_error', + 'thread', + 'tuple', + 'typeindex', + 'typeinfo', + 'type_traits', + 'unordered_map', + 'unordered_set', + 'utility', + 'valarray', + 'vector', + # 17.6.1.2 C++ headers for C library facilities + 'cassert', + 'ccomplex', + 'cctype', + 'cerrno', + 'cfenv', + 'cfloat', + 'cinttypes', + 'ciso646', + 'climits', + 'clocale', + 'cmath', + 'csetjmp', + 'csignal', + 'cstdalign', + 'cstdarg', + 'cstdbool', + 'cstddef', + 'cstdint', + 'cstdio', + 'cstdlib', + 'cstring', + 'ctgmath', + 'ctime', + 'cuchar', + 'cwchar', + 'cwctype', + ]) + + +# These headers are excluded from [build/include] and [build/include_order] +# checks: +# - Anything not following google file name conventions (containing an +# uppercase character, such as Python.h or nsStringAPI.h, for example). +# - Lua headers. +# - rdkafka.cpp header, because it would be located in different directories depending +# on whether it's pulled from librdkafka sources or librdkafka-dev package. +_THIRD_PARTY_HEADERS_PATTERN = re.compile( + r'^(?:[^/]*[A-Z][^/]*\.h|lua\.h|lauxlib\.h|lualib\.h|rdkafkacpp\.h)$') + + +# Assertion macros. These are defined in base/logging.h and +# testing/base/gunit.h. Note that the _M versions need to come first +# for substring matching to work. +_CHECK_MACROS = [ + 'DCHECK', 'CHECK', + 'EXPECT_TRUE_M', 'EXPECT_TRUE', + 'ASSERT_TRUE_M', 'ASSERT_TRUE', + 'EXPECT_FALSE_M', 'EXPECT_FALSE', + 'ASSERT_FALSE_M', 'ASSERT_FALSE', + ] + +# Replacement macros for CHECK/DCHECK/EXPECT_TRUE/EXPECT_FALSE +_CHECK_REPLACEMENT = dict([(m, {}) for m in _CHECK_MACROS]) + +for op, replacement in [('==', 'EQ'), ('!=', 'NE'), + ('>=', 'GE'), ('>', 'GT'), + ('<=', 'LE'), ('<', 'LT')]: + _CHECK_REPLACEMENT['DCHECK'][op] = 'DCHECK_%s' % replacement + _CHECK_REPLACEMENT['CHECK'][op] = 'CHECK_%s' % replacement + _CHECK_REPLACEMENT['EXPECT_TRUE'][op] = 'EXPECT_%s' % replacement + _CHECK_REPLACEMENT['ASSERT_TRUE'][op] = 'ASSERT_%s' % replacement + _CHECK_REPLACEMENT['EXPECT_TRUE_M'][op] = 'EXPECT_%s_M' % replacement + _CHECK_REPLACEMENT['ASSERT_TRUE_M'][op] = 'ASSERT_%s_M' % replacement + +for op, inv_replacement in [('==', 'NE'), ('!=', 'EQ'), + ('>=', 'LT'), ('>', 'LE'), + ('<=', 'GT'), ('<', 'GE')]: + _CHECK_REPLACEMENT['EXPECT_FALSE'][op] = 'EXPECT_%s' % inv_replacement + _CHECK_REPLACEMENT['ASSERT_FALSE'][op] = 'ASSERT_%s' % inv_replacement + _CHECK_REPLACEMENT['EXPECT_FALSE_M'][op] = 'EXPECT_%s_M' % inv_replacement + _CHECK_REPLACEMENT['ASSERT_FALSE_M'][op] = 'ASSERT_%s_M' % inv_replacement + +# Alternative tokens and their replacements. For full list, see section 2.5 +# Alternative tokens [lex.digraph] in the C++ standard. +# +# Digraphs (such as '%:') are not included here since it's a mess to +# match those on a word boundary. +_ALT_TOKEN_REPLACEMENT = { + 'and': '&&', + 'bitor': '|', + 'or': '||', + 'xor': '^', + 'compl': '~', + 'bitand': '&', + 'and_eq': '&=', + 'or_eq': '|=', + 'xor_eq': '^=', + 'not': '!', + 'not_eq': '!=' + } + +# Compile regular expression that matches all the above keywords. The "[ =()]" +# bit is meant to avoid matching these keywords outside of boolean expressions. +# +# False positives include C-style multi-line comments and multi-line strings +# but those have always been troublesome for cpplint. +_ALT_TOKEN_REPLACEMENT_PATTERN = re.compile( + r'[ =()](' + ('|'.join(_ALT_TOKEN_REPLACEMENT.keys())) + r')(?=[ (]|$)') + + +# These constants define types of headers for use with +# _IncludeState.CheckNextIncludeOrder(). +_C_SYS_HEADER = 1 +_CPP_SYS_HEADER = 2 +_LIKELY_MY_HEADER = 3 +_POSSIBLE_MY_HEADER = 4 +_OTHER_HEADER = 5 + +# These constants define the current inline assembly state +_NO_ASM = 0 # Outside of inline assembly block +_INSIDE_ASM = 1 # Inside inline assembly block +_END_ASM = 2 # Last line of inline assembly block +_BLOCK_ASM = 3 # The whole block is an inline assembly block + +# Match start of assembly blocks +_MATCH_ASM = re.compile(r'^\s*(?:asm|_asm|__asm|__asm__)' + r'(?:\s+(volatile|__volatile__))?' + r'\s*[{(]') + + +_regexp_compile_cache = {} + +# {str, set(int)}: a map from error categories to sets of linenumbers +# on which those errors are expected and should be suppressed. +_error_suppressions = {} + +# The root directory used for deriving header guard CPP variable. +# This is set by --root flag. +_root = None + +# The allowed line length of files. +# This is set by --linelength flag. +_line_length = 80 + +# The allowed extensions for file names +# This is set by --extensions flag. +_valid_extensions = set(['cc', 'h', 'cpp', 'cu', 'cuh']) + +def ParseNolintSuppressions(filename, raw_line, linenum, error): + """Updates the global list of error-suppressions. + + Parses any NOLINT comments on the current line, updating the global + error_suppressions store. Reports an error if the NOLINT comment + was malformed. + + Args: + filename: str, the name of the input file. + raw_line: str, the line of input text, with comments. + linenum: int, the number of the current line. + error: function, an error handler. + """ + matched = Search(r'\bNOLINT(NEXTLINE)?\b(\([^)]+\))?', raw_line) + if matched: + if matched.group(1): + suppressed_line = linenum + 1 + else: + suppressed_line = linenum + category = matched.group(2) + if category in (None, '(*)'): # => "suppress all" + _error_suppressions.setdefault(None, set()).add(suppressed_line) + else: + if category.startswith('(') and category.endswith(')'): + category = category[1:-1] + if category in _ERROR_CATEGORIES: + _error_suppressions.setdefault(category, set()).add(suppressed_line) + elif category not in _LEGACY_ERROR_CATEGORIES: + error(filename, linenum, 'readability/nolint', 5, + 'Unknown NOLINT error category: %s' % category) + + +def ResetNolintSuppressions(): + """Resets the set of NOLINT suppressions to empty.""" + _error_suppressions.clear() + + +def IsErrorSuppressedByNolint(category, linenum): + """Returns true if the specified error category is suppressed on this line. + + Consults the global error_suppressions map populated by + ParseNolintSuppressions/ResetNolintSuppressions. + + Args: + category: str, the category of the error. + linenum: int, the current line number. + Returns: + bool, True iff the error should be suppressed due to a NOLINT comment. + """ + return (linenum in _error_suppressions.get(category, set()) or + linenum in _error_suppressions.get(None, set())) + + +def Match(pattern, s): + """Matches the string with the pattern, caching the compiled regexp.""" + # The regexp compilation caching is inlined in both Match and Search for + # performance reasons; factoring it out into a separate function turns out + # to be noticeably expensive. + if pattern not in _regexp_compile_cache: + _regexp_compile_cache[pattern] = sre_compile.compile(pattern) + return _regexp_compile_cache[pattern].match(s) + + +def ReplaceAll(pattern, rep, s): + """Replaces instances of pattern in a string with a replacement. + + The compiled regex is kept in a cache shared by Match and Search. + + Args: + pattern: regex pattern + rep: replacement text + s: search string + + Returns: + string with replacements made (or original string if no replacements) + """ + if pattern not in _regexp_compile_cache: + _regexp_compile_cache[pattern] = sre_compile.compile(pattern) + return _regexp_compile_cache[pattern].sub(rep, s) + + +def Search(pattern, s): + """Searches the string for the pattern, caching the compiled regexp.""" + if pattern not in _regexp_compile_cache: + _regexp_compile_cache[pattern] = sre_compile.compile(pattern) + return _regexp_compile_cache[pattern].search(s) + + +class _IncludeState(object): + """Tracks line numbers for includes, and the order in which includes appear. + + include_list contains list of lists of (header, line number) pairs. + It's a lists of lists rather than just one flat list to make it + easier to update across preprocessor boundaries. + + Call CheckNextIncludeOrder() once for each header in the file, passing + in the type constants defined above. Calls in an illegal order will + raise an _IncludeError with an appropriate error message. + + """ + # self._section will move monotonically through this set. If it ever + # needs to move backwards, CheckNextIncludeOrder will raise an error. + _INITIAL_SECTION = 0 + _MY_H_SECTION = 1 + _C_SECTION = 2 + _CPP_SECTION = 3 + _OTHER_H_SECTION = 4 + + _TYPE_NAMES = { + _C_SYS_HEADER: 'C system header', + _CPP_SYS_HEADER: 'C++ system header', + _LIKELY_MY_HEADER: 'header this file implements', + _POSSIBLE_MY_HEADER: 'header this file may implement', + _OTHER_HEADER: 'other header', + } + _SECTION_NAMES = { + _INITIAL_SECTION: "... nothing. (This can't be an error.)", + _MY_H_SECTION: 'a header this file implements', + _C_SECTION: 'C system header', + _CPP_SECTION: 'C++ system header', + _OTHER_H_SECTION: 'other header', + } + + def __init__(self): + self.include_list = [[]] + self.ResetSection('') + + def FindHeader(self, header): + """Check if a header has already been included. + + Args: + header: header to check. + Returns: + Line number of previous occurrence, or -1 if the header has not + been seen before. + """ + for section_list in self.include_list: + for f in section_list: + if f[0] == header: + return f[1] + return -1 + + def ResetSection(self, directive): + """Reset section checking for preprocessor directive. + + Args: + directive: preprocessor directive (e.g. "if", "else"). + """ + # The name of the current section. + self._section = self._INITIAL_SECTION + # The path of last found header. + self._last_header = '' + + # Update list of includes. Note that we never pop from the + # include list. + if directive in ('if', 'ifdef', 'ifndef'): + self.include_list.append([]) + elif directive in ('else', 'elif'): + self.include_list[-1] = [] + + def SetLastHeader(self, header_path): + self._last_header = header_path + + def CanonicalizeAlphabeticalOrder(self, header_path): + """Returns a path canonicalized for alphabetical comparison. + + - replaces "-" with "_" so they both cmp the same. + - removes '-inl' since we don't require them to be after the main header. + - lowercase everything, just in case. + + Args: + header_path: Path to be canonicalized. + + Returns: + Canonicalized path. + """ + return header_path.replace('-inl.h', '.h').replace('-', '_').lower() + + def IsInAlphabeticalOrder(self, clean_lines, linenum, header_path): + """Check if a header is in alphabetical order with the previous header. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + header_path: Canonicalized header to be checked. + + Returns: + Returns true if the header is in alphabetical order. + """ + # If previous section is different from current section, _last_header will + # be reset to empty string, so it's always less than current header. + # + # If previous line was a blank line, assume that the headers are + # intentionally sorted the way they are. + if (self._last_header > header_path and + Match(r'^\s*#\s*include\b', clean_lines.elided[linenum - 1])): + return False + return True + + def CheckNextIncludeOrder(self, header_type): + """Returns a non-empty error message if the next header is out of order. + + This function also updates the internal state to be ready to check + the next include. + + Args: + header_type: One of the _XXX_HEADER constants defined above. + + Returns: + The empty string if the header is in the right order, or an + error message describing what's wrong. + + """ + error_message = ('Found %s after %s' % + (self._TYPE_NAMES[header_type], + self._SECTION_NAMES[self._section])) + + last_section = self._section + + if header_type == _C_SYS_HEADER: + if self._section <= self._C_SECTION: + self._section = self._C_SECTION + else: + self._last_header = '' + return error_message + elif header_type == _CPP_SYS_HEADER: + if self._section <= self._CPP_SECTION: + self._section = self._CPP_SECTION + else: + self._last_header = '' + return error_message + elif header_type == _LIKELY_MY_HEADER: + if self._section <= self._MY_H_SECTION: + self._section = self._MY_H_SECTION + else: + self._section = self._OTHER_H_SECTION + elif header_type == _POSSIBLE_MY_HEADER: + if self._section <= self._MY_H_SECTION: + self._section = self._MY_H_SECTION + else: + # This will always be the fallback because we're not sure + # enough that the header is associated with this file. + self._section = self._OTHER_H_SECTION + else: + assert header_type == _OTHER_HEADER + self._section = self._OTHER_H_SECTION + + if last_section != self._section: + self._last_header = '' + + return '' + + +class _CppLintState(object): + """Maintains module-wide state..""" + + def __init__(self): + self.verbose_level = 1 # global setting. + self.error_count = 0 # global count of reported errors + # filters to apply when emitting error messages + self.filters = _DEFAULT_FILTERS[:] + # backup of filter list. Used to restore the state after each file. + self._filters_backup = self.filters[:] + self.counting = 'total' # In what way are we counting errors? + self.errors_by_category = {} # string to int dict storing error counts + + # output format: + # "emacs" - format that emacs can parse (default) + # "vs7" - format that Microsoft Visual Studio 7 can parse + self.output_format = 'emacs' + + def SetOutputFormat(self, output_format): + """Sets the output format for errors.""" + self.output_format = output_format + + def SetVerboseLevel(self, level): + """Sets the module's verbosity, and returns the previous setting.""" + last_verbose_level = self.verbose_level + self.verbose_level = level + return last_verbose_level + + def SetCountingStyle(self, counting_style): + """Sets the module's counting options.""" + self.counting = counting_style + + def SetFilters(self, filters): + """Sets the error-message filters. + + These filters are applied when deciding whether to emit a given + error message. + + Args: + filters: A string of comma-separated filters (eg "+whitespace/indent"). + Each filter should start with + or -; else we die. + + Raises: + ValueError: The comma-separated filters did not all start with '+' or '-'. + E.g. "-,+whitespace,-whitespace/indent,whitespace/badfilter" + """ + # Default filters always have less priority than the flag ones. + self.filters = _DEFAULT_FILTERS[:] + self.AddFilters(filters) + + def AddFilters(self, filters): + """ Adds more filters to the existing list of error-message filters. """ + for filt in filters.split(','): + clean_filt = filt.strip() + if clean_filt: + self.filters.append(clean_filt) + for filt in self.filters: + if not (filt.startswith('+') or filt.startswith('-')): + raise ValueError('Every filter in --filters must start with + or -' + ' (%s does not)' % filt) + + def BackupFilters(self): + """ Saves the current filter list to backup storage.""" + self._filters_backup = self.filters[:] + + def RestoreFilters(self): + """ Restores filters previously backed up.""" + self.filters = self._filters_backup[:] + + def ResetErrorCounts(self): + """Sets the module's error statistic back to zero.""" + self.error_count = 0 + self.errors_by_category = {} + + def IncrementErrorCount(self, category): + """Bumps the module's error statistic.""" + self.error_count += 1 + if self.counting in ('toplevel', 'detailed'): + if self.counting != 'detailed': + category = category.split('/')[0] + if category not in self.errors_by_category: + self.errors_by_category[category] = 0 + self.errors_by_category[category] += 1 + + def PrintErrorCounts(self): + """Print a summary of errors by category, and the total.""" + for category, count in self.errors_by_category.iteritems(): + sys.stderr.write('Category \'%s\' errors found: %d\n' % + (category, count)) + sys.stderr.write('Total errors found: %d\n' % self.error_count) + +_cpplint_state = _CppLintState() + + +def _OutputFormat(): + """Gets the module's output format.""" + return _cpplint_state.output_format + + +def _SetOutputFormat(output_format): + """Sets the module's output format.""" + _cpplint_state.SetOutputFormat(output_format) + + +def _VerboseLevel(): + """Returns the module's verbosity setting.""" + return _cpplint_state.verbose_level + + +def _SetVerboseLevel(level): + """Sets the module's verbosity, and returns the previous setting.""" + return _cpplint_state.SetVerboseLevel(level) + + +def _SetCountingStyle(level): + """Sets the module's counting options.""" + _cpplint_state.SetCountingStyle(level) + + +def _Filters(): + """Returns the module's list of output filters, as a list.""" + return _cpplint_state.filters + + +def _SetFilters(filters): + """Sets the module's error-message filters. + + These filters are applied when deciding whether to emit a given + error message. + + Args: + filters: A string of comma-separated filters (eg "whitespace/indent"). + Each filter should start with + or -; else we die. + """ + _cpplint_state.SetFilters(filters) + +def _AddFilters(filters): + """Adds more filter overrides. + + Unlike _SetFilters, this function does not reset the current list of filters + available. + + Args: + filters: A string of comma-separated filters (eg "whitespace/indent"). + Each filter should start with + or -; else we die. + """ + _cpplint_state.AddFilters(filters) + +def _BackupFilters(): + """ Saves the current filter list to backup storage.""" + _cpplint_state.BackupFilters() + +def _RestoreFilters(): + """ Restores filters previously backed up.""" + _cpplint_state.RestoreFilters() + +class _FunctionState(object): + """Tracks current function name and the number of lines in its body.""" + + _NORMAL_TRIGGER = 250 # for --v=0, 500 for --v=1, etc. + _TEST_TRIGGER = 400 # about 50% more than _NORMAL_TRIGGER. + + def __init__(self): + self.in_a_function = False + self.lines_in_function = 0 + self.current_function = '' + + def Begin(self, function_name): + """Start analyzing function body. + + Args: + function_name: The name of the function being tracked. + """ + self.in_a_function = True + self.lines_in_function = 0 + self.current_function = function_name + + def Count(self): + """Count line in current function body.""" + if self.in_a_function: + self.lines_in_function += 1 + + def Check(self, error, filename, linenum): + """Report if too many lines in function body. + + Args: + error: The function to call with any errors found. + filename: The name of the current file. + linenum: The number of the line to check. + """ + if Match(r'T(EST|est)', self.current_function): + base_trigger = self._TEST_TRIGGER + else: + base_trigger = self._NORMAL_TRIGGER + trigger = base_trigger * 2**_VerboseLevel() + + if self.lines_in_function > trigger: + error_level = int(math.log(self.lines_in_function / base_trigger, 2)) + # 50 => 0, 100 => 1, 200 => 2, 400 => 3, 800 => 4, 1600 => 5, ... + if error_level > 5: + error_level = 5 + error(filename, linenum, 'readability/fn_size', error_level, + 'Small and focused functions are preferred:' + ' %s has %d non-comment lines' + ' (error triggered by exceeding %d lines).' % ( + self.current_function, self.lines_in_function, trigger)) + + def End(self): + """Stop analyzing function body.""" + self.in_a_function = False + + +class _IncludeError(Exception): + """Indicates a problem with the include order in a file.""" + pass + + +class FileInfo(object): + """Provides utility functions for filenames. + + FileInfo provides easy access to the components of a file's path + relative to the project root. + """ + + def __init__(self, filename): + self._filename = filename + + def FullName(self): + """Make Windows paths like Unix.""" + return os.path.abspath(self._filename).replace('\\', '/') + + def RepositoryName(self): + """FullName after removing the local path to the repository. + + If we have a real absolute path name here we can try to do something smart: + detecting the root of the checkout and truncating /path/to/checkout from + the name so that we get header guards that don't include things like + "C:\Documents and Settings\..." or "/home/username/..." in them and thus + people on different computers who have checked the source out to different + locations won't see bogus errors. + """ + fullname = self.FullName() + + if os.path.exists(fullname): + project_dir = os.path.dirname(fullname) + + if os.path.exists(os.path.join(project_dir, ".svn")): + # If there's a .svn file in the current directory, we recursively look + # up the directory tree for the top of the SVN checkout + root_dir = project_dir + one_up_dir = os.path.dirname(root_dir) + while os.path.exists(os.path.join(one_up_dir, ".svn")): + root_dir = os.path.dirname(root_dir) + one_up_dir = os.path.dirname(one_up_dir) + + prefix = os.path.commonprefix([root_dir, project_dir]) + return fullname[len(prefix) + 1:] + + # Not SVN <= 1.6? Try to find a git, hg, or svn top level directory by + # searching up from the current path. + root_dir = os.path.dirname(fullname) + while (root_dir != os.path.dirname(root_dir) and + not os.path.exists(os.path.join(root_dir, ".git")) and + not os.path.exists(os.path.join(root_dir, ".hg")) and + not os.path.exists(os.path.join(root_dir, ".svn"))): + root_dir = os.path.dirname(root_dir) + + if (os.path.exists(os.path.join(root_dir, ".git")) or + os.path.exists(os.path.join(root_dir, ".hg")) or + os.path.exists(os.path.join(root_dir, ".svn"))): + prefix = os.path.commonprefix([root_dir, project_dir]) + return fullname[len(prefix) + 1:] + + # Don't know what to do; header guard warnings may be wrong... + return fullname + + def Split(self): + """Splits the file into the directory, basename, and extension. + + For 'chrome/browser/browser.cc', Split() would + return ('chrome/browser', 'browser', '.cc') + + Returns: + A tuple of (directory, basename, extension). + """ + + googlename = self.RepositoryName() + project, rest = os.path.split(googlename) + return (project,) + os.path.splitext(rest) + + def BaseName(self): + """File base name - text after the final slash, before the final period.""" + return self.Split()[1] + + def Extension(self): + """File extension - text following the final period.""" + return self.Split()[2] + + def NoExtension(self): + """File has no source file extension.""" + return '/'.join(self.Split()[0:2]) + + def IsSource(self): + """File has a source file extension.""" + return self.Extension()[1:] in ('c', 'cc', 'cpp', 'cxx') + + +def _ShouldPrintError(category, confidence, linenum): + """If confidence >= verbose, category passes filter and is not suppressed.""" + + # There are three ways we might decide not to print an error message: + # a "NOLINT(category)" comment appears in the source, + # the verbosity level isn't high enough, or the filters filter it out. + if IsErrorSuppressedByNolint(category, linenum): + return False + + if confidence < _cpplint_state.verbose_level: + return False + + is_filtered = False + for one_filter in _Filters(): + if one_filter.startswith('-'): + if category.startswith(one_filter[1:]): + is_filtered = True + elif one_filter.startswith('+'): + if category.startswith(one_filter[1:]): + is_filtered = False + else: + assert False # should have been checked for in SetFilter. + if is_filtered: + return False + + return True + + +def Error(filename, linenum, category, confidence, message): + """Logs the fact we've found a lint error. + + We log where the error was found, and also our confidence in the error, + that is, how certain we are this is a legitimate style regression, and + not a misidentification or a use that's sometimes justified. + + False positives can be suppressed by the use of + "cpplint(category)" comments on the offending line. These are + parsed into _error_suppressions. + + Args: + filename: The name of the file containing the error. + linenum: The number of the line containing the error. + category: A string used to describe the "category" this bug + falls under: "whitespace", say, or "runtime". Categories + may have a hierarchy separated by slashes: "whitespace/indent". + confidence: A number from 1-5 representing a confidence score for + the error, with 5 meaning that we are certain of the problem, + and 1 meaning that it could be a legitimate construct. + message: The error message. + """ + if _ShouldPrintError(category, confidence, linenum): + _cpplint_state.IncrementErrorCount(category) + if _cpplint_state.output_format == 'vs7': + sys.stderr.write('%s(%s): %s [%s] [%d]\n' % ( + filename, linenum, message, category, confidence)) + elif _cpplint_state.output_format == 'eclipse': + sys.stderr.write('%s:%s: warning: %s [%s] [%d]\n' % ( + filename, linenum, message, category, confidence)) + else: + sys.stderr.write('%s:%s: %s [%s] [%d]\n' % ( + filename, linenum, message, category, confidence)) + + +# Matches standard C++ escape sequences per 2.13.2.3 of the C++ standard. +_RE_PATTERN_CLEANSE_LINE_ESCAPES = re.compile( + r'\\([abfnrtv?"\\\']|\d+|x[0-9a-fA-F]+)') +# Match a single C style comment on the same line. +_RE_PATTERN_C_COMMENTS = r'/\*(?:[^*]|\*(?!/))*\*/' +# Matches multi-line C style comments. +# This RE is a little bit more complicated than one might expect, because we +# have to take care of space removals tools so we can handle comments inside +# statements better. +# The current rule is: We only clear spaces from both sides when we're at the +# end of the line. Otherwise, we try to remove spaces from the right side, +# if this doesn't work we try on left side but only if there's a non-character +# on the right. +_RE_PATTERN_CLEANSE_LINE_C_COMMENTS = re.compile( + r'(\s*' + _RE_PATTERN_C_COMMENTS + r'\s*$|' + + _RE_PATTERN_C_COMMENTS + r'\s+|' + + r'\s+' + _RE_PATTERN_C_COMMENTS + r'(?=\W)|' + + _RE_PATTERN_C_COMMENTS + r')') + + +def IsCppString(line): + """Does line terminate so, that the next symbol is in string constant. + + This function does not consider single-line nor multi-line comments. + + Args: + line: is a partial line of code starting from the 0..n. + + Returns: + True, if next character appended to 'line' is inside a + string constant. + """ + + line = line.replace(r'\\', 'XX') # after this, \\" does not match to \" + return ((line.count('"') - line.count(r'\"') - line.count("'\"'")) & 1) == 1 + + +def CleanseRawStrings(raw_lines): + """Removes C++11 raw strings from lines. + + Before: + static const char kData[] = R"( + multi-line string + )"; + + After: + static const char kData[] = "" + (replaced by blank line) + ""; + + Args: + raw_lines: list of raw lines. + + Returns: + list of lines with C++11 raw strings replaced by empty strings. + """ + + delimiter = None + lines_without_raw_strings = [] + for line in raw_lines: + if delimiter: + # Inside a raw string, look for the end + end = line.find(delimiter) + if end >= 0: + # Found the end of the string, match leading space for this + # line and resume copying the original lines, and also insert + # a "" on the last line. + leading_space = Match(r'^(\s*)\S', line) + line = leading_space.group(1) + '""' + line[end + len(delimiter):] + delimiter = None + else: + # Haven't found the end yet, append a blank line. + line = '""' + + # Look for beginning of a raw string, and replace them with + # empty strings. This is done in a loop to handle multiple raw + # strings on the same line. + while delimiter is None: + # Look for beginning of a raw string. + # See 2.14.15 [lex.string] for syntax. + matched = Match(r'^(.*)\b(?:R|u8R|uR|UR|LR)"([^\s\\()]*)\((.*)$', line) + if matched: + delimiter = ')' + matched.group(2) + '"' + + end = matched.group(3).find(delimiter) + if end >= 0: + # Raw string ended on same line + line = (matched.group(1) + '""' + + matched.group(3)[end + len(delimiter):]) + delimiter = None + else: + # Start of a multi-line raw string + line = matched.group(1) + '""' + else: + break + + lines_without_raw_strings.append(line) + + # TODO(unknown): if delimiter is not None here, we might want to + # emit a warning for unterminated string. + return lines_without_raw_strings + + +def FindNextMultiLineCommentStart(lines, lineix): + """Find the beginning marker for a multiline comment.""" + while lineix < len(lines): + if lines[lineix].strip().startswith('/*'): + # Only return this marker if the comment goes beyond this line + if lines[lineix].strip().find('*/', 2) < 0: + return lineix + lineix += 1 + return len(lines) + + +def FindNextMultiLineCommentEnd(lines, lineix): + """We are inside a comment, find the end marker.""" + while lineix < len(lines): + if lines[lineix].strip().endswith('*/'): + return lineix + lineix += 1 + return len(lines) + + +def RemoveMultiLineCommentsFromRange(lines, begin, end): + """Clears a range of lines for multi-line comments.""" + # Having // dummy comments makes the lines non-empty, so we will not get + # unnecessary blank line warnings later in the code. + for i in range(begin, end): + lines[i] = '/**/' + + +def RemoveMultiLineComments(filename, lines, error): + """Removes multiline (c-style) comments from lines.""" + lineix = 0 + while lineix < len(lines): + lineix_begin = FindNextMultiLineCommentStart(lines, lineix) + if lineix_begin >= len(lines): + return + lineix_end = FindNextMultiLineCommentEnd(lines, lineix_begin) + if lineix_end >= len(lines): + error(filename, lineix_begin + 1, 'readability/multiline_comment', 5, + 'Could not find end of multi-line comment') + return + RemoveMultiLineCommentsFromRange(lines, lineix_begin, lineix_end + 1) + lineix = lineix_end + 1 + + +def CleanseComments(line): + """Removes //-comments and single-line C-style /* */ comments. + + Args: + line: A line of C++ source. + + Returns: + The line with single-line comments removed. + """ + commentpos = line.find('//') + if commentpos != -1 and not IsCppString(line[:commentpos]): + line = line[:commentpos].rstrip() + # get rid of /* ... */ + return _RE_PATTERN_CLEANSE_LINE_C_COMMENTS.sub('', line) + + +class CleansedLines(object): + """Holds 4 copies of all lines with different preprocessing applied to them. + + 1) elided member contains lines without strings and comments. + 2) lines member contains lines without comments. + 3) raw_lines member contains all the lines without processing. + 4) lines_without_raw_strings member is same as raw_lines, but with C++11 raw + strings removed. + All these members are of , and of the same length. + """ + + def __init__(self, lines): + self.elided = [] + self.lines = [] + self.raw_lines = lines + self.num_lines = len(lines) + self.lines_without_raw_strings = CleanseRawStrings(lines) + for linenum in range(len(self.lines_without_raw_strings)): + self.lines.append(CleanseComments( + self.lines_without_raw_strings[linenum])) + elided = self._CollapseStrings(self.lines_without_raw_strings[linenum]) + self.elided.append(CleanseComments(elided)) + + def NumLines(self): + """Returns the number of lines represented.""" + return self.num_lines + + @staticmethod + def _CollapseStrings(elided): + """Collapses strings and chars on a line to simple "" or '' blocks. + + We nix strings first so we're not fooled by text like '"http://"' + + Args: + elided: The line being processed. + + Returns: + The line with collapsed strings. + """ + if _RE_PATTERN_INCLUDE.match(elided): + return elided + + # Remove escaped characters first to make quote/single quote collapsing + # basic. Things that look like escaped characters shouldn't occur + # outside of strings and chars. + elided = _RE_PATTERN_CLEANSE_LINE_ESCAPES.sub('', elided) + + # Replace quoted strings and digit separators. Both single quotes + # and double quotes are processed in the same loop, otherwise + # nested quotes wouldn't work. + collapsed = '' + while True: + # Find the first quote character + match = Match(r'^([^\'"]*)([\'"])(.*)$', elided) + if not match: + collapsed += elided + break + head, quote, tail = match.groups() + + if quote == '"': + # Collapse double quoted strings + second_quote = tail.find('"') + if second_quote >= 0: + collapsed += head + '""' + elided = tail[second_quote + 1:] + else: + # Unmatched double quote, don't bother processing the rest + # of the line since this is probably a multiline string. + collapsed += elided + break + else: + # Found single quote, check nearby text to eliminate digit separators. + # + # There is no special handling for floating point here, because + # the integer/fractional/exponent parts would all be parsed + # correctly as long as there are digits on both sides of the + # separator. So we are fine as long as we don't see something + # like "0.'3" (gcc 4.9.0 will not allow this literal). + if Search(r'\b(?:0[bBxX]?|[1-9])[0-9a-fA-F]*$', head): + match_literal = Match(r'^((?:\'?[0-9a-zA-Z_])*)(.*)$', "'" + tail) + collapsed += head + match_literal.group(1).replace("'", '') + elided = match_literal.group(2) + else: + second_quote = tail.find('\'') + if second_quote >= 0: + collapsed += head + "''" + elided = tail[second_quote + 1:] + else: + # Unmatched single quote + collapsed += elided + break + + return collapsed + + +def FindEndOfExpressionInLine(line, startpos, stack): + """Find the position just after the end of current parenthesized expression. + + Args: + line: a CleansedLines line. + startpos: start searching at this position. + stack: nesting stack at startpos. + + Returns: + On finding matching end: (index just after matching end, None) + On finding an unclosed expression: (-1, None) + Otherwise: (-1, new stack at end of this line) + """ + for i in xrange(startpos, len(line)): + char = line[i] + if char in '([{': + # Found start of parenthesized expression, push to expression stack + stack.append(char) + elif char == '<': + # Found potential start of template argument list + if i > 0 and line[i - 1] == '<': + # Left shift operator + if stack and stack[-1] == '<': + stack.pop() + if not stack: + return (-1, None) + elif i > 0 and Search(r'\boperator\s*$', line[0:i]): + # operator<, don't add to stack + continue + else: + # Tentative start of template argument list + stack.append('<') + elif char in ')]}': + # Found end of parenthesized expression. + # + # If we are currently expecting a matching '>', the pending '<' + # must have been an operator. Remove them from expression stack. + while stack and stack[-1] == '<': + stack.pop() + if not stack: + return (-1, None) + if ((stack[-1] == '(' and char == ')') or + (stack[-1] == '[' and char == ']') or + (stack[-1] == '{' and char == '}')): + stack.pop() + if not stack: + return (i + 1, None) + else: + # Mismatched parentheses + return (-1, None) + elif char == '>': + # Found potential end of template argument list. + + # Ignore "->" and operator functions + if (i > 0 and + (line[i - 1] == '-' or Search(r'\boperator\s*$', line[0:i - 1]))): + continue + + # Pop the stack if there is a matching '<'. Otherwise, ignore + # this '>' since it must be an operator. + if stack: + if stack[-1] == '<': + stack.pop() + if not stack: + return (i + 1, None) + elif char == ';': + # Found something that look like end of statements. If we are currently + # expecting a '>', the matching '<' must have been an operator, since + # template argument list should not contain statements. + while stack and stack[-1] == '<': + stack.pop() + if not stack: + return (-1, None) + + # Did not find end of expression or unbalanced parentheses on this line + return (-1, stack) + + +def CloseExpression(clean_lines, linenum, pos): + """If input points to ( or { or [ or <, finds the position that closes it. + + If lines[linenum][pos] points to a '(' or '{' or '[' or '<', finds the + linenum/pos that correspond to the closing of the expression. + + TODO(unknown): cpplint spends a fair bit of time matching parentheses. + Ideally we would want to index all opening and closing parentheses once + and have CloseExpression be just a simple lookup, but due to preprocessor + tricks, this is not so easy. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + pos: A position on the line. + + Returns: + A tuple (line, linenum, pos) pointer *past* the closing brace, or + (line, len(lines), -1) if we never find a close. Note we ignore + strings and comments when matching; and the line we return is the + 'cleansed' line at linenum. + """ + + line = clean_lines.elided[linenum] + if (line[pos] not in '({[<') or Match(r'<[<=]', line[pos:]): + return (line, clean_lines.NumLines(), -1) + + # Check first line + (end_pos, stack) = FindEndOfExpressionInLine(line, pos, []) + if end_pos > -1: + return (line, linenum, end_pos) + + # Continue scanning forward + while stack and linenum < clean_lines.NumLines() - 1: + linenum += 1 + line = clean_lines.elided[linenum] + (end_pos, stack) = FindEndOfExpressionInLine(line, 0, stack) + if end_pos > -1: + return (line, linenum, end_pos) + + # Did not find end of expression before end of file, give up + return (line, clean_lines.NumLines(), -1) + + +def FindStartOfExpressionInLine(line, endpos, stack): + """Find position at the matching start of current expression. + + This is almost the reverse of FindEndOfExpressionInLine, but note + that the input position and returned position differs by 1. + + Args: + line: a CleansedLines line. + endpos: start searching at this position. + stack: nesting stack at endpos. + + Returns: + On finding matching start: (index at matching start, None) + On finding an unclosed expression: (-1, None) + Otherwise: (-1, new stack at beginning of this line) + """ + i = endpos + while i >= 0: + char = line[i] + if char in ')]}': + # Found end of expression, push to expression stack + stack.append(char) + elif char == '>': + # Found potential end of template argument list. + # + # Ignore it if it's a "->" or ">=" or "operator>" + if (i > 0 and + (line[i - 1] == '-' or + Match(r'\s>=\s', line[i - 1:]) or + Search(r'\boperator\s*$', line[0:i]))): + i -= 1 + else: + stack.append('>') + elif char == '<': + # Found potential start of template argument list + if i > 0 and line[i - 1] == '<': + # Left shift operator + i -= 1 + else: + # If there is a matching '>', we can pop the expression stack. + # Otherwise, ignore this '<' since it must be an operator. + if stack and stack[-1] == '>': + stack.pop() + if not stack: + return (i, None) + elif char in '([{': + # Found start of expression. + # + # If there are any unmatched '>' on the stack, they must be + # operators. Remove those. + while stack and stack[-1] == '>': + stack.pop() + if not stack: + return (-1, None) + if ((char == '(' and stack[-1] == ')') or + (char == '[' and stack[-1] == ']') or + (char == '{' and stack[-1] == '}')): + stack.pop() + if not stack: + return (i, None) + else: + # Mismatched parentheses + return (-1, None) + elif char == ';': + # Found something that look like end of statements. If we are currently + # expecting a '<', the matching '>' must have been an operator, since + # template argument list should not contain statements. + while stack and stack[-1] == '>': + stack.pop() + if not stack: + return (-1, None) + + i -= 1 + + return (-1, stack) + + +def ReverseCloseExpression(clean_lines, linenum, pos): + """If input points to ) or } or ] or >, finds the position that opens it. + + If lines[linenum][pos] points to a ')' or '}' or ']' or '>', finds the + linenum/pos that correspond to the opening of the expression. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + pos: A position on the line. + + Returns: + A tuple (line, linenum, pos) pointer *at* the opening brace, or + (line, 0, -1) if we never find the matching opening brace. Note + we ignore strings and comments when matching; and the line we + return is the 'cleansed' line at linenum. + """ + line = clean_lines.elided[linenum] + if line[pos] not in ')}]>': + return (line, 0, -1) + + # Check last line + (start_pos, stack) = FindStartOfExpressionInLine(line, pos, []) + if start_pos > -1: + return (line, linenum, start_pos) + + # Continue scanning backward + while stack and linenum > 0: + linenum -= 1 + line = clean_lines.elided[linenum] + (start_pos, stack) = FindStartOfExpressionInLine(line, len(line) - 1, stack) + if start_pos > -1: + return (line, linenum, start_pos) + + # Did not find start of expression before beginning of file, give up + return (line, 0, -1) + + +def CheckForCopyright(filename, lines, error): + """Logs an error if no Copyright message appears at the top of the file.""" + + # We'll say it should occur by line 10. Don't forget there's a + # dummy line at the front. + for line in xrange(1, min(len(lines), 11)): + if re.search(r'Copyright', lines[line], re.I): break + else: # means no copyright line was found + error(filename, 0, 'legal/copyright', 5, + 'No copyright message found. ' + 'You should have a line: "Copyright [year] "') + + +def GetIndentLevel(line): + """Return the number of leading spaces in line. + + Args: + line: A string to check. + + Returns: + An integer count of leading spaces, possibly zero. + """ + indent = Match(r'^( *)\S', line) + if indent: + return len(indent.group(1)) + else: + return 0 + + +def GetHeaderGuardCPPVariable(filename): + """Returns the CPP variable that should be used as a header guard. + + Args: + filename: The name of a C++ header file. + + Returns: + The CPP variable that should be used as a header guard in the + named file. + + """ + + # Restores original filename in case that cpplint is invoked from Emacs's + # flymake. + filename = re.sub(r'_flymake\.h$', '.h', filename) + filename = re.sub(r'/\.flymake/([^/]*)$', r'/\1', filename) + # Replace 'c++' with 'cpp'. + filename = filename.replace('C++', 'cpp').replace('c++', 'cpp') + + fileinfo = FileInfo(filename) + file_path_from_root = fileinfo.RepositoryName() + if _root: + file_path_from_root = re.sub('^' + _root + os.sep, '', file_path_from_root) + return re.sub(r'[^a-zA-Z0-9]', '_', file_path_from_root).upper() + '_' + + +def CheckForHeaderGuard(filename, clean_lines, error): + """Checks that the file contains a header guard. + + Logs an error if no #ifndef header guard is present. For other + headers, checks that the full pathname is used. + + Args: + filename: The name of the C++ header file. + clean_lines: A CleansedLines instance containing the file. + error: The function to call with any errors found. + """ + + # Don't check for header guards if there are error suppression + # comments somewhere in this file. + # + # Because this is silencing a warning for a nonexistent line, we + # only support the very specific NOLINT(build/header_guard) syntax, + # and not the general NOLINT or NOLINT(*) syntax. + raw_lines = clean_lines.lines_without_raw_strings + for i in raw_lines: + if Search(r'//\s*NOLINT\(build/header_guard\)', i): + return + + cppvar = GetHeaderGuardCPPVariable(filename) + + ifndef = '' + ifndef_linenum = 0 + define = '' + endif = '' + endif_linenum = 0 + for linenum, line in enumerate(raw_lines): + linesplit = line.split() + if len(linesplit) >= 2: + # find the first occurrence of #ifndef and #define, save arg + if not ifndef and linesplit[0] == '#ifndef': + # set ifndef to the header guard presented on the #ifndef line. + ifndef = linesplit[1] + ifndef_linenum = linenum + if not define and linesplit[0] == '#define': + define = linesplit[1] + # find the last occurrence of #endif, save entire line + if line.startswith('#endif'): + endif = line + endif_linenum = linenum + + if not ifndef or not define or ifndef != define: + error(filename, 0, 'build/header_guard', 5, + 'No #ifndef header guard found, suggested CPP variable is: %s' % + cppvar) + return + + # The guard should be PATH_FILE_H_, but we also allow PATH_FILE_H__ + # for backward compatibility. + if ifndef != cppvar: + error_level = 0 + if ifndef != cppvar + '_': + error_level = 5 + + ParseNolintSuppressions(filename, raw_lines[ifndef_linenum], ifndef_linenum, + error) + error(filename, ifndef_linenum, 'build/header_guard', error_level, + '#ifndef header guard has wrong style, please use: %s' % cppvar) + + # Check for "//" comments on endif line. + ParseNolintSuppressions(filename, raw_lines[endif_linenum], endif_linenum, + error) + match = Match(r'#endif\s*//\s*' + cppvar + r'(_)?\b', endif) + if match: + if match.group(1) == '_': + # Issue low severity warning for deprecated double trailing underscore + error(filename, endif_linenum, 'build/header_guard', 0, + '#endif line should be "#endif // %s"' % cppvar) + return + + # Didn't find the corresponding "//" comment. If this file does not + # contain any "//" comments at all, it could be that the compiler + # only wants "/**/" comments, look for those instead. + no_single_line_comments = True + for i in xrange(1, len(raw_lines) - 1): + line = raw_lines[i] + if Match(r'^(?:(?:\'(?:\.|[^\'])*\')|(?:"(?:\.|[^"])*")|[^\'"])*//', line): + no_single_line_comments = False + break + + if no_single_line_comments: + match = Match(r'#endif\s*/\*\s*' + cppvar + r'(_)?\s*\*/', endif) + if match: + if match.group(1) == '_': + # Low severity warning for double trailing underscore + error(filename, endif_linenum, 'build/header_guard', 0, + '#endif line should be "#endif /* %s */"' % cppvar) + return + + # Didn't find anything + error(filename, endif_linenum, 'build/header_guard', 5, + '#endif line should be "#endif // %s"' % cppvar) + + +def CheckHeaderFileIncluded(filename, include_state, error): + """Logs an error if a .cc file does not include its header.""" + + # Do not check test files + if filename.endswith('_test.cc') or filename.endswith('_unittest.cc'): + return + + fileinfo = FileInfo(filename) + headerfile = filename[0:len(filename) - 2] + 'h' + if not os.path.exists(headerfile): + return + headername = FileInfo(headerfile).RepositoryName() + first_include = 0 + for section_list in include_state.include_list: + for f in section_list: + if headername in f[0] or f[0] in headername: + return + if not first_include: + first_include = f[1] + + error(filename, first_include, 'build/include', 5, + '%s should include its header file %s' % (fileinfo.RepositoryName(), + headername)) + + +def CheckForBadCharacters(filename, lines, error): + """Logs an error for each line containing bad characters. + + Two kinds of bad characters: + + 1. Unicode replacement characters: These indicate that either the file + contained invalid UTF-8 (likely) or Unicode replacement characters (which + it shouldn't). Note that it's possible for this to throw off line + numbering if the invalid UTF-8 occurred adjacent to a newline. + + 2. NUL bytes. These are problematic for some tools. + + Args: + filename: The name of the current file. + lines: An array of strings, each representing a line of the file. + error: The function to call with any errors found. + """ + for linenum, line in enumerate(lines): + if u'\ufffd' in line: + error(filename, linenum, 'readability/utf8', 5, + 'Line contains invalid UTF-8 (or Unicode replacement character).') + if '\0' in line: + error(filename, linenum, 'readability/nul', 5, 'Line contains NUL byte.') + + +def CheckForNewlineAtEOF(filename, lines, error): + """Logs an error if there is no newline char at the end of the file. + + Args: + filename: The name of the current file. + lines: An array of strings, each representing a line of the file. + error: The function to call with any errors found. + """ + + # The array lines() was created by adding two newlines to the + # original file (go figure), then splitting on \n. + # To verify that the file ends in \n, we just have to make sure the + # last-but-two element of lines() exists and is empty. + if len(lines) < 3 or lines[-2]: + error(filename, len(lines) - 2, 'whitespace/ending_newline', 5, + 'Could not find a newline character at the end of the file.') + + +def CheckForMultilineCommentsAndStrings(filename, clean_lines, linenum, error): + """Logs an error if we see /* ... */ or "..." that extend past one line. + + /* ... */ comments are legit inside macros, for one line. + Otherwise, we prefer // comments, so it's ok to warn about the + other. Likewise, it's ok for strings to extend across multiple + lines, as long as a line continuation character (backslash) + terminates each line. Although not currently prohibited by the C++ + style guide, it's ugly and unnecessary. We don't do well with either + in this lint program, so we warn about both. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Remove all \\ (escaped backslashes) from the line. They are OK, and the + # second (escaped) slash may trigger later \" detection erroneously. + line = line.replace('\\\\', '') + + if line.count('/*') > line.count('*/'): + error(filename, linenum, 'readability/multiline_comment', 5, + 'Complex multi-line /*...*/-style comment found. ' + 'Lint may give bogus warnings. ' + 'Consider replacing these with //-style comments, ' + 'with #if 0...#endif, ' + 'or with more clearly structured multi-line comments.') + + if (line.count('"') - line.count('\\"')) % 2: + error(filename, linenum, 'readability/multiline_string', 5, + 'Multi-line string ("...") found. This lint script doesn\'t ' + 'do well with such strings, and may give bogus warnings. ' + 'Use C++11 raw strings or concatenation instead.') + + +# (non-threadsafe name, thread-safe alternative, validation pattern) +# +# The validation pattern is used to eliminate false positives such as: +# _rand(); // false positive due to substring match. +# ->rand(); // some member function rand(). +# ACMRandom rand(seed); // some variable named rand. +# ISAACRandom rand(); // another variable named rand. +# +# Basically we require the return value of these functions to be used +# in some expression context on the same line by matching on some +# operator before the function name. This eliminates constructors and +# member function calls. +_UNSAFE_FUNC_PREFIX = r'(?:[-+*/=%^&|(<]\s*|>\s+)' +_THREADING_LIST = ( + ('asctime(', 'asctime_r(', _UNSAFE_FUNC_PREFIX + r'asctime\([^)]+\)'), + ('ctime(', 'ctime_r(', _UNSAFE_FUNC_PREFIX + r'ctime\([^)]+\)'), + ('getgrgid(', 'getgrgid_r(', _UNSAFE_FUNC_PREFIX + r'getgrgid\([^)]+\)'), + ('getgrnam(', 'getgrnam_r(', _UNSAFE_FUNC_PREFIX + r'getgrnam\([^)]+\)'), + ('getlogin(', 'getlogin_r(', _UNSAFE_FUNC_PREFIX + r'getlogin\(\)'), + ('getpwnam(', 'getpwnam_r(', _UNSAFE_FUNC_PREFIX + r'getpwnam\([^)]+\)'), + ('getpwuid(', 'getpwuid_r(', _UNSAFE_FUNC_PREFIX + r'getpwuid\([^)]+\)'), + ('gmtime(', 'gmtime_r(', _UNSAFE_FUNC_PREFIX + r'gmtime\([^)]+\)'), + ('localtime(', 'localtime_r(', _UNSAFE_FUNC_PREFIX + r'localtime\([^)]+\)'), + ('rand(', 'rand_r(', _UNSAFE_FUNC_PREFIX + r'rand\(\)'), + ('strtok(', 'strtok_r(', + _UNSAFE_FUNC_PREFIX + r'strtok\([^)]+\)'), + ('ttyname(', 'ttyname_r(', _UNSAFE_FUNC_PREFIX + r'ttyname\([^)]+\)'), + ) + + +def CheckPosixThreading(filename, clean_lines, linenum, error): + """Checks for calls to thread-unsafe functions. + + Much code has been originally written without consideration of + multi-threading. Also, engineers are relying on their old experience; + they have learned posix before threading extensions were added. These + tests guide the engineers to use thread-safe functions (when using + posix directly). + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + for single_thread_func, multithread_safe_func, pattern in _THREADING_LIST: + # Additional pattern matching check to confirm that this is the + # function we are looking for + if Search(pattern, line): + error(filename, linenum, 'runtime/threadsafe_fn', 2, + 'Consider using ' + multithread_safe_func + + '...) instead of ' + single_thread_func + + '...) for improved thread safety.') + + +def CheckVlogArguments(filename, clean_lines, linenum, error): + """Checks that VLOG() is only used for defining a logging level. + + For example, VLOG(2) is correct. VLOG(INFO), VLOG(WARNING), VLOG(ERROR), and + VLOG(FATAL) are not. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + if Search(r'\bVLOG\((INFO|ERROR|WARNING|DFATAL|FATAL)\)', line): + error(filename, linenum, 'runtime/vlog', 5, + 'VLOG() should be used with numeric verbosity level. ' + 'Use LOG() if you want symbolic severity levels.') + +# Matches invalid increment: *count++, which moves pointer instead of +# incrementing a value. +_RE_PATTERN_INVALID_INCREMENT = re.compile( + r'^\s*\*\w+(\+\+|--);') + + +def CheckInvalidIncrement(filename, clean_lines, linenum, error): + """Checks for invalid increment *count++. + + For example following function: + void increment_counter(int* count) { + *count++; + } + is invalid, because it effectively does count++, moving pointer, and should + be replaced with ++*count, (*count)++ or *count += 1. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + if _RE_PATTERN_INVALID_INCREMENT.match(line): + error(filename, linenum, 'runtime/invalid_increment', 5, + 'Changing pointer instead of value (or unused value of operator*).') + + +def IsMacroDefinition(clean_lines, linenum): + if Search(r'^#define', clean_lines[linenum]): + return True + + if linenum > 0 and Search(r'\\$', clean_lines[linenum - 1]): + return True + + return False + + +def IsForwardClassDeclaration(clean_lines, linenum): + return Match(r'^\s*(\btemplate\b)*.*class\s+\w+;\s*$', clean_lines[linenum]) + + +class _BlockInfo(object): + """Stores information about a generic block of code.""" + + def __init__(self, seen_open_brace): + self.seen_open_brace = seen_open_brace + self.open_parentheses = 0 + self.inline_asm = _NO_ASM + self.check_namespace_indentation = False + + def CheckBegin(self, filename, clean_lines, linenum, error): + """Run checks that applies to text up to the opening brace. + + This is mostly for checking the text after the class identifier + and the "{", usually where the base class is specified. For other + blocks, there isn't much to check, so we always pass. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + pass + + def CheckEnd(self, filename, clean_lines, linenum, error): + """Run checks that applies to text after the closing brace. + + This is mostly used for checking end of namespace comments. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + pass + + def IsBlockInfo(self): + """Returns true if this block is a _BlockInfo. + + This is convenient for verifying that an object is an instance of + a _BlockInfo, but not an instance of any of the derived classes. + + Returns: + True for this class, False for derived classes. + """ + return self.__class__ == _BlockInfo + + +class _ExternCInfo(_BlockInfo): + """Stores information about an 'extern "C"' block.""" + + def __init__(self): + _BlockInfo.__init__(self, True) + + +class _ClassInfo(_BlockInfo): + """Stores information about a class.""" + + def __init__(self, name, class_or_struct, clean_lines, linenum): + _BlockInfo.__init__(self, False) + self.name = name + self.starting_linenum = linenum + self.is_derived = False + self.check_namespace_indentation = True + if class_or_struct == 'struct': + self.access = 'public' + self.is_struct = True + else: + self.access = 'private' + self.is_struct = False + + # Remember initial indentation level for this class. Using raw_lines here + # instead of elided to account for leading comments. + self.class_indent = GetIndentLevel(clean_lines.raw_lines[linenum]) + + # Try to find the end of the class. This will be confused by things like: + # class A { + # } *x = { ... + # + # But it's still good enough for CheckSectionSpacing. + self.last_line = 0 + depth = 0 + for i in range(linenum, clean_lines.NumLines()): + line = clean_lines.elided[i] + depth += line.count('{') - line.count('}') + if not depth: + self.last_line = i + break + + def CheckBegin(self, filename, clean_lines, linenum, error): + # Look for a bare ':' + if Search('(^|[^:]):($|[^:])', clean_lines.elided[linenum]): + self.is_derived = True + + def CheckEnd(self, filename, clean_lines, linenum, error): + # If there is a DISALLOW macro, it should appear near the end of + # the class. + seen_last_thing_in_class = False + for i in xrange(linenum - 1, self.starting_linenum, -1): + match = Search( + r'\b(DISALLOW_COPY_AND_ASSIGN|DISALLOW_IMPLICIT_CONSTRUCTORS)\(' + + self.name + r'\)', + clean_lines.elided[i]) + if match: + if seen_last_thing_in_class: + error(filename, i, 'readability/constructors', 3, + match.group(1) + ' should be the last thing in the class') + break + + if not Match(r'^\s*$', clean_lines.elided[i]): + seen_last_thing_in_class = True + + # Check that closing brace is aligned with beginning of the class. + # Only do this if the closing brace is indented by only whitespaces. + # This means we will not check single-line class definitions. + indent = Match(r'^( *)\}', clean_lines.elided[linenum]) + if indent and len(indent.group(1)) != self.class_indent: + if self.is_struct: + parent = 'struct ' + self.name + else: + parent = 'class ' + self.name + error(filename, linenum, 'whitespace/indent', 3, + 'Closing brace should be aligned with beginning of %s' % parent) + + +class _NamespaceInfo(_BlockInfo): + """Stores information about a namespace.""" + + def __init__(self, name, linenum): + _BlockInfo.__init__(self, False) + self.name = name or '' + self.starting_linenum = linenum + self.check_namespace_indentation = True + + def CheckEnd(self, filename, clean_lines, linenum, error): + """Check end of namespace comments.""" + line = clean_lines.raw_lines[linenum] + + # Check how many lines is enclosed in this namespace. Don't issue + # warning for missing namespace comments if there aren't enough + # lines. However, do apply checks if there is already an end of + # namespace comment and it's incorrect. + # + # TODO(unknown): We always want to check end of namespace comments + # if a namespace is large, but sometimes we also want to apply the + # check if a short namespace contained nontrivial things (something + # other than forward declarations). There is currently no logic on + # deciding what these nontrivial things are, so this check is + # triggered by namespace size only, which works most of the time. + if (linenum - self.starting_linenum < 10 + and not Match(r'};*\s*(//|/\*).*\bnamespace\b', line)): + return + + # Look for matching comment at end of namespace. + # + # Note that we accept C style "/* */" comments for terminating + # namespaces, so that code that terminate namespaces inside + # preprocessor macros can be cpplint clean. + # + # We also accept stuff like "// end of namespace ." with the + # period at the end. + # + # Besides these, we don't accept anything else, otherwise we might + # get false negatives when existing comment is a substring of the + # expected namespace. + if self.name: + # Named namespace + if not Match((r'};*\s*(//|/\*).*\bnamespace\s+' + re.escape(self.name) + + r'[\*/\.\\\s]*$'), + line): + error(filename, linenum, 'readability/namespace', 5, + 'Namespace should be terminated with "// namespace %s"' % + self.name) + else: + # Anonymous namespace + if not Match(r'};*\s*(//|/\*).*\bnamespace[\*/\.\\\s]*$', line): + # If "// namespace anonymous" or "// anonymous namespace (more text)", + # mention "// anonymous namespace" as an acceptable form + if Match(r'}.*\b(namespace anonymous|anonymous namespace)\b', line): + error(filename, linenum, 'readability/namespace', 5, + 'Anonymous namespace should be terminated with "// namespace"' + ' or "// anonymous namespace"') + else: + error(filename, linenum, 'readability/namespace', 5, + 'Anonymous namespace should be terminated with "// namespace"') + + +class _PreprocessorInfo(object): + """Stores checkpoints of nesting stacks when #if/#else is seen.""" + + def __init__(self, stack_before_if): + # The entire nesting stack before #if + self.stack_before_if = stack_before_if + + # The entire nesting stack up to #else + self.stack_before_else = [] + + # Whether we have already seen #else or #elif + self.seen_else = False + + +class NestingState(object): + """Holds states related to parsing braces.""" + + def __init__(self): + # Stack for tracking all braces. An object is pushed whenever we + # see a "{", and popped when we see a "}". Only 3 types of + # objects are possible: + # - _ClassInfo: a class or struct. + # - _NamespaceInfo: a namespace. + # - _BlockInfo: some other type of block. + self.stack = [] + + # Top of the previous stack before each Update(). + # + # Because the nesting_stack is updated at the end of each line, we + # had to do some convoluted checks to find out what is the current + # scope at the beginning of the line. This check is simplified by + # saving the previous top of nesting stack. + # + # We could save the full stack, but we only need the top. Copying + # the full nesting stack would slow down cpplint by ~10%. + self.previous_stack_top = [] + + # Stack of _PreprocessorInfo objects. + self.pp_stack = [] + + def SeenOpenBrace(self): + """Check if we have seen the opening brace for the innermost block. + + Returns: + True if we have seen the opening brace, False if the innermost + block is still expecting an opening brace. + """ + return (not self.stack) or self.stack[-1].seen_open_brace + + def InNamespaceBody(self): + """Check if we are currently one level inside a namespace body. + + Returns: + True if top of the stack is a namespace block, False otherwise. + """ + return self.stack and isinstance(self.stack[-1], _NamespaceInfo) + + def InExternC(self): + """Check if we are currently one level inside an 'extern "C"' block. + + Returns: + True if top of the stack is an extern block, False otherwise. + """ + return self.stack and isinstance(self.stack[-1], _ExternCInfo) + + def InClassDeclaration(self): + """Check if we are currently one level inside a class or struct declaration. + + Returns: + True if top of the stack is a class/struct, False otherwise. + """ + return self.stack and isinstance(self.stack[-1], _ClassInfo) + + def InAsmBlock(self): + """Check if we are currently one level inside an inline ASM block. + + Returns: + True if the top of the stack is a block containing inline ASM. + """ + return self.stack and self.stack[-1].inline_asm != _NO_ASM + + def InTemplateArgumentList(self, clean_lines, linenum, pos): + """Check if current position is inside template argument list. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + pos: position just after the suspected template argument. + Returns: + True if (linenum, pos) is inside template arguments. + """ + while linenum < clean_lines.NumLines(): + # Find the earliest character that might indicate a template argument + line = clean_lines.elided[linenum] + match = Match(r'^[^{};=\[\]\.<>]*(.)', line[pos:]) + if not match: + linenum += 1 + pos = 0 + continue + token = match.group(1) + pos += len(match.group(0)) + + # These things do not look like template argument list: + # class Suspect { + # class Suspect x; } + if token in ('{', '}', ';'): return False + + # These things look like template argument list: + # template + # template + # template + # template + if token in ('>', '=', '[', ']', '.'): return True + + # Check if token is an unmatched '<'. + # If not, move on to the next character. + if token != '<': + pos += 1 + if pos >= len(line): + linenum += 1 + pos = 0 + continue + + # We can't be sure if we just find a single '<', and need to + # find the matching '>'. + (_, end_line, end_pos) = CloseExpression(clean_lines, linenum, pos - 1) + if end_pos < 0: + # Not sure if template argument list or syntax error in file + return False + linenum = end_line + pos = end_pos + return False + + def UpdatePreprocessor(self, line): + """Update preprocessor stack. + + We need to handle preprocessors due to classes like this: + #ifdef SWIG + struct ResultDetailsPageElementExtensionPoint { + #else + struct ResultDetailsPageElementExtensionPoint : public Extension { + #endif + + We make the following assumptions (good enough for most files): + - Preprocessor condition evaluates to true from #if up to first + #else/#elif/#endif. + + - Preprocessor condition evaluates to false from #else/#elif up + to #endif. We still perform lint checks on these lines, but + these do not affect nesting stack. + + Args: + line: current line to check. + """ + if Match(r'^\s*#\s*(if|ifdef|ifndef)\b', line): + # Beginning of #if block, save the nesting stack here. The saved + # stack will allow us to restore the parsing state in the #else case. + self.pp_stack.append(_PreprocessorInfo(copy.deepcopy(self.stack))) + elif Match(r'^\s*#\s*(else|elif)\b', line): + # Beginning of #else block + if self.pp_stack: + if not self.pp_stack[-1].seen_else: + # This is the first #else or #elif block. Remember the + # whole nesting stack up to this point. This is what we + # keep after the #endif. + self.pp_stack[-1].seen_else = True + self.pp_stack[-1].stack_before_else = copy.deepcopy(self.stack) + + # Restore the stack to how it was before the #if + self.stack = copy.deepcopy(self.pp_stack[-1].stack_before_if) + else: + # TODO(unknown): unexpected #else, issue warning? + pass + elif Match(r'^\s*#\s*endif\b', line): + # End of #if or #else blocks. + if self.pp_stack: + # If we saw an #else, we will need to restore the nesting + # stack to its former state before the #else, otherwise we + # will just continue from where we left off. + if self.pp_stack[-1].seen_else: + # Here we can just use a shallow copy since we are the last + # reference to it. + self.stack = self.pp_stack[-1].stack_before_else + # Drop the corresponding #if + self.pp_stack.pop() + else: + # TODO(unknown): unexpected #endif, issue warning? + pass + + # TODO(unknown): Update() is too long, but we will refactor later. + def Update(self, filename, clean_lines, linenum, error): + """Update nesting state with current line. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Remember top of the previous nesting stack. + # + # The stack is always pushed/popped and not modified in place, so + # we can just do a shallow copy instead of copy.deepcopy. Using + # deepcopy would slow down cpplint by ~28%. + if self.stack: + self.previous_stack_top = self.stack[-1] + else: + self.previous_stack_top = None + + # Update pp_stack + self.UpdatePreprocessor(line) + + # Count parentheses. This is to avoid adding struct arguments to + # the nesting stack. + if self.stack: + inner_block = self.stack[-1] + depth_change = line.count('(') - line.count(')') + inner_block.open_parentheses += depth_change + + # Also check if we are starting or ending an inline assembly block. + if inner_block.inline_asm in (_NO_ASM, _END_ASM): + if (depth_change != 0 and + inner_block.open_parentheses == 1 and + _MATCH_ASM.match(line)): + # Enter assembly block + inner_block.inline_asm = _INSIDE_ASM + else: + # Not entering assembly block. If previous line was _END_ASM, + # we will now shift to _NO_ASM state. + inner_block.inline_asm = _NO_ASM + elif (inner_block.inline_asm == _INSIDE_ASM and + inner_block.open_parentheses == 0): + # Exit assembly block + inner_block.inline_asm = _END_ASM + + # Consume namespace declaration at the beginning of the line. Do + # this in a loop so that we catch same line declarations like this: + # namespace proto2 { namespace bridge { class MessageSet; } } + while True: + # Match start of namespace. The "\b\s*" below catches namespace + # declarations even if it weren't followed by a whitespace, this + # is so that we don't confuse our namespace checker. The + # missing spaces will be flagged by CheckSpacing. + namespace_decl_match = Match(r'^\s*namespace\b\s*([:\w]+)?(.*)$', line) + if not namespace_decl_match: + break + + new_namespace = _NamespaceInfo(namespace_decl_match.group(1), linenum) + self.stack.append(new_namespace) + + line = namespace_decl_match.group(2) + if line.find('{') != -1: + new_namespace.seen_open_brace = True + line = line[line.find('{') + 1:] + + # Look for a class declaration in whatever is left of the line + # after parsing namespaces. The regexp accounts for decorated classes + # such as in: + # class LOCKABLE API Object { + # }; + class_decl_match = Match( + r'^(\s*(?:template\s*<[\w\s<>,:]*>\s*)?' + r'(class|struct)\s+(?:[A-Z_]+\s+)*(\w+(?:::\w+)*))' + r'(.*)$', line) + if (class_decl_match and + (not self.stack or self.stack[-1].open_parentheses == 0)): + # We do not want to accept classes that are actually template arguments: + # template , + # template class Ignore3> + # void Function() {}; + # + # To avoid template argument cases, we scan forward and look for + # an unmatched '>'. If we see one, assume we are inside a + # template argument list. + end_declaration = len(class_decl_match.group(1)) + if not self.InTemplateArgumentList(clean_lines, linenum, end_declaration): + self.stack.append(_ClassInfo( + class_decl_match.group(3), class_decl_match.group(2), + clean_lines, linenum)) + line = class_decl_match.group(4) + + # If we have not yet seen the opening brace for the innermost block, + # run checks here. + if not self.SeenOpenBrace(): + self.stack[-1].CheckBegin(filename, clean_lines, linenum, error) + + # Update access control if we are inside a class/struct + if self.stack and isinstance(self.stack[-1], _ClassInfo): + classinfo = self.stack[-1] + access_match = Match( + r'^(.*)\b(public|private|protected|signals)(\s+(?:slots\s*)?)?' + r':(?:[^:]|$)', + line) + if access_match: + classinfo.access = access_match.group(2) + + # Check that access keywords are indented +1 space. Skip this + # check if the keywords are not preceded by whitespaces. + indent = access_match.group(1) + if (len(indent) != classinfo.class_indent + 1 and + Match(r'^\s*$', indent)): + if classinfo.is_struct: + parent = 'struct ' + classinfo.name + else: + parent = 'class ' + classinfo.name + slots = '' + if access_match.group(3): + slots = access_match.group(3) + error(filename, linenum, 'whitespace/indent', 3, + '%s%s: should be indented +1 space inside %s' % ( + access_match.group(2), slots, parent)) + + # Consume braces or semicolons from what's left of the line + while True: + # Match first brace, semicolon, or closed parenthesis. + matched = Match(r'^[^{;)}]*([{;)}])(.*)$', line) + if not matched: + break + + token = matched.group(1) + if token == '{': + # If namespace or class hasn't seen a opening brace yet, mark + # namespace/class head as complete. Push a new block onto the + # stack otherwise. + if not self.SeenOpenBrace(): + self.stack[-1].seen_open_brace = True + elif Match(r'^extern\s*"[^"]*"\s*\{', line): + self.stack.append(_ExternCInfo()) + else: + self.stack.append(_BlockInfo(True)) + if _MATCH_ASM.match(line): + self.stack[-1].inline_asm = _BLOCK_ASM + + elif token == ';' or token == ')': + # If we haven't seen an opening brace yet, but we already saw + # a semicolon, this is probably a forward declaration. Pop + # the stack for these. + # + # Similarly, if we haven't seen an opening brace yet, but we + # already saw a closing parenthesis, then these are probably + # function arguments with extra "class" or "struct" keywords. + # Also pop these stack for these. + if not self.SeenOpenBrace(): + self.stack.pop() + else: # token == '}' + # Perform end of block checks and pop the stack. + if self.stack: + self.stack[-1].CheckEnd(filename, clean_lines, linenum, error) + self.stack.pop() + line = matched.group(2) + + def InnermostClass(self): + """Get class info on the top of the stack. + + Returns: + A _ClassInfo object if we are inside a class, or None otherwise. + """ + for i in range(len(self.stack), 0, -1): + classinfo = self.stack[i - 1] + if isinstance(classinfo, _ClassInfo): + return classinfo + return None + + def CheckCompletedBlocks(self, filename, error): + """Checks that all classes and namespaces have been completely parsed. + + Call this when all lines in a file have been processed. + Args: + filename: The name of the current file. + error: The function to call with any errors found. + """ + # Note: This test can result in false positives if #ifdef constructs + # get in the way of brace matching. See the testBuildClass test in + # cpplint_unittest.py for an example of this. + for obj in self.stack: + if isinstance(obj, _ClassInfo): + error(filename, obj.starting_linenum, 'build/class', 5, + 'Failed to find complete declaration of class %s' % + obj.name) + elif isinstance(obj, _NamespaceInfo): + error(filename, obj.starting_linenum, 'build/namespaces', 5, + 'Failed to find complete declaration of namespace %s' % + obj.name) + + +def CheckForNonStandardConstructs(filename, clean_lines, linenum, + nesting_state, error): + r"""Logs an error if we see certain non-ANSI constructs ignored by gcc-2. + + Complain about several constructs which gcc-2 accepts, but which are + not standard C++. Warning about these in lint is one way to ease the + transition to new compilers. + - put storage class first (e.g. "static const" instead of "const static"). + - "%lld" instead of %qd" in printf-type functions. + - "%1$d" is non-standard in printf-type functions. + - "\%" is an undefined character escape sequence. + - text after #endif is not allowed. + - invalid inner-style forward declaration. + - >? and ?= and )\?=?\s*(\w+|[+-]?\d+)(\.\d*)?', + line): + error(filename, linenum, 'build/deprecated', 3, + '>? and ))?' + # r'\s*const\s*' + type_name + '\s*&\s*\w+\s*;' + error(filename, linenum, 'runtime/member_string_references', 2, + 'const string& members are dangerous. It is much better to use ' + 'alternatives, such as pointers or simple constants.') + + # Everything else in this function operates on class declarations. + # Return early if the top of the nesting stack is not a class, or if + # the class head is not completed yet. + classinfo = nesting_state.InnermostClass() + if not classinfo or not classinfo.seen_open_brace: + return + + # The class may have been declared with namespace or classname qualifiers. + # The constructor and destructor will not have those qualifiers. + base_classname = classinfo.name.split('::')[-1] + + # Look for single-argument constructors that aren't marked explicit. + # Technically a valid construct, but against style. Also look for + # non-single-argument constructors which are also technically valid, but + # strongly suggest something is wrong. + explicit_constructor_match = Match( + r'\s+(?:inline\s+)?(explicit\s+)?(?:inline\s+)?%s\s*' + r'\(((?:[^()]|\([^()]*\))*)\)' + % re.escape(base_classname), + line) + + if explicit_constructor_match: + is_marked_explicit = explicit_constructor_match.group(1) + + if not explicit_constructor_match.group(2): + constructor_args = [] + else: + constructor_args = explicit_constructor_match.group(2).split(',') + + # collapse arguments so that commas in template parameter lists and function + # argument parameter lists don't split arguments in two + i = 0 + while i < len(constructor_args): + constructor_arg = constructor_args[i] + while (constructor_arg.count('<') > constructor_arg.count('>') or + constructor_arg.count('(') > constructor_arg.count(')')): + constructor_arg += ',' + constructor_args[i + 1] + del constructor_args[i + 1] + constructor_args[i] = constructor_arg + i += 1 + + defaulted_args = [arg for arg in constructor_args if '=' in arg] + noarg_constructor = (not constructor_args or # empty arg list + # 'void' arg specifier + (len(constructor_args) == 1 and + constructor_args[0].strip() == 'void')) + onearg_constructor = ((len(constructor_args) == 1 and # exactly one arg + not noarg_constructor) or + # all but at most one arg defaulted + (len(constructor_args) >= 1 and + not noarg_constructor and + len(defaulted_args) >= len(constructor_args) - 1)) + initializer_list_constructor = bool( + onearg_constructor and + Search(r'\bstd\s*::\s*initializer_list\b', constructor_args[0])) + copy_constructor = bool( + onearg_constructor and + Match(r'(const\s+)?%s(\s*<[^>]*>)?(\s+const)?\s*(?:<\w+>\s*)?&' + % re.escape(base_classname), constructor_args[0].strip())) + + if (not is_marked_explicit and + onearg_constructor and + not initializer_list_constructor and + not copy_constructor): + if defaulted_args: + error(filename, linenum, 'runtime/explicit', 5, + 'Constructors callable with one argument ' + 'should be marked explicit.') + else: + error(filename, linenum, 'runtime/explicit', 5, + 'Single-parameter constructors should be marked explicit.') + elif is_marked_explicit and not onearg_constructor: + if noarg_constructor: + error(filename, linenum, 'runtime/explicit', 5, + 'Zero-parameter constructors should not be marked explicit.') + else: + error(filename, linenum, 'runtime/explicit', 0, + 'Constructors that require multiple arguments ' + 'should not be marked explicit.') + + +def CheckSpacingForFunctionCall(filename, clean_lines, linenum, error): + """Checks for the correctness of various spacing around function calls. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Since function calls often occur inside if/for/while/switch + # expressions - which have their own, more liberal conventions - we + # first see if we should be looking inside such an expression for a + # function call, to which we can apply more strict standards. + fncall = line # if there's no control flow construct, look at whole line + for pattern in (r'\bif\s*\((.*)\)\s*{', + r'\bfor\s*\((.*)\)\s*{', + r'\bwhile\s*\((.*)\)\s*[{;]', + r'\bswitch\s*\((.*)\)\s*{'): + match = Search(pattern, line) + if match: + fncall = match.group(1) # look inside the parens for function calls + break + + # Except in if/for/while/switch, there should never be space + # immediately inside parens (eg "f( 3, 4 )"). We make an exception + # for nested parens ( (a+b) + c ). Likewise, there should never be + # a space before a ( when it's a function argument. I assume it's a + # function argument when the char before the whitespace is legal in + # a function name (alnum + _) and we're not starting a macro. Also ignore + # pointers and references to arrays and functions coz they're too tricky: + # we use a very simple way to recognize these: + # " (something)(maybe-something)" or + # " (something)(maybe-something," or + # " (something)[something]" + # Note that we assume the contents of [] to be short enough that + # they'll never need to wrap. + if ( # Ignore control structures. + not Search(r'\b(if|for|while|switch|return|new|delete|catch|sizeof)\b', + fncall) and + # Ignore pointers/references to functions. + not Search(r' \([^)]+\)\([^)]*(\)|,$)', fncall) and + # Ignore pointers/references to arrays. + not Search(r' \([^)]+\)\[[^\]]+\]', fncall)): + if Search(r'\w\s*\(\s(?!\s*\\$)', fncall): # a ( used for a fn call + error(filename, linenum, 'whitespace/parens', 4, + 'Extra space after ( in function call') + elif Search(r'\(\s+(?!(\s*\\)|\()', fncall): + error(filename, linenum, 'whitespace/parens', 2, + 'Extra space after (') + if (Search(r'\w\s+\(', fncall) and + not Search(r'#\s*define|typedef|using\s+\w+\s*=', fncall) and + not Search(r'\w\s+\((\w+::)*\*\w+\)\(', fncall) and + not Search(r'\bcase\s+\(', fncall)): + # TODO(unknown): Space after an operator function seem to be a common + # error, silence those for now by restricting them to highest verbosity. + if Search(r'\boperator_*\b', line): + error(filename, linenum, 'whitespace/parens', 0, + 'Extra space before ( in function call') + else: + error(filename, linenum, 'whitespace/parens', 4, + 'Extra space before ( in function call') + # If the ) is followed only by a newline or a { + newline, assume it's + # part of a control statement (if/while/etc), and don't complain + if Search(r'[^)]\s+\)\s*[^{\s]', fncall): + # If the closing parenthesis is preceded by only whitespaces, + # try to give a more descriptive error message. + if Search(r'^\s+\)', fncall): + error(filename, linenum, 'whitespace/parens', 2, + 'Closing ) should be moved to the previous line') + else: + error(filename, linenum, 'whitespace/parens', 2, + 'Extra space before )') + + +def IsBlankLine(line): + """Returns true if the given line is blank. + + We consider a line to be blank if the line is empty or consists of + only white spaces. + + Args: + line: A line of a string. + + Returns: + True, if the given line is blank. + """ + return not line or line.isspace() + + +def CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line, + error): + is_namespace_indent_item = ( + len(nesting_state.stack) > 1 and + nesting_state.stack[-1].check_namespace_indentation and + isinstance(nesting_state.previous_stack_top, _NamespaceInfo) and + nesting_state.previous_stack_top == nesting_state.stack[-2]) + + if ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item, + clean_lines.elided, line): + CheckItemIndentationInNamespace(filename, clean_lines.elided, + line, error) + + +def CheckForFunctionLengths(filename, clean_lines, linenum, + function_state, error): + """Reports for long function bodies. + + For an overview why this is done, see: + http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Write_Short_Functions + + Uses a simplistic algorithm assuming other style guidelines + (especially spacing) are followed. + Only checks unindented functions, so class members are unchecked. + Trivial bodies are unchecked, so constructors with huge initializer lists + may be missed. + Blank/comment lines are not counted so as to avoid encouraging the removal + of vertical space and comments just to get through a lint check. + NOLINT *on the last line of a function* disables this check. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + function_state: Current function name and lines in body so far. + error: The function to call with any errors found. + """ + lines = clean_lines.lines + line = lines[linenum] + joined_line = '' + + starting_func = False + regexp = r'(\w(\w|::|\*|\&|\s)*)\(' # decls * & space::name( ... + match_result = Match(regexp, line) + if match_result: + # If the name is all caps and underscores, figure it's a macro and + # ignore it, unless it's TEST or TEST_F. + function_name = match_result.group(1).split()[-1] + if function_name == 'TEST' or function_name == 'TEST_F' or ( + not Match(r'[A-Z_]+$', function_name)): + starting_func = True + + if starting_func: + body_found = False + for start_linenum in xrange(linenum, clean_lines.NumLines()): + start_line = lines[start_linenum] + joined_line += ' ' + start_line.lstrip() + if Search(r'(;|})', start_line): # Declarations and trivial functions + body_found = True + break # ... ignore + elif Search(r'{', start_line): + body_found = True + function = Search(r'((\w|:)*)\(', line).group(1) + if Match(r'TEST', function): # Handle TEST... macros + parameter_regexp = Search(r'(\(.*\))', joined_line) + if parameter_regexp: # Ignore bad syntax + function += parameter_regexp.group(1) + else: + function += '()' + function_state.Begin(function) + break + if not body_found: + # No body for the function (or evidence of a non-function) was found. + error(filename, linenum, 'readability/fn_size', 5, + 'Lint failed to find start of function body.') + elif Match(r'^\}\s*$', line): # function end + function_state.Check(error, filename, linenum) + function_state.End() + elif not Match(r'^\s*$', line): + function_state.Count() # Count non-blank/non-comment lines. + + +_RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?') + + +def CheckComment(line, filename, linenum, next_line_start, error): + """Checks for common mistakes in comments. + + Args: + line: The line in question. + filename: The name of the current file. + linenum: The number of the line to check. + next_line_start: The first non-whitespace column of the next line. + error: The function to call with any errors found. + """ + commentpos = line.find('//') + if commentpos != -1: + # Check if the // may be in quotes. If so, ignore it + # Comparisons made explicit for clarity -- pylint: disable=g-explicit-bool-comparison + if (line.count('"', 0, commentpos) - + line.count('\\"', 0, commentpos)) % 2 == 0: # not in quotes + # Allow one space for new scopes, two spaces otherwise: + if (not (Match(r'^.*{ *//', line) and next_line_start == commentpos) and + ((commentpos >= 1 and + line[commentpos-1] not in string.whitespace) or + (commentpos >= 2 and + line[commentpos-2] not in string.whitespace))): + error(filename, linenum, 'whitespace/comments', 2, + 'At least two spaces is best between code and comments') + + # Checks for common mistakes in TODO comments. + comment = line[commentpos:] + match = _RE_PATTERN_TODO.match(comment) + if match: + # One whitespace is correct; zero whitespace is handled elsewhere. + leading_whitespace = match.group(1) + if len(leading_whitespace) > 1: + error(filename, linenum, 'whitespace/todo', 2, + 'Too many spaces before TODO') + + username = match.group(2) + if not username: + error(filename, linenum, 'readability/todo', 2, + 'Missing username in TODO; it should look like ' + '"// TODO(my_username): Stuff."') + + middle_whitespace = match.group(3) + # Comparisons made explicit for correctness -- pylint: disable=g-explicit-bool-comparison + if middle_whitespace != ' ' and middle_whitespace != '': + error(filename, linenum, 'whitespace/todo', 2, + 'TODO(my_username) should be followed by a space') + + # If the comment contains an alphanumeric character, there + # should be a space somewhere between it and the // unless + # it's a /// or //! Doxygen comment. + if (Match(r'//[^ ]*\w', comment) and + not Match(r'(///|//\!)(\s+|$)', comment)): + error(filename, linenum, 'whitespace/comments', 4, + 'Should have a space between // and comment') + + +def CheckAccess(filename, clean_lines, linenum, nesting_state, error): + """Checks for improper use of DISALLOW* macros. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + nesting_state: A NestingState instance which maintains information about + the current stack of nested blocks being parsed. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] # get rid of comments and strings + + matched = Match((r'\s*(DISALLOW_COPY_AND_ASSIGN|' + r'DISALLOW_IMPLICIT_CONSTRUCTORS)'), line) + if not matched: + return + if nesting_state.stack and isinstance(nesting_state.stack[-1], _ClassInfo): + if nesting_state.stack[-1].access != 'private': + error(filename, linenum, 'readability/constructors', 3, + '%s must be in the private: section' % matched.group(1)) + + else: + # Found DISALLOW* macro outside a class declaration, or perhaps it + # was used inside a function when it should have been part of the + # class declaration. We could issue a warning here, but it + # probably resulted in a compiler error already. + pass + + +def CheckSpacing(filename, clean_lines, linenum, nesting_state, error): + """Checks for the correctness of various spacing issues in the code. + + Things we check for: spaces around operators, spaces after + if/for/while/switch, no spaces around parens in function calls, two + spaces between code and comment, don't start a block with a blank + line, don't end a function with a blank line, don't add a blank line + after public/protected/private, don't have too many blank lines in a row. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + nesting_state: A NestingState instance which maintains information about + the current stack of nested blocks being parsed. + error: The function to call with any errors found. + """ + + # Don't use "elided" lines here, otherwise we can't check commented lines. + # Don't want to use "raw" either, because we don't want to check inside C++11 + # raw strings, + raw = clean_lines.lines_without_raw_strings + line = raw[linenum] + + # Before nixing comments, check if the line is blank for no good + # reason. This includes the first line after a block is opened, and + # blank lines at the end of a function (ie, right before a line like '}' + # + # Skip all the blank line checks if we are immediately inside a + # namespace body. In other words, don't issue blank line warnings + # for this block: + # namespace { + # + # } + # + # A warning about missing end of namespace comments will be issued instead. + # + # Also skip blank line checks for 'extern "C"' blocks, which are formatted + # like namespaces. + if (IsBlankLine(line) and + not nesting_state.InNamespaceBody() and + not nesting_state.InExternC()): + elided = clean_lines.elided + prev_line = elided[linenum - 1] + prevbrace = prev_line.rfind('{') + # TODO(unknown): Don't complain if line before blank line, and line after, + # both start with alnums and are indented the same amount. + # This ignores whitespace at the start of a namespace block + # because those are not usually indented. + if prevbrace != -1 and prev_line[prevbrace:].find('}') == -1: + # OK, we have a blank line at the start of a code block. Before we + # complain, we check if it is an exception to the rule: The previous + # non-empty line has the parameters of a function header that are indented + # 4 spaces (because they did not fit in a 80 column line when placed on + # the same line as the function name). We also check for the case where + # the previous line is indented 6 spaces, which may happen when the + # initializers of a constructor do not fit into a 80 column line. + exception = False + if Match(r' {6}\w', prev_line): # Initializer list? + # We are looking for the opening column of initializer list, which + # should be indented 4 spaces to cause 6 space indentation afterwards. + search_position = linenum-2 + while (search_position >= 0 + and Match(r' {6}\w', elided[search_position])): + search_position -= 1 + exception = (search_position >= 0 + and elided[search_position][:5] == ' :') + else: + # Search for the function arguments or an initializer list. We use a + # simple heuristic here: If the line is indented 4 spaces; and we have a + # closing paren, without the opening paren, followed by an opening brace + # or colon (for initializer lists) we assume that it is the last line of + # a function header. If we have a colon indented 4 spaces, it is an + # initializer list. + exception = (Match(r' {4}\w[^\(]*\)\s*(const\s*)?(\{\s*$|:)', + prev_line) + or Match(r' {4}:', prev_line)) + + if not exception: + error(filename, linenum, 'whitespace/blank_line', 2, + 'Redundant blank line at the start of a code block ' + 'should be deleted.') + # Ignore blank lines at the end of a block in a long if-else + # chain, like this: + # if (condition1) { + # // Something followed by a blank line + # + # } else if (condition2) { + # // Something else + # } + if linenum + 1 < clean_lines.NumLines(): + next_line = raw[linenum + 1] + if (next_line + and Match(r'\s*}', next_line) + and next_line.find('} else ') == -1): + error(filename, linenum, 'whitespace/blank_line', 3, + 'Redundant blank line at the end of a code block ' + 'should be deleted.') + + matched = Match(r'\s*(public|protected|private):', prev_line) + if matched: + error(filename, linenum, 'whitespace/blank_line', 3, + 'Do not leave a blank line after "%s:"' % matched.group(1)) + + # Next, check comments + next_line_start = 0 + if linenum + 1 < clean_lines.NumLines(): + next_line = raw[linenum + 1] + next_line_start = len(next_line) - len(next_line.lstrip()) + CheckComment(line, filename, linenum, next_line_start, error) + + # get rid of comments and strings + line = clean_lines.elided[linenum] + + # You shouldn't have spaces before your brackets, except maybe after + # 'delete []' or 'return []() {};' + if Search(r'\w\s+\[', line) and not Search(r'(?:delete|return)\s+\[', line): + error(filename, linenum, 'whitespace/braces', 5, + 'Extra space before [') + + # In range-based for, we wanted spaces before and after the colon, but + # not around "::" tokens that might appear. + if (Search(r'for *\(.*[^:]:[^: ]', line) or + Search(r'for *\(.*[^: ]:[^:]', line)): + error(filename, linenum, 'whitespace/forcolon', 2, + 'Missing space around colon in range-based for loop') + + +def CheckOperatorSpacing(filename, clean_lines, linenum, error): + """Checks for horizontal spacing around operators. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Don't try to do spacing checks for operator methods. Do this by + # replacing the troublesome characters with something else, + # preserving column position for all other characters. + # + # The replacement is done repeatedly to avoid false positives from + # operators that call operators. + while True: + match = Match(r'^(.*\boperator\b)(\S+)(\s*\(.*)$', line) + if match: + line = match.group(1) + ('_' * len(match.group(2))) + match.group(3) + else: + break + + # We allow no-spaces around = within an if: "if ( (a=Foo()) == 0 )". + # Otherwise not. Note we only check for non-spaces on *both* sides; + # sometimes people put non-spaces on one side when aligning ='s among + # many lines (not that this is behavior that I approve of...) + if ((Search(r'[\w.]=', line) or + Search(r'=[\w.]', line)) + and not Search(r'\b(if|while|for) ', line) + # Operators taken from [lex.operators] in C++11 standard. + and not Search(r'(>=|<=|==|!=|&=|\^=|\|=|\+=|\*=|\/=|\%=)', line) + and not Search(r'operator=', line)): + error(filename, linenum, 'whitespace/operators', 4, + 'Missing spaces around =') + + # It's ok not to have spaces around binary operators like + - * /, but if + # there's too little whitespace, we get concerned. It's hard to tell, + # though, so we punt on this one for now. TODO. + + # You should always have whitespace around binary operators. + # + # Check <= and >= first to avoid false positives with < and >, then + # check non-include lines for spacing around < and >. + # + # If the operator is followed by a comma, assume it's be used in a + # macro context and don't do any checks. This avoids false + # positives. + # + # Note that && is not included here. Those are checked separately + # in CheckRValueReference + match = Search(r'[^<>=!\s](==|!=|<=|>=|\|\|)[^<>=!\s,;\)]', line) + if match: + error(filename, linenum, 'whitespace/operators', 3, + 'Missing spaces around %s' % match.group(1)) + elif not Match(r'#.*include', line): + # Look for < that is not surrounded by spaces. This is only + # triggered if both sides are missing spaces, even though + # technically should should flag if at least one side is missing a + # space. This is done to avoid some false positives with shifts. + match = Match(r'^(.*[^\s<])<[^\s=<,]', line) + if match: + (_, _, end_pos) = CloseExpression( + clean_lines, linenum, len(match.group(1))) + if end_pos <= -1: + error(filename, linenum, 'whitespace/operators', 3, + 'Missing spaces around <') + + # Look for > that is not surrounded by spaces. Similar to the + # above, we only trigger if both sides are missing spaces to avoid + # false positives with shifts. + match = Match(r'^(.*[^-\s>])>[^\s=>,]', line) + if match: + (_, _, start_pos) = ReverseCloseExpression( + clean_lines, linenum, len(match.group(1))) + if start_pos <= -1: + error(filename, linenum, 'whitespace/operators', 3, + 'Missing spaces around >') + + # We allow no-spaces around << when used like this: 10<<20, but + # not otherwise (particularly, not when used as streams) + # + # We also allow operators following an opening parenthesis, since + # those tend to be macros that deal with operators. + match = Search(r'(operator|[^\s(<])(?:L|UL|ULL|l|ul|ull)?<<([^\s,=<])', line) + if (match and not (match.group(1).isdigit() and match.group(2).isdigit()) and + not (match.group(1) == 'operator' and match.group(2) == ';')): + error(filename, linenum, 'whitespace/operators', 3, + 'Missing spaces around <<') + + # We allow no-spaces around >> for almost anything. This is because + # C++11 allows ">>" to close nested templates, which accounts for + # most cases when ">>" is not followed by a space. + # + # We still warn on ">>" followed by alpha character, because that is + # likely due to ">>" being used for right shifts, e.g.: + # value >> alpha + # + # When ">>" is used to close templates, the alphanumeric letter that + # follows would be part of an identifier, and there should still be + # a space separating the template type and the identifier. + # type> alpha + match = Search(r'>>[a-zA-Z_]', line) + if match: + error(filename, linenum, 'whitespace/operators', 3, + 'Missing spaces around >>') + + # There shouldn't be space around unary operators + match = Search(r'(!\s|~\s|[\s]--[\s;]|[\s]\+\+[\s;])', line) + if match: + error(filename, linenum, 'whitespace/operators', 4, + 'Extra space for operator %s' % match.group(1)) + + +def CheckParenthesisSpacing(filename, clean_lines, linenum, error): + """Checks for horizontal spacing around parentheses. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # No spaces after an if, while, switch, or for + match = Search(r' (if\(|for\(|while\(|switch\()', line) + if match: + error(filename, linenum, 'whitespace/parens', 5, + 'Missing space before ( in %s' % match.group(1)) + + # For if/for/while/switch, the left and right parens should be + # consistent about how many spaces are inside the parens, and + # there should either be zero or one spaces inside the parens. + # We don't want: "if ( foo)" or "if ( foo )". + # Exception: "for ( ; foo; bar)" and "for (foo; bar; )" are allowed. + match = Search(r'\b(if|for|while|switch)\s*' + r'\(([ ]*)(.).*[^ ]+([ ]*)\)\s*{\s*$', + line) + if match: + if len(match.group(2)) != len(match.group(4)): + if not (match.group(3) == ';' and + len(match.group(2)) == 1 + len(match.group(4)) or + not match.group(2) and Search(r'\bfor\s*\(.*; \)', line)): + error(filename, linenum, 'whitespace/parens', 5, + 'Mismatching spaces inside () in %s' % match.group(1)) + if len(match.group(2)) not in [0, 1]: + error(filename, linenum, 'whitespace/parens', 5, + 'Should have zero or one spaces inside ( and ) in %s' % + match.group(1)) + + +def CheckCommaSpacing(filename, clean_lines, linenum, error): + """Checks for horizontal spacing near commas and semicolons. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + raw = clean_lines.lines_without_raw_strings + line = clean_lines.elided[linenum] + + # You should always have a space after a comma (either as fn arg or operator) + # + # This does not apply when the non-space character following the + # comma is another comma, since the only time when that happens is + # for empty macro arguments. + # + # We run this check in two passes: first pass on elided lines to + # verify that lines contain missing whitespaces, second pass on raw + # lines to confirm that those missing whitespaces are not due to + # elided comments. + if (Search(r',[^,\s]', ReplaceAll(r'\boperator\s*,\s*\(', 'F(', line)) and + Search(r',[^,\s]', raw[linenum])): + error(filename, linenum, 'whitespace/comma', 3, + 'Missing space after ,') + + # You should always have a space after a semicolon + # except for few corner cases + # TODO(unknown): clarify if 'if (1) { return 1;}' is requires one more + # space after ; + if Search(r';[^\s};\\)/]', line): + error(filename, linenum, 'whitespace/semicolon', 3, + 'Missing space after ;') + + +def CheckBracesSpacing(filename, clean_lines, linenum, error): + """Checks for horizontal spacing near commas. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Except after an opening paren, or after another opening brace (in case of + # an initializer list, for instance), you should have spaces before your + # braces. And since you should never have braces at the beginning of a line, + # this is an easy test. + match = Match(r'^(.*[^ ({>]){', line) + if match: + # Try a bit harder to check for brace initialization. This + # happens in one of the following forms: + # Constructor() : initializer_list_{} { ... } + # Constructor{}.MemberFunction() + # Type variable{}; + # FunctionCall(type{}, ...); + # LastArgument(..., type{}); + # LOG(INFO) << type{} << " ..."; + # map_of_type[{...}] = ...; + # ternary = expr ? new type{} : nullptr; + # OuterTemplate{}> + # + # We check for the character following the closing brace, and + # silence the warning if it's one of those listed above, i.e. + # "{.;,)<>]:". + # + # To account for nested initializer list, we allow any number of + # closing braces up to "{;,)<". We can't simply silence the + # warning on first sight of closing brace, because that would + # cause false negatives for things that are not initializer lists. + # Silence this: But not this: + # Outer{ if (...) { + # Inner{...} if (...){ // Missing space before { + # }; } + # + # There is a false negative with this approach if people inserted + # spurious semicolons, e.g. "if (cond){};", but we will catch the + # spurious semicolon with a separate check. + (endline, endlinenum, endpos) = CloseExpression( + clean_lines, linenum, len(match.group(1))) + trailing_text = '' + if endpos > -1: + trailing_text = endline[endpos:] + for offset in xrange(endlinenum + 1, + min(endlinenum + 3, clean_lines.NumLines() - 1)): + trailing_text += clean_lines.elided[offset] + if not Match(r'^[\s}]*[{.;,)<>\]:]', trailing_text): + error(filename, linenum, 'whitespace/braces', 5, + 'Missing space before {') + + # Make sure '} else {' has spaces. + if Search(r'}else', line): + error(filename, linenum, 'whitespace/braces', 5, + 'Missing space before else') + + # You shouldn't have a space before a semicolon at the end of the line. + # There's a special case for "for" since the style guide allows space before + # the semicolon there. + if Search(r':\s*;\s*$', line): + error(filename, linenum, 'whitespace/semicolon', 5, + 'Semicolon defining empty statement. Use {} instead.') + elif Search(r'^\s*;\s*$', line): + error(filename, linenum, 'whitespace/semicolon', 5, + 'Line contains only semicolon. If this should be an empty statement, ' + 'use {} instead.') + elif (Search(r'\s+;\s*$', line) and + not Search(r'\bfor\b', line)): + error(filename, linenum, 'whitespace/semicolon', 5, + 'Extra space before last semicolon. If this should be an empty ' + 'statement, use {} instead.') + + +def IsDecltype(clean_lines, linenum, column): + """Check if the token ending on (linenum, column) is decltype(). + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: the number of the line to check. + column: end column of the token to check. + Returns: + True if this token is decltype() expression, False otherwise. + """ + (text, _, start_col) = ReverseCloseExpression(clean_lines, linenum, column) + if start_col < 0: + return False + if Search(r'\bdecltype\s*$', text[0:start_col]): + return True + return False + + +def IsTemplateParameterList(clean_lines, linenum, column): + """Check if the token ending on (linenum, column) is the end of template<>. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: the number of the line to check. + column: end column of the token to check. + Returns: + True if this token is end of a template parameter list, False otherwise. + """ + (_, startline, startpos) = ReverseCloseExpression( + clean_lines, linenum, column) + if (startpos > -1 and + Search(r'\btemplate\s*$', clean_lines.elided[startline][0:startpos])): + return True + return False + + +def IsRValueType(typenames, clean_lines, nesting_state, linenum, column): + """Check if the token ending on (linenum, column) is a type. + + Assumes that text to the right of the column is "&&" or a function + name. + + Args: + typenames: set of type names from template-argument-list. + clean_lines: A CleansedLines instance containing the file. + nesting_state: A NestingState instance which maintains information about + the current stack of nested blocks being parsed. + linenum: the number of the line to check. + column: end column of the token to check. + Returns: + True if this token is a type, False if we are not sure. + """ + prefix = clean_lines.elided[linenum][0:column] + + # Get one word to the left. If we failed to do so, this is most + # likely not a type, since it's unlikely that the type name and "&&" + # would be split across multiple lines. + match = Match(r'^(.*)(\b\w+|[>*)&])\s*$', prefix) + if not match: + return False + + # Check text following the token. If it's "&&>" or "&&," or "&&...", it's + # most likely a rvalue reference used inside a template. + suffix = clean_lines.elided[linenum][column:] + if Match(r'&&\s*(?:[>,]|\.\.\.)', suffix): + return True + + # Check for known types and end of templates: + # int&& variable + # vector&& variable + # + # Because this function is called recursively, we also need to + # recognize pointer and reference types: + # int* Function() + # int& Function() + if (match.group(2) in typenames or + match.group(2) in ['char', 'char16_t', 'char32_t', 'wchar_t', 'bool', + 'short', 'int', 'long', 'signed', 'unsigned', + 'float', 'double', 'void', 'auto', '>', '*', '&']): + return True + + # If we see a close parenthesis, look for decltype on the other side. + # decltype would unambiguously identify a type, anything else is + # probably a parenthesized expression and not a type. + if match.group(2) == ')': + return IsDecltype( + clean_lines, linenum, len(match.group(1)) + len(match.group(2)) - 1) + + # Check for casts and cv-qualifiers. + # match.group(1) remainder + # -------------- --------- + # const_cast< type&& + # const type&& + # type const&& + if Search(r'\b(?:const_cast\s*<|static_cast\s*<|dynamic_cast\s*<|' + r'reinterpret_cast\s*<|\w+\s)\s*$', + match.group(1)): + return True + + # Look for a preceding symbol that might help differentiate the context. + # These are the cases that would be ambiguous: + # match.group(1) remainder + # -------------- --------- + # Call ( expression && + # Declaration ( type&& + # sizeof ( type&& + # if ( expression && + # while ( expression && + # for ( type&& + # for( ; expression && + # statement ; type&& + # block { type&& + # constructor { expression && + start = linenum + line = match.group(1) + match_symbol = None + while start >= 0: + # We want to skip over identifiers and commas to get to a symbol. + # Commas are skipped so that we can find the opening parenthesis + # for function parameter lists. + match_symbol = Match(r'^(.*)([^\w\s,])[\w\s,]*$', line) + if match_symbol: + break + start -= 1 + line = clean_lines.elided[start] + + if not match_symbol: + # Probably the first statement in the file is an rvalue reference + return True + + if match_symbol.group(2) == '}': + # Found closing brace, probably an indicate of this: + # block{} type&& + return True + + if match_symbol.group(2) == ';': + # Found semicolon, probably one of these: + # for(; expression && + # statement; type&& + + # Look for the previous 'for(' in the previous lines. + before_text = match_symbol.group(1) + for i in xrange(start - 1, max(start - 6, 0), -1): + before_text = clean_lines.elided[i] + before_text + if Search(r'for\s*\([^{};]*$', before_text): + # This is the condition inside a for-loop + return False + + # Did not find a for-init-statement before this semicolon, so this + # is probably a new statement and not a condition. + return True + + if match_symbol.group(2) == '{': + # Found opening brace, probably one of these: + # block{ type&& = ... ; } + # constructor{ expression && expression } + + # Look for a closing brace or a semicolon. If we see a semicolon + # first, this is probably a rvalue reference. + line = clean_lines.elided[start][0:len(match_symbol.group(1)) + 1] + end = start + depth = 1 + while True: + for ch in line: + if ch == ';': + return True + elif ch == '{': + depth += 1 + elif ch == '}': + depth -= 1 + if depth == 0: + return False + end += 1 + if end >= clean_lines.NumLines(): + break + line = clean_lines.elided[end] + # Incomplete program? + return False + + if match_symbol.group(2) == '(': + # Opening parenthesis. Need to check what's to the left of the + # parenthesis. Look back one extra line for additional context. + before_text = match_symbol.group(1) + if linenum > 1: + before_text = clean_lines.elided[linenum - 1] + before_text + before_text = match_symbol.group(1) + + # Patterns that are likely to be types: + # [](type&& + # for (type&& + # sizeof(type&& + # operator=(type&& + # + if Search(r'(?:\]|\bfor|\bsizeof|\boperator\s*\S+\s*)\s*$', before_text): + return True + + # Patterns that are likely to be expressions: + # if (expression && + # while (expression && + # : initializer(expression && + # , initializer(expression && + # ( FunctionCall(expression && + # + FunctionCall(expression && + # + (expression && + # + # The last '+' represents operators such as '+' and '-'. + if Search(r'(?:\bif|\bwhile|[-+=%^(]*>)?\s*$', + match_symbol.group(1)) + if match_func: + # Check for constructors, which don't have return types. + if Search(r'\b(?:explicit|inline)$', match_func.group(1)): + return True + implicit_constructor = Match(r'\s*(\w+)\((?:const\s+)?(\w+)', prefix) + if (implicit_constructor and + implicit_constructor.group(1) == implicit_constructor.group(2)): + return True + return IsRValueType(typenames, clean_lines, nesting_state, linenum, + len(match_func.group(1))) + + # Nothing before the function name. If this is inside a block scope, + # this is probably a function call. + return not (nesting_state.previous_stack_top and + nesting_state.previous_stack_top.IsBlockInfo()) + + if match_symbol.group(2) == '>': + # Possibly a closing bracket, check that what's on the other side + # looks like the start of a template. + return IsTemplateParameterList( + clean_lines, start, len(match_symbol.group(1))) + + # Some other symbol, usually something like "a=b&&c". This is most + # likely not a type. + return False + + +def IsDeletedOrDefault(clean_lines, linenum): + """Check if current constructor or operator is deleted or default. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + Returns: + True if this is a deleted or default constructor. + """ + open_paren = clean_lines.elided[linenum].find('(') + if open_paren < 0: + return False + (close_line, _, close_paren) = CloseExpression( + clean_lines, linenum, open_paren) + if close_paren < 0: + return False + return Match(r'\s*=\s*(?:delete|default)\b', close_line[close_paren:]) + + +def IsRValueAllowed(clean_lines, linenum, typenames): + """Check if RValue reference is allowed on a particular line. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + typenames: set of type names from template-argument-list. + Returns: + True if line is within the region where RValue references are allowed. + """ + # Allow region marked by PUSH/POP macros + for i in xrange(linenum, 0, -1): + line = clean_lines.elided[i] + if Match(r'GOOGLE_ALLOW_RVALUE_REFERENCES_(?:PUSH|POP)', line): + if not line.endswith('PUSH'): + return False + for j in xrange(linenum, clean_lines.NumLines(), 1): + line = clean_lines.elided[j] + if Match(r'GOOGLE_ALLOW_RVALUE_REFERENCES_(?:PUSH|POP)', line): + return line.endswith('POP') + + # Allow operator= + line = clean_lines.elided[linenum] + if Search(r'\boperator\s*=\s*\(', line): + return IsDeletedOrDefault(clean_lines, linenum) + + # Allow constructors + match = Match(r'\s*(?:[\w<>]+::)*([\w<>]+)\s*::\s*([\w<>]+)\s*\(', line) + if match and match.group(1) == match.group(2): + return IsDeletedOrDefault(clean_lines, linenum) + if Search(r'\b(?:explicit|inline)\s+[\w<>]+\s*\(', line): + return IsDeletedOrDefault(clean_lines, linenum) + + if Match(r'\s*[\w<>]+\s*\(', line): + previous_line = 'ReturnType' + if linenum > 0: + previous_line = clean_lines.elided[linenum - 1] + if Match(r'^\s*$', previous_line) or Search(r'[{}:;]\s*$', previous_line): + return IsDeletedOrDefault(clean_lines, linenum) + + # Reject types not mentioned in template-argument-list + while line: + match = Match(r'^.*?(\w+)\s*&&(.*)$', line) + if not match: + break + if match.group(1) not in typenames: + return False + line = match.group(2) + + # All RValue types that were in template-argument-list should have + # been removed by now. Those were allowed, assuming that they will + # be forwarded. + # + # If there are no remaining RValue types left (i.e. types that were + # not found in template-argument-list), flag those as not allowed. + return line.find('&&') < 0 + + +def GetTemplateArgs(clean_lines, linenum): + """Find list of template arguments associated with this function declaration. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: Line number containing the start of the function declaration, + usually one line after the end of the template-argument-list. + Returns: + Set of type names, or empty set if this does not appear to have + any template parameters. + """ + # Find start of function + func_line = linenum + while func_line > 0: + line = clean_lines.elided[func_line] + if Match(r'^\s*$', line): + return set() + if line.find('(') >= 0: + break + func_line -= 1 + if func_line == 0: + return set() + + # Collapse template-argument-list into a single string + argument_list = '' + match = Match(r'^(\s*template\s*)<', clean_lines.elided[func_line]) + if match: + # template-argument-list on the same line as function name + start_col = len(match.group(1)) + _, end_line, end_col = CloseExpression(clean_lines, func_line, start_col) + if end_col > -1 and end_line == func_line: + start_col += 1 # Skip the opening bracket + argument_list = clean_lines.elided[func_line][start_col:end_col] + + elif func_line > 1: + # template-argument-list one line before function name + match = Match(r'^(.*)>\s*$', clean_lines.elided[func_line - 1]) + if match: + end_col = len(match.group(1)) + _, start_line, start_col = ReverseCloseExpression( + clean_lines, func_line - 1, end_col) + if start_col > -1: + start_col += 1 # Skip the opening bracket + while start_line < func_line - 1: + argument_list += clean_lines.elided[start_line][start_col:] + start_col = 0 + start_line += 1 + argument_list += clean_lines.elided[func_line - 1][start_col:end_col] + + if not argument_list: + return set() + + # Extract type names + typenames = set() + while True: + match = Match(r'^[,\s]*(?:typename|class)(?:\.\.\.)?\s+(\w+)(.*)$', + argument_list) + if not match: + break + typenames.add(match.group(1)) + argument_list = match.group(2) + return typenames + + +def CheckRValueReference(filename, clean_lines, linenum, nesting_state, error): + """Check for rvalue references. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + nesting_state: A NestingState instance which maintains information about + the current stack of nested blocks being parsed. + error: The function to call with any errors found. + """ + # Find lines missing spaces around &&. + # TODO(unknown): currently we don't check for rvalue references + # with spaces surrounding the && to avoid false positives with + # boolean expressions. + line = clean_lines.elided[linenum] + match = Match(r'^(.*\S)&&', line) + if not match: + match = Match(r'(.*)&&\S', line) + if (not match) or '(&&)' in line or Search(r'\boperator\s*$', match.group(1)): + return + + # Either poorly formed && or an rvalue reference, check the context + # to get a more accurate error message. Mostly we want to determine + # if what's to the left of "&&" is a type or not. + typenames = GetTemplateArgs(clean_lines, linenum) + and_pos = len(match.group(1)) + if IsRValueType(typenames, clean_lines, nesting_state, linenum, and_pos): + if not IsRValueAllowed(clean_lines, linenum, typenames): + error(filename, linenum, 'build/c++11', 3, + 'RValue references are an unapproved C++ feature.') + else: + error(filename, linenum, 'whitespace/operators', 3, + 'Missing spaces around &&') + + +def CheckSectionSpacing(filename, clean_lines, class_info, linenum, error): + """Checks for additional blank line issues related to sections. + + Currently the only thing checked here is blank line before protected/private. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + class_info: A _ClassInfo objects. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + # Skip checks if the class is small, where small means 25 lines or less. + # 25 lines seems like a good cutoff since that's the usual height of + # terminals, and any class that can't fit in one screen can't really + # be considered "small". + # + # Also skip checks if we are on the first line. This accounts for + # classes that look like + # class Foo { public: ... }; + # + # If we didn't find the end of the class, last_line would be zero, + # and the check will be skipped by the first condition. + if (class_info.last_line - class_info.starting_linenum <= 24 or + linenum <= class_info.starting_linenum): + return + + matched = Match(r'\s*(public|protected|private):', clean_lines.lines[linenum]) + if matched: + # Issue warning if the line before public/protected/private was + # not a blank line, but don't do this if the previous line contains + # "class" or "struct". This can happen two ways: + # - We are at the beginning of the class. + # - We are forward-declaring an inner class that is semantically + # private, but needed to be public for implementation reasons. + # Also ignores cases where the previous line ends with a backslash as can be + # common when defining classes in C macros. + prev_line = clean_lines.lines[linenum - 1] + if (not IsBlankLine(prev_line) and + not Search(r'\b(class|struct)\b', prev_line) and + not Search(r'\\$', prev_line)): + # Try a bit harder to find the beginning of the class. This is to + # account for multi-line base-specifier lists, e.g.: + # class Derived + # : public Base { + end_class_head = class_info.starting_linenum + for i in range(class_info.starting_linenum, linenum): + if Search(r'\{\s*$', clean_lines.lines[i]): + end_class_head = i + break + if end_class_head < linenum - 1: + error(filename, linenum, 'whitespace/blank_line', 3, + '"%s:" should be preceded by a blank line' % matched.group(1)) + + +def GetPreviousNonBlankLine(clean_lines, linenum): + """Return the most recent non-blank line and its line number. + + Args: + clean_lines: A CleansedLines instance containing the file contents. + linenum: The number of the line to check. + + Returns: + A tuple with two elements. The first element is the contents of the last + non-blank line before the current line, or the empty string if this is the + first non-blank line. The second is the line number of that line, or -1 + if this is the first non-blank line. + """ + + prevlinenum = linenum - 1 + while prevlinenum >= 0: + prevline = clean_lines.elided[prevlinenum] + if not IsBlankLine(prevline): # if not a blank line... + return (prevline, prevlinenum) + prevlinenum -= 1 + return ('', -1) + + +def CheckBraces(filename, clean_lines, linenum, error): + """Looks for misplaced braces (e.g. at the end of line). + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + + line = clean_lines.elided[linenum] # get rid of comments and strings + + if Match(r'\s*{\s*$', line): + # We allow an open brace to start a line in the case where someone is using + # braces in a block to explicitly create a new scope, which is commonly used + # to control the lifetime of stack-allocated variables. Braces are also + # used for brace initializers inside function calls. We don't detect this + # perfectly: we just don't complain if the last non-whitespace character on + # the previous non-blank line is ',', ';', ':', '(', '{', or '}', or if the + # previous line starts a preprocessor block. + prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] + if (not Search(r'[,;:}{(]\s*$', prevline) and + not Match(r'\s*#', prevline)): + error(filename, linenum, 'whitespace/braces', 4, + '{ should almost always be at the end of the previous line') + + # An else clause should be on the same line as the preceding closing brace. + if Match(r'\s*else\b\s*(?:if\b|\{|$)', line): + prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] + if Match(r'\s*}\s*$', prevline): + error(filename, linenum, 'whitespace/newline', 4, + 'An else should appear on the same line as the preceding }') + + # If braces come on one side of an else, they should be on both. + # However, we have to worry about "else if" that spans multiple lines! + if Search(r'else if\s*\(', line): # could be multi-line if + brace_on_left = bool(Search(r'}\s*else if\s*\(', line)) + # find the ( after the if + pos = line.find('else if') + pos = line.find('(', pos) + if pos > 0: + (endline, _, endpos) = CloseExpression(clean_lines, linenum, pos) + brace_on_right = endline[endpos:].find('{') != -1 + if brace_on_left != brace_on_right: # must be brace after if + error(filename, linenum, 'readability/braces', 5, + 'If an else has a brace on one side, it should have it on both') + elif Search(r'}\s*else[^{]*$', line) or Match(r'[^}]*else\s*{', line): + error(filename, linenum, 'readability/braces', 5, + 'If an else has a brace on one side, it should have it on both') + + # Likewise, an else should never have the else clause on the same line + if Search(r'\belse [^\s{]', line) and not Search(r'\belse if\b', line): + error(filename, linenum, 'whitespace/newline', 4, + 'Else clause should never be on same line as else (use 2 lines)') + + # In the same way, a do/while should never be on one line + if Match(r'\s*do [^\s{]', line): + error(filename, linenum, 'whitespace/newline', 4, + 'do/while clauses should not be on a single line') + + # Check single-line if/else bodies. The style guide says 'curly braces are not + # required for single-line statements'. We additionally allow multi-line, + # single statements, but we reject anything with more than one semicolon in + # it. This means that the first semicolon after the if should be at the end of + # its line, and the line after that should have an indent level equal to or + # lower than the if. We also check for ambiguous if/else nesting without + # braces. + if_else_match = Search(r'\b(if\s*\(|else\b)', line) + if if_else_match and not Match(r'\s*#', line): + if_indent = GetIndentLevel(line) + endline, endlinenum, endpos = line, linenum, if_else_match.end() + if_match = Search(r'\bif\s*\(', line) + if if_match: + # This could be a multiline if condition, so find the end first. + pos = if_match.end() - 1 + (endline, endlinenum, endpos) = CloseExpression(clean_lines, linenum, pos) + # Check for an opening brace, either directly after the if or on the next + # line. If found, this isn't a single-statement conditional. + if (not Match(r'\s*{', endline[endpos:]) + and not (Match(r'\s*$', endline[endpos:]) + and endlinenum < (len(clean_lines.elided) - 1) + and Match(r'\s*{', clean_lines.elided[endlinenum + 1]))): + while (endlinenum < len(clean_lines.elided) + and ';' not in clean_lines.elided[endlinenum][endpos:]): + endlinenum += 1 + endpos = 0 + if endlinenum < len(clean_lines.elided): + endline = clean_lines.elided[endlinenum] + # We allow a mix of whitespace and closing braces (e.g. for one-liner + # methods) and a single \ after the semicolon (for macros) + endpos = endline.find(';') + if not Match(r';[\s}]*(\\?)$', endline[endpos:]): + # Semicolon isn't the last character, there's something trailing. + # Output a warning if the semicolon is not contained inside + # a lambda expression. + if not Match(r'^[^{};]*\[[^\[\]]*\][^{}]*\{[^{}]*\}\s*\)*[;,]\s*$', + endline): + error(filename, linenum, 'readability/braces', 4, + 'If/else bodies with multiple statements require braces') + elif endlinenum < len(clean_lines.elided) - 1: + # Make sure the next line is dedented + next_line = clean_lines.elided[endlinenum + 1] + next_indent = GetIndentLevel(next_line) + # With ambiguous nested if statements, this will error out on the + # if that *doesn't* match the else, regardless of whether it's the + # inner one or outer one. + if (if_match and Match(r'\s*else\b', next_line) + and next_indent != if_indent): + error(filename, linenum, 'readability/braces', 4, + 'Else clause should be indented at the same level as if. ' + 'Ambiguous nested if/else chains require braces.') + elif next_indent > if_indent: + error(filename, linenum, 'readability/braces', 4, + 'If/else bodies with multiple statements require braces') + + +def CheckTrailingSemicolon(filename, clean_lines, linenum, error): + """Looks for redundant trailing semicolon. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + + line = clean_lines.elided[linenum] + + # Block bodies should not be followed by a semicolon. Due to C++11 + # brace initialization, there are more places where semicolons are + # required than not, so we use a whitelist approach to check these + # rather than a blacklist. These are the places where "};" should + # be replaced by just "}": + # 1. Some flavor of block following closing parenthesis: + # for (;;) {}; + # while (...) {}; + # switch (...) {}; + # Function(...) {}; + # if (...) {}; + # if (...) else if (...) {}; + # + # 2. else block: + # if (...) else {}; + # + # 3. const member function: + # Function(...) const {}; + # + # 4. Block following some statement: + # x = 42; + # {}; + # + # 5. Block at the beginning of a function: + # Function(...) { + # {}; + # } + # + # Note that naively checking for the preceding "{" will also match + # braces inside multi-dimensional arrays, but this is fine since + # that expression will not contain semicolons. + # + # 6. Block following another block: + # while (true) {} + # {}; + # + # 7. End of namespaces: + # namespace {}; + # + # These semicolons seems far more common than other kinds of + # redundant semicolons, possibly due to people converting classes + # to namespaces. For now we do not warn for this case. + # + # Try matching case 1 first. + match = Match(r'^(.*\)\s*)\{', line) + if match: + # Matched closing parenthesis (case 1). Check the token before the + # matching opening parenthesis, and don't warn if it looks like a + # macro. This avoids these false positives: + # - macro that defines a base class + # - multi-line macro that defines a base class + # - macro that defines the whole class-head + # + # But we still issue warnings for macros that we know are safe to + # warn, specifically: + # - TEST, TEST_F, TEST_P, MATCHER, MATCHER_P + # - TYPED_TEST + # - INTERFACE_DEF + # - EXCLUSIVE_LOCKS_REQUIRED, SHARED_LOCKS_REQUIRED, LOCKS_EXCLUDED: + # + # We implement a whitelist of safe macros instead of a blacklist of + # unsafe macros, even though the latter appears less frequently in + # google code and would have been easier to implement. This is because + # the downside for getting the whitelist wrong means some extra + # semicolons, while the downside for getting the blacklist wrong + # would result in compile errors. + # + # In addition to macros, we also don't want to warn on + # - Compound literals + # - Lambdas + # - alignas specifier with anonymous structs: + closing_brace_pos = match.group(1).rfind(')') + opening_parenthesis = ReverseCloseExpression( + clean_lines, linenum, closing_brace_pos) + if opening_parenthesis[2] > -1: + line_prefix = opening_parenthesis[0][0:opening_parenthesis[2]] + macro = Search(r'\b([A-Z_]+)\s*$', line_prefix) + func = Match(r'^(.*\])\s*$', line_prefix) + if ((macro and + macro.group(1) not in ( + 'TEST', 'TEST_F', 'MATCHER', 'MATCHER_P', 'TYPED_TEST', + 'EXCLUSIVE_LOCKS_REQUIRED', 'SHARED_LOCKS_REQUIRED', + 'LOCKS_EXCLUDED', 'INTERFACE_DEF')) or + (func and not Search(r'\boperator\s*\[\s*\]', func.group(1))) or + Search(r'\b(?:struct|union)\s+alignas\s*$', line_prefix) or + Search(r'\s+=\s*$', line_prefix)): + match = None + if (match and + opening_parenthesis[1] > 1 and + Search(r'\]\s*$', clean_lines.elided[opening_parenthesis[1] - 1])): + # Multi-line lambda-expression + match = None + + else: + # Try matching cases 2-3. + match = Match(r'^(.*(?:else|\)\s*const)\s*)\{', line) + if not match: + # Try matching cases 4-6. These are always matched on separate lines. + # + # Note that we can't simply concatenate the previous line to the + # current line and do a single match, otherwise we may output + # duplicate warnings for the blank line case: + # if (cond) { + # // blank line + # } + prevline = GetPreviousNonBlankLine(clean_lines, linenum)[0] + if prevline and Search(r'[;{}]\s*$', prevline): + match = Match(r'^(\s*)\{', line) + + # Check matching closing brace + if match: + (endline, endlinenum, endpos) = CloseExpression( + clean_lines, linenum, len(match.group(1))) + if endpos > -1 and Match(r'^\s*;', endline[endpos:]): + # Current {} pair is eligible for semicolon check, and we have found + # the redundant semicolon, output warning here. + # + # Note: because we are scanning forward for opening braces, and + # outputting warnings for the matching closing brace, if there are + # nested blocks with trailing semicolons, we will get the error + # messages in reversed order. + error(filename, endlinenum, 'readability/braces', 4, + "You don't need a ; after a }") + + +def CheckEmptyBlockBody(filename, clean_lines, linenum, error): + """Look for empty loop/conditional body with only a single semicolon. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + + # Search for loop keywords at the beginning of the line. Because only + # whitespaces are allowed before the keywords, this will also ignore most + # do-while-loops, since those lines should start with closing brace. + # + # We also check "if" blocks here, since an empty conditional block + # is likely an error. + line = clean_lines.elided[linenum] + matched = Match(r'\s*(for|while|if)\s*\(', line) + if matched: + # Find the end of the conditional expression + (end_line, end_linenum, end_pos) = CloseExpression( + clean_lines, linenum, line.find('(')) + + # Output warning if what follows the condition expression is a semicolon. + # No warning for all other cases, including whitespace or newline, since we + # have a separate check for semicolons preceded by whitespace. + if end_pos >= 0 and Match(r';', end_line[end_pos:]): + if matched.group(1) == 'if': + error(filename, end_linenum, 'whitespace/empty_conditional_body', 5, + 'Empty conditional bodies should use {}') + else: + error(filename, end_linenum, 'whitespace/empty_loop_body', 5, + 'Empty loop bodies should use {} or continue') + + +def FindCheckMacro(line): + """Find a replaceable CHECK-like macro. + + Args: + line: line to search on. + Returns: + (macro name, start position), or (None, -1) if no replaceable + macro is found. + """ + for macro in _CHECK_MACROS: + i = line.find(macro) + if i >= 0: + # Find opening parenthesis. Do a regular expression match here + # to make sure that we are matching the expected CHECK macro, as + # opposed to some other macro that happens to contain the CHECK + # substring. + matched = Match(r'^(.*\b' + macro + r'\s*)\(', line) + if not matched: + continue + return (macro, len(matched.group(1))) + return (None, -1) + + +def CheckCheck(filename, clean_lines, linenum, error): + """Checks the use of CHECK and EXPECT macros. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + + # Decide the set of replacement macros that should be suggested + lines = clean_lines.elided + (check_macro, start_pos) = FindCheckMacro(lines[linenum]) + if not check_macro: + return + + # Find end of the boolean expression by matching parentheses + (last_line, end_line, end_pos) = CloseExpression( + clean_lines, linenum, start_pos) + if end_pos < 0: + return + + # If the check macro is followed by something other than a + # semicolon, assume users will log their own custom error messages + # and don't suggest any replacements. + if not Match(r'\s*;', last_line[end_pos:]): + return + + if linenum == end_line: + expression = lines[linenum][start_pos + 1:end_pos - 1] + else: + expression = lines[linenum][start_pos + 1:] + for i in xrange(linenum + 1, end_line): + expression += lines[i] + expression += last_line[0:end_pos - 1] + + # Parse expression so that we can take parentheses into account. + # This avoids false positives for inputs like "CHECK((a < 4) == b)", + # which is not replaceable by CHECK_LE. + lhs = '' + rhs = '' + operator = None + while expression: + matched = Match(r'^\s*(<<|<<=|>>|>>=|->\*|->|&&|\|\||' + r'==|!=|>=|>|<=|<|\()(.*)$', expression) + if matched: + token = matched.group(1) + if token == '(': + # Parenthesized operand + expression = matched.group(2) + (end, _) = FindEndOfExpressionInLine(expression, 0, ['(']) + if end < 0: + return # Unmatched parenthesis + lhs += '(' + expression[0:end] + expression = expression[end:] + elif token in ('&&', '||'): + # Logical and/or operators. This means the expression + # contains more than one term, for example: + # CHECK(42 < a && a < b); + # + # These are not replaceable with CHECK_LE, so bail out early. + return + elif token in ('<<', '<<=', '>>', '>>=', '->*', '->'): + # Non-relational operator + lhs += token + expression = matched.group(2) + else: + # Relational operator + operator = token + rhs = matched.group(2) + break + else: + # Unparenthesized operand. Instead of appending to lhs one character + # at a time, we do another regular expression match to consume several + # characters at once if possible. Trivial benchmark shows that this + # is more efficient when the operands are longer than a single + # character, which is generally the case. + matched = Match(r'^([^-=!<>()&|]+)(.*)$', expression) + if not matched: + matched = Match(r'^(\s*\S)(.*)$', expression) + if not matched: + break + lhs += matched.group(1) + expression = matched.group(2) + + # Only apply checks if we got all parts of the boolean expression + if not (lhs and operator and rhs): + return + + # Check that rhs do not contain logical operators. We already know + # that lhs is fine since the loop above parses out && and ||. + if rhs.find('&&') > -1 or rhs.find('||') > -1: + return + + # At least one of the operands must be a constant literal. This is + # to avoid suggesting replacements for unprintable things like + # CHECK(variable != iterator) + # + # The following pattern matches decimal, hex integers, strings, and + # characters (in that order). + lhs = lhs.strip() + rhs = rhs.strip() + match_constant = r'^([-+]?(\d+|0[xX][0-9a-fA-F]+)[lLuU]{0,3}|".*"|\'.*\')$' + if Match(match_constant, lhs) or Match(match_constant, rhs): + # Note: since we know both lhs and rhs, we can provide a more + # descriptive error message like: + # Consider using CHECK_EQ(x, 42) instead of CHECK(x == 42) + # Instead of: + # Consider using CHECK_EQ instead of CHECK(a == b) + # + # We are still keeping the less descriptive message because if lhs + # or rhs gets long, the error message might become unreadable. + error(filename, linenum, 'readability/check', 2, + 'Consider using %s instead of %s(a %s b)' % ( + _CHECK_REPLACEMENT[check_macro][operator], + check_macro, operator)) + + +def CheckAltTokens(filename, clean_lines, linenum, error): + """Check alternative keywords being used in boolean expressions. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Avoid preprocessor lines + if Match(r'^\s*#', line): + return + + # Last ditch effort to avoid multi-line comments. This will not help + # if the comment started before the current line or ended after the + # current line, but it catches most of the false positives. At least, + # it provides a way to workaround this warning for people who use + # multi-line comments in preprocessor macros. + # + # TODO(unknown): remove this once cpplint has better support for + # multi-line comments. + if line.find('/*') >= 0 or line.find('*/') >= 0: + return + + for match in _ALT_TOKEN_REPLACEMENT_PATTERN.finditer(line): + error(filename, linenum, 'readability/alt_tokens', 2, + 'Use operator %s instead of %s' % ( + _ALT_TOKEN_REPLACEMENT[match.group(1)], match.group(1))) + + +def GetLineWidth(line): + """Determines the width of the line in column positions. + + Args: + line: A string, which may be a Unicode string. + + Returns: + The width of the line in column positions, accounting for Unicode + combining characters and wide characters. + """ + if isinstance(line, unicode): + width = 0 + for uc in unicodedata.normalize('NFC', line): + if unicodedata.east_asian_width(uc) in ('W', 'F'): + width += 2 + elif not unicodedata.combining(uc): + width += 1 + return width + else: + return len(line) + + +def CheckStyle(filename, clean_lines, linenum, file_extension, nesting_state, + error): + """Checks rules from the 'C++ style rules' section of cppguide.html. + + Most of these rules are hard to test (naming, comment style), but we + do what we can. In particular we check for 2-space indents, line lengths, + tab usage, spaces inside code, etc. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + file_extension: The extension (without the dot) of the filename. + nesting_state: A NestingState instance which maintains information about + the current stack of nested blocks being parsed. + error: The function to call with any errors found. + """ + + # Don't use "elided" lines here, otherwise we can't check commented lines. + # Don't want to use "raw" either, because we don't want to check inside C++11 + # raw strings, + raw_lines = clean_lines.lines_without_raw_strings + line = raw_lines[linenum] + + if line.find('\t') != -1: + error(filename, linenum, 'whitespace/tab', 1, + 'Tab found; better to use spaces') + + # One or three blank spaces at the beginning of the line is weird; it's + # hard to reconcile that with 2-space indents. + # NOTE: here are the conditions rob pike used for his tests. Mine aren't + # as sophisticated, but it may be worth becoming so: RLENGTH==initial_spaces + # if(RLENGTH > 20) complain = 0; + # if(match($0, " +(error|private|public|protected):")) complain = 0; + # if(match(prev, "&& *$")) complain = 0; + # if(match(prev, "\\|\\| *$")) complain = 0; + # if(match(prev, "[\",=><] *$")) complain = 0; + # if(match($0, " <<")) complain = 0; + # if(match(prev, " +for \\(")) complain = 0; + # if(prevodd && match(prevprev, " +for \\(")) complain = 0; + scope_or_label_pattern = r'\s*\w+\s*:\s*\\?$' + classinfo = nesting_state.InnermostClass() + initial_spaces = 0 + cleansed_line = clean_lines.elided[linenum] + while initial_spaces < len(line) and line[initial_spaces] == ' ': + initial_spaces += 1 + if line and line[-1].isspace(): + error(filename, linenum, 'whitespace/end_of_line', 4, + 'Line ends in whitespace. Consider deleting these extra spaces.') + # There are certain situations we allow one space, notably for + # section labels, and also lines containing multi-line raw strings. + elif ((initial_spaces == 1 or initial_spaces == 3) and + not Match(scope_or_label_pattern, cleansed_line) and + not (clean_lines.raw_lines[linenum] != line and + Match(r'^\s*""', line))): + error(filename, linenum, 'whitespace/indent', 3, + 'Weird number of spaces at line-start. ' + 'Are you using a 2-space indent?') + + # Check if the line is a header guard. + is_header_guard = False + if file_extension == 'h': + cppvar = GetHeaderGuardCPPVariable(filename) + if (line.startswith('#ifndef %s' % cppvar) or + line.startswith('#define %s' % cppvar) or + line.startswith('#endif // %s' % cppvar)): + is_header_guard = True + # #include lines and header guards can be long, since there's no clean way to + # split them. + # + # URLs can be long too. It's possible to split these, but it makes them + # harder to cut&paste. + # + # The "$Id:...$" comment may also get very long without it being the + # developers fault. + if (not line.startswith('#include') and not is_header_guard and + not Match(r'^\s*//.*http(s?)://\S*$', line) and + not Match(r'^// \$Id:.*#[0-9]+ \$$', line)): + line_width = GetLineWidth(line) + extended_length = int((_line_length * 1.25)) + if line_width > extended_length: + error(filename, linenum, 'whitespace/line_length', 4, + 'Lines should very rarely be longer than %i characters' % + extended_length) + elif line_width > _line_length: + error(filename, linenum, 'whitespace/line_length', 2, + 'Lines should be <= %i characters long' % _line_length) + + if (cleansed_line.count(';') > 1 and + # for loops are allowed two ;'s (and may run over two lines). + cleansed_line.find('for') == -1 and + (GetPreviousNonBlankLine(clean_lines, linenum)[0].find('for') == -1 or + GetPreviousNonBlankLine(clean_lines, linenum)[0].find(';') != -1) and + # It's ok to have many commands in a switch case that fits in 1 line + not ((cleansed_line.find('case ') != -1 or + cleansed_line.find('default:') != -1) and + cleansed_line.find('break;') != -1)): + error(filename, linenum, 'whitespace/newline', 0, + 'More than one command on the same line') + + # Some more style checks + CheckBraces(filename, clean_lines, linenum, error) + CheckTrailingSemicolon(filename, clean_lines, linenum, error) + CheckEmptyBlockBody(filename, clean_lines, linenum, error) + CheckAccess(filename, clean_lines, linenum, nesting_state, error) + CheckSpacing(filename, clean_lines, linenum, nesting_state, error) + CheckOperatorSpacing(filename, clean_lines, linenum, error) + CheckParenthesisSpacing(filename, clean_lines, linenum, error) + CheckCommaSpacing(filename, clean_lines, linenum, error) + CheckBracesSpacing(filename, clean_lines, linenum, error) + CheckSpacingForFunctionCall(filename, clean_lines, linenum, error) + CheckRValueReference(filename, clean_lines, linenum, nesting_state, error) + CheckCheck(filename, clean_lines, linenum, error) + CheckAltTokens(filename, clean_lines, linenum, error) + classinfo = nesting_state.InnermostClass() + if classinfo: + CheckSectionSpacing(filename, clean_lines, classinfo, linenum, error) + + +_RE_PATTERN_INCLUDE = re.compile(r'^\s*#\s*include\s*([<"])([^>"]*)[>"].*$') +# Matches the first component of a filename delimited by -s and _s. That is: +# _RE_FIRST_COMPONENT.match('foo').group(0) == 'foo' +# _RE_FIRST_COMPONENT.match('foo.cc').group(0) == 'foo' +# _RE_FIRST_COMPONENT.match('foo-bar_baz.cc').group(0) == 'foo' +# _RE_FIRST_COMPONENT.match('foo_bar-baz.cc').group(0) == 'foo' +_RE_FIRST_COMPONENT = re.compile(r'^[^-_.]+') + + +def _DropCommonSuffixes(filename): + """Drops common suffixes like _test.cc or -inl.h from filename. + + For example: + >>> _DropCommonSuffixes('foo/foo-inl.h') + 'foo/foo' + >>> _DropCommonSuffixes('foo/bar/foo.cc') + 'foo/bar/foo' + >>> _DropCommonSuffixes('foo/foo_internal.h') + 'foo/foo' + >>> _DropCommonSuffixes('foo/foo_unusualinternal.h') + 'foo/foo_unusualinternal' + + Args: + filename: The input filename. + + Returns: + The filename with the common suffix removed. + """ + for suffix in ('test.cc', 'regtest.cc', 'unittest.cc', + 'inl.h', 'impl.h', 'internal.h'): + if (filename.endswith(suffix) and len(filename) > len(suffix) and + filename[-len(suffix) - 1] in ('-', '_')): + return filename[:-len(suffix) - 1] + return os.path.splitext(filename)[0] + + +def _IsTestFilename(filename): + """Determines if the given filename has a suffix that identifies it as a test. + + Args: + filename: The input filename. + + Returns: + True if 'filename' looks like a test, False otherwise. + """ + if (filename.endswith('_test.cc') or + filename.endswith('_unittest.cc') or + filename.endswith('_regtest.cc')): + return True + else: + return False + + +def _ClassifyInclude(fileinfo, include, is_system): + """Figures out what kind of header 'include' is. + + Args: + fileinfo: The current file cpplint is running over. A FileInfo instance. + include: The path to a #included file. + is_system: True if the #include used <> rather than "". + + Returns: + One of the _XXX_HEADER constants. + + For example: + >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'stdio.h', True) + _C_SYS_HEADER + >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'string', True) + _CPP_SYS_HEADER + >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/foo.h', False) + _LIKELY_MY_HEADER + >>> _ClassifyInclude(FileInfo('foo/foo_unknown_extension.cc'), + ... 'bar/foo_other_ext.h', False) + _POSSIBLE_MY_HEADER + >>> _ClassifyInclude(FileInfo('foo/foo.cc'), 'foo/bar.h', False) + _OTHER_HEADER + """ + # This is a list of all standard c++ header files, except + # those already checked for above. + is_cpp_h = include in _CPP_HEADERS + + if is_system: + if is_cpp_h: + return _CPP_SYS_HEADER + else: + return _C_SYS_HEADER + + # If the target file and the include we're checking share a + # basename when we drop common extensions, and the include + # lives in . , then it's likely to be owned by the target file. + target_dir, target_base = ( + os.path.split(_DropCommonSuffixes(fileinfo.RepositoryName()))) + include_dir, include_base = os.path.split(_DropCommonSuffixes(include)) + if target_base == include_base and ( + include_dir == target_dir or + include_dir == os.path.normpath(target_dir + '/../public')): + return _LIKELY_MY_HEADER + + # If the target and include share some initial basename + # component, it's possible the target is implementing the + # include, so it's allowed to be first, but we'll never + # complain if it's not there. + target_first_component = _RE_FIRST_COMPONENT.match(target_base) + include_first_component = _RE_FIRST_COMPONENT.match(include_base) + if (target_first_component and include_first_component and + target_first_component.group(0) == + include_first_component.group(0)): + return _POSSIBLE_MY_HEADER + + return _OTHER_HEADER + + + +def CheckIncludeLine(filename, clean_lines, linenum, include_state, error): + """Check rules that are applicable to #include lines. + + Strings on #include lines are NOT removed from elided line, to make + certain tasks easier. However, to prevent false positives, checks + applicable to #include lines in CheckLanguage must be put here. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + include_state: An _IncludeState instance in which the headers are inserted. + error: The function to call with any errors found. + """ + fileinfo = FileInfo(filename) + line = clean_lines.lines[linenum] + + # "include" should use the new style "foo/bar.h" instead of just "bar.h" + # Only do this check if the included header follows google naming + # conventions. If not, assume that it's a 3rd party API that + # requires special include conventions. + # + # We also make an exception for Lua headers, which follow google + # naming convention but not the include convention. + match = Match(r'#include\s*"([^/]+\.h)"', line) + if match and not _THIRD_PARTY_HEADERS_PATTERN.match(match.group(1)): + error(filename, linenum, 'build/include', 4, + 'Include the directory when naming .h files') + + # we shouldn't include a file more than once. actually, there are a + # handful of instances where doing so is okay, but in general it's + # not. + match = _RE_PATTERN_INCLUDE.search(line) + if match: + include = match.group(2) + is_system = (match.group(1) == '<') + duplicate_line = include_state.FindHeader(include) + if duplicate_line >= 0: + error(filename, linenum, 'build/include', 4, + '"%s" already included at %s:%s' % + (include, filename, duplicate_line)) + elif (include.endswith('.cc') and + os.path.dirname(fileinfo.RepositoryName()) != os.path.dirname(include)): + error(filename, linenum, 'build/include', 4, + 'Do not include .cc files from other packages') + elif not _THIRD_PARTY_HEADERS_PATTERN.match(include): + include_state.include_list[-1].append((include, linenum)) + + # We want to ensure that headers appear in the right order: + # 1) for foo.cc, foo.h (preferred location) + # 2) c system files + # 3) cpp system files + # 4) for foo.cc, foo.h (deprecated location) + # 5) other google headers + # + # We classify each include statement as one of those 5 types + # using a number of techniques. The include_state object keeps + # track of the highest type seen, and complains if we see a + # lower type after that. + error_message = include_state.CheckNextIncludeOrder( + _ClassifyInclude(fileinfo, include, is_system)) + if error_message: + error(filename, linenum, 'build/include_order', 4, + '%s. Should be: %s.h, c system, c++ system, other.' % + (error_message, fileinfo.BaseName())) + canonical_include = include_state.CanonicalizeAlphabeticalOrder(include) + if not include_state.IsInAlphabeticalOrder( + clean_lines, linenum, canonical_include): + error(filename, linenum, 'build/include_alpha', 4, + 'Include "%s" not in alphabetical order' % include) + include_state.SetLastHeader(canonical_include) + + + +def _GetTextInside(text, start_pattern): + r"""Retrieves all the text between matching open and close parentheses. + + Given a string of lines and a regular expression string, retrieve all the text + following the expression and between opening punctuation symbols like + (, [, or {, and the matching close-punctuation symbol. This properly nested + occurrences of the punctuations, so for the text like + printf(a(), b(c())); + a call to _GetTextInside(text, r'printf\(') will return 'a(), b(c())'. + start_pattern must match string having an open punctuation symbol at the end. + + Args: + text: The lines to extract text. Its comments and strings must be elided. + It can be single line and can span multiple lines. + start_pattern: The regexp string indicating where to start extracting + the text. + Returns: + The extracted text. + None if either the opening string or ending punctuation could not be found. + """ + # TODO(unknown): Audit cpplint.py to see what places could be profitably + # rewritten to use _GetTextInside (and use inferior regexp matching today). + + # Give opening punctuations to get the matching close-punctuations. + matching_punctuation = {'(': ')', '{': '}', '[': ']'} + closing_punctuation = set(matching_punctuation.itervalues()) + + # Find the position to start extracting text. + match = re.search(start_pattern, text, re.M) + if not match: # start_pattern not found in text. + return None + start_position = match.end(0) + + assert start_position > 0, ( + 'start_pattern must ends with an opening punctuation.') + assert text[start_position - 1] in matching_punctuation, ( + 'start_pattern must ends with an opening punctuation.') + # Stack of closing punctuations we expect to have in text after position. + punctuation_stack = [matching_punctuation[text[start_position - 1]]] + position = start_position + while punctuation_stack and position < len(text): + if text[position] == punctuation_stack[-1]: + punctuation_stack.pop() + elif text[position] in closing_punctuation: + # A closing punctuation without matching opening punctuations. + return None + elif text[position] in matching_punctuation: + punctuation_stack.append(matching_punctuation[text[position]]) + position += 1 + if punctuation_stack: + # Opening punctuations left without matching close-punctuations. + return None + # punctuations match. + return text[start_position:position - 1] + + +# Patterns for matching call-by-reference parameters. +# +# Supports nested templates up to 2 levels deep using this messy pattern: +# < (?: < (?: < [^<>]* +# > +# | [^<>] )* +# > +# | [^<>] )* +# > +_RE_PATTERN_IDENT = r'[_a-zA-Z]\w*' # =~ [[:alpha:]][[:alnum:]]* +_RE_PATTERN_TYPE = ( + r'(?:const\s+)?(?:typename\s+|class\s+|struct\s+|union\s+|enum\s+)?' + r'(?:\w|' + r'\s*<(?:<(?:<[^<>]*>|[^<>])*>|[^<>])*>|' + r'::)+') +# A call-by-reference parameter ends with '& identifier'. +_RE_PATTERN_REF_PARAM = re.compile( + r'(' + _RE_PATTERN_TYPE + r'(?:\s*(?:\bconst\b|[*]))*\s*' + r'&\s*' + _RE_PATTERN_IDENT + r')\s*(?:=[^,()]+)?[,)]') +# A call-by-const-reference parameter either ends with 'const& identifier' +# or looks like 'const type& identifier' when 'type' is atomic. +_RE_PATTERN_CONST_REF_PARAM = ( + r'(?:.*\s*\bconst\s*&\s*' + _RE_PATTERN_IDENT + + r'|const\s+' + _RE_PATTERN_TYPE + r'\s*&\s*' + _RE_PATTERN_IDENT + r')') + + +def CheckLanguage(filename, clean_lines, linenum, file_extension, + include_state, nesting_state, error): + """Checks rules from the 'C++ language rules' section of cppguide.html. + + Some of these rules are hard to test (function overloading, using + uint32 inappropriately), but we do the best we can. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + file_extension: The extension (without the dot) of the filename. + include_state: An _IncludeState instance in which the headers are inserted. + nesting_state: A NestingState instance which maintains information about + the current stack of nested blocks being parsed. + error: The function to call with any errors found. + """ + # If the line is empty or consists of entirely a comment, no need to + # check it. + line = clean_lines.elided[linenum] + if not line: + return + + match = _RE_PATTERN_INCLUDE.search(line) + if match: + CheckIncludeLine(filename, clean_lines, linenum, include_state, error) + return + + # Reset include state across preprocessor directives. This is meant + # to silence warnings for conditional includes. + match = Match(r'^\s*#\s*(if|ifdef|ifndef|elif|else|endif)\b', line) + if match: + include_state.ResetSection(match.group(1)) + + # Make Windows paths like Unix. + fullname = os.path.abspath(filename).replace('\\', '/') + + # Perform other checks now that we are sure that this is not an include line + CheckCasts(filename, clean_lines, linenum, error) + CheckGlobalStatic(filename, clean_lines, linenum, error) + CheckPrintf(filename, clean_lines, linenum, error) + + if file_extension == 'h': + # TODO(unknown): check that 1-arg constructors are explicit. + # How to tell it's a constructor? + # (handled in CheckForNonStandardConstructs for now) + # TODO(unknown): check that classes declare or disable copy/assign + # (level 1 error) + pass + + # Check if people are using the verboten C basic types. The only exception + # we regularly allow is "unsigned short port" for port. + if Search(r'\bshort port\b', line): + if not Search(r'\bunsigned short port\b', line): + error(filename, linenum, 'runtime/int', 4, + 'Use "unsigned short" for ports, not "short"') + else: + match = Search(r'\b(short|long(?! +double)|long long)\b', line) + if match: + error(filename, linenum, 'runtime/int', 4, + 'Use int16/int64/etc, rather than the C type %s' % match.group(1)) + + # Check if some verboten operator overloading is going on + # TODO(unknown): catch out-of-line unary operator&: + # class X {}; + # int operator&(const X& x) { return 42; } // unary operator& + # The trick is it's hard to tell apart from binary operator&: + # class Y { int operator&(const Y& x) { return 23; } }; // binary operator& + if Search(r'\boperator\s*&\s*\(\s*\)', line): + error(filename, linenum, 'runtime/operator', 4, + 'Unary operator& is dangerous. Do not use it.') + + # Check for suspicious usage of "if" like + # } if (a == b) { + if Search(r'\}\s*if\s*\(', line): + error(filename, linenum, 'readability/braces', 4, + 'Did you mean "else if"? If not, start a new line for "if".') + + # Check for potential format string bugs like printf(foo). + # We constrain the pattern not to pick things like DocidForPrintf(foo). + # Not perfect but it can catch printf(foo.c_str()) and printf(foo->c_str()) + # TODO(unknown): Catch the following case. Need to change the calling + # convention of the whole function to process multiple line to handle it. + # printf( + # boy_this_is_a_really_long_variable_that_cannot_fit_on_the_prev_line); + printf_args = _GetTextInside(line, r'(?i)\b(string)?printf\s*\(') + if printf_args: + match = Match(r'([\w.\->()]+)$', printf_args) + if match and match.group(1) != '__VA_ARGS__': + function_name = re.search(r'\b((?:string)?printf)\s*\(', + line, re.I).group(1) + error(filename, linenum, 'runtime/printf', 4, + 'Potential format string bug. Do %s("%%s", %s) instead.' + % (function_name, match.group(1))) + + # Check for potential memset bugs like memset(buf, sizeof(buf), 0). + match = Search(r'memset\s*\(([^,]*),\s*([^,]*),\s*0\s*\)', line) + if match and not Match(r"^''|-?[0-9]+|0x[0-9A-Fa-f]$", match.group(2)): + error(filename, linenum, 'runtime/memset', 4, + 'Did you mean "memset(%s, 0, %s)"?' + % (match.group(1), match.group(2))) + + if Search(r'\busing namespace\b', line): + error(filename, linenum, 'build/namespaces', 5, + 'Do not use namespace using-directives. ' + 'Use using-declarations instead.') + + # Detect variable-length arrays. + match = Match(r'\s*(.+::)?(\w+) [a-z]\w*\[(.+)];', line) + if (match and match.group(2) != 'return' and match.group(2) != 'delete' and + match.group(3).find(']') == -1): + # Split the size using space and arithmetic operators as delimiters. + # If any of the resulting tokens are not compile time constants then + # report the error. + tokens = re.split(r'\s|\+|\-|\*|\/|<<|>>]', match.group(3)) + is_const = True + skip_next = False + for tok in tokens: + if skip_next: + skip_next = False + continue + + if Search(r'sizeof\(.+\)', tok): continue + if Search(r'arraysize\(\w+\)', tok): continue + + tok = tok.lstrip('(') + tok = tok.rstrip(')') + if not tok: continue + if Match(r'\d+', tok): continue + if Match(r'0[xX][0-9a-fA-F]+', tok): continue + if Match(r'k[A-Z0-9]\w*', tok): continue + if Match(r'(.+::)?k[A-Z0-9]\w*', tok): continue + if Match(r'(.+::)?[A-Z][A-Z0-9_]*', tok): continue + # A catch all for tricky sizeof cases, including 'sizeof expression', + # 'sizeof(*type)', 'sizeof(const type)', 'sizeof(struct StructName)' + # requires skipping the next token because we split on ' ' and '*'. + if tok.startswith('sizeof'): + skip_next = True + continue + is_const = False + break + if not is_const: + error(filename, linenum, 'runtime/arrays', 1, + 'Do not use variable-length arrays. Use an appropriately named ' + "('k' followed by CamelCase) compile-time constant for the size.") + + # Check for use of unnamed namespaces in header files. Registration + # macros are typically OK, so we allow use of "namespace {" on lines + # that end with backslashes. + if (file_extension == 'h' + and Search(r'\bnamespace\s*{', line) + and line[-1] != '\\'): + error(filename, linenum, 'build/namespaces', 4, + 'Do not use unnamed namespaces in header files. See ' + 'http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Namespaces' + ' for more information.') + + +def CheckGlobalStatic(filename, clean_lines, linenum, error): + """Check for unsafe global or static objects. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Match two lines at a time to support multiline declarations + if linenum + 1 < clean_lines.NumLines() and not Search(r'[;({]', line): + line += clean_lines.elided[linenum + 1].strip() + + # Check for people declaring static/global STL strings at the top level. + # This is dangerous because the C++ language does not guarantee that + # globals with constructors are initialized before the first access. + match = Match( + r'((?:|static +)(?:|const +))string +([a-zA-Z0-9_:]+)\b(.*)', + line) + + # Remove false positives: + # - String pointers (as opposed to values). + # string *pointer + # const string *pointer + # string const *pointer + # string *const pointer + # + # - Functions and template specializations. + # string Function(... + # string Class::Method(... + # + # - Operators. These are matched separately because operator names + # cross non-word boundaries, and trying to match both operators + # and functions at the same time would decrease accuracy of + # matching identifiers. + # string Class::operator*() + if (match and + not Search(r'\bstring\b(\s+const)?\s*\*\s*(const\s+)?\w', line) and + not Search(r'\boperator\W', line) and + not Match(r'\s*(<.*>)?(::[a-zA-Z0-9_]+)*\s*\(([^"]|$)', match.group(3))): + error(filename, linenum, 'runtime/string', 4, + 'For a static/global string constant, use a C style string instead: ' + '"%schar %s[]".' % + (match.group(1), match.group(2))) + + if Search(r'\b([A-Za-z0-9_]*_)\(\1\)', line): + error(filename, linenum, 'runtime/init', 4, + 'You seem to be initializing a member variable with itself.') + + +def CheckPrintf(filename, clean_lines, linenum, error): + """Check for printf related issues. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # When snprintf is used, the second argument shouldn't be a literal. + match = Search(r'snprintf\s*\(([^,]*),\s*([0-9]*)\s*,', line) + if match and match.group(2) != '0': + # If 2nd arg is zero, snprintf is used to calculate size. + error(filename, linenum, 'runtime/printf', 3, + 'If you can, use sizeof(%s) instead of %s as the 2nd arg ' + 'to snprintf.' % (match.group(1), match.group(2))) + + # Check if some verboten C functions are being used. + if Search(r'\bsprintf\s*\(', line): + error(filename, linenum, 'runtime/printf', 5, + 'Never use sprintf. Use snprintf instead.') + match = Search(r'\b(strcpy|strcat)\s*\(', line) + if match: + error(filename, linenum, 'runtime/printf', 4, + 'Almost always, snprintf is better than %s' % match.group(1)) + + +def IsDerivedFunction(clean_lines, linenum): + """Check if current line contains an inherited function. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + Returns: + True if current line contains a function with "override" + virt-specifier. + """ + # Scan back a few lines for start of current function + for i in xrange(linenum, max(-1, linenum - 10), -1): + match = Match(r'^([^()]*\w+)\(', clean_lines.elided[i]) + if match: + # Look for "override" after the matching closing parenthesis + line, _, closing_paren = CloseExpression( + clean_lines, i, len(match.group(1))) + return (closing_paren >= 0 and + Search(r'\boverride\b', line[closing_paren:])) + return False + + +def IsOutOfLineMethodDefinition(clean_lines, linenum): + """Check if current line contains an out-of-line method definition. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + Returns: + True if current line contains an out-of-line method definition. + """ + # Scan back a few lines for start of current function + for i in xrange(linenum, max(-1, linenum - 10), -1): + if Match(r'^([^()]*\w+)\(', clean_lines.elided[i]): + return Match(r'^[^()]*\w+::\w+\(', clean_lines.elided[i]) is not None + return False + + +def IsInitializerList(clean_lines, linenum): + """Check if current line is inside constructor initializer list. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + Returns: + True if current line appears to be inside constructor initializer + list, False otherwise. + """ + for i in xrange(linenum, 1, -1): + line = clean_lines.elided[i] + if i == linenum: + remove_function_body = Match(r'^(.*)\{\s*$', line) + if remove_function_body: + line = remove_function_body.group(1) + + if Search(r'\s:\s*\w+[({]', line): + # A lone colon tend to indicate the start of a constructor + # initializer list. It could also be a ternary operator, which + # also tend to appear in constructor initializer lists as + # opposed to parameter lists. + return True + if Search(r'\}\s*,\s*$', line): + # A closing brace followed by a comma is probably the end of a + # brace-initialized member in constructor initializer list. + return True + if Search(r'[{};]\s*$', line): + # Found one of the following: + # - A closing brace or semicolon, probably the end of the previous + # function. + # - An opening brace, probably the start of current class or namespace. + # + # Current line is probably not inside an initializer list since + # we saw one of those things without seeing the starting colon. + return False + + # Got to the beginning of the file without seeing the start of + # constructor initializer list. + return False + + +def CheckForNonConstReference(filename, clean_lines, linenum, + nesting_state, error): + """Check for non-const references. + + Separate from CheckLanguage since it scans backwards from current + line, instead of scanning forward. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + nesting_state: A NestingState instance which maintains information about + the current stack of nested blocks being parsed. + error: The function to call with any errors found. + """ + # Do nothing if there is no '&' on current line. + line = clean_lines.elided[linenum] + if '&' not in line: + return + + # If a function is inherited, current function doesn't have much of + # a choice, so any non-const references should not be blamed on + # derived function. + if IsDerivedFunction(clean_lines, linenum): + return + + # Don't warn on out-of-line method definitions, as we would warn on the + # in-line declaration, if it isn't marked with 'override'. + if IsOutOfLineMethodDefinition(clean_lines, linenum): + return + + # Long type names may be broken across multiple lines, usually in one + # of these forms: + # LongType + # ::LongTypeContinued &identifier + # LongType:: + # LongTypeContinued &identifier + # LongType< + # ...>::LongTypeContinued &identifier + # + # If we detected a type split across two lines, join the previous + # line to current line so that we can match const references + # accordingly. + # + # Note that this only scans back one line, since scanning back + # arbitrary number of lines would be expensive. If you have a type + # that spans more than 2 lines, please use a typedef. + if linenum > 1: + previous = None + if Match(r'\s*::(?:[\w<>]|::)+\s*&\s*\S', line): + # previous_line\n + ::current_line + previous = Search(r'\b((?:const\s*)?(?:[\w<>]|::)+[\w<>])\s*$', + clean_lines.elided[linenum - 1]) + elif Match(r'\s*[a-zA-Z_]([\w<>]|::)+\s*&\s*\S', line): + # previous_line::\n + current_line + previous = Search(r'\b((?:const\s*)?(?:[\w<>]|::)+::)\s*$', + clean_lines.elided[linenum - 1]) + if previous: + line = previous.group(1) + line.lstrip() + else: + # Check for templated parameter that is split across multiple lines + endpos = line.rfind('>') + if endpos > -1: + (_, startline, startpos) = ReverseCloseExpression( + clean_lines, linenum, endpos) + if startpos > -1 and startline < linenum: + # Found the matching < on an earlier line, collect all + # pieces up to current line. + line = '' + for i in xrange(startline, linenum + 1): + line += clean_lines.elided[i].strip() + + # Check for non-const references in function parameters. A single '&' may + # found in the following places: + # inside expression: binary & for bitwise AND + # inside expression: unary & for taking the address of something + # inside declarators: reference parameter + # We will exclude the first two cases by checking that we are not inside a + # function body, including one that was just introduced by a trailing '{'. + # TODO(unknown): Doesn't account for 'catch(Exception& e)' [rare]. + if (nesting_state.previous_stack_top and + not (isinstance(nesting_state.previous_stack_top, _ClassInfo) or + isinstance(nesting_state.previous_stack_top, _NamespaceInfo))): + # Not at toplevel, not within a class, and not within a namespace + return + + # Avoid initializer lists. We only need to scan back from the + # current line for something that starts with ':'. + # + # We don't need to check the current line, since the '&' would + # appear inside the second set of parentheses on the current line as + # opposed to the first set. + if linenum > 0: + for i in xrange(linenum - 1, max(0, linenum - 10), -1): + previous_line = clean_lines.elided[i] + if not Search(r'[),]\s*$', previous_line): + break + if Match(r'^\s*:\s+\S', previous_line): + return + + # Avoid preprocessors + if Search(r'\\\s*$', line): + return + + # Avoid constructor initializer lists + if IsInitializerList(clean_lines, linenum): + return + + # We allow non-const references in a few standard places, like functions + # called "swap()" or iostream operators like "<<" or ">>". Do not check + # those function parameters. + # + # We also accept & in static_assert, which looks like a function but + # it's actually a declaration expression. + whitelisted_functions = (r'(?:[sS]wap(?:<\w:+>)?|' + r'operator\s*[<>][<>]|' + r'static_assert|COMPILE_ASSERT' + r')\s*\(') + if Search(whitelisted_functions, line): + return + elif not Search(r'\S+\([^)]*$', line): + # Don't see a whitelisted function on this line. Actually we + # didn't see any function name on this line, so this is likely a + # multi-line parameter list. Try a bit harder to catch this case. + for i in xrange(2): + if (linenum > i and + Search(whitelisted_functions, clean_lines.elided[linenum - i - 1])): + return + + decls = ReplaceAll(r'{[^}]*}', ' ', line) # exclude function body + for parameter in re.findall(_RE_PATTERN_REF_PARAM, decls): + if not Match(_RE_PATTERN_CONST_REF_PARAM, parameter): + error(filename, linenum, 'runtime/references', 2, + 'Is this a non-const reference? ' + 'If so, make const or use a pointer: ' + + ReplaceAll(' *<', '<', parameter)) + + +def CheckCasts(filename, clean_lines, linenum, error): + """Various cast related checks. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Check to see if they're using an conversion function cast. + # I just try to capture the most common basic types, though there are more. + # Parameterless conversion functions, such as bool(), are allowed as they are + # probably a member operator declaration or default constructor. + match = Search( + r'(\bnew\s+|\S<\s*(?:const\s+)?)?\b' + r'(int|float|double|bool|char|int32|uint32|int64|uint64)' + r'(\([^)].*)', line) + expecting_function = ExpectingFunctionArgs(clean_lines, linenum) + if match and not expecting_function: + matched_type = match.group(2) + + # matched_new_or_template is used to silence two false positives: + # - New operators + # - Template arguments with function types + # + # For template arguments, we match on types immediately following + # an opening bracket without any spaces. This is a fast way to + # silence the common case where the function type is the first + # template argument. False negative with less-than comparison is + # avoided because those operators are usually followed by a space. + # + # function // bracket + no space = false positive + # value < double(42) // bracket + space = true positive + matched_new_or_template = match.group(1) + + # Avoid arrays by looking for brackets that come after the closing + # parenthesis. + if Match(r'\([^()]+\)\s*\[', match.group(3)): + return + + # Other things to ignore: + # - Function pointers + # - Casts to pointer types + # - Placement new + # - Alias declarations + matched_funcptr = match.group(3) + if (matched_new_or_template is None and + not (matched_funcptr and + (Match(r'\((?:[^() ]+::\s*\*\s*)?[^() ]+\)\s*\(', + matched_funcptr) or + matched_funcptr.startswith('(*)'))) and + not Match(r'\s*using\s+\S+\s*=\s*' + matched_type, line) and + not Search(r'new\(\S+\)\s*' + matched_type, line)): + error(filename, linenum, 'readability/casting', 4, + 'Using deprecated casting style. ' + 'Use static_cast<%s>(...) instead' % + matched_type) + + if not expecting_function: + CheckCStyleCast(filename, clean_lines, linenum, 'static_cast', + r'\((int|float|double|bool|char|u?int(16|32|64))\)', error) + + # This doesn't catch all cases. Consider (const char * const)"hello". + # + # (char *) "foo" should always be a const_cast (reinterpret_cast won't + # compile). + if CheckCStyleCast(filename, clean_lines, linenum, 'const_cast', + r'\((char\s?\*+\s?)\)\s*"', error): + pass + else: + # Check pointer casts for other than string constants + CheckCStyleCast(filename, clean_lines, linenum, 'reinterpret_cast', + r'\((\w+\s?\*+\s?)\)', error) + + # In addition, we look for people taking the address of a cast. This + # is dangerous -- casts can assign to temporaries, so the pointer doesn't + # point where you think. + # + # Some non-identifier character is required before the '&' for the + # expression to be recognized as a cast. These are casts: + # expression = &static_cast(temporary()); + # function(&(int*)(temporary())); + # + # This is not a cast: + # reference_type&(int* function_param); + match = Search( + r'(?:[^\w]&\(([^)*][^)]*)\)[\w(])|' + r'(?:[^\w]&(static|dynamic|down|reinterpret)_cast\b)', line) + if match: + # Try a better error message when the & is bound to something + # dereferenced by the casted pointer, as opposed to the casted + # pointer itself. + parenthesis_error = False + match = Match(r'^(.*&(?:static|dynamic|down|reinterpret)_cast\b)<', line) + if match: + _, y1, x1 = CloseExpression(clean_lines, linenum, len(match.group(1))) + if x1 >= 0 and clean_lines.elided[y1][x1] == '(': + _, y2, x2 = CloseExpression(clean_lines, y1, x1) + if x2 >= 0: + extended_line = clean_lines.elided[y2][x2:] + if y2 < clean_lines.NumLines() - 1: + extended_line += clean_lines.elided[y2 + 1] + if Match(r'\s*(?:->|\[)', extended_line): + parenthesis_error = True + + if parenthesis_error: + error(filename, linenum, 'readability/casting', 4, + ('Are you taking an address of something dereferenced ' + 'from a cast? Wrapping the dereferenced expression in ' + 'parentheses will make the binding more obvious')) + else: + error(filename, linenum, 'runtime/casting', 4, + ('Are you taking an address of a cast? ' + 'This is dangerous: could be a temp var. ' + 'Take the address before doing the cast, rather than after')) + + +def CheckCStyleCast(filename, clean_lines, linenum, cast_type, pattern, error): + """Checks for a C-style cast by looking for the pattern. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + cast_type: The string for the C++ cast to recommend. This is either + reinterpret_cast, static_cast, or const_cast, depending. + pattern: The regular expression used to find C-style casts. + error: The function to call with any errors found. + + Returns: + True if an error was emitted. + False otherwise. + """ + line = clean_lines.elided[linenum] + match = Search(pattern, line) + if not match: + return False + + # Exclude lines with keywords that tend to look like casts + context = line[0:match.start(1) - 1] + if Match(r'.*\b(?:sizeof|alignof|alignas|[_A-Z][_A-Z0-9]*)\s*$', context): + return False + + # Try expanding current context to see if we one level of + # parentheses inside a macro. + if linenum > 0: + for i in xrange(linenum - 1, max(0, linenum - 5), -1): + context = clean_lines.elided[i] + context + if Match(r'.*\b[_A-Z][_A-Z0-9]*\s*\((?:\([^()]*\)|[^()])*$', context): + return False + + # operator++(int) and operator--(int) + if context.endswith(' operator++') or context.endswith(' operator--'): + return False + + # A single unnamed argument for a function tends to look like old + # style cast. If we see those, don't issue warnings for deprecated + # casts, instead issue warnings for unnamed arguments where + # appropriate. + # + # These are things that we want warnings for, since the style guide + # explicitly require all parameters to be named: + # Function(int); + # Function(int) { + # ConstMember(int) const; + # ConstMember(int) const { + # ExceptionMember(int) throw (...); + # ExceptionMember(int) throw (...) { + # PureVirtual(int) = 0; + # [](int) -> bool { + # + # These are functions of some sort, where the compiler would be fine + # if they had named parameters, but people often omit those + # identifiers to reduce clutter: + # (FunctionPointer)(int); + # (FunctionPointer)(int) = value; + # Function((function_pointer_arg)(int)) + # Function((function_pointer_arg)(int), int param) + # ; + # <(FunctionPointerTemplateArgument)(int)>; + remainder = line[match.end(0):] + if Match(r'^\s*(?:;|const\b|throw\b|final\b|override\b|[=>{),]|->)', + remainder): + # Looks like an unnamed parameter. + + # Don't warn on any kind of template arguments. + if Match(r'^\s*>', remainder): + return False + + # Don't warn on assignments to function pointers, but keep warnings for + # unnamed parameters to pure virtual functions. Note that this pattern + # will also pass on assignments of "0" to function pointers, but the + # preferred values for those would be "nullptr" or "NULL". + matched_zero = Match(r'^\s=\s*(\S+)\s*;', remainder) + if matched_zero and matched_zero.group(1) != '0': + return False + + # Don't warn on function pointer declarations. For this we need + # to check what came before the "(type)" string. + if Match(r'.*\)\s*$', line[0:match.start(0)]): + return False + + # Don't warn if the parameter is named with block comments, e.g.: + # Function(int /*unused_param*/); + raw_line = clean_lines.raw_lines[linenum] + if '/*' in raw_line: + return False + + # Passed all filters, issue warning here. + error(filename, linenum, 'readability/function', 3, + 'All parameters should be named in a function') + return True + + # At this point, all that should be left is actual casts. + error(filename, linenum, 'readability/casting', 4, + 'Using C-style cast. Use %s<%s>(...) instead' % + (cast_type, match.group(1))) + + return True + + +def ExpectingFunctionArgs(clean_lines, linenum): + """Checks whether where function type arguments are expected. + + Args: + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + + Returns: + True if the line at 'linenum' is inside something that expects arguments + of function types. + """ + line = clean_lines.elided[linenum] + return (Match(r'^\s*MOCK_(CONST_)?METHOD\d+(_T)?\(', line) or + (linenum >= 2 and + (Match(r'^\s*MOCK_(?:CONST_)?METHOD\d+(?:_T)?\((?:\S+,)?\s*$', + clean_lines.elided[linenum - 1]) or + Match(r'^\s*MOCK_(?:CONST_)?METHOD\d+(?:_T)?\(\s*$', + clean_lines.elided[linenum - 2]) or + Search(r'\bstd::m?function\s*\<\s*$', + clean_lines.elided[linenum - 1])))) + + +_HEADERS_CONTAINING_TEMPLATES = ( + ('', ('deque',)), + ('', ('unary_function', 'binary_function', + 'plus', 'minus', 'multiplies', 'divides', 'modulus', + 'negate', + 'equal_to', 'not_equal_to', 'greater', 'less', + 'greater_equal', 'less_equal', + 'logical_and', 'logical_or', 'logical_not', + 'unary_negate', 'not1', 'binary_negate', 'not2', + 'bind1st', 'bind2nd', + 'pointer_to_unary_function', + 'pointer_to_binary_function', + 'ptr_fun', + 'mem_fun_t', 'mem_fun', 'mem_fun1_t', 'mem_fun1_ref_t', + 'mem_fun_ref_t', + 'const_mem_fun_t', 'const_mem_fun1_t', + 'const_mem_fun_ref_t', 'const_mem_fun1_ref_t', + 'mem_fun_ref', + )), + ('', ('numeric_limits',)), + ('', ('list',)), + ('', ('map', 'multimap',)), + ('', ('allocator',)), + ('', ('queue', 'priority_queue',)), + ('', ('set', 'multiset',)), + ('', ('stack',)), + ('', ('char_traits', 'basic_string',)), + ('', ('tuple',)), + ('', ('pair',)), + ('', ('vector',)), + + # gcc extensions. + # Note: std::hash is their hash, ::hash is our hash + ('', ('hash_map', 'hash_multimap',)), + ('', ('hash_set', 'hash_multiset',)), + ('', ('slist',)), + ) + +_RE_PATTERN_STRING = re.compile(r'\bstring\b') + +_re_pattern_algorithm_header = [] +for _template in ('copy', 'max', 'min', 'min_element', 'sort', 'swap', + 'transform'): + # Match max(..., ...), max(..., ...), but not foo->max, foo.max or + # type::max(). + _re_pattern_algorithm_header.append( + (re.compile(r'[^>.]\b' + _template + r'(<.*?>)?\([^\)]'), + _template, + '')) + +_re_pattern_templates = [] +for _header, _templates in _HEADERS_CONTAINING_TEMPLATES: + for _template in _templates: + _re_pattern_templates.append( + (re.compile(r'(\<|\b)' + _template + r'\s*\<'), + _template + '<>', + _header)) + + +def FilesBelongToSameModule(filename_cc, filename_h): + """Check if these two filenames belong to the same module. + + The concept of a 'module' here is a as follows: + foo.h, foo-inl.h, foo.cc, foo_test.cc and foo_unittest.cc belong to the + same 'module' if they are in the same directory. + some/path/public/xyzzy and some/path/internal/xyzzy are also considered + to belong to the same module here. + + If the filename_cc contains a longer path than the filename_h, for example, + '/absolute/path/to/base/sysinfo.cc', and this file would include + 'base/sysinfo.h', this function also produces the prefix needed to open the + header. This is used by the caller of this function to more robustly open the + header file. We don't have access to the real include paths in this context, + so we need this guesswork here. + + Known bugs: tools/base/bar.cc and base/bar.h belong to the same module + according to this implementation. Because of this, this function gives + some false positives. This should be sufficiently rare in practice. + + Args: + filename_cc: is the path for the .cc file + filename_h: is the path for the header path + + Returns: + Tuple with a bool and a string: + bool: True if filename_cc and filename_h belong to the same module. + string: the additional prefix needed to open the header file. + """ + + if not filename_cc.endswith('.cc'): + return (False, '') + filename_cc = filename_cc[:-len('.cc')] + if filename_cc.endswith('_unittest'): + filename_cc = filename_cc[:-len('_unittest')] + elif filename_cc.endswith('_test'): + filename_cc = filename_cc[:-len('_test')] + filename_cc = filename_cc.replace('/public/', '/') + filename_cc = filename_cc.replace('/internal/', '/') + + if not filename_h.endswith('.h'): + return (False, '') + filename_h = filename_h[:-len('.h')] + if filename_h.endswith('-inl'): + filename_h = filename_h[:-len('-inl')] + filename_h = filename_h.replace('/public/', '/') + filename_h = filename_h.replace('/internal/', '/') + + files_belong_to_same_module = filename_cc.endswith(filename_h) + common_path = '' + if files_belong_to_same_module: + common_path = filename_cc[:-len(filename_h)] + return files_belong_to_same_module, common_path + + +def UpdateIncludeState(filename, include_dict, io=codecs): + """Fill up the include_dict with new includes found from the file. + + Args: + filename: the name of the header to read. + include_dict: a dictionary in which the headers are inserted. + io: The io factory to use to read the file. Provided for testability. + + Returns: + True if a header was successfully added. False otherwise. + """ + headerfile = None + try: + headerfile = io.open(filename, 'r', 'utf8', 'replace') + except IOError: + return False + linenum = 0 + for line in headerfile: + linenum += 1 + clean_line = CleanseComments(line) + match = _RE_PATTERN_INCLUDE.search(clean_line) + if match: + include = match.group(2) + include_dict.setdefault(include, linenum) + return True + + +def CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error, + io=codecs): + """Reports for missing stl includes. + + This function will output warnings to make sure you are including the headers + necessary for the stl containers and functions that you use. We only give one + reason to include a header. For example, if you use both equal_to<> and + less<> in a .h file, only one (the latter in the file) of these will be + reported as a reason to include the . + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + include_state: An _IncludeState instance. + error: The function to call with any errors found. + io: The IO factory to use to read the header file. Provided for unittest + injection. + """ + required = {} # A map of header name to linenumber and the template entity. + # Example of required: { '': (1219, 'less<>') } + + for linenum in xrange(clean_lines.NumLines()): + line = clean_lines.elided[linenum] + if not line or line[0] == '#': + continue + + # String is special -- it is a non-templatized type in STL. + matched = _RE_PATTERN_STRING.search(line) + if matched: + # Don't warn about strings in non-STL namespaces: + # (We check only the first match per line; good enough.) + prefix = line[:matched.start()] + if prefix.endswith('std::') or not prefix.endswith('::'): + required[''] = (linenum, 'string') + + for pattern, template, header in _re_pattern_algorithm_header: + if pattern.search(line): + required[header] = (linenum, template) + + # The following function is just a speed up, no semantics are changed. + if not '<' in line: # Reduces the cpu time usage by skipping lines. + continue + + for pattern, template, header in _re_pattern_templates: + if pattern.search(line): + required[header] = (linenum, template) + + # The policy is that if you #include something in foo.h you don't need to + # include it again in foo.cc. Here, we will look at possible includes. + # Let's flatten the include_state include_list and copy it into a dictionary. + include_dict = dict([item for sublist in include_state.include_list + for item in sublist]) + + # Did we find the header for this file (if any) and successfully load it? + header_found = False + + # Use the absolute path so that matching works properly. + abs_filename = FileInfo(filename).FullName() + + # For Emacs's flymake. + # If cpplint is invoked from Emacs's flymake, a temporary file is generated + # by flymake and that file name might end with '_flymake.cc'. In that case, + # restore original file name here so that the corresponding header file can be + # found. + # e.g. If the file name is 'foo_flymake.cc', we should search for 'foo.h' + # instead of 'foo_flymake.h' + abs_filename = re.sub(r'_flymake\.cc$', '.cc', abs_filename) + + # include_dict is modified during iteration, so we iterate over a copy of + # the keys. + header_keys = include_dict.keys() + for header in header_keys: + (same_module, common_path) = FilesBelongToSameModule(abs_filename, header) + fullpath = common_path + header + if same_module and UpdateIncludeState(fullpath, include_dict, io): + header_found = True + + # If we can't find the header file for a .cc, assume it's because we don't + # know where to look. In that case we'll give up as we're not sure they + # didn't include it in the .h file. + # TODO(unknown): Do a better job of finding .h files so we are confident that + # not having the .h file means there isn't one. + if filename.endswith('.cc') and not header_found: + return + + # All the lines have been processed, report the errors found. + for required_header_unstripped in required: + template = required[required_header_unstripped][1] + if required_header_unstripped.strip('<>"') not in include_dict: + error(filename, required[required_header_unstripped][0], + 'build/include_what_you_use', 4, + 'Add #include ' + required_header_unstripped + ' for ' + template) + + +_RE_PATTERN_EXPLICIT_MAKEPAIR = re.compile(r'\bmake_pair\s*<') + + +def CheckMakePairUsesDeduction(filename, clean_lines, linenum, error): + """Check that make_pair's template arguments are deduced. + + G++ 4.6 in C++11 mode fails badly if make_pair's template arguments are + specified explicitly, and such use isn't intended in any case. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + match = _RE_PATTERN_EXPLICIT_MAKEPAIR.search(line) + if match: + error(filename, linenum, 'build/explicit_make_pair', + 4, # 4 = high confidence + 'For C++11-compatibility, omit template arguments from make_pair' + ' OR use pair directly OR if appropriate, construct a pair directly') + + +def CheckDefaultLambdaCaptures(filename, clean_lines, linenum, error): + """Check that default lambda captures are not used. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # A lambda introducer specifies a default capture if it starts with "[=" + # or if it starts with "[&" _not_ followed by an identifier. + match = Match(r'^(.*)\[\s*(?:=|&[^\w])', line) + if match: + # Found a potential error, check what comes after the lambda-introducer. + # If it's not open parenthesis (for lambda-declarator) or open brace + # (for compound-statement), it's not a lambda. + line, _, pos = CloseExpression(clean_lines, linenum, len(match.group(1))) + if pos >= 0 and Match(r'^\s*[{(]', line[pos:]): + error(filename, linenum, 'build/c++11', + 4, # 4 = high confidence + 'Default lambda captures are an unapproved C++ feature.') + + +def CheckRedundantVirtual(filename, clean_lines, linenum, error): + """Check if line contains a redundant "virtual" function-specifier. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + # Look for "virtual" on current line. + line = clean_lines.elided[linenum] + virtual = Match(r'^(.*)(\bvirtual\b)(.*)$', line) + if not virtual: return + + # Ignore "virtual" keywords that are near access-specifiers. These + # are only used in class base-specifier and do not apply to member + # functions. + if (Search(r'\b(public|protected|private)\s+$', virtual.group(1)) or + Match(r'^\s+(public|protected|private)\b', virtual.group(3))): + return + + # Ignore the "virtual" keyword from virtual base classes. Usually + # there is a column on the same line in these cases (virtual base + # classes are rare in google3 because multiple inheritance is rare). + if Match(r'^.*[^:]:[^:].*$', line): return + + # Look for the next opening parenthesis. This is the start of the + # parameter list (possibly on the next line shortly after virtual). + # TODO(unknown): doesn't work if there are virtual functions with + # decltype() or other things that use parentheses, but csearch suggests + # that this is rare. + end_col = -1 + end_line = -1 + start_col = len(virtual.group(2)) + for start_line in xrange(linenum, min(linenum + 3, clean_lines.NumLines())): + line = clean_lines.elided[start_line][start_col:] + parameter_list = Match(r'^([^(]*)\(', line) + if parameter_list: + # Match parentheses to find the end of the parameter list + (_, end_line, end_col) = CloseExpression( + clean_lines, start_line, start_col + len(parameter_list.group(1))) + break + start_col = 0 + + if end_col < 0: + return # Couldn't find end of parameter list, give up + + # Look for "override" or "final" after the parameter list + # (possibly on the next few lines). + for i in xrange(end_line, min(end_line + 3, clean_lines.NumLines())): + line = clean_lines.elided[i][end_col:] + match = Search(r'\b(override|final)\b', line) + if match: + error(filename, linenum, 'readability/inheritance', 4, + ('"virtual" is redundant since function is ' + 'already declared as "%s"' % match.group(1))) + + # Set end_col to check whole lines after we are done with the + # first line. + end_col = 0 + if Search(r'[^\w]\s*$', line): + break + + +def CheckRedundantOverrideOrFinal(filename, clean_lines, linenum, error): + """Check if line contains a redundant "override" or "final" virt-specifier. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + # Look for closing parenthesis nearby. We need one to confirm where + # the declarator ends and where the virt-specifier starts to avoid + # false positives. + line = clean_lines.elided[linenum] + declarator_end = line.rfind(')') + if declarator_end >= 0: + fragment = line[declarator_end:] + else: + if linenum > 1 and clean_lines.elided[linenum - 1].rfind(')') >= 0: + fragment = line + else: + return + + # Check that at most one of "override" or "final" is present, not both + if Search(r'\boverride\b', fragment) and Search(r'\bfinal\b', fragment): + error(filename, linenum, 'readability/inheritance', 4, + ('"override" is redundant since function is ' + 'already declared as "final"')) + + + + +# Returns true if we are at a new block, and it is directly +# inside of a namespace. +def IsBlockInNameSpace(nesting_state, is_forward_declaration): + """Checks that the new block is directly in a namespace. + + Args: + nesting_state: The _NestingState object that contains info about our state. + is_forward_declaration: If the class is a forward declared class. + Returns: + Whether or not the new block is directly in a namespace. + """ + if is_forward_declaration: + if len(nesting_state.stack) >= 1 and ( + isinstance(nesting_state.stack[-1], _NamespaceInfo)): + return True + else: + return False + + return (len(nesting_state.stack) > 1 and + nesting_state.stack[-1].check_namespace_indentation and + isinstance(nesting_state.stack[-2], _NamespaceInfo)) + + +def ShouldCheckNamespaceIndentation(nesting_state, is_namespace_indent_item, + raw_lines_no_comments, linenum): + """This method determines if we should apply our namespace indentation check. + + Args: + nesting_state: The current nesting state. + is_namespace_indent_item: If we just put a new class on the stack, True. + If the top of the stack is not a class, or we did not recently + add the class, False. + raw_lines_no_comments: The lines without the comments. + linenum: The current line number we are processing. + + Returns: + True if we should apply our namespace indentation check. Currently, it + only works for classes and namespaces inside of a namespace. + """ + + is_forward_declaration = IsForwardClassDeclaration(raw_lines_no_comments, + linenum) + + if not (is_namespace_indent_item or is_forward_declaration): + return False + + # If we are in a macro, we do not want to check the namespace indentation. + if IsMacroDefinition(raw_lines_no_comments, linenum): + return False + + return IsBlockInNameSpace(nesting_state, is_forward_declaration) + + +# Call this method if the line is directly inside of a namespace. +# If the line above is blank (excluding comments) or the start of +# an inner namespace, it cannot be indented. +def CheckItemIndentationInNamespace(filename, raw_lines_no_comments, linenum, + error): + line = raw_lines_no_comments[linenum] + if Match(r'^\s+', line): + error(filename, linenum, 'runtime/indentation_namespace', 4, + 'Do not indent within a namespace') + + +def ProcessLine(filename, file_extension, clean_lines, line, + include_state, function_state, nesting_state, error, + extra_check_functions=[]): + """Processes a single line in the file. + + Args: + filename: Filename of the file that is being processed. + file_extension: The extension (dot not included) of the file. + clean_lines: An array of strings, each representing a line of the file, + with comments stripped. + line: Number of line being processed. + include_state: An _IncludeState instance in which the headers are inserted. + function_state: A _FunctionState instance which counts function lines, etc. + nesting_state: A NestingState instance which maintains information about + the current stack of nested blocks being parsed. + error: A callable to which errors are reported, which takes 4 arguments: + filename, line number, error level, and message + extra_check_functions: An array of additional check functions that will be + run on each source line. Each function takes 4 + arguments: filename, clean_lines, line, error + """ + raw_lines = clean_lines.raw_lines + ParseNolintSuppressions(filename, raw_lines[line], line, error) + nesting_state.Update(filename, clean_lines, line, error) + CheckForNamespaceIndentation(filename, nesting_state, clean_lines, line, + error) + if nesting_state.InAsmBlock(): return + CheckForFunctionLengths(filename, clean_lines, line, function_state, error) + CheckForMultilineCommentsAndStrings(filename, clean_lines, line, error) + CheckStyle(filename, clean_lines, line, file_extension, nesting_state, error) + CheckLanguage(filename, clean_lines, line, file_extension, include_state, + nesting_state, error) + CheckForNonConstReference(filename, clean_lines, line, nesting_state, error) + CheckForNonStandardConstructs(filename, clean_lines, line, + nesting_state, error) + CheckVlogArguments(filename, clean_lines, line, error) + CheckPosixThreading(filename, clean_lines, line, error) + CheckInvalidIncrement(filename, clean_lines, line, error) + CheckMakePairUsesDeduction(filename, clean_lines, line, error) + CheckDefaultLambdaCaptures(filename, clean_lines, line, error) + CheckRedundantVirtual(filename, clean_lines, line, error) + CheckRedundantOverrideOrFinal(filename, clean_lines, line, error) + for check_fn in extra_check_functions: + check_fn(filename, clean_lines, line, error) + +def FlagCxx11Features(filename, clean_lines, linenum, error): + """Flag those c++11 features that we only allow in certain places. + + Args: + filename: The name of the current file. + clean_lines: A CleansedLines instance containing the file. + linenum: The number of the line to check. + error: The function to call with any errors found. + """ + line = clean_lines.elided[linenum] + + # Flag unapproved C++11 headers. + include = Match(r'\s*#\s*include\s+[<"]([^<"]+)[">]', line) + if include and include.group(1) in ('cfenv', + 'condition_variable', + 'fenv.h', + 'future', + 'mutex', + 'thread', + 'chrono', + 'ratio', + 'regex', + 'system_error', + ): + error(filename, linenum, 'build/c++11', 5, + ('<%s> is an unapproved C++11 header.') % include.group(1)) + + # The only place where we need to worry about C++11 keywords and library + # features in preprocessor directives is in macro definitions. + if Match(r'\s*#', line) and not Match(r'\s*#\s*define\b', line): return + + # These are classes and free functions. The classes are always + # mentioned as std::*, but we only catch the free functions if + # they're not found by ADL. They're alphabetical by header. + for top_name in ( + # type_traits + 'alignment_of', + 'aligned_union', + ): + if Search(r'\bstd::%s\b' % top_name, line): + error(filename, linenum, 'build/c++11', 5, + ('std::%s is an unapproved C++11 class or function. Send c-style ' + 'an example of where it would make your code more readable, and ' + 'they may let you use it.') % top_name) + + +def ProcessFileData(filename, file_extension, lines, error, + extra_check_functions=[]): + """Performs lint checks and reports any errors to the given error function. + + Args: + filename: Filename of the file that is being processed. + file_extension: The extension (dot not included) of the file. + lines: An array of strings, each representing a line of the file, with the + last element being empty if the file is terminated with a newline. + error: A callable to which errors are reported, which takes 4 arguments: + filename, line number, error level, and message + extra_check_functions: An array of additional check functions that will be + run on each source line. Each function takes 4 + arguments: filename, clean_lines, line, error + """ + lines = (['// marker so line numbers and indices both start at 1'] + lines + + ['// marker so line numbers end in a known way']) + + include_state = _IncludeState() + function_state = _FunctionState() + nesting_state = NestingState() + + ResetNolintSuppressions() + + CheckForCopyright(filename, lines, error) + + RemoveMultiLineComments(filename, lines, error) + clean_lines = CleansedLines(lines) + + if file_extension == 'h': + CheckForHeaderGuard(filename, clean_lines, error) + + for line in xrange(clean_lines.NumLines()): + ProcessLine(filename, file_extension, clean_lines, line, + include_state, function_state, nesting_state, error, + extra_check_functions) + FlagCxx11Features(filename, clean_lines, line, error) + nesting_state.CheckCompletedBlocks(filename, error) + + CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error) + + # Check that the .cc file has included its header if it exists. + if file_extension == 'cc': + CheckHeaderFileIncluded(filename, include_state, error) + + # We check here rather than inside ProcessLine so that we see raw + # lines rather than "cleaned" lines. + CheckForBadCharacters(filename, lines, error) + + CheckForNewlineAtEOF(filename, lines, error) + +def ProcessConfigOverrides(filename): + """ Loads the configuration files and processes the config overrides. + + Args: + filename: The name of the file being processed by the linter. + + Returns: + False if the current |filename| should not be processed further. + """ + + abs_filename = os.path.abspath(filename) + cfg_filters = [] + keep_looking = True + while keep_looking: + abs_path, base_name = os.path.split(abs_filename) + if not base_name: + break # Reached the root directory. + + cfg_file = os.path.join(abs_path, "CPPLINT.cfg") + abs_filename = abs_path + if not os.path.isfile(cfg_file): + continue + + try: + with open(cfg_file) as file_handle: + for line in file_handle: + line, _, _ = line.partition('#') # Remove comments. + if not line.strip(): + continue + + name, _, val = line.partition('=') + name = name.strip() + val = val.strip() + if name == 'set noparent': + keep_looking = False + elif name == 'filter': + cfg_filters.append(val) + elif name == 'exclude_files': + # When matching exclude_files pattern, use the base_name of + # the current file name or the directory name we are processing. + # For example, if we are checking for lint errors in /foo/bar/baz.cc + # and we found the .cfg file at /foo/CPPLINT.cfg, then the config + # file's "exclude_files" filter is meant to be checked against "bar" + # and not "baz" nor "bar/baz.cc". + if base_name: + pattern = re.compile(val) + if pattern.match(base_name): + sys.stderr.write('Ignoring "%s": file excluded by "%s". ' + 'File path component "%s" matches ' + 'pattern "%s"\n' % + (filename, cfg_file, base_name, val)) + return False + elif name == 'linelength': + global _line_length + try: + _line_length = int(val) + except ValueError: + sys.stderr.write('Line length must be numeric.') + else: + sys.stderr.write( + 'Invalid configuration option (%s) in file %s\n' % + (name, cfg_file)) + + except IOError: + sys.stderr.write( + "Skipping config file '%s': Can't open for reading\n" % cfg_file) + keep_looking = False + + # Apply all the accumulated filters in reverse order (top-level directory + # config options having the least priority). + for filter in reversed(cfg_filters): + _AddFilters(filter) + + return True + + +def ProcessFile(filename, vlevel, extra_check_functions=[]): + """Does google-lint on a single file. + + Args: + filename: The name of the file to parse. + + vlevel: The level of errors to report. Every error of confidence + >= verbose_level will be reported. 0 is a good default. + + extra_check_functions: An array of additional check functions that will be + run on each source line. Each function takes 4 + arguments: filename, clean_lines, line, error + """ + + _SetVerboseLevel(vlevel) + _BackupFilters() + + if not ProcessConfigOverrides(filename): + _RestoreFilters() + return + + lf_lines = [] + crlf_lines = [] + try: + # Support the UNIX convention of using "-" for stdin. Note that + # we are not opening the file with universal newline support + # (which codecs doesn't support anyway), so the resulting lines do + # contain trailing '\r' characters if we are reading a file that + # has CRLF endings. + # If after the split a trailing '\r' is present, it is removed + # below. + if filename == '-': + lines = codecs.StreamReaderWriter(sys.stdin, + codecs.getreader('utf8'), + codecs.getwriter('utf8'), + 'replace').read().split('\n') + else: + lines = codecs.open(filename, 'r', 'utf8', 'replace').read().split('\n') + + # Remove trailing '\r'. + # The -1 accounts for the extra trailing blank line we get from split() + for linenum in range(len(lines) - 1): + if lines[linenum].endswith('\r'): + lines[linenum] = lines[linenum].rstrip('\r') + crlf_lines.append(linenum + 1) + else: + lf_lines.append(linenum + 1) + + except IOError: + sys.stderr.write( + "Skipping input '%s': Can't open for reading\n" % filename) + _RestoreFilters() + return + + # Note, if no dot is found, this will give the entire filename as the ext. + file_extension = filename[filename.rfind('.') + 1:] + + # When reading from stdin, the extension is unknown, so no cpplint tests + # should rely on the extension. + if filename != '-' and file_extension not in _valid_extensions: + sys.stderr.write('Ignoring %s; not a valid file name ' + '(%s)\n' % (filename, ', '.join(_valid_extensions))) + else: + ProcessFileData(filename, file_extension, lines, Error, + extra_check_functions) + + # If end-of-line sequences are a mix of LF and CR-LF, issue + # warnings on the lines with CR. + # + # Don't issue any warnings if all lines are uniformly LF or CR-LF, + # since critique can handle these just fine, and the style guide + # doesn't dictate a particular end of line sequence. + # + # We can't depend on os.linesep to determine what the desired + # end-of-line sequence should be, since that will return the + # server-side end-of-line sequence. + if lf_lines and crlf_lines: + # Warn on every line with CR. An alternative approach might be to + # check whether the file is mostly CRLF or just LF, and warn on the + # minority, we bias toward LF here since most tools prefer LF. + for linenum in crlf_lines: + Error(filename, linenum, 'whitespace/newline', 1, + 'Unexpected \\r (^M) found; better to use only \\n') + + sys.stderr.write('Done processing %s\n' % filename) + _RestoreFilters() + + +def PrintUsage(message): + """Prints a brief usage string and exits, optionally with an error message. + + Args: + message: The optional error message. + """ + sys.stderr.write(_USAGE) + if message: + sys.exit('\nFATAL ERROR: ' + message) + else: + sys.exit(1) + + +def PrintCategories(): + """Prints a list of all the error-categories used by error messages. + + These are the categories used to filter messages via --filter. + """ + sys.stderr.write(''.join(' %s\n' % cat for cat in _ERROR_CATEGORIES)) + sys.exit(0) + + +def ParseArguments(args): + """Parses the command line arguments. + + This may set the output format and verbosity level as side-effects. + + Args: + args: The command line arguments: + + Returns: + The list of filenames to lint. + """ + try: + (opts, filenames) = getopt.getopt(args, '', ['help', 'output=', 'verbose=', + 'counting=', + 'filter=', + 'root=', + 'linelength=', + 'extensions=']) + except getopt.GetoptError: + PrintUsage('Invalid arguments.') + + verbosity = _VerboseLevel() + output_format = _OutputFormat() + filters = '' + counting_style = '' + + for (opt, val) in opts: + if opt == '--help': + PrintUsage(None) + elif opt == '--output': + if val not in ('emacs', 'vs7', 'eclipse'): + PrintUsage('The only allowed output formats are emacs, vs7 and eclipse.') + output_format = val + elif opt == '--verbose': + verbosity = int(val) + elif opt == '--filter': + filters = val + if not filters: + PrintCategories() + elif opt == '--counting': + if val not in ('total', 'toplevel', 'detailed'): + PrintUsage('Valid counting options are total, toplevel, and detailed') + counting_style = val + elif opt == '--root': + global _root + _root = val + elif opt == '--linelength': + global _line_length + try: + _line_length = int(val) + except ValueError: + PrintUsage('Line length must be digits.') + elif opt == '--extensions': + global _valid_extensions + try: + _valid_extensions = set(val.split(',')) + except ValueError: + PrintUsage('Extensions must be comma seperated list.') + + if not filenames: + PrintUsage('No files were specified.') + + _SetOutputFormat(output_format) + _SetVerboseLevel(verbosity) + _SetFilters(filters) + _SetCountingStyle(counting_style) + + return filenames + + +def main(): + filenames = ParseArguments(sys.argv[1:]) + + # Change stderr to write with replacement characters so we don't die + # if we try to print something containing non-ASCII characters. + sys.stderr = codecs.StreamReaderWriter(sys.stderr, + codecs.getreader('utf8'), + codecs.getwriter('utf8'), + 'replace') + + _cpplint_state.ResetErrorCounts() + for filename in filenames: + ProcessFile(filename, _cpplint_state.verbose_level) + _cpplint_state.PrintErrorCounts() + + sys.exit(_cpplint_state.error_count > 0) + + +if __name__ == '__main__': + main() diff --git a/deploy.enc b/deploy.enc new file mode 100644 index 0000000..d419f9e Binary files /dev/null and b/deploy.enc differ diff --git a/deps/librdkafka.gyp b/deps/librdkafka.gyp new file mode 100644 index 0000000..6093d7b --- /dev/null +++ b/deps/librdkafka.gyp @@ -0,0 +1,62 @@ +{ + 'targets': [ + { + "target_name": "librdkafka", + "type": "none", + "conditions": [ + [ + 'OS=="win"', + { + }, + { + "actions": [ + { + "action_name": "configure", + "inputs": [], + "outputs": [ + "librdkafka/config.h", + ], + "action": [ + "node", "../util/configure" + ] + }, + { + "action_name": "build_dependencies", + "inputs": [ + "librdkafka/config.h", + ], + "action": [ + "make", "-C", "librdkafka", "libs", "install" + ], + "conditions": [ + [ + 'OS=="mac"', + { + 'outputs': [ + 'deps/librdkafka/src-cpp/librdkafka++.dylib', + 'deps/librdkafka/src-cpp/librdkafka++.1.dylib', + 'deps/librdkafka/src/librdkafka.dylib', + 'deps/librdkafka/src/librdkafka.1.dylib' + ], + }, + { + 'outputs': [ + 'deps/librdkafka/src-cpp/librdkafka++.so', + 'deps/librdkafka/src-cpp/librdkafka++.so.1', + 'deps/librdkafka/src/librdkafka.so', + 'deps/librdkafka/src/librdkafka.so.1', + 'deps/librdkafka/src-cpp/librdkafka++.a', + 'deps/librdkafka/src/librdkafka.a', + ], + } + ] + ], + } + ] + } + + ] + ] + } + ] +} diff --git a/deps/windows-install.py b/deps/windows-install.py new file mode 100644 index 0000000..65888b1 --- /dev/null +++ b/deps/windows-install.py @@ -0,0 +1,75 @@ +librdkafkaVersion = '' +# read librdkafka version from package.json +import json +import os +with open('../package.json') as f: + librdkafkaVersion = json.load(f)['librdkafka'] +librdkafkaWinSufix = '7' if librdkafkaVersion == '0.11.5' else ''; + +depsPrecompiledDir = '../deps/precompiled' +depsIncludeDir = '../deps/include' +buildReleaseDir = 'Release' + +# alternative: 'https://api.nuget.org/v3-flatcontainer/librdkafka.redist/{}/librdkafka.redist.{}.nupkg'.format(librdkafkaVersion, librdkafkaVersion) +env_dist = os.environ +downloadBaseUrl = env_dist['NODE_RDKAFKA_NUGET_BASE_URL'] if 'NODE_RDKAFKA_NUGET_BASE_URL' in env_dist else 'https://globalcdn.nuget.org/packages/' +librdkafkaNugetUrl = downloadBaseUrl + 'librdkafka.redist.{}.nupkg'.format(librdkafkaVersion) +print('download librdkafka form ' + librdkafkaNugetUrl) +outputDir = 'librdkafka.redist' +outputFile = outputDir + '.zip' +dllPath = outputDir + '/runtimes/win{}-x64/native'.format(librdkafkaWinSufix) +libPath = outputDir + '/build/native/lib/win{}/x64/win{}-x64-Release/v120'.format(librdkafkaWinSufix, librdkafkaWinSufix) +includePath = outputDir + '/build/native/include/librdkafka' + +# download librdkafka from nuget +try: + # For Python 3.0 and later + from urllib.request import urlopen +except ImportError: + # Fall back to Python 2's urllib2 + from urllib2 import urlopen +import ssl + +filedata = urlopen(librdkafkaNugetUrl, context=ssl._create_unverified_context()) + +datatowrite = filedata.read() +with open(outputFile, 'wb') as f: + f.write(datatowrite) + +# extract package +import zipfile +zip_ref = zipfile.ZipFile(outputFile, 'r') +zip_ref.extractall(outputDir) +zip_ref.close() + +# copy files +import shutil, os, errno + +def createdir(dir): + try: + os.makedirs(dir) + except OSError as e: + if errno.EEXIST != e.errno: + raise + +createdir(depsPrecompiledDir) +createdir(depsIncludeDir) +createdir(buildReleaseDir) + +shutil.copy2(libPath + '/librdkafka.lib', depsPrecompiledDir) +shutil.copy2(libPath + '/librdkafkacpp.lib', depsPrecompiledDir) + +shutil.copy2(includePath + '/rdkafka.h', depsIncludeDir) +shutil.copy2(includePath + '/rdkafkacpp.h', depsIncludeDir) + +shutil.copy2(dllPath + '/zlib.dll', buildReleaseDir) +shutil.copy2(dllPath + '/msvcr120.dll', buildReleaseDir) +shutil.copy2(dllPath + '/librdkafka.dll', buildReleaseDir) +shutil.copy2(dllPath + '/librdkafkacpp.dll', buildReleaseDir) +if not librdkafkaVersion.startswith('0.'): + shutil.copy2(dllPath + '/libzstd.dll', buildReleaseDir) + shutil.copy2(dllPath + '/msvcp120.dll', buildReleaseDir) + +# clean up +os.remove(outputFile) +shutil.rmtree(outputDir) diff --git a/docker-compose.yml b/docker-compose.yml new file mode 100644 index 0000000..abe29df --- /dev/null +++ b/docker-compose.yml @@ -0,0 +1,23 @@ +--- +zookeeper: + image: confluentinc/cp-zookeeper + ports: + - "2181:2181" + environment: + ZOOKEEPER_CLIENT_PORT: 2181 + ZOOKEEPER_TICK_TIME: 2000 +kafka: + image: confluentinc/cp-kafka + links: + - zookeeper + ports: + - "9092:9092" + environment: + KAFKA_BROKER_ID: 1 + KAFKA_ZOOKEEPER_CONNECT: 'zookeeper:2181' + KAFKA_ADVERTISED_LISTENERS: 'PLAINTEXT://localhost:9092' + KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: 1 + KAFKA_GROUP_INITIAL_REBALANCE_DELAY_MS: 0 + KAFKA_TRANSACTION_STATE_LOG_REPLICATION_FACTOR: 1 + KAFKA_DEFAULT_REPLICATION_FACTOR: 1 + KAFKA_TRANSACTION_STATE_LOG_MIN_ISR: 1 diff --git a/e2e/admin.spec.js b/e2e/admin.spec.js new file mode 100644 index 0000000..83d65e4 --- /dev/null +++ b/e2e/admin.spec.js @@ -0,0 +1,193 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Kafka = require('../'); +var t = require('assert'); + +var eventListener = require('./listener'); +var kafkaBrokerList = process.env.KAFKA_HOST || 'localhost:9092'; +var time = Date.now(); + +function pollForTopic(client, topicName, maxTries, tryDelay, cb, customCondition) { + var tries = 0; + + function getTopicIfExists(innerCb) { + client.getMetadata({ + topic: topicName, + }, function(metadataErr, metadata) { + if (metadataErr) { + cb(metadataErr); + return; + } + + var topicFound = metadata.topics.filter(function(topicObj) { + var foundTopic = topicObj.name === topicName; + + // If we have a custom condition for "foundedness", do it here after + // we make sure we are operating on the correct topic + if (foundTopic && customCondition) { + return customCondition(topicObj); + } + return foundTopic; + }); + + if (topicFound.length >= 1) { + innerCb(null, topicFound[0]); + return; + } + + innerCb(new Error('Could not find topic ' + topicName)); + }); + } + + function maybeFinish(err, obj) { + if (err) { + queueNextTry(); + return; + } + + cb(null, obj); + } + + function queueNextTry() { + tries += 1; + if (tries < maxTries) { + setTimeout(function() { + getTopicIfExists(maybeFinish); + }, tryDelay); + } else { + cb(new Error('Exceeded max tries of ' + maxTries)); + } + } + + queueNextTry(); +} + +describe('Admin', function() { + var client; + var producer; + + before(function(done) { + producer = new Kafka.Producer({ + 'metadata.broker.list': kafkaBrokerList, + }); + producer.connect(null, function(err) { + t.ifError(err); + done(); + }); + }); + + after(function(done) { + producer.disconnect(function() { + done(); + }); + }); + + beforeEach(function() { + this.timeout(10000); + client = Kafka.AdminClient.create({ + 'client.id': 'kafka-test', + 'metadata.broker.list': kafkaBrokerList + }); + }); + + describe('createTopic', function() { + it('should create topic sucessfully', function(done) { + var topicName = 'admin-test-topic-' + time; + this.timeout(30000); + client.createTopic({ + topic: topicName, + num_partitions: 1, + replication_factor: 1 + }, function(err) { + pollForTopic(producer, topicName, 10, 1000, function(err) { + t.ifError(err); + done(); + }); + }); + }); + + it('should raise an error when replication_factor is larger than number of brokers', function(done) { + var topicName = 'admin-test-topic-bad-' + time; + this.timeout(30000); + client.createTopic({ + topic: topicName, + num_partitions: 9999, + replication_factor: 9999 + }, function(err) { + t.equal(typeof err, 'object', 'an error should be returned'); + done(); + }); + }); + }); + + describe('deleteTopic', function() { + it('should be able to delete a topic after creation', function(done) { + var topicName = 'admin-test-topic-2bdeleted-' + time; + this.timeout(30000); + client.createTopic({ + topic: topicName, + num_partitions: 1, + replication_factor: 1 + }, function(err) { + pollForTopic(producer, topicName, 10, 1000, function(err) { + t.ifError(err); + client.deleteTopic(topicName, function(deleteErr) { + // Fail if we got an error + t.ifError(deleteErr); + done(); + }); + }); + }); + }); + }); + + describe('createPartitions', function() { + it('should be able to add partitions to a topic after creation', function(done) { + var topicName = 'admin-test-topic-newparts-' + time; + this.timeout(30000); + client.createTopic({ + topic: topicName, + num_partitions: 1, + replication_factor: 1 + }, function(err) { + pollForTopic(producer, topicName, 10, 1000, function(err) { + t.ifError(err); + client.createPartitions(topicName, 20, function(createErr) { + pollForTopic(producer, topicName, 10, 1000, function(pollErr) { + t.ifError(pollErr); + done(); + }, function(topic) { + return topic.partitions.length === 20; + }); + }); + }); + }); + }); + + it('should NOT be able to reduce partitions to a topic after creation', function(done) { + var topicName = 'admin-test-topic-newparts2-' + time; + this.timeout(30000); + client.createTopic({ + topic: topicName, + num_partitions: 4, + replication_factor: 1 + }, function(err) { + pollForTopic(producer, topicName, 10, 1000, function(err) { + t.ifError(err); + client.createPartitions(topicName, 1, function(createErr) { + t.equal(typeof createErr, 'object', 'an error should be returned'); + done(); + }); + }); + }); + }); + }); + +}); diff --git a/e2e/both.spec.js b/e2e/both.spec.js new file mode 100644 index 0000000..5871c7c --- /dev/null +++ b/e2e/both.spec.js @@ -0,0 +1,690 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var crypto = require('crypto'); +var t = require('assert'); + +var Kafka = require('../'); +var kafkaBrokerList = process.env.KAFKA_HOST || 'localhost:9092'; +var eventListener = require('./listener'); +var topic = 'test'; +var topic2 = 'test2'; + +describe('Consumer/Producer', function() { + + var producer; + var consumer; + + beforeEach(function(done) { + var finished = 0; + var called = false; + + function maybeDone(err) { + if (called) { + return; + } + + finished++; + if (err) { + called = true; + return done(err); + } + + if (finished === 2) { + done(); + } + } + + var grp = 'kafka-mocha-grp-' + crypto.randomBytes(20).toString('hex'); + + consumer = new Kafka.KafkaConsumer({ + 'metadata.broker.list': kafkaBrokerList, + 'group.id': grp, + 'fetch.wait.max.ms': 1000, + 'session.timeout.ms': 10000, + 'enable.auto.commit': true, + 'enable.partition.eof': true, + 'debug': 'all' + // paused: true, + }, { + 'auto.offset.reset': 'largest' + }); + + consumer.connect({}, function(err, d) { + t.ifError(err); + t.equal(typeof d, 'object', 'metadata should be returned'); + maybeDone(err); + }); + + eventListener(consumer); + + producer = new Kafka.Producer({ + 'client.id': 'kafka-mocha', + 'metadata.broker.list': kafkaBrokerList, + 'fetch.wait.max.ms': 1, + 'debug': 'all', + 'dr_cb': true + }, { + 'produce.offset.report': true + }); + + producer.connect({}, function(err, d) { + t.ifError(err); + t.equal(typeof d, 'object', 'metadata should be returned'); + maybeDone(err); + }); + + eventListener(producer); + + }); + + afterEach(function(done) { + this.timeout(6000); + var finished = 0; + var called = false; + + function maybeDone(err) { + if (called) { + return; + } + + finished++; + if (err) { + called = true; + return done(err); + } + + if (finished === 2) { + done(); + } + } + + consumer.disconnect(function(err) { + maybeDone(err); + }); + + producer.disconnect(function(err) { + maybeDone(err); + }); + }); + + it('should be able to produce, consume messages, read position: subscribe/consumeOnce', function(done) { + this.timeout(8000); + crypto.randomBytes(4096, function(ex, buffer) { + producer.setPollInterval(10); + + var offset; + + producer.once('delivery-report', function(err, report) { + t.ifError(err); + offset = report.offset; + }); + + consumer.setDefaultConsumeTimeout(10); + consumer.subscribe([topic]); + + var ct; + + var consumeOne = function() { + consumer.consume(1, function(err, messages) { + if (err && err.code === -185) { + ct = setTimeout(consumeOne, 100); + return; + } else if (messages.length === 0 || (err && err.code === -191)) { + producer.produce(topic, null, buffer, null); + ct = setTimeout(consumeOne, 100); + return; + } else if (err) { + return; + } + + var message = messages[0]; + + t.equal(Array.isArray(consumer.assignments()), true, 'Assignments should be an array'); + t.equal(consumer.assignments().length > 0, true, 'Should have at least one assignment'); + t.equal(buffer.toString(), message.value.toString(), + 'message is not equal to buffer'); + + // test consumer.position as we have consumed + var position = consumer.position(); + t.equal(position.length, 1); + t.deepStrictEqual(position[0].partition, 0); + t.ok(position[0].offset >= 0); + done(); + }); + }; + + // Consume until we get it or time out + consumeOne(); + + }); + }); + + it('should return ready messages on partition EOF', function(done) { + this.timeout(8000); + crypto.randomBytes(4096, function(ex, buffer) { + producer.setPollInterval(10); + + producer.once('delivery-report', function(err, report) { + t.ifError(err); + }); + + consumer.subscribe([topic]); + + var consumeAll = function() { + // Make sure we get the message fast when consuming with large timeout + consumer.setDefaultConsumeTimeout(1000000); + consumer.consume(100000, function(err, messages) { + t.ifError(err); + t.equal(messages.length, 1); + done(); + }); + }; + + var consumeNone = function() { + // With no new messages, the consume should wait whole timeout + var start = Date.now(); + // Set the timeout to 2000ms to see that it actually waits the whole time + // (Needs to be higher than fetch.max.wait.ms which is 1000 here + // to ensure we don't only wait that long) + consumer.setDefaultConsumeTimeout(2000); + consumer.consume(100000, function(err, messages) { + t.ifError(err); + t.ok(Date.now() - start >= 1998); + t.equal(messages.length, 0); + + // Produce one message to cause EOF with waiting message when consuming all + producer.produce(topic, null, buffer, null); + consumeAll(); + }); + }; + + consumeNone(); + }); + }); + + it('should emit partition.eof event when reaching end of partition', function(done) { + this.timeout(8000); + crypto.randomBytes(4096, function(ex, buffer) { + producer.setPollInterval(10); + + producer.once('delivery-report', function(err, report) { + t.ifError(err); + }); + + consumer.subscribe([topic]); + + var events = []; + + consumer.once('data', function(msg) { + events.push("data"); + }); + + consumer.once('partition.eof', function(eof) { + events.push("partition.eof"); + }); + + setTimeout(function() { + producer.produce(topic, null, buffer, null); + }, 500) + consumer.setDefaultConsumeTimeout(2000); + consumer.consume(1000, function(err, messages) { + t.ifError(err); + t.equal(messages.length, 1); + t.deepStrictEqual(events, ["data", "partition.eof"]); + done(); + }); + }); + }); + + it('should emit partition.eof when already at end of partition', function(done) { + this.timeout(8000); + crypto.randomBytes(4096, function(ex, buffer) { + producer.setPollInterval(10); + + producer.once('delivery-report', function(err, report) { + t.ifError(err); + }); + + consumer.subscribe([topic]); + + var events = []; + + consumer.once('data', function(msg) { + events.push("data"); + }); + + consumer.on('partition.eof', function(eof) { + events.push("partition.eof"); + }); + + setTimeout(function() { + producer.produce(topic, null, buffer, null); + }, 2000) + consumer.setDefaultConsumeTimeout(3000); + consumer.consume(1000, function(err, messages) { + t.ifError(err); + t.equal(messages.length, 1); + t.deepStrictEqual(events, ["partition.eof", "data", "partition.eof"]); + done(); + }); + }); + }); + + it('should be able to produce and consume messages: consumeLoop', function(done) { + var key = 'key'; + + this.timeout(5000); + + crypto.randomBytes(4096, function(ex, buffer) { + + producer.setPollInterval(10); + + producer.once('delivery-report', function(err, report) { + if (!err) { + t.equal(topic, report.topic, 'invalid delivery-report topic'); + t.equal(key, report.key, 'invalid delivery-report key'); + t.ok(report.offset >= 0, 'invalid delivery-report offset'); + } + }); + + consumer.on('data', function(message) { + t.equal(buffer.toString(), message.value.toString(), 'invalid message value'); + t.equal(key, message.key, 'invalid message key'); + t.equal(topic, message.topic, 'invalid message topic'); + t.ok(message.offset >= 0, 'invalid message offset'); + done(); + }); + + consumer.subscribe([topic]); + consumer.consume(); + + setTimeout(function() { + producer.produce(topic, null, buffer, key); + }, 2000); + + }); + }); + + it('should emit \'partition.eof\' events in consumeLoop', function(done) { + this.timeout(7000); + + crypto.randomBytes(4096, function(ex, buffer) { + producer.setPollInterval(10); + + producer.once('delivery-report', function(err, report) { + t.ifError(err); + }); + + + var events = []; + var offsets = []; + + consumer.on('data', function(message) { + t.equal(message.topic, topic); + t.equal(message.partition, 0); + offsets.push(message.offset); + events.push('data'); + }); + + consumer.on('partition.eof', function(eofEvent) { + t.equal(eofEvent.topic, topic); + t.equal(eofEvent.partition, 0); + offsets.push(eofEvent.offset); + events.push('partition.eof'); + }); + + consumer.subscribe([topic]); + consumer.consume(); + + setTimeout(function() { + producer.produce(topic, null, buffer); + }, 2000); + + setTimeout(function() { + producer.produce(topic, null, buffer); + }, 4000); + + setTimeout(function() { + t.deepStrictEqual(events, ['partition.eof', 'data', 'partition.eof', 'data', 'partition.eof']); + var startOffset = offsets[0]; + t.deepStrictEqual(offsets, + [ startOffset, + startOffset, + startOffset + 1, + startOffset + 1, + startOffset + 2 ]); + done(); + }, 6000); + }); + }); + + it('should emit [warning] event on UNKNOWN_TOPIC_OR_PART error: consumeLoop', function(done) { + consumer.on('warning', function (err) { + if (err.code === Kafka.CODES.ERRORS.ERR_UNKNOWN_TOPIC_OR_PART) { + consumer.disconnect(function() { + done(); + }); + } else { + t.ifError(err); + } + }); + + consumer.subscribe(['non_existing_topic']); + consumer.consume(); + }); + + it('should be able to produce and consume messages with one header value as string: consumeLoop', function(done) { + var headers = [ + { key: "value" } + ]; + this.timeout(5000); + run_headers_test(done, headers); + }); + + it('should be able to produce and consume messages with one header value as buffer: consumeLoop', function(done) { + var headers = [ + { key: Buffer.from('value') } + ]; + this.timeout(5000); + run_headers_test(done, headers); + }); + + it('should be able to produce and consume messages with one header value as int: consumeLoop', function(done) { + var headers = [ + { key: 10 } + ]; + this.timeout(5000); + run_headers_test(done, headers); + }); + + it('should be able to produce and consume messages with one header value as float: consumeLoop', function(done) { + var headers = [ + { key: 1.11 } + ]; + this.timeout(5000); + run_headers_test(done, headers); + }); + + it('should be able to produce and consume messages with multiple headers value as buffer: consumeLoop', function(done) { + var headers = [ + { key1: Buffer.from('value1') }, + { key2: Buffer.from('value2') }, + { key3: Buffer.from('value3') }, + { key4: Buffer.from('value4') }, + ]; + this.timeout(5000); + run_headers_test(done, headers); + }); + + it('should be able to produce and consume messages with multiple headers value as string: consumeLoop', function(done) { + var headers = [ + { key1: 'value1' }, + { key2: 'value2' }, + { key3: 'value3' }, + { key4: 'value4' }, + ]; + this.timeout(5000); + run_headers_test(done, headers); + }); + + it('should be able to produce and consume messages with multiple headers with mixed values: consumeLoop', function(done) { + var headers = [ + { key1: 'value1' }, + { key2: Buffer.from('value2') }, + { key3: 100 }, + { key4: 10.1 }, + ]; + this.timeout(5000); + run_headers_test(done, headers); + }); + + it('should be able to produce and consume messages: empty buffer key and empty value', function(done) { + this.timeout(20000); + var emptyString = ''; + var key = Buffer.from(emptyString); + var value = Buffer.from(''); + + producer.setPollInterval(10); + + consumer.once('data', function(message) { + t.notEqual(message.value, null, 'message should not be null'); + t.equal(value.toString(), message.value.toString(), 'invalid message value'); + t.equal(emptyString, message.key, 'invalid message key'); + done(); + }); + + consumer.subscribe([topic]); + consumer.consume(); + + setTimeout(function() { + producer.produce(topic, null, value, key); + }, 2000); + }); + + it('should be able to produce and consume messages: empty key and empty value', function(done) { + this.timeout(20000); + var key = ''; + var value = Buffer.from(''); + + producer.setPollInterval(10); + + consumer.once('data', function(message) { + t.notEqual(message.value, null, 'message should not be null'); + t.equal(value.toString(), message.value.toString(), 'invalid message value'); + t.equal(key, message.key, 'invalid message key'); + done(); + }); + + consumer.subscribe([topic]); + consumer.consume(); + + setTimeout(function() { + producer.produce(topic, null, value, key); + }, 2000); + }); + + it('should be able to produce and consume messages: null key and null value', function(done) { + this.timeout(20000); + var key = null; + var value = null; + + producer.setPollInterval(10); + + consumer.once('data', function(message) { + t.equal(value, message.value, 'invalid message value'); + t.equal(key, message.key, 'invalid message key'); + done(); + }); + + consumer.subscribe([topic]); + consumer.consume(); + + setTimeout(function() { + producer.produce(topic, null, value, key); + }, 2000); + }); + + describe('Exceptional case - offset_commit_cb true', function() { + var grp = 'kafka-mocha-grp-' + crypto.randomBytes(20).toString('hex'); + var consumerOpts = { + 'metadata.broker.list': kafkaBrokerList, + 'group.id': grp, + 'fetch.wait.max.ms': 1000, + 'session.timeout.ms': 10000, + 'enable.auto.commit': false, + 'debug': 'all', + 'offset_commit_cb': true + }; + + beforeEach(function(done) { + consumer = new Kafka.KafkaConsumer(consumerOpts, { + 'auto.offset.reset': 'largest', + }); + + consumer.connect({}, function(err, d) { + t.ifError(err); + t.equal(typeof d, 'object', 'metadata should be returned'); + done(); + }); + + eventListener(consumer); + }); + + afterEach(function(done) { + this.timeout(10000); + consumer.disconnect(function() { + done(); + }); + }); + + it('should async commit after consuming', function(done) { + this.timeout(25000); + var key = ''; + var value = Buffer.from(''); + + var lastOffset = null; + + consumer.once('data', function(message) { + lastOffset = message.offset; + + // disconnect in offset commit callback + consumer.on('offset.commit', function(err, offsets) { + t.ifError(err); + t.equal(typeof offsets, 'object', 'offsets should be returned'); + + consumer.disconnect(function() { + // reconnect in disconnect callback + consumer.connect({}, function(err, d) { + t.ifError(err); + t.equal(typeof d, 'object', 'metadata should be returned'); + + // check that no new messages arrive, as the offset was committed + consumer.once('data', function(message) { + done(new Error('Should never be here')); + }); + + consumer.subscribe([topic]); + consumer.consume(); + + setTimeout(function() { + done(); + }, 5000); + }); + }); + }); + + consumer.commitMessage(message); + }); + + consumer.subscribe([topic]); + consumer.consume(); + + setTimeout(function() { + producer.produce(topic, null, value, key); + }, 2000); + }); + }); + + describe('Exceptional case - offset_commit_cb function', function() { + var grp = 'kafka-mocha-grp-' + crypto.randomBytes(20).toString('hex'); + + afterEach(function(done) { + this.timeout(10000); + consumer.disconnect(function() { + done(); + }); + }); + + it('should callback offset_commit_cb after commit', function(done) { + this.timeout(20000); + + var consumerOpts = { + 'metadata.broker.list': kafkaBrokerList, + 'group.id': grp, + 'fetch.wait.max.ms': 1000, + 'session.timeout.ms': 10000, + 'enable.auto.commit': false, + 'debug': 'all', + 'offset_commit_cb': function(offset) { + done(); + } + }; + consumer = new Kafka.KafkaConsumer(consumerOpts, { + 'auto.offset.reset': 'largest', + }); + eventListener(consumer); + + consumer.connect({}, function(err, d) { + t.ifError(err); + t.equal(typeof d, 'object', 'metadata should be returned'); + consumer.subscribe([topic]); + consumer.consume(); + setTimeout(function() { + producer.produce(topic, null, Buffer.from(''), ''); + }, 2000); + }); + + consumer.once('data', function(message) { + consumer.commitMessage(message); + }); + }); + }); + + function assert_headers_match(expectedHeaders, messageHeaders) { + t.equal(expectedHeaders.length, messageHeaders.length, 'Headers length does not match expected length'); + for (var i = 0; i < expectedHeaders.length; i++) { + var expectedKey = Object.keys(expectedHeaders[i])[0]; + var messageKey = Object.keys(messageHeaders[i]); + t.equal(messageKey.length, 1, 'Expected only one Header key'); + t.equal(expectedKey, messageKey[0], 'Expected key does not match message key'); + var expectedValue = Buffer.isBuffer(expectedHeaders[i][expectedKey]) ? + expectedHeaders[i][expectedKey].toString() : + expectedHeaders[i][expectedKey]; + var actualValue = messageHeaders[i][expectedKey].toString(); + t.equal(expectedValue, actualValue, 'invalid message header'); + } + } + + function run_headers_test(done, headers) { + var key = 'key'; + + crypto.randomBytes(4096, function(ex, buffer) { + + producer.setPollInterval(10); + + producer.once('delivery-report', function(err, report) { + if (!err) { + t.equal(topic, report.topic, 'invalid delivery-report topic'); + t.equal(key, report.key, 'invalid delivery-report key'); + t.ok(report.offset >= 0, 'invalid delivery-report offset'); + } + }); + + consumer.on('data', function(message) { + t.equal(buffer.toString(), message.value.toString(), 'invalid message value'); + t.equal(key, message.key, 'invalid message key'); + t.equal(topic, message.topic, 'invalid message topic'); + t.ok(message.offset >= 0, 'invalid message offset'); + assert_headers_match(headers, message.headers); + done(); + }); + + consumer.subscribe([topic]); + consumer.consume(); + + setTimeout(function() { + var timestamp = new Date().getTime(); + producer.produce(topic, null, buffer, key, timestamp, "", headers); + }, 2000); + + }); + } + +}); diff --git a/e2e/consumer.spec.js b/e2e/consumer.spec.js new file mode 100644 index 0000000..e2505bb --- /dev/null +++ b/e2e/consumer.spec.js @@ -0,0 +1,350 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var t = require('assert'); +var crypto = require('crypto'); + +var eventListener = require('./listener'); + +var KafkaConsumer = require('../').KafkaConsumer; + +var kafkaBrokerList = process.env.KAFKA_HOST || 'localhost:9092'; +var topic = 'test'; + +describe('Consumer', function() { + var gcfg; + + beforeEach(function() { + var grp = 'kafka-mocha-grp-' + crypto.randomBytes(20).toString('hex'); + gcfg = { + 'bootstrap.servers': kafkaBrokerList, + 'group.id': grp, + 'debug': 'all', + 'rebalance_cb': true, + 'enable.auto.commit': false + }; + }); + + describe('commit', function() { + var consumer; + beforeEach(function(done) { + consumer = new KafkaConsumer(gcfg, {}); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + done(); + }); + + eventListener(consumer); + }); + + it('should allow commit with an array', function(done) { + consumer.commit([{ topic: topic, partition: 0, offset: -1 }]); + done(); + }); + + it('should allow commit without an array', function(done) { + consumer.commit({ topic: topic, partition: 0, offset: -1 }); + done(); + }); + + afterEach(function(done) { + consumer.disconnect(function() { + done(); + }); + }); + }); + + describe('committed and position', function() { + var consumer; + beforeEach(function(done) { + consumer = new KafkaConsumer(gcfg, {}); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + done(); + }); + + eventListener(consumer); + }); + + afterEach(function(done) { + consumer.disconnect(function() { + done(); + }); + }); + + it('before assign, committed offsets are empty', function(done) { + consumer.committed(null, 1000, function(err, committed) { + t.ifError(err); + t.equal(Array.isArray(committed), true, 'Committed offsets should be an array'); + t.equal(committed.length, 0); + done(); + }); + }); + + it('before assign, position returns an empty array', function() { + var position = consumer.position(); + t.equal(Array.isArray(position), true, 'Position should be an array'); + t.equal(position.length, 0); + }); + + it('after assign, should get committed array without offsets ', function(done) { + consumer.assign([{topic:topic, partition:0}]); + // Defer this for a second + setTimeout(function() { + consumer.committed(null, 1000, function(err, committed) { + t.ifError(err); + t.equal(committed.length, 1); + t.equal(typeof committed[0], 'object', 'TopicPartition should be an object'); + t.deepStrictEqual(committed[0].partition, 0); + t.equal(committed[0].offset, undefined); + done(); + }); + }, 1000); + }); + + it('after assign and commit, should get committed offsets', function(done) { + this.timeout(6000); + consumer.assign([{topic:topic, partition:0}]); + consumer.commitSync({topic:topic, partition:0, offset:1000}); + consumer.committed(null, 1000, function(err, committed) { + t.ifError(err); + t.equal(committed.length, 1); + t.equal(typeof committed[0], 'object', 'TopicPartition should be an object'); + t.deepStrictEqual(committed[0].partition, 0); + t.deepStrictEqual(committed[0].offset, 1000); + done(); + }); + }); + + it('after assign, before consume, position should return an array without offsets', function(done) { + consumer.assign([{topic:topic, partition:0}]); + var position = consumer.position(); + t.equal(Array.isArray(position), true, 'Position should be an array'); + t.equal(position.length, 1); + t.equal(typeof position[0], 'object', 'TopicPartition should be an object'); + t.deepStrictEqual(position[0].partition, 0); + t.equal(position[0].offset, undefined, 'before consuming, offset is undefined'); + // see both.spec.js 'should be able to produce, consume messages, read position...' + // for checking of offset numeric value + done(); + }); + + it('should obey the timeout', function(done) { + consumer.committed(null, 0, function(err, committed) { + if (!err) { + t.fail(err, 'not null', 'Error should be set for a timeout'); + } + done(); + }); + }); + + }); + + describe('seek and positioning', function() { + var consumer; + beforeEach(function(done) { + consumer = new KafkaConsumer(gcfg, {}); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + consumer.assign([{ + topic: 'test', + partition: 0, + offset: 0 + }]); + done(); + }); + + eventListener(consumer); + }); + + afterEach(function(done) { + consumer.disconnect(function() { + done(); + }); + }); + + it('should be able to seek', function(cb) { + consumer.seek({ + topic: 'test', + partition: 0, + offset: 0 + }, 1, function(err) { + t.ifError(err); + cb(); + }); + }); + + it('should be able to seek with a timeout of 0', function(cb) { + consumer.seek({ + topic: 'test', + partition: 0, + offset: 0 + }, 0, function(err) { + t.ifError(err); + cb(); + }); + }); + }); + + describe('subscribe', function() { + + var consumer; + beforeEach(function(done) { + consumer = new KafkaConsumer(gcfg, {}); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + done(); + }); + + eventListener(consumer); + }); + + afterEach(function(done) { + consumer.disconnect(function() { + done(); + }); + }); + + it('should be able to subscribe', function() { + t.equal(0, consumer.subscription().length); + consumer.subscribe([topic]); + t.equal(1, consumer.subscription().length); + t.equal('test', consumer.subscription()[0]); + t.equal(0, consumer.assignments().length); + }); + + it('should be able to unsubscribe', function() { + consumer.subscribe([topic]); + t.equal(1, consumer.subscription().length); + consumer.unsubscribe(); + t.equal(0, consumer.subscription().length); + t.equal(0, consumer.assignments().length); + }); + }); + + describe('assign', function() { + + var consumer; + beforeEach(function(done) { + consumer = new KafkaConsumer(gcfg, {}); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + done(); + }); + + eventListener(consumer); + }); + + afterEach(function(done) { + consumer.disconnect(function() { + done(); + }); + }); + + it('should be able to take an assignment', function() { + t.equal(0, consumer.assignments().length); + consumer.assign([{ topic:topic, partition:0 }]); + t.equal(1, consumer.assignments().length); + t.equal(topic, consumer.assignments()[0].topic); + t.equal(0, consumer.subscription().length); + }); + + it('should be able to take an empty assignment', function() { + consumer.assign([{ topic:topic, partition:0 }]); + t.equal(1, consumer.assignments().length); + consumer.assign([]); + t.equal(0, consumer.assignments().length); + }); + }); + + describe('disconnect', function() { + var tcfg = { 'auto.offset.reset': 'earliest' }; + + it('should happen gracefully', function(cb) { + var consumer = new KafkaConsumer(gcfg, tcfg); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + + consumer.disconnect(function() { + cb(); + }); + + }); + + }); + + it('should happen without issue after subscribing', function(cb) { + var consumer = new KafkaConsumer(gcfg, tcfg); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + + consumer.subscribe([topic]); + + consumer.disconnect(function() { + cb(); + }); + + }); + + }); + + it('should happen without issue after consuming', function(cb) { + this.timeout(11000); + + var consumer = new KafkaConsumer(gcfg, tcfg); + consumer.setDefaultConsumeTimeout(10000); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + + consumer.subscribe([topic]); + + consumer.consume(1, function(err, messages) { + t.ifError(err); + + consumer.disconnect(function() { + cb(); + }); + }); + + }); + + }); + + it('should happen without issue after consuming an error', function(cb) { + var consumer = new KafkaConsumer(gcfg, tcfg); + + consumer.setDefaultConsumeTimeout(1); + + consumer.connect({ timeout: 2000 }, function(err, info) { + t.ifError(err); + + consumer.subscribe([topic]); + + consumer.consume(1, function(err, messages) { + + // Timeouts do not classify as errors anymore + t.equal(messages[0], undefined, 'Message should not be set'); + + consumer.disconnect(function() { + cb(); + }); + }); + + }); + }); + + }); +}); diff --git a/e2e/groups.spec.js b/e2e/groups.spec.js new file mode 100644 index 0000000..a05a48e --- /dev/null +++ b/e2e/groups.spec.js @@ -0,0 +1,133 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var crypto = require('crypto'); +var t = require('assert'); + +var Kafka = require('../'); +var kafkaBrokerList = process.env.KAFKA_HOST || 'localhost:9092'; +var eventListener = require('./listener'); + +describe('Consumer group/Producer', function() { + + var producer; + var consumer; + var grp = 'kafka-mocha-grp-' + crypto.randomBytes(20).toString('hex'); + + var config = { + 'metadata.broker.list': kafkaBrokerList, + 'group.id': grp, + 'fetch.wait.max.ms': 1000, + 'session.timeout.ms': 10000, + 'enable.auto.commit': false, + 'debug': 'all' + }; + + beforeEach(function(done) { + producer = new Kafka.Producer({ + 'client.id': 'kafka-mocha', + 'metadata.broker.list': kafkaBrokerList, + 'fetch.wait.max.ms': 1, + 'debug': 'all', + 'dr_cb': true + }); + + producer.connect({}, function(err, d) { + t.ifError(err); + t.equal(typeof d, 'object', 'metadata should be returned'); + done(); + }); + + eventListener(producer); + }); + + beforeEach(function(done) { + consumer = new Kafka.KafkaConsumer(config, { + 'auto.offset.reset': 'largest' + }); + + consumer.connect({}, function(err, d) { + t.ifError(err); + t.equal(typeof d, 'object', 'metadata should be returned'); + done(); + }); + + eventListener(consumer); + }); + + afterEach(function(done) { + producer.disconnect(function() { + done(); + }); + }); + + it('should be able to commit, read committed and restart from the committed offset', function(done) { + this.timeout(30000); + var topic = 'test'; + var key = 'key'; + var payload = Buffer.from('value'); + var count = 0; + var offsets = { + 'first': true + }; + + var tt = setInterval(function() { + try { + producer.produce(topic, null, payload, key); + } catch (e) { + clearInterval(tt); + } + }, 100); + + consumer.on('disconnected', function() { + var consumer2 = new Kafka.KafkaConsumer(config, { + 'auto.offset.reset': 'largest' + }); + + consumer2.on('data', function(message) { + if (offsets.first) { + offsets.first = false; + t.deepStrictEqual(offsets.committed, message.offset, 'Offset read by consumer 2 incorrect'); + clearInterval(tt); + consumer2.unsubscribe(); + consumer2.disconnect(function() { + done(); + }); + } + }); + + consumer2.on('ready', function() { + consumer2.subscribe([topic]); + consumer2.consume(); + }); + consumer2.connect(); + }); + + consumer.on('data', function(message) { + count++; + if (count === 3) { + consumer.commitMessageSync(message); + // test consumer.committed( ) API + consumer.committed(null, 5000, function(err, topicPartitions) { + t.ifError(err); + t.deepStrictEqual(topicPartitions.length, 1); + t.deepStrictEqual(topicPartitions[0].offset, message.offset + 1, 'Offset read by consumer 1 incorrect'); + offsets.committed = message.offset + 1; + consumer.unsubscribe(); + consumer.disconnect(); + }); + + } + }); + + consumer.subscribe([topic]); + consumer.consume(); + }); + +}); diff --git a/e2e/listener.js b/e2e/listener.js new file mode 100644 index 0000000..581baab --- /dev/null +++ b/e2e/listener.js @@ -0,0 +1,51 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +module.exports = eventListener; + +function eventListener(client) { + if (!process.env.DEBUG) { + return; + } + + client + .on('event.error', function (err) { + console.error(err); + }) + .on('event.log', function(event) { + var info = { + severity: event.severity, + fac: event.fac, + }; + if (event.severity >= 7) { + console.error(info, event.message); + } else if (event.severity === 6 || event.severity === 5) { + console.error(info, event.message); + } else if (event.severity === 4) { + console.error(info, event.message); + } else if (event.severity > 0) { + console.error(info, event.message); + } else { + console.error(info, event.message); + } + }) + .on('event.stats', function(event) { + console.log(event, event.message); + }) + .on('event.throttle', function(event) { + console.log(event, '%s#%d throttled.', event.brokerName, event.brokerId); + // event.throttleTime; + }) + .on('event.event', function(event) { + console.log(event, event.message); + }) + .on('ready', function(info) { + console.log('%s connected to kafka server', info.name); + }); + +} diff --git a/e2e/producer-transaction.spec.js b/e2e/producer-transaction.spec.js new file mode 100644 index 0000000..70ed833 --- /dev/null +++ b/e2e/producer-transaction.spec.js @@ -0,0 +1,334 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Kafka = require('../'); + +var kafkaBrokerList = process.env.KAFKA_HOST || 'localhost:9092'; + +describe('Transactional Producer', function () { + this.timeout(5000); + var TRANSACTIONS_TIMEOUT_MS = 30000; + var r = Date.now() + '_' + Math.round(Math.random() * 1000); + var topicIn = 'transaction_input_' + r; + var topicOut = 'transaction_output_' + r; + + var producerTras; + var consumerTrans; + + before(function (done) { + /* + prepare: + transactional consumer (read from input topic) + transactional producer (write to output topic) + write 3 messages to input topic: A, B, C + A will be skipped, B will be committed, C will be aborted + */ + var connecting = 3; + var producerInput; + function connectedCb(err) { + if (err) { + done(err); + return; + } + connecting--; + if (connecting === 0) { + producerInput.produce(topicIn, -1, Buffer.from('A')); + producerInput.produce(topicIn, -1, Buffer.from('B')); + producerInput.produce(topicIn, -1, Buffer.from('C')); + producerInput.disconnect(function (err) { + consumerTrans.subscribe([topicIn]); + done(err); + }) + } + } + producerInput = Kafka.Producer({ + 'client.id': 'kafka-test', + 'metadata.broker.list': kafkaBrokerList, + 'enable.idempotence': true + }); + producerInput.setPollInterval(100); + producerInput.connect({}, connectedCb); + + producerTras = new Kafka.Producer({ + 'client.id': 'kafka-test', + 'metadata.broker.list': kafkaBrokerList, + 'dr_cb': true, + 'debug': 'all', + 'transactional.id': 'noderdkafka_transactions_send_offset', + 'enable.idempotence': true + }); + producerTras.setPollInterval(100); + producerTras.connect({}, connectedCb); + + consumerTrans = new Kafka.KafkaConsumer({ + 'metadata.broker.list': kafkaBrokerList, + 'group.id': 'gropu_transaction_consumer', + 'enable.auto.commit': false + }, { + 'auto.offset.reset': 'earliest', + }); + consumerTrans.connect({}, connectedCb); + }); + + after(function (done) { + let connected = 2; + function execDisconnect(client) { + if (!client.isConnected) { + connected--; + if (connected === 0) { + done(); + } + } else { + client.disconnect(function() { + connected--; + if (connected === 0) { + done(); + } + }); + } + } + execDisconnect(producerTras); + execDisconnect(consumerTrans); + }); + + it('should init transactions', function(done) { + producerTras.initTransactions(TRANSACTIONS_TIMEOUT_MS, function (err) { + done(err); + }); + }); + + it('should complete transaction', function(done) { + function readMessage() { + consumerTrans.consume(1, function(err, m) { + if (err) { + done(err); + return; + } + if (m.length === 0) { + readMessage(); + } else { + var v = m[0].value.toString(); + if (v === 'A') { // skip first message + readMessage(); + return; + } + if (v !== 'B') { + done('Expected B'); + return; + } + producerTras.beginTransaction(function (err) { + if (err) { + done(err); + return; + } + producerTras.produce(topicOut, -1, Buffer.from(v)); + var position = consumerTrans.position(); + producerTras.sendOffsetsToTransaction(position, consumerTrans, function(err) { + if (err) { + done(err); + return; + } + producerTras.commitTransaction(function(err) { + if (err) { + done(err); + return; + } + consumerTrans.committed(5000, function(err, tpo) { + if (err) { + done(err); + return; + } + if (JSON.stringify(position) !== JSON.stringify(tpo)) { + done('Committed mismatch'); + return; + } + done(); + }); + }); + }); + }); + } + }); + } + readMessage(); + }); + + describe('abort transaction', function() { + var lastConsumerTransPosition; + before(function(done) { + function readMessage() { + consumerTrans.consume(1, function(err, m) { + if (err) { + done(err); + return; + } + if (m.length === 0) { + readMessage(); + } else { + var v = m[0].value.toString(); + if (v !== 'C') { + done('Expected C'); + return; + } + producerTras.beginTransaction(function (err) { + if (err) { + done(err); + return; + } + producerTras.produce(topicOut, -1, Buffer.from(v)); + lastConsumerTransPosition = consumerTrans.position(); + producerTras.sendOffsetsToTransaction(lastConsumerTransPosition, consumerTrans, function(err) { + if (err) { + done(err); + return; + } + done(); + }); + }); + } + }); + } + readMessage(); + }); + + it ('should consume committed and uncommitted for read_uncommitted', function(done) { + var allMsgs = []; + var consumer = new Kafka.KafkaConsumer({ + 'metadata.broker.list': kafkaBrokerList, + 'group.id': 'group_read_uncommitted', + 'enable.auto.commit': false, + 'isolation.level': 'read_uncommitted' + }, { + 'auto.offset.reset': 'earliest', + }); + consumer.connect({}, function(err) { + if (err) { + done(err); + return; + } + consumer.subscribe([topicOut]); + consumer.consume(); + }); + consumer.on('data', function(msg) { + var v = msg.value.toString(); + allMsgs.push(v); + // both B and C must be consumed + if (allMsgs.length === 2 && allMsgs[0] === 'B' && allMsgs[1] === 'C') { + consumer.disconnect(function(err) { + if (err) { + done(err); + return; + } + done(); + }) + } + }); + }); + + it ('should consume only committed for read_committed', function(done) { + var allMsgs = []; + var consumer = new Kafka.KafkaConsumer({ + 'metadata.broker.list': kafkaBrokerList, + 'group.id': 'group_read_committed', + 'enable.partition.eof': true, + 'enable.auto.commit': false, + 'isolation.level': 'read_committed' + }, { + 'auto.offset.reset': 'earliest', + }); + consumer.connect({}, function(err) { + if (err) { + done(err); + return; + } + consumer.subscribe([topicOut]); + consumer.consume(); + }); + consumer.on('data', function(msg) { + var v = msg.value.toString(); + allMsgs.push(v); + }); + consumer.on('partition.eof', function(eof) { + if (allMsgs.length === 1 && allMsgs[0] === 'B') { + consumer.disconnect(function(err) { + if (err) { + done(err); + return; + } + done(); + }) + } else { + done('Expected only B'); + return; + } + }); + }); + + it('should abort transaction', function(done) { + producerTras.abortTransaction(function(err) { + if (err) { + done(err); + return; + } + consumerTrans.committed(5000, function(err, tpo) { + if (err) { + done(err); + return; + } + if (lastConsumerTransPosition[0].offset <= tpo[0].offset) { + done('Committed mismatch'); + return; + } + done(); + }); + }); + }); + + it('should consume only committed', function(done) { + var gotB = false; + var consumer = new Kafka.KafkaConsumer({ + 'metadata.broker.list': kafkaBrokerList, + 'group.id': 'group_default', + 'enable.partition.eof': true, + 'enable.auto.commit': false, + }, { + 'auto.offset.reset': 'earliest', + }); + consumer.connect({}, function(err) { + if (err) { + done(err); + return; + } + consumer.subscribe([topicOut]); + consumer.consume(); + }); + consumer.on('data', function(msg) { + var v = msg.value.toString(); + if (v !== 'B') { + done('Expected B'); + return; + } + gotB = true; + }); + consumer.on('partition.eof', function(eof) { + if (!gotB) { + done('Expected B'); + return; + } + consumer.disconnect(function(err) { + if (err) { + done(err); + return; + } + done(); + }); + }); + }); + }); +}); diff --git a/e2e/producer.spec.js b/e2e/producer.spec.js new file mode 100644 index 0000000..8e91479 --- /dev/null +++ b/e2e/producer.spec.js @@ -0,0 +1,300 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Kafka = require('../'); +var t = require('assert'); +var crypto = require('crypto'); + +var eventListener = require('./listener'); + +var kafkaBrokerList = process.env.KAFKA_HOST || 'localhost:9092'; + +var serviceStopped = false; + +describe('Producer', function() { + + var producer; + + describe('with dr_cb', function() { + beforeEach(function(done) { + producer = new Kafka.Producer({ + 'client.id': 'kafka-test', + 'metadata.broker.list': kafkaBrokerList, + 'dr_cb': true, + 'debug': 'all' + }); + producer.connect({}, function(err) { + t.ifError(err); + done(); + }); + + eventListener(producer); + }); + + afterEach(function(done) { + producer.disconnect(function() { + done(); + }); + }); + + it('should connect to Kafka', function(done) { + producer.getMetadata({}, function(err, metadata) { + t.ifError(err); + t.ok(metadata); + + // Ensure it is in the correct format + t.ok(metadata.orig_broker_name, 'Broker name is not set'); + t.notStrictEqual(metadata.orig_broker_id, undefined, 'Broker id is not set'); + t.equal(Array.isArray(metadata.brokers), true); + t.equal(Array.isArray(metadata.topics), true); + + done(); + }); + }); + + it('should produce a message with a null payload and null key', function(done) { + this.timeout(3000); + + var tt = setInterval(function() { + producer.poll(); + }, 200); + + producer.once('delivery-report', function(err, report) { + clearInterval(tt); + t.ifError(err); + t.notStrictEqual(report, undefined); + t.strictEqual(typeof report.topic, 'string'); + t.strictEqual(typeof report.partition, 'number'); + t.strictEqual(typeof report.offset, 'number'); + t.strictEqual( report.key, null); + done(); + }); + + producer.produce('test', null, null, null); + }); + + it('should produce a message with a payload and key', function(done) { + this.timeout(3000); + + var tt = setInterval(function() { + producer.poll(); + }, 200); + + producer.once('delivery-report', function(err, report) { + clearInterval(tt); + t.ifError(err); + t.notStrictEqual(report, undefined); + t.strictEqual(report.value, undefined); + t.strictEqual(typeof report.topic, 'string'); + t.strictEqual(typeof report.partition, 'number'); + t.strictEqual(typeof report.offset, 'number'); + t.equal(report.key, 'key'); + done(); + }); + + producer.produce('test', null, Buffer.from('value'), 'key'); + }); + + it('should produce a message with a payload and key buffer', function(done) { + this.timeout(3000); + + var tt = setInterval(function() { + producer.poll(); + }, 200); + + producer.once('delivery-report', function(err, report) { + clearInterval(tt); + t.ifError(err); + t.notStrictEqual(report, undefined); + t.strictEqual(report.value, undefined); + t.strictEqual(typeof report.topic, 'string'); + t.strictEqual(typeof report.partition, 'number'); + t.strictEqual(typeof report.offset, 'number'); + t.equal(report.key.length > 3, true); + done(); + }); + + producer.produce('test', null, Buffer.from('value'), Buffer.from('key\0s')); + }); + + it('should produce a message with an opaque', function(done) { + this.timeout(3000); + + var tt = setInterval(function() { + producer.poll(); + }, 200); + + producer.once('delivery-report', function(err, report) { + clearInterval(tt); + t.ifError(err); + t.notStrictEqual(report, undefined); + t.strictEqual(typeof report.topic, 'string'); + t.strictEqual(typeof report.partition, 'number'); + t.strictEqual(typeof report.offset, 'number'); + t.equal(report.opaque, 'opaque'); + done(); + }); + + producer.produce('test', null, Buffer.from('value'), null, null, 'opaque'); + }); + + + it('should get 100% deliverability', function(done) { + this.timeout(3000); + + var total = 0; + var max = 10000; + var verified_received = 0; + + var tt = setInterval(function() { + producer.poll(); + }, 200); + + producer + .on('delivery-report', function(err, report) { + t.ifError(err); + t.notStrictEqual(report, undefined); + t.strictEqual(typeof report.topic, 'string'); + t.strictEqual(typeof report.partition, 'number'); + t.strictEqual(typeof report.offset, 'number'); + verified_received++; + if (verified_received === max) { + clearInterval(tt); + done(); + } + }); + + // Produce + for (total = 0; total <= max; total++) { + producer.produce('test', null, Buffer.from('message ' + total), null); + } + + }); + + }); + + describe('with_dr_msg_cb', function() { + beforeEach(function(done) { + producer = new Kafka.Producer({ + 'client.id': 'kafka-test', + 'metadata.broker.list': kafkaBrokerList, + 'dr_msg_cb': true, + 'debug': 'all' + }); + producer.connect({}, function(err) { + t.ifError(err); + done(); + }); + + eventListener(producer); + }); + + afterEach(function(done) { + producer.disconnect(function() { + done(); + }); + }); + + it('should produce a message with a payload and key', function(done) { + this.timeout(3000); + + var tt = setInterval(function() { + producer.poll(); + }, 200); + + producer.once('delivery-report', function(err, report) { + clearInterval(tt); + t.ifError(err); + t.notStrictEqual(report, undefined); + t.strictEqual(typeof report.topic, 'string'); + t.strictEqual(typeof report.partition, 'number'); + t.strictEqual(typeof report.offset, 'number'); + t.ok(report.key.toString(), 'key'); + t.equal(report.value.toString(), 'hai'); + done(); + }); + + producer.produce('test', null, Buffer.from('hai'), 'key'); + }); + + it('should produce a message with an empty payload and empty key (https://github.com/Blizzard/node-rdkafka/issues/117)', function(done) { + this.timeout(3000); + + var tt = setInterval(function() { + producer.poll(); + }, 200); + + producer.once('delivery-report', function(err, report) { + clearInterval(tt); + t.ifError(err); + t.notStrictEqual(report, undefined); + + t.strictEqual(typeof report.topic, 'string'); + t.strictEqual(typeof report.partition, 'number'); + t.strictEqual(typeof report.offset, 'number'); + t.equal(report.key.toString(), '', 'key should be an empty string'); + t.strictEqual(report.value.toString(), '', 'payload should be an empty string'); + done(); + }); + + producer.produce('test', null, Buffer.from(''), ''); + }); + + it('should produce a message with a null payload and null key (https://github.com/Blizzard/node-rdkafka/issues/117)', function(done) { + this.timeout(3000); + + producer.setPollInterval(10); + + producer.once('delivery-report', function(err, report) { + t.ifError(err); + t.notStrictEqual(report, undefined); + + t.strictEqual(typeof report.topic, 'string'); + t.strictEqual(typeof report.partition, 'number'); + t.strictEqual(typeof report.offset, 'number'); + t.strictEqual(report.key, null, 'key should be null'); + t.strictEqual(report.value, null, 'payload should be null'); + done(); + }); + + producer.produce('test', null, null, null); + }); + + it('should produce an int64 key (https://github.com/Blizzard/node-rdkafka/issues/208)', function(done) { + + var v1 = 0x0000000000000084; + var arr = new Uint8Array(8); + arr[0] = 0x00; + arr[1] = 0x00; + arr[2] = 0x00; + arr[3] = 0x00; + arr[4] = 0x00; + arr[5] = 0x00; + arr[6] = 0x00; + arr[7] = 84; + var buf = Buffer.from(arr.buffer); + + producer.setPollInterval(10); + + producer.once('delivery-report', function(err, report) { + t.ifError(err); + t.notStrictEqual(report, undefined); + + t.deepEqual(buf, report.key); + done(); + }); + + producer.produce('test', null, null, Buffer.from(arr.buffer)); + + this.timeout(3000); + }); + + }); + +}); diff --git a/errors.d.ts b/errors.d.ts new file mode 100644 index 0000000..c95bc05 --- /dev/null +++ b/errors.d.ts @@ -0,0 +1,345 @@ +// ====== Generated from librdkafka 1.6.1 file src-cpp/rdkafkacpp.h ====== +export const CODES: { ERRORS: { + /* Internal errors to rdkafka: */ + /** Begin internal error codes (**-200**) */ + ERR__BEGIN: number, + /** Received message is incorrect (**-199**) */ + ERR__BAD_MSG: number, + /** Bad/unknown compression (**-198**) */ + ERR__BAD_COMPRESSION: number, + /** Broker is going away (**-197**) */ + ERR__DESTROY: number, + /** Generic failure (**-196**) */ + ERR__FAIL: number, + /** Broker transport failure (**-195**) */ + ERR__TRANSPORT: number, + /** Critical system resource (**-194**) */ + ERR__CRIT_SYS_RESOURCE: number, + /** Failed to resolve broker (**-193**) */ + ERR__RESOLVE: number, + /** Produced message timed out (**-192**) */ + ERR__MSG_TIMED_OUT: number, + /** Reached the end of the topic+partition queue on + * the broker. Not really an error. + * This event is disabled by default, + * see the `enable.partition.eof` configuration property (**-191**) */ + ERR__PARTITION_EOF: number, + /** Permanent: Partition does not exist in cluster (**-190**) */ + ERR__UNKNOWN_PARTITION: number, + /** File or filesystem error (**-189**) */ + ERR__FS: number, + /** Permanent: Topic does not exist in cluster (**-188**) */ + ERR__UNKNOWN_TOPIC: number, + /** All broker connections are down (**-187**) */ + ERR__ALL_BROKERS_DOWN: number, + /** Invalid argument, or invalid configuration (**-186**) */ + ERR__INVALID_ARG: number, + /** Operation timed out (**-185**) */ + ERR__TIMED_OUT: number, + /** Queue is full (**-184**) */ + ERR__QUEUE_FULL: number, + /** ISR count < required.acks (**-183**) */ + ERR__ISR_INSUFF: number, + /** Broker node update (**-182**) */ + ERR__NODE_UPDATE: number, + /** SSL error (**-181**) */ + ERR__SSL: number, + /** Waiting for coordinator to become available (**-180**) */ + ERR__WAIT_COORD: number, + /** Unknown client group (**-179**) */ + ERR__UNKNOWN_GROUP: number, + /** Operation in progress (**-178**) */ + ERR__IN_PROGRESS: number, + /** Previous operation in progress, wait for it to finish (**-177**) */ + ERR__PREV_IN_PROGRESS: number, + /** This operation would interfere with an existing subscription (**-176**) */ + ERR__EXISTING_SUBSCRIPTION: number, + /** Assigned partitions (rebalance_cb) (**-175**) */ + ERR__ASSIGN_PARTITIONS: number, + /** Revoked partitions (rebalance_cb) (**-174**) */ + ERR__REVOKE_PARTITIONS: number, + /** Conflicting use (**-173**) */ + ERR__CONFLICT: number, + /** Wrong state (**-172**) */ + ERR__STATE: number, + /** Unknown protocol (**-171**) */ + ERR__UNKNOWN_PROTOCOL: number, + /** Not implemented (**-170**) */ + ERR__NOT_IMPLEMENTED: number, + /** Authentication failure (**-169**) */ + ERR__AUTHENTICATION: number, + /** No stored offset (**-168**) */ + ERR__NO_OFFSET: number, + /** Outdated (**-167**) */ + ERR__OUTDATED: number, + /** Timed out in queue (**-166**) */ + ERR__TIMED_OUT_QUEUE: number, + /** Feature not supported by broker (**-165**) */ + ERR__UNSUPPORTED_FEATURE: number, + /** Awaiting cache update (**-164**) */ + ERR__WAIT_CACHE: number, + /** Operation interrupted (**-163**) */ + ERR__INTR: number, + /** Key serialization error (**-162**) */ + ERR__KEY_SERIALIZATION: number, + /** Value serialization error (**-161**) */ + ERR__VALUE_SERIALIZATION: number, + /** Key deserialization error (**-160**) */ + ERR__KEY_DESERIALIZATION: number, + /** Value deserialization error (**-159**) */ + ERR__VALUE_DESERIALIZATION: number, + /** Partial response (**-158**) */ + ERR__PARTIAL: number, + /** Modification attempted on read-only object (**-157**) */ + ERR__READ_ONLY: number, + /** No such entry / item not found (**-156**) */ + ERR__NOENT: number, + /** Read underflow (**-155**) */ + ERR__UNDERFLOW: number, + /** Invalid type (**-154**) */ + ERR__INVALID_TYPE: number, + /** Retry operation (**-153**) */ + ERR__RETRY: number, + /** Purged in queue (**-152**) */ + ERR__PURGE_QUEUE: number, + /** Purged in flight (**-151**) */ + ERR__PURGE_INFLIGHT: number, + /** Fatal error: see RdKafka::Handle::fatal_error() (**-150**) */ + ERR__FATAL: number, + /** Inconsistent state (**-149**) */ + ERR__INCONSISTENT: number, + /** Gap-less ordering would not be guaranteed if proceeding (**-148**) */ + ERR__GAPLESS_GUARANTEE: number, + /** Maximum poll interval exceeded (**-147**) */ + ERR__MAX_POLL_EXCEEDED: number, + /** Unknown broker (**-146**) */ + ERR__UNKNOWN_BROKER: number, + /** Functionality not configured (**-145**) */ + ERR__NOT_CONFIGURED: number, + /** Instance has been fenced (**-144**) */ + ERR__FENCED: number, + /** Application generated error (**-143**) */ + ERR__APPLICATION: number, + /** Assignment lost (**-142**) */ + ERR__ASSIGNMENT_LOST: number, + /** No operation performed (**-141**) */ + ERR__NOOP: number, + /** No offset to automatically reset to (**-140**) */ + ERR__AUTO_OFFSET_RESET: number, + /** End internal error codes (**-100**) */ + ERR__END: number, + /* Kafka broker errors: */ + /** Unknown broker error (**-1**) */ + ERR_UNKNOWN: number, + /** Success (**0**) */ + ERR_NO_ERROR: number, + /** Offset out of range (**1**) */ + ERR_OFFSET_OUT_OF_RANGE: number, + /** Invalid message (**2**) */ + ERR_INVALID_MSG: number, + /** Unknown topic or partition (**3**) */ + ERR_UNKNOWN_TOPIC_OR_PART: number, + /** Invalid message size (**4**) */ + ERR_INVALID_MSG_SIZE: number, + /** Leader not available (**5**) */ + ERR_LEADER_NOT_AVAILABLE: number, + /** Not leader for partition (**6**) */ + ERR_NOT_LEADER_FOR_PARTITION: number, + /** Request timed out (**7**) */ + ERR_REQUEST_TIMED_OUT: number, + /** Broker not available (**8**) */ + ERR_BROKER_NOT_AVAILABLE: number, + /** Replica not available (**9**) */ + ERR_REPLICA_NOT_AVAILABLE: number, + /** Message size too large (**10**) */ + ERR_MSG_SIZE_TOO_LARGE: number, + /** StaleControllerEpochCode (**11**) */ + ERR_STALE_CTRL_EPOCH: number, + /** Offset metadata string too large (**12**) */ + ERR_OFFSET_METADATA_TOO_LARGE: number, + /** Broker disconnected before response received (**13**) */ + ERR_NETWORK_EXCEPTION: number, + /** Coordinator load in progress (**14**) */ + ERR_COORDINATOR_LOAD_IN_PROGRESS: number, + /** Group coordinator load in progress (**14**) */ + ERR_GROUP_LOAD_IN_PROGRESS: number, + /** Coordinator not available (**15**) */ + ERR_COORDINATOR_NOT_AVAILABLE: number, + /** Group coordinator not available (**15**) */ + ERR_GROUP_COORDINATOR_NOT_AVAILABLE: number, + /** Not coordinator (**16**) */ + ERR_NOT_COORDINATOR: number, + /** Not coordinator for group (**16**) */ + ERR_NOT_COORDINATOR_FOR_GROUP: number, + /** Invalid topic (**17**) */ + ERR_TOPIC_EXCEPTION: number, + /** Message batch larger than configured server segment size (**18**) */ + ERR_RECORD_LIST_TOO_LARGE: number, + /** Not enough in-sync replicas (**19**) */ + ERR_NOT_ENOUGH_REPLICAS: number, + /** Message(s) written to insufficient number of in-sync replicas (**20**) */ + ERR_NOT_ENOUGH_REPLICAS_AFTER_APPEND: number, + /** Invalid required acks value (**21**) */ + ERR_INVALID_REQUIRED_ACKS: number, + /** Specified group generation id is not valid (**22**) */ + ERR_ILLEGAL_GENERATION: number, + /** Inconsistent group protocol (**23**) */ + ERR_INCONSISTENT_GROUP_PROTOCOL: number, + /** Invalid group.id (**24**) */ + ERR_INVALID_GROUP_ID: number, + /** Unknown member (**25**) */ + ERR_UNKNOWN_MEMBER_ID: number, + /** Invalid session timeout (**26**) */ + ERR_INVALID_SESSION_TIMEOUT: number, + /** Group rebalance in progress (**27**) */ + ERR_REBALANCE_IN_PROGRESS: number, + /** Commit offset data size is not valid (**28**) */ + ERR_INVALID_COMMIT_OFFSET_SIZE: number, + /** Topic authorization failed (**29**) */ + ERR_TOPIC_AUTHORIZATION_FAILED: number, + /** Group authorization failed (**30**) */ + ERR_GROUP_AUTHORIZATION_FAILED: number, + /** Cluster authorization failed (**31**) */ + ERR_CLUSTER_AUTHORIZATION_FAILED: number, + /** Invalid timestamp (**32**) */ + ERR_INVALID_TIMESTAMP: number, + /** Unsupported SASL mechanism (**33**) */ + ERR_UNSUPPORTED_SASL_MECHANISM: number, + /** Illegal SASL state (**34**) */ + ERR_ILLEGAL_SASL_STATE: number, + /** Unuspported version (**35**) */ + ERR_UNSUPPORTED_VERSION: number, + /** Topic already exists (**36**) */ + ERR_TOPIC_ALREADY_EXISTS: number, + /** Invalid number of partitions (**37**) */ + ERR_INVALID_PARTITIONS: number, + /** Invalid replication factor (**38**) */ + ERR_INVALID_REPLICATION_FACTOR: number, + /** Invalid replica assignment (**39**) */ + ERR_INVALID_REPLICA_ASSIGNMENT: number, + /** Invalid config (**40**) */ + ERR_INVALID_CONFIG: number, + /** Not controller for cluster (**41**) */ + ERR_NOT_CONTROLLER: number, + /** Invalid request (**42**) */ + ERR_INVALID_REQUEST: number, + /** Message format on broker does not support request (**43**) */ + ERR_UNSUPPORTED_FOR_MESSAGE_FORMAT: number, + /** Policy violation (**44**) */ + ERR_POLICY_VIOLATION: number, + /** Broker received an out of order sequence number (**45**) */ + ERR_OUT_OF_ORDER_SEQUENCE_NUMBER: number, + /** Broker received a duplicate sequence number (**46**) */ + ERR_DUPLICATE_SEQUENCE_NUMBER: number, + /** Producer attempted an operation with an old epoch (**47**) */ + ERR_INVALID_PRODUCER_EPOCH: number, + /** Producer attempted a transactional operation in an invalid state (**48**) */ + ERR_INVALID_TXN_STATE: number, + /** Producer attempted to use a producer id which is not + * currently assigned to its transactional id (**49**) */ + ERR_INVALID_PRODUCER_ID_MAPPING: number, + /** Transaction timeout is larger than the maximum + * value allowed by the broker's max.transaction.timeout.ms (**50**) */ + ERR_INVALID_TRANSACTION_TIMEOUT: number, + /** Producer attempted to update a transaction while another + * concurrent operation on the same transaction was ongoing (**51**) */ + ERR_CONCURRENT_TRANSACTIONS: number, + /** Indicates that the transaction coordinator sending a + * WriteTxnMarker is no longer the current coordinator for a + * given producer (**52**) */ + ERR_TRANSACTION_COORDINATOR_FENCED: number, + /** Transactional Id authorization failed (**53**) */ + ERR_TRANSACTIONAL_ID_AUTHORIZATION_FAILED: number, + /** Security features are disabled (**54**) */ + ERR_SECURITY_DISABLED: number, + /** Operation not attempted (**55**) */ + ERR_OPERATION_NOT_ATTEMPTED: number, + /** Disk error when trying to access log file on the disk (**56**) */ + ERR_KAFKA_STORAGE_ERROR: number, + /** The user-specified log directory is not found in the broker config (**57**) */ + ERR_LOG_DIR_NOT_FOUND: number, + /** SASL Authentication failed (**58**) */ + ERR_SASL_AUTHENTICATION_FAILED: number, + /** Unknown Producer Id (**59**) */ + ERR_UNKNOWN_PRODUCER_ID: number, + /** Partition reassignment is in progress (**60**) */ + ERR_REASSIGNMENT_IN_PROGRESS: number, + /** Delegation Token feature is not enabled (**61**) */ + ERR_DELEGATION_TOKEN_AUTH_DISABLED: number, + /** Delegation Token is not found on server (**62**) */ + ERR_DELEGATION_TOKEN_NOT_FOUND: number, + /** Specified Principal is not valid Owner/Renewer (**63**) */ + ERR_DELEGATION_TOKEN_OWNER_MISMATCH: number, + /** Delegation Token requests are not allowed on this connection (**64**) */ + ERR_DELEGATION_TOKEN_REQUEST_NOT_ALLOWED: number, + /** Delegation Token authorization failed (**65**) */ + ERR_DELEGATION_TOKEN_AUTHORIZATION_FAILED: number, + /** Delegation Token is expired (**66**) */ + ERR_DELEGATION_TOKEN_EXPIRED: number, + /** Supplied principalType is not supported (**67**) */ + ERR_INVALID_PRINCIPAL_TYPE: number, + /** The group is not empty (**68**) */ + ERR_NON_EMPTY_GROUP: number, + /** The group id does not exist (**69**) */ + ERR_GROUP_ID_NOT_FOUND: number, + /** The fetch session ID was not found (**70**) */ + ERR_FETCH_SESSION_ID_NOT_FOUND: number, + /** The fetch session epoch is invalid (**71**) */ + ERR_INVALID_FETCH_SESSION_EPOCH: number, + /** No matching listener (**72**) */ + ERR_LISTENER_NOT_FOUND: number, + /** Topic deletion is disabled (**73**) */ + ERR_TOPIC_DELETION_DISABLED: number, + /** Leader epoch is older than broker epoch (**74**) */ + ERR_FENCED_LEADER_EPOCH: number, + /** Leader epoch is newer than broker epoch (**75**) */ + ERR_UNKNOWN_LEADER_EPOCH: number, + /** Unsupported compression type (**76**) */ + ERR_UNSUPPORTED_COMPRESSION_TYPE: number, + /** Broker epoch has changed (**77**) */ + ERR_STALE_BROKER_EPOCH: number, + /** Leader high watermark is not caught up (**78**) */ + ERR_OFFSET_NOT_AVAILABLE: number, + /** Group member needs a valid member ID (**79**) */ + ERR_MEMBER_ID_REQUIRED: number, + /** Preferred leader was not available (**80**) */ + ERR_PREFERRED_LEADER_NOT_AVAILABLE: number, + /** Consumer group has reached maximum size (**81**) */ + ERR_GROUP_MAX_SIZE_REACHED: number, + /** Static consumer fenced by other consumer with same + * group.instance.id (**82**) */ + ERR_FENCED_INSTANCE_ID: number, + /** Eligible partition leaders are not available (**83**) */ + ERR_ELIGIBLE_LEADERS_NOT_AVAILABLE: number, + /** Leader election not needed for topic partition (**84**) */ + ERR_ELECTION_NOT_NEEDED: number, + /** No partition reassignment is in progress (**85**) */ + ERR_NO_REASSIGNMENT_IN_PROGRESS: number, + /** Deleting offsets of a topic while the consumer group is + * subscribed to it (**86**) */ + ERR_GROUP_SUBSCRIBED_TO_TOPIC: number, + /** Broker failed to validate record (**87**) */ + ERR_INVALID_RECORD: number, + /** There are unstable offsets that need to be cleared (**88**) */ + ERR_UNSTABLE_OFFSET_COMMIT: number, + /** Throttling quota has been exceeded (**89**) */ + ERR_THROTTLING_QUOTA_EXCEEDED: number, + /** There is a newer producer with the same transactionalId + * which fences the current one (**90**) */ + ERR_PRODUCER_FENCED: number, + /** Request illegally referred to resource that does not exist (**91**) */ + ERR_RESOURCE_NOT_FOUND: number, + /** Request illegally referred to the same resource twice (**92**) */ + ERR_DUPLICATE_RESOURCE: number, + /** Requested credential would not meet criteria for acceptability (**93**) */ + ERR_UNACCEPTABLE_CREDENTIAL: number, + /** Indicates that the either the sender or recipient of a + * voter-only request is not one of the expected voters (**94**) */ + ERR_INCONSISTENT_VOTER_SET: number, + /** Invalid update version (**95**) */ + ERR_INVALID_UPDATE_VERSION: number, + /** Unable to update finalized features due to server error (**96**) */ + ERR_FEATURE_UPDATE_FAILED: number, + /** Request principal deserialization failed during forwarding (**97**) */ + ERR_PRINCIPAL_DESERIALIZATION_FAILURE: number, +}} \ No newline at end of file diff --git a/index.d.ts b/index.d.ts new file mode 100644 index 0000000..d7ce7e6 --- /dev/null +++ b/index.d.ts @@ -0,0 +1,347 @@ +import { Readable, ReadableOptions, Writable, WritableOptions } from 'stream'; +import { EventEmitter } from 'events'; +import { + GlobalConfig, + TopicConfig, + ConsumerGlobalConfig, + ConsumerTopicConfig, + ProducerGlobalConfig, + ProducerTopicConfig, +} from './config'; + +export * from './config'; +export * from './errors'; + +export interface LibrdKafkaError { + message: string; + code: number; + errno: number; + origin: string; + stack?: string; + isFatal?: boolean; + isRetriable?: boolean; + isTxnRequiresAbort?: boolean; +} + +export interface ReadyInfo { + name: string; +} + +export interface ClientMetrics { + connectionOpened: number; +} + +export interface MetadataOptions { + topic?: string; + allTopics?: boolean; + timeout?: number; +} + +export interface BrokerMetadata { + id: number; + host: string; + port: number; +} + +export interface PartitionMetadata { + id: number; + leader: number; + replicas: number[]; + isrs: number[]; +} + +export interface TopicMetadata { + name: string; + partitions: PartitionMetadata[]; +} + +export interface Metadata { + orig_broker_id: number; + orig_broker_name: string; + topics: TopicMetadata[]; + brokers: BrokerMetadata[]; +} + +export interface WatermarkOffsets{ + lowOffset: number; + highOffset: number; +} + +export interface TopicPartition { + topic: string; + partition: number; +} + +export interface TopicPartitionOffset extends TopicPartition{ + offset: number; +} + +export type TopicPartitionTime = TopicPartitionOffset; + +export type EofEvent = TopicPartitionOffset; + +export type Assignment = TopicPartition | TopicPartitionOffset; + +export interface DeliveryReport extends TopicPartitionOffset { + value?: MessageValue; + size: number; + key?: MessageKey; + timestamp?: number; + opaque?: any; +} + +export type NumberNullUndefined = number | null | undefined; + +export type MessageKey = Buffer | string | null | undefined; +export type MessageHeader = { [key: string]: string | Buffer }; +export type MessageValue = Buffer | null; +export type SubscribeTopic = string | RegExp; +export type SubscribeTopicList = SubscribeTopic[]; + +export interface Message extends TopicPartitionOffset { + value: MessageValue; + size: number; + topic: string; + key?: MessageKey; + timestamp?: number; + headers?: MessageHeader[]; + opaque?: any; +} + +export interface ReadStreamOptions extends ReadableOptions { + topics: SubscribeTopicList | SubscribeTopic | ((metadata: Metadata) => SubscribeTopicList); + waitInterval?: number; + fetchSize?: number; + objectMode?: boolean; + highWaterMark?: number; + autoClose?: boolean; + streamAsBatch?: boolean; + connectOptions?: any; +} + +export interface WriteStreamOptions extends WritableOptions { + encoding?: string; + objectMode?: boolean; + topic?: string; + autoClose?: boolean; + pollInterval?: number; + connectOptions?: any; +} + +export interface ProducerStream extends Writable { + producer: Producer; + connect(metadataOptions?: MetadataOptions): void; + close(cb?: () => void): void; +} + +export interface ConsumerStream extends Readable { + consumer: KafkaConsumer; + connect(options: ConsumerGlobalConfig): void; + close(cb?: () => void): void; +} + +type KafkaClientEvents = 'disconnected' | 'ready' | 'connection.failure' | 'event.error' | 'event.stats' | 'event.log' | 'event.event' | 'event.throttle'; +type KafkaConsumerEvents = 'data' | 'partition.eof' | 'rebalance' | 'rebalance.error' | 'subscribed' | 'unsubscribed' | 'unsubscribe' | 'offset.commit' | KafkaClientEvents; +type KafkaProducerEvents = 'delivery-report' | KafkaClientEvents; + +type EventListenerMap = { + // ### Client + // connectivity events + 'disconnected': (metrics: ClientMetrics) => void, + 'ready': (info: ReadyInfo, metadata: Metadata) => void, + 'connection.failure': (error: LibrdKafkaError, metrics: ClientMetrics) => void, + // event messages + 'event.error': (error: LibrdKafkaError) => void, + 'event.stats': (eventData: any) => void, + 'event.log': (eventData: any) => void, + 'event.event': (eventData: any) => void, + 'event.throttle': (eventData: any) => void, + // ### Consumer only + // domain events + 'data': (arg: Message) => void, + 'partition.eof': (arg: EofEvent) => void, + 'rebalance': (err: LibrdKafkaError, assignments: TopicPartition[]) => void, + 'rebalance.error': (err: Error) => void, + // connectivity events + 'subscribed': (topics: SubscribeTopicList) => void, + 'unsubscribe': () => void, + 'unsubscribed': () => void, + // offsets + 'offset.commit': (error: LibrdKafkaError, topicPartitions: TopicPartitionOffset[]) => void, + // ### Producer only + // delivery + 'delivery-report': (error: LibrdKafkaError, report: DeliveryReport) => void, +} + +type EventListener = K extends keyof EventListenerMap ? EventListenerMap[K] : never; + +export abstract class Client extends EventEmitter { + constructor(globalConf: GlobalConfig, SubClientType: any, topicConf: TopicConfig); + + connect(metadataOptions?: MetadataOptions, cb?: (err: LibrdKafkaError, data: Metadata) => any): this; + + getClient(): any; + + connectedTime(): number; + + getLastError(): LibrdKafkaError; + + disconnect(cb?: (err: any, data: ClientMetrics) => any): this; + disconnect(timeout: number, cb?: (err: any, data: ClientMetrics) => any): this; + + isConnected(): boolean; + + getMetadata(metadataOptions?: MetadataOptions, cb?: (err: LibrdKafkaError, data: Metadata) => any): any; + + queryWatermarkOffsets(topic: string, partition: number, timeout: number, cb?: (err: LibrdKafkaError, offsets: WatermarkOffsets) => any): any; + queryWatermarkOffsets(topic: string, partition: number, cb?: (err: LibrdKafkaError, offsets: WatermarkOffsets) => any): any; + + on(event: E, listener: EventListener): this; + once(event: E, listener: EventListener): this; +} + +export class KafkaConsumer extends Client { + constructor(conf: ConsumerGlobalConfig, topicConf: ConsumerTopicConfig); + + assign(assignments: Assignment[]): this; + + assignments(): Assignment[]; + + commit(topicPartition: TopicPartitionOffset | TopicPartitionOffset[]): this; + commit(): this; + + commitMessage(msg: TopicPartitionOffset): this; + + commitMessageSync(msg: TopicPartitionOffset): this; + + commitSync(topicPartition: TopicPartitionOffset | TopicPartitionOffset[]): this; + + committed(toppars: TopicPartition[], timeout: number, cb: (err: LibrdKafkaError, topicPartitions: TopicPartitionOffset[]) => void): this; + committed(timeout: number, cb: (err: LibrdKafkaError, topicPartitions: TopicPartitionOffset[]) => void): this; + + consume(number: number, cb?: (err: LibrdKafkaError, messages: Message[]) => void): void; + consume(cb: (err: LibrdKafkaError, messages: Message[]) => void): void; + consume(): void; + + getWatermarkOffsets(topic: string, partition: number): WatermarkOffsets; + + offsetsStore(topicPartitions: TopicPartitionOffset[]): any; + + pause(topicPartitions: TopicPartition[]): any; + + position(toppars?: TopicPartition[]): TopicPartitionOffset[]; + + resume(topicPartitions: TopicPartition[]): any; + + seek(toppar: TopicPartitionOffset, timeout: number | null, cb: (err: LibrdKafkaError) => void): this; + + setDefaultConsumeTimeout(timeoutMs: number): void; + + setDefaultConsumeLoopTimeoutDelay(timeoutMs: number): void; + + subscribe(topics: SubscribeTopicList): this; + + subscription(): string[]; + + unassign(): this; + + unsubscribe(): this; + + offsetsForTimes(topicPartitions: TopicPartitionTime[], timeout: number, cb?: (err: LibrdKafkaError, offsets: TopicPartitionOffset[]) => any): void; + offsetsForTimes(topicPartitions: TopicPartitionTime[], cb?: (err: LibrdKafkaError, offsets: TopicPartitionOffset[]) => any): void; + + static createReadStream(conf: ConsumerGlobalConfig, topicConfig: ConsumerTopicConfig, streamOptions: ReadStreamOptions | number): ConsumerStream; +} + +export class Producer extends Client { + constructor(conf: ProducerGlobalConfig, topicConf?: ProducerTopicConfig); + + flush(timeout?: NumberNullUndefined, cb?: (err: LibrdKafkaError) => void): this; + + poll(): this; + + produce(topic: string, partition: NumberNullUndefined, message: MessageValue, key?: MessageKey, timestamp?: NumberNullUndefined, opaque?: any, headers?: MessageHeader[]): any; + + setPollInterval(interval: number): this; + + static createWriteStream(conf: ProducerGlobalConfig, topicConf: ProducerTopicConfig, streamOptions: WriteStreamOptions): ProducerStream; + + initTransactions(cb: (err: LibrdKafkaError) => void): void; + initTransactions(timeout: number, cb: (err: LibrdKafkaError) => void): void; + beginTransaction(cb: (err: LibrdKafkaError) => void): void; + commitTransaction(cb: (err: LibrdKafkaError) => void): void; + commitTransaction(timeout: number, cb: (err: LibrdKafkaError) => void): void; + abortTransaction(cb: (err: LibrdKafkaError) => void): void; + abortTransaction(timeout: number, cb: (err: LibrdKafkaError) => void): void; + sendOffsetsToTransaction(offsets: TopicPartitionOffset[], consumer: KafkaConsumer, cb: (err: LibrdKafkaError) => void): void; + sendOffsetsToTransaction(offsets: TopicPartitionOffset[], consumer: KafkaConsumer, timeout: number, cb: (err: LibrdKafkaError) => void): void; +} + +export class HighLevelProducer extends Producer { + produce(topic: string, partition: NumberNullUndefined, message: any, key: any, timestamp: NumberNullUndefined, callback: (err: any, offset?: NumberNullUndefined) => void): any; + produce(topic: string, partition: NumberNullUndefined, message: any, key: any, timestamp: NumberNullUndefined, headers: MessageHeader[], callback: (err: any, offset?: NumberNullUndefined) => void): any; + + setKeySerializer(serializer: (key: any, cb: (err: any, key: MessageKey) => void) => void): void; + setKeySerializer(serializer: (key: any) => MessageKey | Promise): void; + setValueSerializer(serializer: (value: any, cb: (err: any, value: MessageValue) => void) => void): void; + setValueSerializer(serializer: (value: any) => MessageValue | Promise): void; +} + +export const features: string[]; + +export const librdkafkaVersion: string; + +export function createReadStream(conf: ConsumerGlobalConfig, topicConf: ConsumerTopicConfig, streamOptions: ReadStreamOptions | number): ConsumerStream; + +export function createWriteStream(conf: ProducerGlobalConfig, topicConf: ProducerTopicConfig, streamOptions: WriteStreamOptions): ProducerStream; + +export interface NewTopic { + topic: string; + num_partitions: number; + replication_factor: number; + config?: { + 'cleanup.policy'?: 'delete' | 'compact' | 'delete,compact' | 'compact,delete'; + 'compression.type'?: 'gzip' | 'snappy' | 'lz4' | 'zstd' | 'uncompressed' | 'producer'; + 'delete.retention.ms'?: string; + 'file.delete.delay.ms'?: string; + 'flush.messages'?: string; + 'flush.ms'?: string; + 'follower.replication.throttled.replicas'?: string; + 'index.interval.bytes'?: string; + 'leader.replication.throttled.replicas'?: string; + 'max.compaction.lag.ms'?: string; + 'max.message.bytes'?: string; + 'message.format.version'?: string; + 'message.timestamp.difference.max.ms'?: string; + 'message.timestamp.type'?: string; + 'min.cleanable.dirty.ratio'?: string; + 'min.compaction.lag.ms'?: string; + 'min.insync.replicas'?: string; + 'preallocate'?: string; + 'retention.bytes'?: string; + 'retention.ms'?: string; + 'segment.bytes'?: string; + 'segment.index.bytes'?: string; + 'segment.jitter.ms'?: string; + 'segment.ms'?: string; + 'unclean.leader.election.enable'?: string; + 'message.downconversion.enable'?: string; + } | { [cfg: string]: string; }; +} + +export interface IAdminClient { + createTopic(topic: NewTopic, cb?: (err: LibrdKafkaError) => void): void; + createTopic(topic: NewTopic, timeout?: number, cb?: (err: LibrdKafkaError) => void): void; + + deleteTopic(topic: string, cb?: (err: LibrdKafkaError) => void): void; + deleteTopic(topic: string, timeout?: number, cb?: (err: LibrdKafkaError) => void): void; + + createPartitions(topic: string, desiredPartitions: number, cb?: (err: LibrdKafkaError) => void): void; + createPartitions(topic: string, desiredPartitions: number, timeout?: number, cb?: (err: LibrdKafkaError) => void): void; + + disconnect(): void; +} + +export abstract class AdminClient { + static create(conf: GlobalConfig): IAdminClient; +} diff --git a/lib/admin.js b/lib/admin.js new file mode 100644 index 0000000..773dc95 --- /dev/null +++ b/lib/admin.js @@ -0,0 +1,213 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ +'use strict'; + +module.exports = { + create: createAdminClient, +}; + +var Client = require('./client'); +var util = require('util'); +var Kafka = require('../librdkafka'); +var LibrdKafkaError = require('./error'); +var shallowCopy = require('./util').shallowCopy; + +/** + * Create a new AdminClient for making topics, partitions, and more. + * + * This is a factory method because it immediately starts an + * active handle with the brokers. + * + */ +function createAdminClient(conf) { + var client = new AdminClient(conf); + + // Wrap the error so we throw if it failed with some context + LibrdKafkaError.wrap(client.connect(), true); + + // Return the client if we succeeded + return client; +} + +/** + * AdminClient class for administering Kafka + * + * This client is the way you can interface with the Kafka Admin APIs. + * This class should not be made using the constructor, but instead + * should be made using the factory method. + * + * + * var client = AdminClient.create({ ... }); + * + * + * Once you instantiate this object, it will have a handle to the kafka broker. + * Unlike the other node-rdkafka classes, this class does not ensure that + * it is connected to the upstream broker. Instead, making an action will + * validate that. + * + * @param {object} conf - Key value pairs to configure the admin client + * topic configuration + * @constructor + */ +function AdminClient(conf) { + if (!(this instanceof AdminClient)) { + return new AdminClient(conf); + } + + conf = shallowCopy(conf); + + /** + * NewTopic model. + * + * This is the representation of a new message that is requested to be made + * using the Admin client. + * + * @typedef {object} AdminClient~NewTopic + * @property {string} topic - the topic name to create + * @property {number} num_partitions - the number of partitions to give the topic + * @property {number} replication_factor - the replication factor of the topic + * @property {object} config - a list of key values to be passed as configuration + * for the topic. + */ + + this._client = new Kafka.AdminClient(conf); + this._isConnected = false; + this.globalConfig = conf; +} + +/** + * Connect using the admin client. + * + * Should be run using the factory method, so should never + * need to be called outside. + * + * Unlike the other connect methods, this one is synchronous. + */ +AdminClient.prototype.connect = function() { + LibrdKafkaError.wrap(this._client.connect(), true); + this._isConnected = true; +}; + +/** + * Disconnect the admin client. + * + * This is a synchronous method, but all it does is clean up + * some memory and shut some threads down + */ +AdminClient.prototype.disconnect = function() { + LibrdKafkaError.wrap(this._client.disconnect(), true); + this._isConnected = false; +}; + +/** + * Create a topic with a given config. + * + * @param {NewTopic} topic - Topic to create. + * @param {number} timeout - Number of milliseconds to wait while trying to create the topic. + * @param {function} cb - The callback to be executed when finished + */ +AdminClient.prototype.createTopic = function(topic, timeout, cb) { + if (!this._isConnected) { + throw new Error('Client is disconnected'); + } + + if (typeof timeout === 'function') { + cb = timeout; + timeout = 5000; + } + + if (!timeout) { + timeout = 5000; + } + + this._client.createTopic(topic, timeout, function(err) { + if (err) { + if (cb) { + cb(LibrdKafkaError.create(err)); + } + return; + } + + if (cb) { + cb(); + } + }); +}; + +/** + * Delete a topic. + * + * @param {string} topic - The topic to delete, by name. + * @param {number} timeout - Number of milliseconds to wait while trying to delete the topic. + * @param {function} cb - The callback to be executed when finished + */ +AdminClient.prototype.deleteTopic = function(topic, timeout, cb) { + if (!this._isConnected) { + throw new Error('Client is disconnected'); + } + + if (typeof timeout === 'function') { + cb = timeout; + timeout = 5000; + } + + if (!timeout) { + timeout = 5000; + } + + this._client.deleteTopic(topic, timeout, function(err) { + if (err) { + if (cb) { + cb(LibrdKafkaError.create(err)); + } + return; + } + + if (cb) { + cb(); + } + }); +}; + +/** + * Create new partitions for a topic. + * + * @param {string} topic - The topic to add partitions to, by name. + * @param {number} totalPartitions - The total number of partitions the topic should have + * after the request + * @param {number} timeout - Number of milliseconds to wait while trying to create the partitions. + * @param {function} cb - The callback to be executed when finished + */ +AdminClient.prototype.createPartitions = function(topic, totalPartitions, timeout, cb) { + if (!this._isConnected) { + throw new Error('Client is disconnected'); + } + + if (typeof timeout === 'function') { + cb = timeout; + timeout = 5000; + } + + if (!timeout) { + timeout = 5000; + } + + this._client.createPartitions(topic, totalPartitions, timeout, function(err) { + if (err) { + if (cb) { + cb(LibrdKafkaError.create(err)); + } + return; + } + + if (cb) { + cb(); + } + }); +}; diff --git a/lib/client.js b/lib/client.js new file mode 100644 index 0000000..9cbd3f9 --- /dev/null +++ b/lib/client.js @@ -0,0 +1,554 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +module.exports = Client; + +var Emitter = require('events').EventEmitter; +var util = require('util'); +var Kafka = require('../librdkafka.js'); +var assert = require('assert'); + +var LibrdKafkaError = require('./error'); + +util.inherits(Client, Emitter); + +/** + * Base class for Consumer and Producer + * + * This should not be created independently, but rather is + * the base class on which both producer and consumer + * get their common functionality. + * + * @param {object} globalConf - Global configuration in key value pairs. + * @param {function} SubClientType - The function representing the subclient + * type. In C++ land this needs to be a class that inherits from Connection. + * @param {object} topicConf - Topic configuration in key value pairs + * @constructor + * @extends Emitter + */ +function Client(globalConf, SubClientType, topicConf) { + if (!(this instanceof Client)) { + return new Client(globalConf, SubClientType, topicConf); + } + + Emitter.call(this); + + // This superclass must be initialized with the Kafka.{Producer,Consumer} + // @example var client = new Client({}, Kafka.Producer); + // remember this is a superclass so this will get taken care of in + // the producer and consumer main wrappers + + var no_event_cb = globalConf.event_cb === false; + topicConf = topicConf || {}; + + // delete this because librdkafka will complain since this particular + // key is a real conf value + delete globalConf.event_cb; + + this._client = new SubClientType(globalConf, topicConf); + + var extractFunctions = function(obj) { + obj = obj || {}; + var obj2 = {}; + for (var p in obj) { + if (typeof obj[p] === "function") { + obj2[p] = obj[p]; + } + } + return obj2; + } + this._cb_configs = { + global: extractFunctions(globalConf), + topic: extractFunctions(topicConf), + event: {}, + } + + if (!no_event_cb) { + this._cb_configs.event.event_cb = function(eventType, eventData) { + switch (eventType) { + case 'error': + this.emit('event.error', LibrdKafkaError.create(eventData)); + break; + case 'stats': + this.emit('event.stats', eventData); + break; + case 'log': + this.emit('event.log', eventData); + break; + default: + this.emit('event.event', eventData); + this.emit('event.' + eventType, eventData); + } + }.bind(this); + } + + this.metrics = {}; + this._isConnected = false; + this.errorCounter = 0; + + /** + * Metadata object. Starts out empty but will be filled with information after + * the initial connect. + * + * @type {Client~Metadata} + */ + this._metadata = {}; + + var self = this; + + this.on('ready', function(info) { + self.metrics.connectionOpened = Date.now(); + self.name = info.name; + }) + .on('disconnected', function() { + // reset metrics + self.metrics = {}; + self._isConnected = false; + // keep the metadata. it still may be useful + }) + .on('event.error', function(err) { + self.lastError = err; + ++self.errorCounter; + }); + +} + +/** + * Connect to the broker and receive its metadata. + * + * Connects to a broker by establishing the client and fetches its metadata. + * + * @param {object} metadataOptions - Options to be sent to the metadata. + * @param {string} metadataOptions.topic - Topic to fetch metadata for. Empty string is treated as empty. + * @param {boolean} metadataOptions.allTopics - Fetch metadata for all topics, not just the ones we know about. + * @param {int} metadataOptions.timeout - The timeout, in ms, to allow for fetching metadata. Defaults to 30000ms + * @param {Client~connectionCallback} cb - Callback that indicates we are + * done connecting. + * @return {Client} - Returns itself. + */ +Client.prototype.connect = function(metadataOptions, cb) { + var self = this; + + var next = function(err, data) { + self._isConnecting = false; + if (cb) { + cb(err, data); + } + }; + + if (this._isConnected) { + setImmediate(next); + return self; + } + + if (this._isConnecting) { + this.once('ready', function() { + next(null, this._metadata); + }); + return self; + } + + this._isConnecting = true; + + var fail = function(err) { + var callbackCalled = false; + var t; + + if (self._isConnected) { + self._isConnected = false; + self._client.disconnect(function() { + if (callbackCalled) { + return; + } + clearTimeout(t); + callbackCalled = true; + + next(err); return; + }); + + // don't take too long. this is a failure, after all + t = setTimeout(function() { + if (callbackCalled) { + return; + } + callbackCalled = true; + + next(err); return; + }, 10000).unref(); + + self.emit('connection.failure', err, self.metrics); + } else { + + next(err); + } + }; + + this._client.configureCallbacks(true, this._cb_configs); + + this._client.connect(function(err, info) { + if (err) { + fail(LibrdKafkaError.create(err)); return; + } + + self._isConnected = true; + + // Otherwise we are successful + self.getMetadata(metadataOptions || {}, function(err, metadata) { + if (err) { + // We are connected so we need to disconnect + fail(LibrdKafkaError.create(err)); return; + } + + self._isConnecting = false; + // We got the metadata otherwise. It is set according to above param + // Set it here as well so subsequent ready callbacks + // can check it + self._isConnected = true; + + /** + * Ready event. Called when the Client connects successfully + * + * @event Client#ready + * @type {object} + * @property {string} name - the name of the broker. + */ + self.emit('ready', info, metadata); + next(null, metadata); return; + + }); + + }); + + return self; + +}; + +/** + * Get the native Kafka client. + * + * You probably shouldn't use this, but if you want to execute methods directly + * on the c++ wrapper you can do it here. + * + * @see connection.cc + * @return {Connection} - The native Kafka client. + */ +Client.prototype.getClient = function() { + return this._client; +}; + +/** + * Find out how long we have been connected to Kafka. + * + * @return {number} - Milliseconds since the connection has been established. + */ +Client.prototype.connectedTime = function() { + if (!this.isConnected()) { + return 0; + } + return Date.now() - this.metrics.connectionOpened; +}; + +/** + * Whether or not we are connected to Kafka. + * + * @return {boolean} - Whether we are connected. + */ +Client.prototype.isConnected = function() { + return !!(this._isConnected && this._client); +}; + +/** + * Get the last error emitted if it exists. + * + * @return {LibrdKafkaError} - Returns the LibrdKafkaError or null if + * one hasn't been thrown. + */ +Client.prototype.getLastError = function() { + return this.lastError || null; +}; + +/** + * Disconnect from the Kafka client. + * + * This method will disconnect us from Kafka unless we are already in a + * disconnecting state. Use this when you're done reading or producing messages + * on a given client. + * + * It will also emit the disconnected event. + * + * @fires Client#disconnected + * @return {function} - Callback to call when disconnection is complete. + */ +Client.prototype.disconnect = function(cb) { + var self = this; + + if (!this._isDisconnecting && this._client) { + this._isDisconnecting = true; + this._client.disconnect(function() { + // this take 5000 milliseconds. Librdkafka needs to make sure the memory + // has been cleaned up before we delete things. @see RdKafka::wait_destroyed + self._client.configureCallbacks(false, self._cb_configs); + + // Broadcast metrics. Gives people one last chance to do something with them + self._isDisconnecting = false; + /** + * Disconnect event. Called after disconnection is finished. + * + * @event Client#disconnected + * @type {object} + * @property {date} connectionOpened - when the connection was opened. + */ + var metricsCopy = Object.assign({}, self.metrics); + self.emit('disconnected', metricsCopy); + if (cb) { + cb(null, metricsCopy); + } + + }); + + } + + return self; +}; + +/** + * Get client metadata. + * + * Note: using a metadataOptions.topic parameter has a potential side-effect. + * A Topic object will be created, if it did not exist yet, with default options + * and it will be cached by librdkafka. + * + * A subsequent call to create the topic object with specific options (e.g. acks) will return + * the previous instance and the specific options will be silently ignored. + * + * To avoid this side effect, the topic object can be created with the expected options before requesting metadata, + * or the metadata request can be performed for all topics (by omitting metadataOptions.topic). + * + * @param {object} metadataOptions - Metadata options to pass to the client. + * @param {string} metadataOptions.topic - Topic string for which to fetch + * metadata + * @param {number} metadataOptions.timeout - Max time, in ms, to try to fetch + * metadata before timing out. Defaults to 3000. + * @param {Client~metadataCallback} cb - Callback to fire with the metadata. + */ +Client.prototype.getMetadata = function(metadataOptions, cb) { + if (!this.isConnected()) { + return cb(new Error('Client is disconnected')); + } + + var self = this; + + this._client.getMetadata(metadataOptions || {}, function(err, metadata) { + if (err) { + if (cb) { + cb(LibrdKafkaError.create(err)); + } + return; + } + + // No error otherwise + self._metadata = metadata; + + if (cb) { + cb(null, metadata); + } + + }); + +}; + +/** + * Query offsets from the broker. + * + * This function makes a call to the broker to get the current low (oldest/beginning) + * and high (newest/end) offsets for a topic partition. + * + * @param {string} topic - Topic to recieve offsets from. + * @param {number} partition - Partition of the provided topic to recieve offsets from + * @param {number} timeout - Number of ms to wait to recieve a response. + * @param {Client~watermarkOffsetsCallback} cb - Callback to fire with the offsets. + */ +Client.prototype.queryWatermarkOffsets = function(topic, partition, timeout, cb) { + if (!this.isConnected()) { + if (cb) { + return cb(new Error('Client is disconnected')); + } else { + return; + } + } + + var self = this; + + if (typeof timeout === 'function') { + cb = timeout; + timeout = 1000; + } + + if (!timeout) { + timeout = 1000; + } + + this._client.queryWatermarkOffsets(topic, partition, timeout, function(err, offsets) { + if (err) { + if (cb) { + cb(LibrdKafkaError.create(err)); + } + return; + } + + if (cb) { + cb(null, offsets); + } + + }); +}; + +/** + * Query offsets for times from the broker. + * + * This function makes a call to the broker to get the offsets for times specified. + * + * @param {TopicPartition[]} toppars - Array of topic partitions. The offset in these + * should instead refer to a timestamp you want + * offsets for + * @param {number} timeout - Number of ms to wait to recieve a response. + * @param {Client~offsetsForTimesCallback} cb - Callback to fire with the filled in offsets. + */ +Client.prototype.offsetsForTimes = function(toppars, timeout, cb) { + if (!this.isConnected()) { + if (cb) { + return cb(new Error('Client is disconnected')); + } else { + return; + } + } + + var self = this; + + if (typeof timeout === 'function') { + cb = timeout; + timeout = 1000; + } + + if (!timeout) { + timeout = 1000; + } + + this._client.offsetsForTimes(toppars, timeout, function(err, toppars) { + if (err) { + if (cb) { + cb(LibrdKafkaError.create(err)); + } + return; + } + + if (cb) { + cb(null, toppars); + } + + }); +}; + +/** + * Wrap a potential RdKafka error. + * + * This internal method is meant to take a return value + * from a function that returns an RdKafka error code, throw if it + * is an error (Making it a proper librdkafka error object), or + * return the appropriate value otherwise. + * + * It is intended to be used in a return statement, + * + * @private + * @param {number} errorCode - Error code returned from a native method + * @param {bool} intIsError - If specified true, any non-number return type will be classified as a success + * @return {boolean} - Returns true or the method return value unless it throws. + */ +Client.prototype._errorWrap = function(errorCode, intIsError) { + var returnValue = true; + if (intIsError) { + returnValue = errorCode; + errorCode = typeof errorCode === 'number' ? errorCode : 0; + } + + if (errorCode !== LibrdKafkaError.codes.ERR_NO_ERROR) { + var e = LibrdKafkaError.create(errorCode); + throw e; + } + + return returnValue; +}; + +/** + * This callback is used to pass metadata or an error after a successful + * connection + * + * @callback Client~connectionCallback + * @param {Error} err - An error, if one occurred while connecting. + * @param {Client~Metadata} metadata - Metadata object. + */ + + /** + * This callback is used to pass offsets or an error after a successful + * query + * + * @callback Client~watermarkOffsetsCallback + * @param {Error} err - An error, if one occurred while connecting. + * @param {Client~watermarkOffsets} offsets - Watermark offsets + */ + + /** + * This callback is used to pass toppars or an error after a successful + * times query + * + * @callback Client~offsetsForTimesCallback + * @param {Error} err - An error, if one occurred while connecting. + * @param {TopicPartition[]} toppars - Topic partitions with offsets filled in + */ + +/** + * @typedef {object} Client~watermarkOffsets + * @property {number} high - High (newest/end) offset + * @property {number} low - Low (oldest/beginning) offset + */ + +/** + * @typedef {object} Client~MetadataBroker + * @property {number} id - Broker ID + * @property {string} host - Broker host + * @property {number} port - Broker port. + */ + +/** + * @typedef {object} Client~MetadataTopic + * @property {string} name - Topic name + * @property {Client~MetadataPartition[]} partitions - Array of partitions + */ + +/** + * @typedef {object} Client~MetadataPartition + * @property {number} id - Partition id + * @property {number} leader - Broker ID for the partition leader + * @property {number[]} replicas - Array of replica IDs + * @property {number[]} isrs - Arrqay of ISRS ids +*/ + +/** + * Metadata object. + * + * This is the representation of Kafka metadata in JavaScript. + * + * @typedef {object} Client~Metadata + * @property {number} orig_broker_id - The broker ID of the original bootstrap + * broker. + * @property {string} orig_broker_name - The name of the original bootstrap + * broker. + * @property {Client~MetadataBroker[]} brokers - An array of broker objects + * @property {Client~MetadataTopic[]} topics - An array of topics. + */ diff --git a/lib/error.js b/lib/error.js new file mode 100644 index 0000000..f726164 --- /dev/null +++ b/lib/error.js @@ -0,0 +1,470 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +module.exports = LibrdKafkaError; + +var util = require('util'); +var librdkafka = require('../librdkafka'); + +util.inherits(LibrdKafkaError, Error); + +LibrdKafkaError.create = createLibrdkafkaError; +LibrdKafkaError.wrap = errorWrap; + +/** + * Enum for identifying errors reported by the library + * + * You can find this list in the C++ code at + * https://github.com/edenhill/librdkafka/blob/master/src-cpp/rdkafkacpp.h#L148 + * + * @readonly + * @enum {number} + * @constant + */ +// ====== Generated from librdkafka 1.6.1 file src-cpp/rdkafkacpp.h ====== +LibrdKafkaError.codes = { + + /* Internal errors to rdkafka: */ + /** Begin internal error codes */ + ERR__BEGIN: -200, + /** Received message is incorrect */ + ERR__BAD_MSG: -199, + /** Bad/unknown compression */ + ERR__BAD_COMPRESSION: -198, + /** Broker is going away */ + ERR__DESTROY: -197, + /** Generic failure */ + ERR__FAIL: -196, + /** Broker transport failure */ + ERR__TRANSPORT: -195, + /** Critical system resource */ + ERR__CRIT_SYS_RESOURCE: -194, + /** Failed to resolve broker */ + ERR__RESOLVE: -193, + /** Produced message timed out*/ + ERR__MSG_TIMED_OUT: -192, + /** Reached the end of the topic+partition queue on + * the broker. Not really an error. + * This event is disabled by default, + * see the `enable.partition.eof` configuration property. */ + ERR__PARTITION_EOF: -191, + /** Permanent: Partition does not exist in cluster. */ + ERR__UNKNOWN_PARTITION: -190, + /** File or filesystem error */ + ERR__FS: -189, + /** Permanent: Topic does not exist in cluster. */ + ERR__UNKNOWN_TOPIC: -188, + /** All broker connections are down. */ + ERR__ALL_BROKERS_DOWN: -187, + /** Invalid argument, or invalid configuration */ + ERR__INVALID_ARG: -186, + /** Operation timed out */ + ERR__TIMED_OUT: -185, + /** Queue is full */ + ERR__QUEUE_FULL: -184, + /** ISR count < required.acks */ + ERR__ISR_INSUFF: -183, + /** Broker node update */ + ERR__NODE_UPDATE: -182, + /** SSL error */ + ERR__SSL: -181, + /** Waiting for coordinator to become available. */ + ERR__WAIT_COORD: -180, + /** Unknown client group */ + ERR__UNKNOWN_GROUP: -179, + /** Operation in progress */ + ERR__IN_PROGRESS: -178, + /** Previous operation in progress, wait for it to finish. */ + ERR__PREV_IN_PROGRESS: -177, + /** This operation would interfere with an existing subscription */ + ERR__EXISTING_SUBSCRIPTION: -176, + /** Assigned partitions (rebalance_cb) */ + ERR__ASSIGN_PARTITIONS: -175, + /** Revoked partitions (rebalance_cb) */ + ERR__REVOKE_PARTITIONS: -174, + /** Conflicting use */ + ERR__CONFLICT: -173, + /** Wrong state */ + ERR__STATE: -172, + /** Unknown protocol */ + ERR__UNKNOWN_PROTOCOL: -171, + /** Not implemented */ + ERR__NOT_IMPLEMENTED: -170, + /** Authentication failure*/ + ERR__AUTHENTICATION: -169, + /** No stored offset */ + ERR__NO_OFFSET: -168, + /** Outdated */ + ERR__OUTDATED: -167, + /** Timed out in queue */ + ERR__TIMED_OUT_QUEUE: -166, + /** Feature not supported by broker */ + ERR__UNSUPPORTED_FEATURE: -165, + /** Awaiting cache update */ + ERR__WAIT_CACHE: -164, + /** Operation interrupted */ + ERR__INTR: -163, + /** Key serialization error */ + ERR__KEY_SERIALIZATION: -162, + /** Value serialization error */ + ERR__VALUE_SERIALIZATION: -161, + /** Key deserialization error */ + ERR__KEY_DESERIALIZATION: -160, + /** Value deserialization error */ + ERR__VALUE_DESERIALIZATION: -159, + /** Partial response */ + ERR__PARTIAL: -158, + /** Modification attempted on read-only object */ + ERR__READ_ONLY: -157, + /** No such entry / item not found */ + ERR__NOENT: -156, + /** Read underflow */ + ERR__UNDERFLOW: -155, + /** Invalid type */ + ERR__INVALID_TYPE: -154, + /** Retry operation */ + ERR__RETRY: -153, + /** Purged in queue */ + ERR__PURGE_QUEUE: -152, + /** Purged in flight */ + ERR__PURGE_INFLIGHT: -151, + /** Fatal error: see RdKafka::Handle::fatal_error() */ + ERR__FATAL: -150, + /** Inconsistent state */ + ERR__INCONSISTENT: -149, + /** Gap-less ordering would not be guaranteed if proceeding */ + ERR__GAPLESS_GUARANTEE: -148, + /** Maximum poll interval exceeded */ + ERR__MAX_POLL_EXCEEDED: -147, + /** Unknown broker */ + ERR__UNKNOWN_BROKER: -146, + /** Functionality not configured */ + ERR__NOT_CONFIGURED: -145, + /** Instance has been fenced */ + ERR__FENCED: -144, + /** Application generated error */ + ERR__APPLICATION: -143, + /** Assignment lost */ + ERR__ASSIGNMENT_LOST: -142, + /** No operation performed */ + ERR__NOOP: -141, + /** No offset to automatically reset to */ + ERR__AUTO_OFFSET_RESET: -140, + /** End internal error codes */ + ERR__END: -100, + /* Kafka broker errors: */ + /** Unknown broker error */ + ERR_UNKNOWN: -1, + /** Success */ + ERR_NO_ERROR: 0, + /** Offset out of range */ + ERR_OFFSET_OUT_OF_RANGE: 1, + /** Invalid message */ + ERR_INVALID_MSG: 2, + /** Unknown topic or partition */ + ERR_UNKNOWN_TOPIC_OR_PART: 3, + /** Invalid message size */ + ERR_INVALID_MSG_SIZE: 4, + /** Leader not available */ + ERR_LEADER_NOT_AVAILABLE: 5, + /** Not leader for partition */ + ERR_NOT_LEADER_FOR_PARTITION: 6, + /** Request timed out */ + ERR_REQUEST_TIMED_OUT: 7, + /** Broker not available */ + ERR_BROKER_NOT_AVAILABLE: 8, + /** Replica not available */ + ERR_REPLICA_NOT_AVAILABLE: 9, + /** Message size too large */ + ERR_MSG_SIZE_TOO_LARGE: 10, + /** StaleControllerEpochCode */ + ERR_STALE_CTRL_EPOCH: 11, + /** Offset metadata string too large */ + ERR_OFFSET_METADATA_TOO_LARGE: 12, + /** Broker disconnected before response received */ + ERR_NETWORK_EXCEPTION: 13, + /** Coordinator load in progress */ + ERR_COORDINATOR_LOAD_IN_PROGRESS: 14, + /** Group coordinator load in progress */ + ERR_GROUP_LOAD_IN_PROGRESS: 14, + /** Coordinator not available */ + ERR_COORDINATOR_NOT_AVAILABLE: 15, + /** Group coordinator not available */ + ERR_GROUP_COORDINATOR_NOT_AVAILABLE: 15, + /** Not coordinator */ + ERR_NOT_COORDINATOR: 16, + /** Not coordinator for group */ + ERR_NOT_COORDINATOR_FOR_GROUP: 16, + /** Invalid topic */ + ERR_TOPIC_EXCEPTION: 17, + /** Message batch larger than configured server segment size */ + ERR_RECORD_LIST_TOO_LARGE: 18, + /** Not enough in-sync replicas */ + ERR_NOT_ENOUGH_REPLICAS: 19, + /** Message(s) written to insufficient number of in-sync replicas */ + ERR_NOT_ENOUGH_REPLICAS_AFTER_APPEND: 20, + /** Invalid required acks value */ + ERR_INVALID_REQUIRED_ACKS: 21, + /** Specified group generation id is not valid */ + ERR_ILLEGAL_GENERATION: 22, + /** Inconsistent group protocol */ + ERR_INCONSISTENT_GROUP_PROTOCOL: 23, + /** Invalid group.id */ + ERR_INVALID_GROUP_ID: 24, + /** Unknown member */ + ERR_UNKNOWN_MEMBER_ID: 25, + /** Invalid session timeout */ + ERR_INVALID_SESSION_TIMEOUT: 26, + /** Group rebalance in progress */ + ERR_REBALANCE_IN_PROGRESS: 27, + /** Commit offset data size is not valid */ + ERR_INVALID_COMMIT_OFFSET_SIZE: 28, + /** Topic authorization failed */ + ERR_TOPIC_AUTHORIZATION_FAILED: 29, + /** Group authorization failed */ + ERR_GROUP_AUTHORIZATION_FAILED: 30, + /** Cluster authorization failed */ + ERR_CLUSTER_AUTHORIZATION_FAILED: 31, + /** Invalid timestamp */ + ERR_INVALID_TIMESTAMP: 32, + /** Unsupported SASL mechanism */ + ERR_UNSUPPORTED_SASL_MECHANISM: 33, + /** Illegal SASL state */ + ERR_ILLEGAL_SASL_STATE: 34, + /** Unuspported version */ + ERR_UNSUPPORTED_VERSION: 35, + /** Topic already exists */ + ERR_TOPIC_ALREADY_EXISTS: 36, + /** Invalid number of partitions */ + ERR_INVALID_PARTITIONS: 37, + /** Invalid replication factor */ + ERR_INVALID_REPLICATION_FACTOR: 38, + /** Invalid replica assignment */ + ERR_INVALID_REPLICA_ASSIGNMENT: 39, + /** Invalid config */ + ERR_INVALID_CONFIG: 40, + /** Not controller for cluster */ + ERR_NOT_CONTROLLER: 41, + /** Invalid request */ + ERR_INVALID_REQUEST: 42, + /** Message format on broker does not support request */ + ERR_UNSUPPORTED_FOR_MESSAGE_FORMAT: 43, + /** Policy violation */ + ERR_POLICY_VIOLATION: 44, + /** Broker received an out of order sequence number */ + ERR_OUT_OF_ORDER_SEQUENCE_NUMBER: 45, + /** Broker received a duplicate sequence number */ + ERR_DUPLICATE_SEQUENCE_NUMBER: 46, + /** Producer attempted an operation with an old epoch */ + ERR_INVALID_PRODUCER_EPOCH: 47, + /** Producer attempted a transactional operation in an invalid state */ + ERR_INVALID_TXN_STATE: 48, + /** Producer attempted to use a producer id which is not + * currently assigned to its transactional id */ + ERR_INVALID_PRODUCER_ID_MAPPING: 49, + /** Transaction timeout is larger than the maximum + * value allowed by the broker's max.transaction.timeout.ms */ + ERR_INVALID_TRANSACTION_TIMEOUT: 50, + /** Producer attempted to update a transaction while another + * concurrent operation on the same transaction was ongoing */ + ERR_CONCURRENT_TRANSACTIONS: 51, + /** Indicates that the transaction coordinator sending a + * WriteTxnMarker is no longer the current coordinator for a + * given producer */ + ERR_TRANSACTION_COORDINATOR_FENCED: 52, + /** Transactional Id authorization failed */ + ERR_TRANSACTIONAL_ID_AUTHORIZATION_FAILED: 53, + /** Security features are disabled */ + ERR_SECURITY_DISABLED: 54, + /** Operation not attempted */ + ERR_OPERATION_NOT_ATTEMPTED: 55, + /** Disk error when trying to access log file on the disk */ + ERR_KAFKA_STORAGE_ERROR: 56, + /** The user-specified log directory is not found in the broker config */ + ERR_LOG_DIR_NOT_FOUND: 57, + /** SASL Authentication failed */ + ERR_SASL_AUTHENTICATION_FAILED: 58, + /** Unknown Producer Id */ + ERR_UNKNOWN_PRODUCER_ID: 59, + /** Partition reassignment is in progress */ + ERR_REASSIGNMENT_IN_PROGRESS: 60, + /** Delegation Token feature is not enabled */ + ERR_DELEGATION_TOKEN_AUTH_DISABLED: 61, + /** Delegation Token is not found on server */ + ERR_DELEGATION_TOKEN_NOT_FOUND: 62, + /** Specified Principal is not valid Owner/Renewer */ + ERR_DELEGATION_TOKEN_OWNER_MISMATCH: 63, + /** Delegation Token requests are not allowed on this connection */ + ERR_DELEGATION_TOKEN_REQUEST_NOT_ALLOWED: 64, + /** Delegation Token authorization failed */ + ERR_DELEGATION_TOKEN_AUTHORIZATION_FAILED: 65, + /** Delegation Token is expired */ + ERR_DELEGATION_TOKEN_EXPIRED: 66, + /** Supplied principalType is not supported */ + ERR_INVALID_PRINCIPAL_TYPE: 67, + /** The group is not empty */ + ERR_NON_EMPTY_GROUP: 68, + /** The group id does not exist */ + ERR_GROUP_ID_NOT_FOUND: 69, + /** The fetch session ID was not found */ + ERR_FETCH_SESSION_ID_NOT_FOUND: 70, + /** The fetch session epoch is invalid */ + ERR_INVALID_FETCH_SESSION_EPOCH: 71, + /** No matching listener */ + ERR_LISTENER_NOT_FOUND: 72, + /** Topic deletion is disabled */ + ERR_TOPIC_DELETION_DISABLED: 73, + /** Leader epoch is older than broker epoch */ + ERR_FENCED_LEADER_EPOCH: 74, + /** Leader epoch is newer than broker epoch */ + ERR_UNKNOWN_LEADER_EPOCH: 75, + /** Unsupported compression type */ + ERR_UNSUPPORTED_COMPRESSION_TYPE: 76, + /** Broker epoch has changed */ + ERR_STALE_BROKER_EPOCH: 77, + /** Leader high watermark is not caught up */ + ERR_OFFSET_NOT_AVAILABLE: 78, + /** Group member needs a valid member ID */ + ERR_MEMBER_ID_REQUIRED: 79, + /** Preferred leader was not available */ + ERR_PREFERRED_LEADER_NOT_AVAILABLE: 80, + /** Consumer group has reached maximum size */ + ERR_GROUP_MAX_SIZE_REACHED: 81, + /** Static consumer fenced by other consumer with same + * group.instance.id. */ + ERR_FENCED_INSTANCE_ID: 82, + /** Eligible partition leaders are not available */ + ERR_ELIGIBLE_LEADERS_NOT_AVAILABLE: 83, + /** Leader election not needed for topic partition */ + ERR_ELECTION_NOT_NEEDED: 84, + /** No partition reassignment is in progress */ + ERR_NO_REASSIGNMENT_IN_PROGRESS: 85, + /** Deleting offsets of a topic while the consumer group is + * subscribed to it */ + ERR_GROUP_SUBSCRIBED_TO_TOPIC: 86, + /** Broker failed to validate record */ + ERR_INVALID_RECORD: 87, + /** There are unstable offsets that need to be cleared */ + ERR_UNSTABLE_OFFSET_COMMIT: 88, + /** Throttling quota has been exceeded */ + ERR_THROTTLING_QUOTA_EXCEEDED: 89, + /** There is a newer producer with the same transactionalId + * which fences the current one */ + ERR_PRODUCER_FENCED: 90, + /** Request illegally referred to resource that does not exist */ + ERR_RESOURCE_NOT_FOUND: 91, + /** Request illegally referred to the same resource twice */ + ERR_DUPLICATE_RESOURCE: 92, + /** Requested credential would not meet criteria for acceptability */ + ERR_UNACCEPTABLE_CREDENTIAL: 93, + /** Indicates that the either the sender or recipient of a + * voter-only request is not one of the expected voters */ + ERR_INCONSISTENT_VOTER_SET: 94, + /** Invalid update version */ + ERR_INVALID_UPDATE_VERSION: 95, + /** Unable to update finalized features due to server error */ + ERR_FEATURE_UPDATE_FAILED: 96, + /** Request principal deserialization failed during forwarding */ + ERR_PRINCIPAL_DESERIALIZATION_FAILURE: 97 +}; + +/** + * Representation of a librdkafka error + * + * This can be created by giving either another error + * to piggy-back on. In this situation it tries to parse + * the error string to figure out the intent. However, more usually, + * it is constructed by an error object created by a C++ Baton. + * + * @param {object|error} e - An object or error to wrap + * @property {string} message - The error message + * @property {number} code - The error code. + * @property {string} origin - The origin, whether it is local or remote + * @constructor + */ +function LibrdKafkaError(e) { + if (!(this instanceof LibrdKafkaError)) { + return new LibrdKafkaError(e); + } + + if (typeof e === 'number') { + this.message = librdkafka.err2str(e); + this.code = e; + this.errno = e; + if (e >= LibrdKafkaError.codes.ERR__END) { + this.origin = 'local'; + } else { + this.origin = 'kafka'; + } + Error.captureStackTrace(this, this.constructor); + } else if (!util.isError(e)) { + // This is the better way + this.message = e.message; + this.code = e.code; + this.errno = e.code; + if (e.code >= LibrdKafkaError.codes.ERR__END) { + this.origin = 'local'; + } else { + this.origin = 'kafka'; + } + Error.captureStackTrace(this, this.constructor); + } else { + var message = e.message; + var parsedMessage = message.split(': '); + + var origin, msg; + + if (parsedMessage.length > 1) { + origin = parsedMessage[0].toLowerCase(); + msg = parsedMessage[1].toLowerCase(); + } else { + origin = 'unknown'; + msg = message.toLowerCase(); + } + + // special cases + if (msg === 'consumer is disconnected' || msg === 'producer is disconnected') { + this.origin = 'local'; + this.code = LibrdKafkaError.codes.ERR__STATE; + this.errno = this.code; + this.message = msg; + } else { + this.origin = origin; + this.message = msg; + this.code = typeof e.code === 'number' ? e.code : -1; + this.errno = this.code; + this.stack = e.stack; + } + + } + + if (e.hasOwnProperty('isFatal')) this.isFatal = e.isFatal; + if (e.hasOwnProperty('isRetriable')) this.isRetriable = e.isRetriable; + if (e.hasOwnProperty('isTxnRequiresAbort')) this.isTxnRequiresAbort = e.isTxnRequiresAbort; + +} + +function createLibrdkafkaError(e) { + return new LibrdKafkaError(e); +} + +function errorWrap(errorCode, intIsError) { + var returnValue = true; + if (intIsError) { + returnValue = errorCode; + errorCode = typeof errorCode === 'number' ? errorCode : 0; + } + + if (errorCode !== LibrdKafkaError.codes.ERR_NO_ERROR) { + var e = LibrdKafkaError.create(errorCode); + throw e; + } + + return returnValue; +} diff --git a/lib/index.js b/lib/index.js new file mode 100644 index 0000000..e2e8a9c --- /dev/null +++ b/lib/index.js @@ -0,0 +1,34 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var KafkaConsumer = require('./kafka-consumer'); +var Producer = require('./producer'); +var HighLevelProducer = require('./producer/high-level-producer'); +var error = require('./error'); +var util = require('util'); +var lib = require('../librdkafka'); +var Topic = require('./topic'); +var Admin = require('./admin'); +var features = lib.features().split(','); + +module.exports = { + Consumer: util.deprecate(KafkaConsumer, 'Use KafkaConsumer instead. This may be changed in a later version'), + Producer: Producer, + HighLevelProducer: HighLevelProducer, + AdminClient: Admin, + KafkaConsumer: KafkaConsumer, + createReadStream: KafkaConsumer.createReadStream, + createWriteStream: Producer.createWriteStream, + CODES: { + ERRORS: error.codes, + }, + Topic: Topic, + features: features, + librdkafkaVersion: lib.librdkafkaVersion +}; diff --git a/lib/kafka-consumer-stream.js b/lib/kafka-consumer-stream.js new file mode 100644 index 0000000..0abb535 --- /dev/null +++ b/lib/kafka-consumer-stream.js @@ -0,0 +1,381 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +'use strict'; + +module.exports = KafkaConsumerStream; + +var Readable = require('stream').Readable; +var util = require('util'); + +util.inherits(KafkaConsumerStream, Readable); + +/** + * ReadableStream integrating with the Kafka Consumer. + * + * This class is used to read data off of Kafka in a streaming way. It is + * useful if you'd like to have a way to pipe Kafka into other systems. You + * should generally not make this class yourself, as it is not even exposed + * as part of module.exports. Instead, you should KafkaConsumer.createReadStream. + * + * The stream implementation is slower than the continuous subscribe callback. + * If you don't care so much about backpressure and would rather squeeze + * out performance, use that method. Using the stream will ensure you read only + * as fast as you write. + * + * The stream detects if Kafka is already connected. If it is, it will begin + * reading. If it is not, it will connect and read when it is ready. + * + * This stream operates in objectMode. It streams {Consumer~Message} + * + * @param {Consumer} consumer - The Kafka Consumer object. + * @param {object} options - Options to configure the stream. + * @param {number} options.waitInterval - Number of ms to wait if Kafka reports + * that it has timed out or that we are out of messages (right now). + * @param {array} options.topics - Array of topics, or a function that parses + * metadata into an array of topics + * @constructor + * @extends stream.Readable + * @see Consumer~Message + */ +function KafkaConsumerStream(consumer, options) { + if (!(this instanceof KafkaConsumerStream)) { + return new KafkaConsumerStream(consumer, options); + } + + if (options === undefined) { + options = { waitInterval: 1000 }; + } else if (typeof options === 'number') { + options = { waitInterval: options }; + } else if (options === null || typeof options !== 'object') { + throw new TypeError('"options" argument must be a number or an object'); + } + + var topics = options.topics; + + if (typeof topics === 'function') { + // Just ignore the rest of the checks here + } else if (!Array.isArray(topics)) { + if (typeof topics !== 'string' && !(topics instanceof RegExp)) { + throw new TypeError('"topics" argument must be a string, regex, or an array'); + } else { + topics = [topics]; + } + } + + options = Object.create(options); + + var fetchSize = options.fetchSize || 1; + + // Run in object mode by default. + if (options.objectMode === null || options.objectMode === undefined) { + options.objectMode = true; + + // If they did not explicitly set high water mark, and we are running + // in object mode, set it to the fetch size + 2 to ensure there is room + // for a standard fetch + if (!options.highWaterMark) { + options.highWaterMark = fetchSize + 2; + } + } + + if (options.objectMode !== true) { + this._read = this._read_buffer; + } else { + this._read = this._read_message; + } + + Readable.call(this, options); + + this.consumer = consumer; + this.topics = topics; + this.autoClose = options.autoClose === undefined ? true : !!options.autoClose; + this.waitInterval = options.waitInterval === undefined ? 1000 : options.waitInterval; + this.fetchSize = fetchSize; + this.connectOptions = options.connectOptions || {}; + this.streamAsBatch = options.streamAsBatch || false; + + // Hold the messages in here + this.messages = []; + + var self = this; + + this.consumer + .on('unsubscribed', function() { + // Invalidate the stream when we unsubscribe + self.push(null); + }); + + // Call connect. Handles potentially being connected already + this.connect(this.connectOptions); + + this.once('end', function() { + if (this.autoClose) { + this.destroy(); + } + }); + +} + +/** + * Internal stream read method. This method reads message objects. + * @param {number} size - This parameter is ignored for our cases. + * @private + */ +KafkaConsumerStream.prototype._read_message = function(size) { + if (this.messages.length > 0) { + return this.push(this.messages.shift()); + } + + if (!this.consumer) { + // This consumer is set to `null` in the close function + return; + } + + if (!this.consumer.isConnected()) { + this.consumer.once('ready', function() { + // This is the way Node.js does it + // https://github.com/nodejs/node/blob/master/lib/fs.js#L1733 + this._read(size); + }.bind(this)); + return; + } + + if (this.destroyed) { + return; + } + + var self = this; + + // If the size (number of messages) we are being advised to fetch is + // greater than or equal to the fetch size, use the fetch size. + // Only opt to use the size in case it is LESS than the fetch size. + // Essentially, we want to use the smaller value here + var fetchSize = size >= this.fetchSize ? this.fetchSize : size; + + this.consumer.consume(fetchSize, onread); + + // Retry function. Will wait up to the wait interval, with some + // random noise if one is provided. Otherwise, will go immediately. + function retry() { + if (!self.waitInterval) { + setImmediate(function() { + self._read(size); + }); + } else { + setTimeout(function() { + self._read(size); + }, self.waitInterval * Math.random()).unref(); + } + } + + function onread(err, messages) { + + // If there was an error we still want to emit it. + // Essentially, if the user does not register an error + // handler, it will still cause the stream to blow up. + // + // But... if one is provided, consumption will move on + // as normal + if (err) { + self.emit('error', err); + } + + // If there are no messages it means we reached EOF or a timeout. + // Do what we used to do + + if (err || messages.length < 1) { + // If we got an error or if there were no messages, initiate a retry + retry(); + return; + } else { + if (self.streamAsBatch) { + self.push(messages); + } else { + for (var i = 0; i < messages.length; i++) { + self.messages.push(messages[i]); + } + + // Now that we have added them all the inner messages buffer, + // we can just push the most recent one + self.push(self.messages.shift()); + } + } + + } +}; + +/** + * Internal stream read method. This method reads message buffers. + * @param {number} size - This parameter is ignored for our cases. + * @private + */ +KafkaConsumerStream.prototype._read_buffer = function(size) { + if (this.messages.length > 0) { + return this.push(this.messages.shift()); + } + + if (!this.consumer) { + // This consumer is set to `null` in the close function + return; + } + + if (!this.consumer.isConnected()) { + this.consumer.once('ready', function() { + // This is the way Node.js does it + // https://github.com/nodejs/node/blob/master/lib/fs.js#L1733 + this._read(size); + }.bind(this)); + return; + } + + if (this.destroyed) { + return; + } + + var self = this; + + // If the size (number of messages) we are being advised to fetch is + // greater than or equal to the fetch size, use the fetch size. + // Only opt to use the size in case it is LESS than the fetch size. + // Essentially, we want to use the smaller value here + var fetchSize = size >= this.fetchSize ? this.fetchSize : size; + + this.consumer.consume(fetchSize, onread); + + // Retry function. Will wait up to the wait interval, with some + // random noise if one is provided. Otherwise, will go immediately. + function retry() { + if (!self.waitInterval) { + setImmediate(function() { + self._read(size); + }); + } else { + setTimeout(function() { + self._read(size); + }, self.waitInterval * Math.random()).unref(); + } + } + + function onread(err, messages) { + // If there was an error we still want to emit it. + // Essentially, if the user does not register an error + // handler, it will still cause the stream to blow up. + // + // But... if one is provided, consumption will move on + // as normal + if (err) { + self.emit('error', err); + } + + // If there are no messages it means we reached EOF or a timeout. + // Do what we used to do + + if (err || messages.length < 1) { + // If we got an error or if there were no messages, initiate a retry + retry(); + return; + } else { + if (self.streamAsBatch) { + self.push(messages); + } else { + for (var i = 0; i < messages.length; i++) { + self.messages.push(messages[i].value); + } + // Now that we have added them all the inner messages buffer, + // we can just push the most recent one + self.push(self.messages.shift()); + } + } + + } +}; + +KafkaConsumerStream.prototype.connect = function(options) { + var self = this; + + function connectCallback(err, metadata) { + if (err) { + self.emit('error', err); + self.destroy(); + return; + } + + try { + // Subscribe to the topics as well so we will be ready + // If this throws the stream is invalid + + // This is the magic part. If topics is a function, before we subscribe, + // pass the metadata in + if (typeof self.topics === 'function') { + var topics = self.topics(metadata); + self.consumer.subscribe(topics); + } else { + self.consumer.subscribe(self.topics); + } + } catch (e) { + self.emit('error', e); + self.destroy(); + return; + } + + // start the flow of data + self.read(); + } + + if (!this.consumer.isConnected()) { + self.consumer.connect(options, connectCallback); + } else { + // Immediately call the connect callback + setImmediate(function() { + connectCallback(null, self.consumer._metadata); + }); + } + +}; + +KafkaConsumerStream.prototype.destroy = function() { + if (this.destroyed) { + return; + } + this.destroyed = true; + this.close(); +}; + +KafkaConsumerStream.prototype.close = function(cb) { + var self = this; + if (cb) { + this.once('close', cb); + } + + if (!self.consumer._isConnecting && !self.consumer._isConnected) { + // If we aren't even connected just exit. We are done. + close(); + return; + } + + if (self.consumer._isConnecting) { + self.consumer.once('ready', function() { + // Don't pass the CB because it has already been passed. + self.close(); + }); + return; + } + + if (self.consumer._isConnected) { + self.consumer.unsubscribe(); + self.consumer.disconnect(function() { + close(); + }); + } + + function close() { + self.emit('close'); + } +}; diff --git a/lib/kafka-consumer.js b/lib/kafka-consumer.js new file mode 100644 index 0000000..be3b639 --- /dev/null +++ b/lib/kafka-consumer.js @@ -0,0 +1,656 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ +'use strict'; + +module.exports = KafkaConsumer; + +var Client = require('./client'); +var util = require('util'); +var Kafka = require('../librdkafka'); +var KafkaConsumerStream = require('./kafka-consumer-stream'); +var LibrdKafkaError = require('./error'); +var TopicPartition = require('./topic-partition'); +var shallowCopy = require('./util').shallowCopy; +var DEFAULT_CONSUME_LOOP_TIMEOUT_DELAY = 500; +var DEFAULT_CONSUME_TIME_OUT = 1000; +util.inherits(KafkaConsumer, Client); + +/** + * KafkaConsumer class for reading messages from Kafka + * + * This is the main entry point for reading data from Kafka. You + * configure this like you do any other client, with a global + * configuration and default topic configuration. + * + * Once you instantiate this object, connecting will open a socket. + * Data will not be read until you tell the consumer what topics + * you want to read from. + * + * @param {object} conf - Key value pairs to configure the consumer + * @param {object} topicConf - Key value pairs to create a default + * topic configuration + * @extends Client + * @constructor + */ +function KafkaConsumer(conf, topicConf) { + if (!(this instanceof KafkaConsumer)) { + return new KafkaConsumer(conf, topicConf); + } + + conf = shallowCopy(conf); + topicConf = shallowCopy(topicConf); + + var onRebalance = conf.rebalance_cb; + + var self = this; + + // If rebalance is undefined we don't want any part of this + if (onRebalance && typeof onRebalance === 'boolean') { + conf.rebalance_cb = function(err, assignment) { + // Create the librdkafka error + err = LibrdKafkaError.create(err); + // Emit the event + self.emit('rebalance', err, assignment); + + // That's it + try { + if (err.code === -175 /*ERR__ASSIGN_PARTITIONS*/) { + self.assign(assignment); + } else if (err.code === -174 /*ERR__REVOKE_PARTITIONS*/) { + self.unassign(); + } + } catch (e) { + // Ignore exceptions if we are not connected + if (self.isConnected()) { + self.emit('rebalance.error', e); + } + } + }; + } else if (onRebalance && typeof onRebalance === 'function') { + /* + * Once this is opted in to, that's it. It's going to manually rebalance + * forever. There is no way to unset config values in librdkafka, just + * a way to override them. + */ + + conf.rebalance_cb = function(err, assignment) { + // Create the librdkafka error + err = err ? LibrdKafkaError.create(err) : undefined; + + self.emit('rebalance', err, assignment); + onRebalance.call(self, err, assignment); + }; + } + + // Same treatment for offset_commit_cb + var onOffsetCommit = conf.offset_commit_cb; + + if (onOffsetCommit && typeof onOffsetCommit === 'boolean') { + conf.offset_commit_cb = function(err, offsets) { + if (err) { + err = LibrdKafkaError.create(err); + } + // Emit the event + self.emit('offset.commit', err, offsets); + }; + } else if (onOffsetCommit && typeof onOffsetCommit === 'function') { + conf.offset_commit_cb = function(err, offsets) { + if (err) { + err = LibrdKafkaError.create(err); + } + // Emit the event + self.emit('offset.commit', err, offsets); + onOffsetCommit.call(self, err, offsets); + }; + } + + /** + * KafkaConsumer message. + * + * This is the representation of a message read from Kafka. + * + * @typedef {object} KafkaConsumer~Message + * @property {buffer} value - the message buffer from Kafka. + * @property {string} topic - the topic name + * @property {number} partition - the partition on the topic the + * message was on + * @property {number} offset - the offset of the message + * @property {string} key - the message key + * @property {number} size - message size, in bytes. + * @property {number} timestamp - message timestamp + */ + + Client.call(this, conf, Kafka.KafkaConsumer, topicConf); + + this.globalConfig = conf; + this.topicConfig = topicConf; + + this._consumeTimeout = DEFAULT_CONSUME_TIME_OUT; + this._consumeLoopTimeoutDelay = DEFAULT_CONSUME_LOOP_TIMEOUT_DELAY; +} + +/** + * Set the default consume timeout provided to c++land + * @param {number} timeoutMs - number of milliseconds to wait for a message to be fetched + */ +KafkaConsumer.prototype.setDefaultConsumeTimeout = function(timeoutMs) { + this._consumeTimeout = timeoutMs; +}; + +/** + * Set the default sleep delay for the next consume loop after the previous one has timed out. + * @param {number} intervalMs - number of milliseconds to sleep after a message fetch has timed out + */ +KafkaConsumer.prototype.setDefaultConsumeLoopTimeoutDelay = function(intervalMs) { + this._consumeLoopTimeoutDelay = intervalMs; +}; + +/** + * Get a stream representation of this KafkaConsumer + * + * @see TopicReadable + * @example + * var consumerStream = Kafka.KafkaConsumer.createReadStream({ + * 'metadata.broker.list': 'localhost:9092', + * 'group.id': 'librd-test', + * 'socket.keepalive.enable': true, + * 'enable.auto.commit': false + * }, {}, { topics: [ 'test' ] }); + * + * @param {object} conf - Key value pairs to configure the consumer + * @param {object} topicConf - Key value pairs to create a default + * topic configuration + * @param {object} streamOptions - Stream options + * @param {array} streamOptions.topics - Array of topics to subscribe to. + * @return {KafkaConsumerStream} - Readable stream that receives messages + * when new ones become available. + */ +KafkaConsumer.createReadStream = function(conf, topicConf, streamOptions) { + var consumer = new KafkaConsumer(conf, topicConf); + return new KafkaConsumerStream(consumer, streamOptions); +}; + +/** + * Get a current list of the committed offsets per topic partition + * + * Returns an array of objects in the form of a topic partition list + * + * @param {TopicPartition[]} toppars - Topic partition list to query committed + * offsets for. Defaults to the current assignment + * @param {number} timeout - Number of ms to block before calling back + * and erroring + * @param {Function} cb - Callback method to execute when finished or timed + * out + * @return {Client} - Returns itself + */ +KafkaConsumer.prototype.committed = function(toppars, timeout, cb) { + // We want to be backwards compatible here, and the previous version of + // this function took two arguments + + // If CB is not set, shift to backwards compatible version + if (!cb) { + cb = arguments[1]; + timeout = arguments[0]; + toppars = this.assignments(); + } else { + toppars = toppars || this.assignments(); + } + + var self = this; + this._client.committed(toppars, timeout, function(err, topicPartitions) { + if (err) { + cb(LibrdKafkaError.create(err)); + return; + } + + cb(null, topicPartitions); + }); + return this; +}; + +/** + * Seek consumer for topic+partition to offset which is either an absolute or + * logical offset. + * + * Does not return anything, as it is asynchronous. There are special cases + * with the timeout parameter. The consumer must have previously been assigned + * to topics and partitions that seek seeks to seek. + * + * @example + * consumer.seek({ topic: 'topic', partition: 0, offset: 1000 }, 0, function(err) { + * if (err) { + * + * } + * }); + * + * @param {TopicPartition} toppar - Topic partition to seek. + * @param {number} timeout - Number of ms to block before calling back + * and erroring. If the parameter is null or 0, the call will not wait + * for the seek to be performed. Essentially, it will happen in the background + * with no notification + * @param {Function} cb - Callback method to execute when finished or timed + * out. If the seek timed out, the internal state of the consumer is unknown. + * @return {Client} - Returns itself + */ +KafkaConsumer.prototype.seek = function(toppar, timeout, cb) { + var self = this; + this._client.seek(TopicPartition.create(toppar), timeout, function(err) { + if (err) { + cb(LibrdKafkaError.create(err)); + return; + } + + cb(); + }); + return this; +}; + +/** + * Assign the consumer specific partitions and topics + * + * @param {array} assignments - Assignments array. Should contain + * objects with topic and partition set. + * @return {Client} - Returns itself + */ + +KafkaConsumer.prototype.assign = function(assignments) { + this._client.assign(TopicPartition.map(assignments)); + return this; +}; + +/** + * Unassign the consumer from its assigned partitions and topics. + * + * @return {Client} - Returns itself + */ + +KafkaConsumer.prototype.unassign = function() { + this._client.unassign(); + return this; +}; + + +/** + * Get the assignments for the consumer + * + * @return {array} assignments - Array of topic partitions + */ + +KafkaConsumer.prototype.assignments = function() { + return this._errorWrap(this._client.assignments(), true); +}; + +/** + * Subscribe to an array of topics (synchronously). + * + * This operation is pretty fast because it just sets + * an assignment in librdkafka. This is the recommended + * way to deal with subscriptions in a situation where you + * will be reading across multiple files or as part of + * your configure-time initialization. + * + * This is also a good way to do it for streams. + * + * @param {array} topics - An array of topics to listen to + * @throws - Throws when an error code came back from native land + * @return {KafkaConsumer} - Returns itself. + */ +KafkaConsumer.prototype.subscribe = function(topics) { + // Will throw if it is a bad error. + this._errorWrap(this._client.subscribe(topics)); + this.emit('subscribed', topics); + return this; +}; + +/** + * Get the current subscription of the KafkaConsumer + * + * Get a list of subscribed topics. Should generally match what you + * passed on via subscribe + * + * @see KafkaConsumer::subscribe + * @throws - Throws when an error code came back from native land + * @return {array} - Array of strings to show the current assignment + */ +KafkaConsumer.prototype.subscription = function() { + return this._errorWrap(this._client.subscription(), true); +}; + +/** + * Get the current offset position of the KafkaConsumer + * + * Returns a list of RdKafka::TopicPartitions on success, or throws + * an error on failure + * + * @param {TopicPartition[]} toppars - List of topic partitions to query + * position for. Defaults to the current assignment + * @throws - Throws when an error code came back from native land + * @return {array} - TopicPartition array. Each item is an object with + * an offset, topic, and partition + */ +KafkaConsumer.prototype.position = function(toppars) { + if (!toppars) { + toppars = this.assignments(); + } + return this._errorWrap(this._client.position(toppars), true); +}; + +/** + * Unsubscribe from all currently subscribed topics + * + * Before you subscribe to new topics you need to unsubscribe + * from the old ones, if there is an active subscription. + * Otherwise, you will get an error because there is an + * existing subscription. + * + * @throws - Throws when an error code comes back from native land + * @return {KafkaConsumer} - Returns itself. + */ +KafkaConsumer.prototype.unsubscribe = function() { + this._errorWrap(this._client.unsubscribe()); + this.emit('unsubscribed', []); + // Backwards compatible change + this.emit('unsubscribe', []); + return this; +}; + +/** + * Read a number of messages from Kafka. + * + * This method is similar to the main one, except that it reads a number + * of messages before calling back. This may get better performance than + * reading a single message each time in stream implementations. + * + * This will keep going until it gets ERR__PARTITION_EOF or ERR__TIMED_OUT + * so the array may not be the same size you ask for. The size is advisory, + * but we will not exceed it. + * + * @param {number} size - Number of messages to read + * @param {KafkaConsumer~readCallback} cb - Callback to return when work is done. + *//** + * Read messages from Kafka as fast as possible + * + * This method keeps a background thread running to fetch the messages + * as quickly as it can, sleeping only in between EOF and broker timeouts. + * + * Use this to get the maximum read performance if you don't care about the + * stream backpressure. + * @param {KafkaConsumer~readCallback} cb - Callback to return when a message + * is fetched. + */ +KafkaConsumer.prototype.consume = function(number, cb) { + var timeoutMs = this._consumeTimeout || DEFAULT_CONSUME_TIME_OUT; + var self = this; + + if ((number && typeof number === 'number') || (number && cb)) { + + if (cb === undefined) { + cb = function() {}; + } else if (typeof cb !== 'function') { + throw new TypeError('Callback must be a function'); + } + + this._consumeNum(timeoutMs, number, cb); + } else { + + // See https://github.com/Blizzard/node-rdkafka/issues/220 + // Docs specify just a callback can be provided but really we needed + // a fallback to the number argument + // @deprecated + if (cb === undefined) { + if (typeof number === 'function') { + cb = number; + } else { + cb = function() {}; + } + } + + this._consumeLoop(timeoutMs, cb); + } +}; + +/** + * Open a background thread and keep getting messages as fast + * as we can. Should not be called directly, and instead should + * be called using consume. + * + * @private + * @see consume + */ +KafkaConsumer.prototype._consumeLoop = function(timeoutMs, cb) { + var self = this; + var retryReadInterval = this._consumeLoopTimeoutDelay; + self._client.consumeLoop(timeoutMs, retryReadInterval, function readCallback(err, message, eofEvent, warning) { + + if (err) { + // A few different types of errors here + // but the two we do NOT care about are + // time outs at least now + // Broker no more messages will also not come here + cb(LibrdKafkaError.create(err)); + } else if (eofEvent) { + self.emit('partition.eof', eofEvent); + } else if (warning) { + self.emit('warning', LibrdKafkaError.create(warning)); + } else { + /** + * Data event. called whenever a message is received. + * + * @event KafkaConsumer#data + * @type {KafkaConsumer~Message} + */ + self.emit('data', message); + cb(err, message); + } + }); + +}; + +/** + * Consume a number of messages and wrap in a try catch with + * proper error reporting. Should not be called directly, + * and instead should be called using consume. + * + * @private + * @see consume + */ +KafkaConsumer.prototype._consumeNum = function(timeoutMs, numMessages, cb) { + var self = this; + + this._client.consume(timeoutMs, numMessages, function(err, messages, eofEvents) { + if (err) { + err = LibrdKafkaError.create(err); + if (cb) { + cb(err); + } + return; + } + + var currentEofEventsIndex = 0; + + function emitEofEventsFor(messageIndex) { + while (currentEofEventsIndex < eofEvents.length && eofEvents[currentEofEventsIndex].messageIndex === messageIndex) { + delete eofEvents[currentEofEventsIndex].messageIndex; + self.emit('partition.eof', eofEvents[currentEofEventsIndex]) + ++currentEofEventsIndex; + } + } + + emitEofEventsFor(-1); + + for (var i = 0; i < messages.length; i++) { + self.emit('data', messages[i]); + emitEofEventsFor(i); + } + + emitEofEventsFor(messages.length); + + if (cb) { + cb(null, messages); + } + + }); + +}; + +/** + * This callback returns the message read from Kafka. + * + * @callback KafkaConsumer~readCallback + * @param {LibrdKafkaError} err - An error, if one occurred while reading + * the data. + * @param {KafkaConsumer~Message} message + */ + +/** + * Commit a topic partition or all topic partitions that have been read + * + * If you provide a topic partition, it will commit that. Otherwise, + * it will commit all read offsets for all topic partitions. + * + * @param {object|array|null} - Topic partition object to commit, list of topic + * partitions, or null if you want to commit all read offsets. + * @throws When commit returns a non 0 error code + * + * @return {KafkaConsumer} - returns itself. + */ +KafkaConsumer.prototype.commit = function(topicPartition) { + this._errorWrap(this._client.commit(topicPartition), true); + return this; +}; + +/** + * Commit a message + * + * This is basically a convenience method to map commit properly. We need to + * add one to the offset in this case + * + * @param {object} - Message object to commit + * @throws When commit returns a non 0 error code + * + * @return {KafkaConsumer} - returns itself. + */ +KafkaConsumer.prototype.commitMessage = function(msg) { + var topicPartition = { + topic: msg.topic, + partition: msg.partition, + offset: msg.offset + 1 + }; + + this._errorWrap(this._client.commit(topicPartition), true); + return this; +}; + +/** + * Commit a topic partition (or all topic partitions) synchronously + * + * @param {object|array|null} - Topic partition object to commit, list of topic + * partitions, or null if you want to commit all read offsets. + * @throws {LibrdKafkaError} - if the commit fails + * + * @return {KafkaConsumer} - returns itself. + */ +KafkaConsumer.prototype.commitSync = function(topicPartition) { + this._errorWrap(this._client.commitSync(topicPartition), true); + return this; +}; + +/** + * Commit a message synchronously + * + * @see KafkaConsumer#commitMessageSync + * @param {object} msg - A message object to commit. + * + * @throws {LibrdKafkaError} - if the commit fails + * + * @return {KafkaConsumer} - returns itself. + */ +KafkaConsumer.prototype.commitMessageSync = function(msg) { + var topicPartition = { + topic: msg.topic, + partition: msg.partition, + offset: msg.offset + 1 + }; + + this._errorWrap(this._client.commitSync(topicPartition), true); + return this; +}; + +/** + * Get last known offsets from the client. + * + * The low offset is updated periodically (if statistics.interval.ms is set) + * while the high offset is updated on each fetched message set from the + * broker. + * + * If there is no cached offset (either low or high, or both), then this will + * throw an error. + * + * @param {string} topic - Topic to recieve offsets from. + * @param {number} partition - Partition of the provided topic to recieve offsets from + * @return {Client~watermarkOffsets} - Returns an object with a high and low property, specifying + * the high and low offsets for the topic partition + * @throws {LibrdKafkaError} - Throws when there is no offset stored + */ +KafkaConsumer.prototype.getWatermarkOffsets = function(topic, partition) { + if (!this.isConnected()) { + throw new Error('Client is disconnected'); + } + + return this._errorWrap(this._client.getWatermarkOffsets(topic, partition), true); +}; + +/** + * Store offset for topic partition. + * + * The offset will be committed (written) to the offset store according to the auto commit interval, + * if auto commit is on, or next manual offset if not. + * + * enable.auto.offset.store must be set to false to use this API, + * + * @see https://github.com/edenhill/librdkafka/blob/261371dc0edef4cea9e58a076c8e8aa7dc50d452/src-cpp/rdkafkacpp.h#L1702 + * + * @param {Array.} topicPartitions - Topic partitions with offsets to store offsets for. + * @throws {LibrdKafkaError} - Throws when there is no offset stored + */ +KafkaConsumer.prototype.offsetsStore = function(topicPartitions) { + if (!this.isConnected()) { + throw new Error('Client is disconnected'); + } + + return this._errorWrap(this._client.offsetsStore(topicPartitions), true); +}; + +/** + * Resume consumption for the provided list of partitions. + * + * @param {Array.} topicPartitions - List of topic partitions to resume consumption on. + * @throws {LibrdKafkaError} - Throws when there is no offset stored + */ +KafkaConsumer.prototype.resume = function(topicPartitions) { + if (!this.isConnected()) { + throw new Error('Client is disconnected'); + } + + return this._errorWrap(this._client.resume(topicPartitions), true); +}; + +/** + * Pause producing or consumption for the provided list of partitions. + * + * @param {Array.} topicPartitions - List of topics to pause consumption on. + * @throws {LibrdKafkaError} - Throws when there is no offset stored + */ +KafkaConsumer.prototype.pause = function(topicPartitions) { + if (!this.isConnected()) { + throw new Error('Client is disconnected'); + } + + return this._errorWrap(this._client.pause(topicPartitions), true); +}; diff --git a/lib/producer-stream.js b/lib/producer-stream.js new file mode 100644 index 0000000..eb53208 --- /dev/null +++ b/lib/producer-stream.js @@ -0,0 +1,307 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +'use strict'; + +module.exports = ProducerStream; + +var Writable = require('stream').Writable; +var util = require('util'); +var ErrorCode = require('./error').codes; + +util.inherits(ProducerStream, Writable); + +/** + * Writable stream integrating with the Kafka Producer. + * + * This class is used to write data to Kafka in a streaming way. It takes + * buffers of data and puts them into the appropriate Kafka topic. If you need + * finer control over partitions or keys, this is probably not the class for + * you. In that situation just use the Producer itself. + * + * The stream detects if Kafka is already connected. You can safely begin + * writing right away. + * + * This stream does not operate in Object mode and can only be given buffers. + * + * @param {Producer} producer - The Kafka Producer object. + * @param {array} topics - Array of topics + * @param {object} options - Topic configuration. + * @constructor + * @extends stream.Writable + */ +function ProducerStream(producer, options) { + if (!(this instanceof ProducerStream)) { + return new ProducerStream(producer, options); + } + + if (options === undefined) { + options = {}; + } else if (typeof options === 'string') { + options = { encoding: options }; + } else if (options === null || typeof options !== 'object') { + throw new TypeError('"streamOptions" argument must be a string or an object'); + } + + if (!options.objectMode && !options.topic) { + throw new TypeError('ProducerStreams not using objectMode must provide a topic to produce to.'); + } + + if (options.objectMode !== true) { + this._write = this._write_buffer; + } else { + this._write = this._write_message; + } + + Writable.call(this, options); + + this.producer = producer; + this.topicName = options.topic; + + this.autoClose = options.autoClose === undefined ? true : !!options.autoClose; + this.connectOptions = options.connectOptions || {}; + + this.producer.setPollInterval(options.pollInterval || 1000); + + if (options.encoding) { + this.setDefaultEncoding(options.encoding); + } + + // Connect to the producer. Unless we are already connected + if (!this.producer.isConnected()) { + this.connect(this.connectOptions); + } + + var self = this; + + this.once('finish', function() { + if (this.autoClose) { + this.close(); + } + }); + +} + +ProducerStream.prototype.connect = function(options) { + this.producer.connect(options, function(err, data) { + if (err) { + this.emit('error', err); + return; + } + + }.bind(this)); +}; + +/** + * Internal stream write method for ProducerStream when writing buffers. + * + * This method should never be called externally. It has some recursion to + * handle cases where the producer is not yet connected. + * + * @param {buffer} chunk - Chunk to write. + * @param {string} encoding - Encoding for the buffer + * @param {Function} cb - Callback to call when the stream is done processing + * the data. + * @private + * @see https://github.com/nodejs/node/blob/master/lib/fs.js#L1901 + */ +ProducerStream.prototype._write_buffer = function(data, encoding, cb) { + if (!(data instanceof Buffer)) { + this.emit('error', new Error('Invalid data. Can only produce buffers')); + return; + } + + var self = this; + + if (!this.producer.isConnected()) { + this.producer.once('ready', function() { + self._write(data, encoding, cb); + }); + return; + } + + try { + this.producer.produce(self.topicName, null, data, null); + setImmediate(cb); + } catch (e) { + if (ErrorCode.ERR__QUEUE_FULL === e.code) { + // Poll for good measure + self.producer.poll(); + + // Just delay this thing a bit and pass the params + // backpressure will get exerted this way. + setTimeout(function() { + self._write(data, encoding, cb); + }, 500); + } else { + if (self.autoClose) { + self.close(); + } + setImmediate(function() { + cb(e); + }); + } + } +}; + +/** + * Internal stream write method for ProducerStream when writing objects. + * + * This method should never be called externally. It has some recursion to + * handle cases where the producer is not yet connected. + * + * @param {object} message - Message to write. + * @param {string} encoding - Encoding for the buffer + * @param {Function} cb - Callback to call when the stream is done processing + * the data. + * @private + * @see https://github.com/nodejs/node/blob/master/lib/fs.js#L1901 + */ +ProducerStream.prototype._write_message = function(message, encoding, cb) { + var self = this; + + if (!this.producer.isConnected()) { + this.producer.once('ready', function() { + self._write(message, encoding, cb); + }); + return; + } + + try { + this.producer.produce(message.topic, message.partition, message.value, message.key, message.timestamp, message.opaque, message.headers); + setImmediate(cb); + } catch (e) { + if (ErrorCode.ERR__QUEUE_FULL === e.code) { + // Poll for good measure + self.producer.poll(); + + // Just delay this thing a bit and pass the params + // backpressure will get exerted this way. + setTimeout(function() { + self._write(message, encoding, cb); + }, 500); + } else { + if (self.autoClose) { + self.close(); + } + setImmediate(function() { + cb(e); + }); + } + } +}; + +function writev(producer, topic, chunks, cb) { + + // @todo maybe a produce batch method? + var doneCount = 0; + var err = null; + var chunk = null; + + function maybeDone(e) { + if (e) { + err = e; + } + doneCount ++; + if (doneCount === chunks.length) { + cb(err); + } + } + + function retry(restChunks) { + // Poll for good measure + producer.poll(); + + // Just delay this thing a bit and pass the params + // backpressure will get exerted this way. + setTimeout(function() { + writev(producer, topic, restChunks, cb); + }, 500); + } + + for (var i = 0; i < chunks.length; i++) { + chunk = chunks[i]; + + try { + if (Buffer.isBuffer(chunk)) { + producer.produce(topic, null, chunk, null); + } else { + producer.produce(chunk.topic, chunk.partition, chunk.value, chunk.key, chunk.timestamp, chunk.opaque, chunk.headers); + } + maybeDone(); + } catch (e) { + if (ErrorCode.ERR__QUEUE_FULL === e.code) { + retry(chunks.slice(i)); + } else { + cb(e); + } + break; + } + } + +} + +ProducerStream.prototype._writev = function(data, cb) { + if (!this.producer.isConnected()) { + this.once('ready', function() { + this._writev(data, cb); + }); + return; + } + + var self = this; + var len = data.length; + var chunks = new Array(len); + var size = 0; + + for (var i = 0; i < len; i++) { + var chunk = data[i].chunk; + + chunks[i] = chunk; + size += chunk.length; + } + + writev(this.producer, this.topicName, chunks, function(err) { + if (err) { + self.close(); + cb(err); + return; + } + cb(); + }); + +}; + +ProducerStream.prototype.close = function(cb) { + var self = this; + if (cb) { + this.once('close', cb); + } + + // Use interval variables in here + if (self.producer._isConnected) { + self.producer.disconnect(function() { + // Previously this set the producer to null. I'm not sure there is any benefit + // to that other than I guess helping flag it for GC? + // https://github.com/Blizzard/node-rdkafka/issues/344 + close(); + }); + } else if (self.producer._isConnecting){ + self.producer.once('ready', function() { + // Don't pass CB this time because it has already been passed + self.close(); + }); + } else { + setImmediate(close); + } + + function close() { + self.emit('close'); + } +}; diff --git a/lib/producer.js b/lib/producer.js new file mode 100644 index 0000000..67f4f67 --- /dev/null +++ b/lib/producer.js @@ -0,0 +1,375 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +module.exports = Producer; + +var Client = require('./client'); + +var util = require('util'); +var Kafka = require('../librdkafka.js'); +var ProducerStream = require('./producer-stream'); +var LibrdKafkaError = require('./error'); +var shallowCopy = require('./util').shallowCopy; + +util.inherits(Producer, Client); + +/** + * Producer class for sending messages to Kafka + * + * This is the main entry point for writing data to Kafka. You + * configure this like you do any other client, with a global + * configuration and default topic configuration. + * + * Once you instantiate this object, you need to connect to it first. + * This allows you to get the metadata and make sure the connection + * can be made before you depend on it. After that, problems with + * the connection will by brought down by using poll, which automatically + * runs when a transaction is made on the object. + * + * @param {object} conf - Key value pairs to configure the producer + * @param {object} topicConf - Key value pairs to create a default + * topic configuration + * @extends Client + * @constructor + */ +function Producer(conf, topicConf) { + if (!(this instanceof Producer)) { + return new Producer(conf, topicConf); + } + + conf = shallowCopy(conf); + topicConf = shallowCopy(topicConf); + + /** + * Producer message. This is sent to the wrapper, not received from it + * + * @typedef {object} Producer~Message + * @property {string|buffer} message - The buffer to send to Kafka. + * @property {Topic} topic - The Kafka topic to produce to. + * @property {number} partition - The partition to produce to. Defaults to + * the partitioner + * @property {string} key - The key string to use for the message. + * @see Consumer~Message + */ + + var gTopic = conf.topic || false; + var gPart = conf.partition || null; + var dr_cb = conf.dr_cb || null; + var dr_msg_cb = conf.dr_msg_cb || null; + + // delete keys we don't want to pass on + delete conf.topic; + delete conf.partition; + + delete conf.dr_cb; + delete conf.dr_msg_cb; + + // client is an initialized consumer object + // @see NodeKafka::Producer::Init + Client.call(this, conf, Kafka.Producer, topicConf); + + // Delete these keys after saving them in vars + this.globalConfig = conf; + this.topicConfig = topicConf; + this.defaultTopic = gTopic || null; + this.defaultPartition = gPart == null ? -1 : gPart; + + this.sentMessages = 0; + + this.pollInterval = undefined; + + if (dr_msg_cb || dr_cb) { + this._cb_configs.event.delivery_cb = function(err, report) { + if (err) { + err = LibrdKafkaError.create(err); + } + this.emit('delivery-report', err, report); + }.bind(this); + this._cb_configs.event.delivery_cb.dr_msg_cb = !!dr_msg_cb; + + if (typeof dr_cb === 'function') { + this.on('delivery-report', dr_cb); + } + + } +} + +/** + * Produce a message to Kafka synchronously. + * + * This is the method mainly used in this class. Use it to produce + * a message to Kafka. + * + * When this is sent off, there is no guarantee it is delivered. If you need + * guaranteed delivery, change your *acks* settings, or use delivery reports. + * + * @param {string} topic - The topic name to produce to. + * @param {number|null} partition - The partition number to produce to. + * @param {Buffer|null} message - The message to produce. + * @param {string} key - The key associated with the message. + * @param {number|null} timestamp - Timestamp to send with the message. + * @param {object} opaque - An object you want passed along with this message, if provided. + * @param {object} headers - A list of custom key value pairs that provide message metadata. + * @throws {LibrdKafkaError} - Throws a librdkafka error if it failed. + * @return {boolean} - returns an error if it failed, or true if not + * @see Producer#produce + */ +Producer.prototype.produce = function(topic, partition, message, key, timestamp, opaque, headers) { + if (!this._isConnected) { + throw new Error('Producer not connected'); + } + + // I have removed support for using a topic object. It is going to be removed + // from librdkafka soon, and it causes issues with shutting down + if (!topic || typeof topic !== 'string') { + throw new TypeError('"topic" must be a string'); + } + + this.sentMessages++; + + partition = partition == null ? this.defaultPartition : partition; + + return this._errorWrap( + this._client.produce(topic, partition, message, key, timestamp, opaque, headers)); + +}; + +/** + * Create a write stream interface for a producer. + * + * This stream does not run in object mode. It only takes buffers of data. + * + * @param {object} conf - Key value pairs to configure the producer + * @param {object} topicConf - Key value pairs to create a default + * topic configuration + * @param {object} streamOptions - Stream options + * @return {ProducerStream} - returns the write stream for writing to Kafka. + */ +Producer.createWriteStream = function(conf, topicConf, streamOptions) { + var producer = new Producer(conf, topicConf); + return new ProducerStream(producer, streamOptions); +}; + +/** + * Poll for events + * + * We need to run poll in order to learn about new events that have occurred. + * This is no longer done automatically when we produce, so we need to run + * it manually, or set the producer to automatically poll. + * + * @return {Producer} - returns itself. + */ +Producer.prototype.poll = function() { + if (!this._isConnected) { + throw new Error('Producer not connected'); + } + this._client.poll(); + return this; +}; + +/** + * Set automatic polling for events. + * + * We need to run poll in order to learn about new events that have occurred. + * If you would like this done on an interval with disconnects and reconnections + * managed, you can do it here + * + * @param {number} interval - Interval, in milliseconds, to poll + * + * @return {Producer} - returns itself. + */ +Producer.prototype.setPollInterval = function(interval) { + // If we already have a poll interval we need to stop it + if (this.pollInterval) { + clearInterval(this.pollInterval); + this.pollInterval = undefined; + } + + if (interval === 0) { + // If the interval was set to 0, bail out. We don't want to process this. + // If there was an interval previously set, it has been removed. + return; + } + + var self = this; + + // Now we want to make sure we are connected. + if (!this._isConnected) { + // If we are not, execute this once the connection goes through. + this.once('ready', function() { + self.setPollInterval(interval); + }); + return; + } + + // We know we are connected at this point. + // Unref this interval + this.pollInterval = setInterval(function() { + try { + self.poll(); + } catch (e) { + // We can probably ignore errors here as far as broadcasting. + // Disconnection issues will get handled below + } + }, interval).unref(); + + // Handle disconnections + this.once('disconnected', function() { + // Just rerun this function with interval 0. If any + // poll interval is set, this will remove it + self.setPollInterval(0); + }); + + return this; +}; + +/** + * Flush the producer + * + * Flush everything on the internal librdkafka producer buffer. Do this before + * disconnects usually + * + * @param {number} timeout - Number of milliseconds to try to flush before giving up. + * @param {function} callback - Callback to fire when the flush is done. + * + * @return {Producer} - returns itself. + */ +Producer.prototype.flush = function(timeout, callback) { + if (!this._isConnected) { + throw new Error('Producer not connected'); + } + + if (timeout === undefined || timeout === null) { + timeout = 500; + } + + this._client.flush(timeout, function(err) { + if (err) { + err = LibrdKafkaError.create(err); + } + + if (callback) { + callback(err); + } + }); + return this; +}; + +/** + * Save the base disconnect method here so we can overwrite it and add a flush + */ +Producer.prototype._disconnect = Producer.prototype.disconnect; + +/** + * Disconnect the producer + * + * Flush everything on the internal librdkafka producer buffer. Then disconnect + * + * @param {number} timeout - Number of milliseconds to try to flush before giving up, defaults to 5 seconds. + * @param {function} cb - The callback to fire when + */ +Producer.prototype.disconnect = function(timeout, cb) { + var self = this; + var timeoutInterval = 5000; + + if (typeof timeout === 'function') { + cb = timeout; + } else { + timeoutInterval = timeout; + } + + this.flush(timeoutInterval, function() { + self._disconnect(cb); + }); +}; + +/** + * Init a transaction. + * + * Initialize transactions, this is only performed once per transactional producer. + * + * @param {number} timeout - Number of milliseconds to try to initialize before giving up, defaults to 5 seconds. + * @param {function} cb - Callback to return when operation is completed + * @return {Producer} - returns itself. + */ +Producer.prototype.initTransactions = function(timeout, cb) { + if (typeof timeout === 'function') { + cb = timeout; + timeout = 5000; + } + this._client.initTransactions(timeout, function(err) { + cb(err ? LibrdKafkaError.create(err) : err); + }); +}; + +/** + * Begin a transaction. + * + * 'initTransaction' must have been called successfully (once) before this function is called. + * + * @return {Producer} - returns itself. + */ +Producer.prototype.beginTransaction = function(cb) { + this._client.beginTransaction(function(err) { + cb(err ? LibrdKafkaError.create(err) : err); + }); +}; + +/** + * Commit the current transaction (as started with 'beginTransaction'). + * + * @param {number} timeout - Number of milliseconds to try to commit before giving up, defaults to 5 seconds + * @param {function} cb - Callback to return when operation is completed + * @return {Producer} - returns itself. + */ +Producer.prototype.commitTransaction = function(timeout, cb) { + if (typeof timeout === 'function') { + cb = timeout; + timeout = 5000; + } + this._client.commitTransaction(timeout, function(err) { + cb(err ? LibrdKafkaError.create(err) : err); + }); +}; + +/** + * Aborts the ongoing transaction. + * + * @param {number} timeout - Number of milliseconds to try to abort, defaults to 5 seconds + * @param {function} cb - Callback to return when operation is completed + * @return {Producer} - returns itself. + */ +Producer.prototype.abortTransaction = function(timeout, cb) { + if (typeof timeout === 'function') { + cb = timeout; + timeout = 5000; + } + this._client.abortTransaction(timeout, function(err) { + cb(err ? LibrdKafkaError.create(err) : err); + }); +}; + +/** + * Send the current offsets of the consumer to the ongoing transaction. + * + * @param {number} offsets - Offsets to send as part of the next commit + * @param {Consumer} consumer - An instance of the consumer + * @param {number} timeout - Number of milliseconds to try to send offsets, defaults to 5 seconds + * @param {function} cb - Callback to return when operation is completed + * @return {Producer} - returns itself. + */ +Producer.prototype.sendOffsetsToTransaction = function(offsets, consumer, timeout, cb) { + if (typeof timeout === 'function') { + cb = timeout; + timeout = 5000; + } + this._client.sendOffsetsToTransaction(offsets, consumer.getClient(), timeout, function(err) { + cb(err ? LibrdKafkaError.create(err) : err); + }); +}; diff --git a/lib/producer/high-level-producer.js b/lib/producer/high-level-producer.js new file mode 100644 index 0000000..5dbda2f --- /dev/null +++ b/lib/producer/high-level-producer.js @@ -0,0 +1,323 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +module.exports = HighLevelProducer; + +var util = require('util'); +var Producer = require('../producer'); +var LibrdKafkaError = require('../error'); +var EventEmitter = require('events').EventEmitter; +var RefCounter = require('../tools/ref-counter'); +var shallowCopy = require('../util').shallowCopy; +var isObject = require('../util').isObject; + +util.inherits(HighLevelProducer, Producer); + +var noopSerializer = createSerializer(function (v) { return v; }); + +/** + * Create a serializer + * + * Method simply wraps a serializer provided by a user + * so it adds context to the error + * + * @returns {function} Serialization function + */ +function createSerializer(serializer) { + var applyFn = function serializationWrapper(v, cb) { + try { + return cb ? serializer(v, cb) : serializer(v); + } catch (e) { + var modifiedError = new Error('Could not serialize value: ' + e.message); + modifiedError.value = v; + modifiedError.serializer = serializer; + throw modifiedError; + } + }; + + // We can check how many parameters the function has and activate the asynchronous + // operation if the number of parameters the function accepts is > 1 + return { + apply: applyFn, + async: serializer.length > 1 + }; +} + +/** + * Producer class for sending messages to Kafka in a higher level fashion + * + * This is the main entry point for writing data to Kafka if you want more + * functionality than librdkafka supports out of the box. You + * configure this like you do any other client, with a global + * configuration and default topic configuration. + * + * Once you instantiate this object, you need to connect to it first. + * This allows you to get the metadata and make sure the connection + * can be made before you depend on it. After that, problems with + * the connection will by brought down by using poll, which automatically + * runs when a transaction is made on the object. + * + * This has a few restrictions, so it is not for free! + * + * 1. You may not define opaque tokens + * The higher level producer is powered by opaque tokens. + * 2. Every message ack will dispatch an event on the node thread. + * 3. Will use a ref counter to determine if there are outgoing produces. + * + * This will return the new object you should use instead when doing your + * produce calls + * + * @param {object} conf - Key value pairs to configure the producer + * @param {object} topicConf - Key value pairs to create a default + * topic configuration + * @extends Producer + * @constructor + */ +function HighLevelProducer(conf, topicConf) { + if (!(this instanceof HighLevelProducer)) { + return new HighLevelProducer(conf, topicConf); + } + + // Force this to be true for the high level producer + conf = shallowCopy(conf); + conf.dr_cb = true; + + // producer is an initialized consumer object + // @see NodeKafka::Producer::Init + Producer.call(this, conf, topicConf); + var self = this; + + // Add a delivery emitter to the producer + this._hl = { + deliveryEmitter: new EventEmitter(), + messageId: 0, + // Special logic for polling. We use a reference counter to know when we need + // to be doing it and when we can stop. This means when we go into fast polling + // mode we don't need to do multiple calls to poll since they all will yield + // the same result + pollingRefTimeout: null, + }; + + // Add the polling ref counter to the class which ensures we poll when we go active + this._hl.pollingRef = new RefCounter(function() { + self._hl.pollingRefTimeout = setInterval(function() { + try { + self.poll(); + } catch (e) { + if (!self._isConnected) { + // If we got disconnected for some reason there is no point + // in polling anymore + clearInterval(self._hl.pollingRefTimeout); + } + } + }, 1); + }, function() { + clearInterval(self._hl.pollingRefTimeout); + }); + + // Default poll interval. More sophisticated polling is also done in create rule method + this.setPollInterval(1000); + + // Listen to all delivery reports to propagate elements with a _message_id to the emitter + this.on('delivery-report', function(err, report) { + if (report.opaque && report.opaque.__message_id !== undefined) { + self._hl.deliveryEmitter.emit(report.opaque.__message_id, err, report.offset); + } + }); + + // Save old produce here since we are making some modifications for it + this._oldProduce = this.produce; + this.produce = this._modifiedProduce; + + // Serializer information + this.keySerializer = noopSerializer; + this.valueSerializer = noopSerializer; +} + +/** + * Produce a message to Kafka asynchronously. + * + * This is the method mainly used in this class. Use it to produce + * a message to Kafka. + * + * When this is sent off, and you recieve your callback, the assurances afforded + * to you will be equal to those provided by your ack level. + * + * @param {string} topic - The topic name to produce to. + * @param {number|null} partition - The partition number to produce to. + * @param {Buffer|null} message - The message to produce. + * @param {string} key - The key associated with the message. + * @param {number|null} timestamp - Timestamp to send with the message. + * @param {object} headers - A list of custom key value pairs that provide message metadata. + * @param {function} callback - Callback to call when the delivery report is recieved. + * @throws {LibrdKafkaError} - Throws a librdkafka error if it failed. + * @return {boolean} - returns an error if it failed, or true if not + * @see Producer#produce + */ +HighLevelProducer.prototype._modifiedProduce = function(topic, partition, message, key, timestamp, headers, callback) { + // headers are optional + if (arguments.length === 6) { + callback = headers; + headers = undefined; + } + + // Add the message id + var opaque = { + __message_id: this._hl.messageId++, + }; + + this._hl.pollingRef.increment(); + + var self = this; + + var resolvedSerializedValue; + var resolvedSerializedKey; + var calledBack = false; + + // Actually do the produce with new key and value based on deserialized + // results + function doProduce(v, k) { + try { + var r = self._oldProduce(topic, partition, + v, k, + timestamp, opaque, headers); + + self._hl.deliveryEmitter.once(opaque.__message_id, function(err, offset) { + self._hl.pollingRef.decrement(); + setImmediate(function() { + // Offset must be greater than or equal to 0 otherwise it is a null offset + // Possibly because we have acks off + callback(err, offset >= 0 ? offset : null); + }); + }); + + return r; + } catch (e) { + callback(e); + } + } + + function produceIfComplete() { + if (resolvedSerializedKey !== undefined && resolvedSerializedValue !== undefined) { + doProduce(resolvedSerializedValue, resolvedSerializedKey); + } + } + + // To run on a promise if returned by the serializer + function finishSerializedValue(v) { + if (!calledBack) { + resolvedSerializedValue = v; + produceIfComplete(); + } + } + + // To run on a promise of returned by the serializer + function finishSerializedKey(k) { + resolvedSerializedKey = k; + + if (!calledBack) { + produceIfComplete(); + } + } + + function failSerializedValue(err) { + if (!calledBack) { + calledBack = true; + callback(err); + } + } + + function failSerializedKey(err) { + if (!calledBack) { + calledBack = true; + callback(err); + } + } + + function valueSerializerCallback(err, v) { + if (err) { + failSerializedValue(err); + } else { + finishSerializedValue(v); + } + } + + function keySerializerCallback(err, v) { + if (err) { + failSerializedKey(err); + } else { + finishSerializedKey(v); + } + } + + try { + if (this.valueSerializer.async) { + // If this is async we need to give it a callback + this.valueSerializer.apply(message, valueSerializerCallback); + } else { + var serializedValue = this.valueSerializer.apply(message); + // Check if we were returned a promise in order to support promise behavior + if (serializedValue && + typeof serializedValue.then === 'function' && + typeof serializedValue.catch === 'function') { + // This is a promise. We need to hook into its then and catch + serializedValue.then(finishSerializedValue).catch(failSerializedValue); + } else { + resolvedSerializedValue = serializedValue; + } + } + + if (this.keySerializer.async) { + // If this is async we need to give it a callback + this.keySerializer.apply(key, keySerializerCallback); + } else { + var serializedKey = this.keySerializer.apply(key); + // Check if we were returned a promise in order to support promise behavior + if (serializedKey && + typeof serializedKey.then === 'function' && + typeof serializedKey.catch === 'function') { + // This is a promise. We need to hook into its then and catch + serializedKey.then(finishSerializedKey).catch(failSerializedKey); + } else { + resolvedSerializedKey = serializedKey; + } + } + + // Only do the produce here if we are complete. That is, if the key + // and value have been serialized. + produceIfComplete(); + } catch (e) { + setImmediate(function() { + calledBack = true; + callback(e); + }); + } +}; + +/** + * Set the key serializer + * + * This allows the value inside the produce call to differ from the value of the + * value actually produced to kafka. Good if, for example, you want to serialize + * it to a particular format. + */ +HighLevelProducer.prototype.setKeySerializer = function(serializer) { + this.keySerializer = createSerializer(serializer); +}; + +/** + * Set the value serializer + * + * This allows the value inside the produce call to differ from the value of the + * value actually produced to kafka. Good if, for example, you want to serialize + * it to a particular format. + */ +HighLevelProducer.prototype.setValueSerializer = function(serializer) { + this.valueSerializer = createSerializer(serializer); +}; diff --git a/lib/tools/ref-counter.js b/lib/tools/ref-counter.js new file mode 100644 index 0000000..ae23078 --- /dev/null +++ b/lib/tools/ref-counter.js @@ -0,0 +1,52 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +module.exports = RefCounter; + +/** + * Ref counter class. + * + * Is used to basically determine active/inactive and allow callbacks that + * hook into each. + * + * For the producer, it is used to begin rapid polling after a produce until + * the delivery report is dispatched. + */ +function RefCounter(onActive, onPassive) { + this.context = {}; + this.onActive = onActive; + this.onPassive = onPassive; + this.currentValue = 0; + this.isRunning = false; +} + +/** + * Increment the ref counter + */ +RefCounter.prototype.increment = function() { + this.currentValue += 1; + + // If current value exceeds 0, activate the start + if (this.currentValue > 0 && !this.isRunning) { + this.isRunning = true; + this.onActive(this.context); + } +}; + +/** + * Decrement the ref counter + */ +RefCounter.prototype.decrement = function() { + this.currentValue -= 1; + + if (this.currentValue <= 0 && this.isRunning) { + this.isRunning = false; + this.onPassive(this.context); + } +}; diff --git a/lib/topic-partition.js b/lib/topic-partition.js new file mode 100644 index 0000000..0961c85 --- /dev/null +++ b/lib/topic-partition.js @@ -0,0 +1,88 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Topic = require('./topic'); + +module.exports = TopicPartition; + +/** + * Map an array of topic partition js objects to real topic partition objects. + * + * @param array The array of topic partition raw objects to map to topic + * partition objects + */ +TopicPartition.map = function(array) { + return array.map(function(element) { + return TopicPartition.create(element); + }); +}; + +/** + * Take a topic partition javascript object and convert it to the class. + * The class will automatically convert offset identifiers to special constants + * + * @param element The topic partition raw javascript object + */ +TopicPartition.create = function(element) { + // Just ensure we take something that can have properties. The topic partition + // class will + element = element || {}; + return new TopicPartition(element.topic, element.partition, element.offset); +}; + +/** + * Create a topic partition. Just does some validation and decoration + * on topic partitions provided. + * + * Goal is still to behave like a plain javascript object but with validation + * and potentially some extra methods + */ +function TopicPartition(topic, partition, offset) { + if (!(this instanceof TopicPartition)) { + return new TopicPartition(topic, partition, offset); + } + + // Validate that the elements we are iterating over are actual topic partition + // js objects. They do not need an offset, but they do need partition + if (!topic) { + throw new TypeError('"topic" must be a string and must be set'); + } + + if (partition === null || partition === undefined) { + throw new TypeError('"partition" must be a number and must set'); + } + + // We can just set topic and partition as they stand. + this.topic = topic; + this.partition = partition; + + if (offset === undefined || offset === null) { + this.offset = Topic.OFFSET_STORED; + } else if (typeof offset === 'string') { + switch (offset.toLowerCase()) { + case 'earliest': + case 'beginning': + this.offset = Topic.OFFSET_BEGINNING; + break; + case 'latest': + case 'end': + this.offset = Topic.OFFSET_END; + break; + case 'stored': + this.offset = Topic.OFFSET_STORED; + break; + default: + throw new TypeError('"offset", if provided as a string, must be beginning, end, or stored.'); + } + } else if (typeof offset === 'number') { + this.offset = offset; + } else { + throw new TypeError('"offset" must be a special string or number if it is set'); + } +} diff --git a/lib/topic.js b/lib/topic.js new file mode 100644 index 0000000..4f6a679 --- /dev/null +++ b/lib/topic.js @@ -0,0 +1,42 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var librdkafka = require('../librdkafka'); + +module.exports = Topic; + +var topicKey = 'RdKafka::Topic::'; +var topicKeyLength = topicKey.length; + +// Take all of the topic special codes from librdkafka and add them +// to the object +// You can find this list in the C++ code at +// https://github.com/edenhill/librdkafka/blob/master/src-cpp/rdkafkacpp.h#L1250 +for (var key in librdkafka.topic) { + // Skip it if it doesn't start with ErrorCode + if (key.indexOf('RdKafka::Topic::') !== 0) { + continue; + } + + // Replace/add it if there are any discrepancies + var newKey = key.substring(topicKeyLength); + Topic[newKey] = librdkafka.topic[key]; +} + +/** + * Create a topic. Just returns the string you gave it right now. + * + * Looks like a class, but all it does is return the topic name. + * This is so that one day if there are interface changes that allow + * different use of topic parameters, we can just add to this constructor and + * have it return something richer + */ +function Topic(topicName) { + return topicName; +} diff --git a/lib/util.js b/lib/util.js new file mode 100644 index 0000000..c3250d0 --- /dev/null +++ b/lib/util.js @@ -0,0 +1,29 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var util = module.exports = {}; + +util.shallowCopy = function (obj) { + + if (!util.isObject(obj)) { return obj; } + + var copy = {}; + + for (var k in obj) { + if (obj.hasOwnProperty(k)) { + copy[k] = obj[k]; + } + } + + return copy; +}; + +util.isObject = function (obj) { + return obj && typeof obj === 'object'; +}; diff --git a/librdkafka.js b/librdkafka.js new file mode 100644 index 0000000..4198da0 --- /dev/null +++ b/librdkafka.js @@ -0,0 +1,12 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var kafka = require('bindings')('node-librdkafka'); + +module.exports = kafka; diff --git a/make_docs.sh b/make_docs.sh new file mode 100755 index 0000000..8008512 --- /dev/null +++ b/make_docs.sh @@ -0,0 +1,57 @@ +#!/bin/bash + +if [[ `git status --porcelain` ]]; then + # changes + >&2 echo "You have unstaged changes. Please commit before you run this." + exit 1 +fi + +# REPO=git@github.com:Blizzard/node-rdkafka.git +REPO=https://github.com/Blizzard/node-rdkafka.git + +git remote add deploy $REPO + +# Get the most recent stuff if we don't have it +git fetch deploy gh-pages || exit $? + +make docs || exit $? + +# Get package version and save to variable + +PACKAGE=$(node -pe 'require("./package.json").name.split("/")[1]') +VERSION=$(node -pe 'require("./package.json").version') + +# Make a temporary folder + +TEMPDIR=$(mktemp -d) + +VERSIONDIR="$TEMPDIR/$VERSION" +cp -r docs $VERSIONDIR + +# Now, checkout the gh-pages, but first get current checked out branch +# + +CURRENT_BRANCH=$(git rev-parse --symbolic-full-name --abbrev-ref HEAD) + +COMMIT_MESSAGE=$(git log --pretty='format:%B' -1) +COMMIT_AUTHOR=$(git log --pretty='format:%aN <%aE>' -1) + +if [[ `git checkout --quiet -b gh-pages deploy/gh-pages` ]]; then + >&2 echo "Could not checkout gh-pages" + exit 1 +fi + +rm -rf current +rm -rf $VERSION + +cp -r $VERSIONDIR $VERSION +cp -r $VERSIONDIR current + +git add --all +git commit --author="$COMMIT_AUTHOR" -m "Updated docs for '$COMMIT_MESSAGE'" + +rm -rf $TEMPDIR + +git push $REPO gh-pages || exit $? + +git checkout $CURRENT_BRANCH diff --git a/package.json b/package.json new file mode 100644 index 0000000..c265ceb --- /dev/null +++ b/package.json @@ -0,0 +1,34 @@ +{ + "name": "node-kafka", + "version": "v0.12.0", + "description": "Node.js bindings for librdkafka", + "librdkafka": "1.6.1", + "main": "lib/index.js", + "author": "sentientgeeks", + "scripts": { + "configure": "node-gyp configure", + "build": "node-gyp build", + "test": "make test", + "install": "node-gyp rebuild", + "prepack": "node ./ci/prepublish.js" + }, + "keywords": [ + "kafka", + "librdkafka" + ], + "devDependencies": { + "bluebird": "^3.5.3", + "jsdoc": "^3.4.0", + "jshint": "^2.10.1", + "mocha": "^5.2.0", + "node-gyp": "^5.1.0", + "toolkit-jsdoc": "^1.0.0" + }, + "dependencies": { + "bindings": "^1.3.1", + "nan": "^2.14.0" + }, + "engines": { + "node": ">=6.0.0" + } +} diff --git a/run_docker.sh b/run_docker.sh new file mode 100755 index 0000000..a6aadbd --- /dev/null +++ b/run_docker.sh @@ -0,0 +1,42 @@ +#!/bin/bash + +COMPOSE_VERSION=$(docker-compose --version) +DOCKER_VERSION=$(docker --version) + +# Start the docker compose file +echo "Running docker compose up. Docker version $DOCKER_VERSION. Compose version $COMPOSE_VERSION. " + +docker-compose up -d + +if [ "$?" == "1" ]; then + echo "Failed to start docker images." + exit 1 +fi + +# List of topics to create in container +topics=( + "test" + "test2" + "test3" + "test4" + "test5" + "test6" +) + +# Run docker-compose exec to make them +for topic in "${topics[@]}" +do + echo "Making topic $topic" + until docker-compose exec kafka \ + kafka-topics --create --topic $topic --partitions 1 --replication-factor 1 --if-not-exists --zookeeper zookeeper:2181 + do + topic_result="$?" + if [ "$topic_result" == "1" ]; then + echo "Bad status code: $topic_result. Trying again." + else + # If it is some unknown status code, die. + exit 1 + fi + done + +done diff --git a/src/admin.cc b/src/admin.cc new file mode 100644 index 0000000..1453ad3 --- /dev/null +++ b/src/admin.cc @@ -0,0 +1,601 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include +#include + +#include "src/workers.h" +#include "src/admin.h" + +using Nan::FunctionCallbackInfo; + +namespace NodeKafka { + +/** + * @brief AdminClient v8 wrapped object. + * + * Specializes the connection to wrap a consumer object through compositional + * inheritence. Establishes its prototype in node through `Init` + * + * @sa RdKafka::Handle + * @sa NodeKafka::Client + */ + +AdminClient::AdminClient(Conf* gconfig): + Connection(gconfig, NULL) { + rkqu = NULL; +} + +AdminClient::~AdminClient() { + Disconnect(); +} + +Baton AdminClient::Connect() { + std::string errstr; + + { + scoped_shared_write_lock lock(m_connection_lock); + m_client = RdKafka::Producer::create(m_gconfig, errstr); + } + + if (!m_client || !errstr.empty()) { + return Baton(RdKafka::ERR__STATE, errstr); + } + + if (rkqu == NULL) { + rkqu = rd_kafka_queue_new(m_client->c_ptr()); + } + + return Baton(RdKafka::ERR_NO_ERROR); +} + +Baton AdminClient::Disconnect() { + if (IsConnected()) { + scoped_shared_write_lock lock(m_connection_lock); + + if (rkqu != NULL) { + rd_kafka_queue_destroy(rkqu); + rkqu = NULL; + } + + delete m_client; + m_client = NULL; + } + + return Baton(RdKafka::ERR_NO_ERROR); +} + +Nan::Persistent AdminClient::constructor; + +void AdminClient::Init(v8::Local exports) { + Nan::HandleScope scope; + + v8::Local tpl = Nan::New(New); + tpl->SetClassName(Nan::New("AdminClient").ToLocalChecked()); + tpl->InstanceTemplate()->SetInternalFieldCount(1); + + // Admin client operations + Nan::SetPrototypeMethod(tpl, "createTopic", NodeCreateTopic); + Nan::SetPrototypeMethod(tpl, "deleteTopic", NodeDeleteTopic); + Nan::SetPrototypeMethod(tpl, "createPartitions", NodeCreatePartitions); + + Nan::SetPrototypeMethod(tpl, "connect", NodeConnect); + Nan::SetPrototypeMethod(tpl, "disconnect", NodeDisconnect); + + constructor.Reset( + (tpl->GetFunction(Nan::GetCurrentContext())).ToLocalChecked()); + Nan::Set(exports, Nan::New("AdminClient").ToLocalChecked(), + tpl->GetFunction(Nan::GetCurrentContext()).ToLocalChecked()); +} + +void AdminClient::New(const Nan::FunctionCallbackInfo& info) { + if (!info.IsConstructCall()) { + return Nan::ThrowError("non-constructor invocation not supported"); + } + + if (info.Length() < 1) { + return Nan::ThrowError("You must supply a global configuration"); + } + + if (!info[0]->IsObject()) { + return Nan::ThrowError("Global configuration data must be specified"); + } + + std::string errstr; + + Conf* gconfig = + Conf::create(RdKafka::Conf::CONF_GLOBAL, + (info[0]->ToObject(Nan::GetCurrentContext())).ToLocalChecked(), errstr); + + if (!gconfig) { + return Nan::ThrowError(errstr.c_str()); + } + + AdminClient* client = new AdminClient(gconfig); + + // Wrap it + client->Wrap(info.This()); + + // Then there is some weird initialization that happens + // basically it sets the configuration data + // we don't need to do that because we lazy load it + + info.GetReturnValue().Set(info.This()); +} + +v8::Local AdminClient::NewInstance(v8::Local arg) { + Nan::EscapableHandleScope scope; + + const unsigned argc = 1; + + v8::Local argv[argc] = { arg }; + v8::Local cons = Nan::New(constructor); + v8::Local instance = + Nan::NewInstance(cons, argc, argv).ToLocalChecked(); + + return scope.Escape(instance); +} + +/** + * Poll for a particular event on a queue. + * + * This will keep polling until it gets an event of that type, + * given the number of tries and a timeout + */ +rd_kafka_event_t* PollForEvent( + rd_kafka_queue_t * topic_rkqu, + rd_kafka_event_type_t event_type, + int timeout_ms) { + // Initiate exponential timeout + int attempts = 1; + int exp_timeout_ms = timeout_ms; + if (timeout_ms > 2000) { + // measure optimal number of attempts + attempts = log10(timeout_ms / 1000) / log10(2) + 1; + // measure initial exponential timeout based on attempts + exp_timeout_ms = timeout_ms / (pow(2, attempts) - 1); + } + + rd_kafka_event_t * event_response = nullptr; + + // Poll the event queue until we get it + do { + // free previously fetched event + rd_kafka_event_destroy(event_response); + // poll and update attempts and exponential timeout + event_response = rd_kafka_queue_poll(topic_rkqu, exp_timeout_ms); + attempts = attempts - 1; + exp_timeout_ms = 2 * exp_timeout_ms; + } while ( + rd_kafka_event_type(event_response) != event_type && + attempts > 0); + + // If this isn't the type of response we want, or if we do not have a response + // type, bail out with a null + if (event_response == NULL || + rd_kafka_event_type(event_response) != event_type) { + rd_kafka_event_destroy(event_response); + return NULL; + } + + return event_response; +} + +Baton AdminClient::CreateTopic(rd_kafka_NewTopic_t* topic, int timeout_ms) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + { + scoped_shared_write_lock lock(m_connection_lock); + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + // Make admin options to establish that we are creating topics + rd_kafka_AdminOptions_t *options = rd_kafka_AdminOptions_new( + m_client->c_ptr(), RD_KAFKA_ADMIN_OP_CREATETOPICS); + + // Create queue just for this operation + rd_kafka_queue_t * topic_rkqu = rd_kafka_queue_new(m_client->c_ptr()); + + rd_kafka_CreateTopics(m_client->c_ptr(), &topic, 1, options, topic_rkqu); + + // Poll for an event by type in that queue + rd_kafka_event_t * event_response = PollForEvent( + topic_rkqu, + RD_KAFKA_EVENT_CREATETOPICS_RESULT, + timeout_ms); + + // Destroy the queue since we are done with it. + rd_kafka_queue_destroy(topic_rkqu); + + // Destroy the options we just made because we polled already + rd_kafka_AdminOptions_destroy(options); + + // If we got no response from that operation, this is a failure + // likely due to time out + if (event_response == NULL) { + return Baton(RdKafka::ERR__TIMED_OUT); + } + + // Now we can get the error code from the event + if (rd_kafka_event_error(event_response)) { + // If we had a special error code, get out of here with it + const rd_kafka_resp_err_t errcode = rd_kafka_event_error(event_response); + rd_kafka_event_destroy(event_response); + return Baton(static_cast(errcode)); + } + + // get the created results + const rd_kafka_CreateTopics_result_t * create_topic_results = + rd_kafka_event_CreateTopics_result(event_response); + + size_t created_topic_count; + const rd_kafka_topic_result_t **restopics = rd_kafka_CreateTopics_result_topics( // NOLINT + create_topic_results, + &created_topic_count); + + for (int i = 0 ; i < static_cast(created_topic_count) ; i++) { + const rd_kafka_topic_result_t *terr = restopics[i]; + const rd_kafka_resp_err_t errcode = rd_kafka_topic_result_error(terr); + const char *errmsg = rd_kafka_topic_result_error_string(terr); + + if (errcode != RD_KAFKA_RESP_ERR_NO_ERROR) { + if (errmsg) { + const std::string errormsg = std::string(errmsg); + rd_kafka_event_destroy(event_response); + return Baton(static_cast(errcode), errormsg); // NOLINT + } else { + rd_kafka_event_destroy(event_response); + return Baton(static_cast(errcode)); + } + } + } + + rd_kafka_event_destroy(event_response); + return Baton(RdKafka::ERR_NO_ERROR); + } +} + +Baton AdminClient::DeleteTopic(rd_kafka_DeleteTopic_t* topic, int timeout_ms) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + { + scoped_shared_write_lock lock(m_connection_lock); + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + // Make admin options to establish that we are deleting topics + rd_kafka_AdminOptions_t *options = rd_kafka_AdminOptions_new( + m_client->c_ptr(), RD_KAFKA_ADMIN_OP_DELETETOPICS); + + // Create queue just for this operation. + // May be worth making a "scoped queue" class or something like a lock + // for RAII + rd_kafka_queue_t * topic_rkqu = rd_kafka_queue_new(m_client->c_ptr()); + + rd_kafka_DeleteTopics(m_client->c_ptr(), &topic, 1, options, topic_rkqu); + + // Poll for an event by type in that queue + rd_kafka_event_t * event_response = PollForEvent( + topic_rkqu, + RD_KAFKA_EVENT_DELETETOPICS_RESULT, + timeout_ms); + + // Destroy the queue since we are done with it. + rd_kafka_queue_destroy(topic_rkqu); + + // Destroy the options we just made because we polled already + rd_kafka_AdminOptions_destroy(options); + + // If we got no response from that operation, this is a failure + // likely due to time out + if (event_response == NULL) { + return Baton(RdKafka::ERR__TIMED_OUT); + } + + // Now we can get the error code from the event + if (rd_kafka_event_error(event_response)) { + // If we had a special error code, get out of here with it + const rd_kafka_resp_err_t errcode = rd_kafka_event_error(event_response); + rd_kafka_event_destroy(event_response); + return Baton(static_cast(errcode)); + } + + // get the created results + const rd_kafka_DeleteTopics_result_t * delete_topic_results = + rd_kafka_event_DeleteTopics_result(event_response); + + size_t deleted_topic_count; + const rd_kafka_topic_result_t **restopics = rd_kafka_DeleteTopics_result_topics( // NOLINT + delete_topic_results, + &deleted_topic_count); + + for (int i = 0 ; i < static_cast(deleted_topic_count) ; i++) { + const rd_kafka_topic_result_t *terr = restopics[i]; + const rd_kafka_resp_err_t errcode = rd_kafka_topic_result_error(terr); + + if (errcode != RD_KAFKA_RESP_ERR_NO_ERROR) { + rd_kafka_event_destroy(event_response); + return Baton(static_cast(errcode)); + } + } + + rd_kafka_event_destroy(event_response); + return Baton(RdKafka::ERR_NO_ERROR); + } +} + +Baton AdminClient::CreatePartitions( + rd_kafka_NewPartitions_t* partitions, + int timeout_ms) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + { + scoped_shared_write_lock lock(m_connection_lock); + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + // Make admin options to establish that we are deleting topics + rd_kafka_AdminOptions_t *options = rd_kafka_AdminOptions_new( + m_client->c_ptr(), RD_KAFKA_ADMIN_OP_CREATEPARTITIONS); + + // Create queue just for this operation. + // May be worth making a "scoped queue" class or something like a lock + // for RAII + rd_kafka_queue_t * topic_rkqu = rd_kafka_queue_new(m_client->c_ptr()); + + rd_kafka_CreatePartitions(m_client->c_ptr(), + &partitions, 1, options, topic_rkqu); + + // Poll for an event by type in that queue + rd_kafka_event_t * event_response = PollForEvent( + topic_rkqu, + RD_KAFKA_EVENT_CREATEPARTITIONS_RESULT, + timeout_ms); + + // Destroy the queue since we are done with it. + rd_kafka_queue_destroy(topic_rkqu); + + // Destroy the options we just made because we polled already + rd_kafka_AdminOptions_destroy(options); + + // If we got no response from that operation, this is a failure + // likely due to time out + if (event_response == NULL) { + return Baton(RdKafka::ERR__TIMED_OUT); + } + + // Now we can get the error code from the event + if (rd_kafka_event_error(event_response)) { + // If we had a special error code, get out of here with it + const rd_kafka_resp_err_t errcode = rd_kafka_event_error(event_response); + rd_kafka_event_destroy(event_response); + return Baton(static_cast(errcode)); + } + + // get the created results + const rd_kafka_CreatePartitions_result_t * create_partitions_results = + rd_kafka_event_CreatePartitions_result(event_response); + + size_t created_partitions_topic_count; + const rd_kafka_topic_result_t **restopics = rd_kafka_CreatePartitions_result_topics( // NOLINT + create_partitions_results, + &created_partitions_topic_count); + + for (int i = 0 ; i < static_cast(created_partitions_topic_count) ; i++) { // NOLINT + const rd_kafka_topic_result_t *terr = restopics[i]; + const rd_kafka_resp_err_t errcode = rd_kafka_topic_result_error(terr); + const char *errmsg = rd_kafka_topic_result_error_string(terr); + + if (errcode != RD_KAFKA_RESP_ERR_NO_ERROR) { + if (errmsg) { + const std::string errormsg = std::string(errmsg); + rd_kafka_event_destroy(event_response); + return Baton(static_cast(errcode), errormsg); // NOLINT + } else { + rd_kafka_event_destroy(event_response); + return Baton(static_cast(errcode)); + } + } + } + + rd_kafka_event_destroy(event_response); + return Baton(RdKafka::ERR_NO_ERROR); + } +} + +void AdminClient::ActivateDispatchers() { + // Listen to global config + m_gconfig->listen(); + + // Listen to non global config + // tconfig->listen(); + + // This should be refactored to config based management + m_event_cb.dispatcher.Activate(); +} +void AdminClient::DeactivateDispatchers() { + // Stop listening to the config dispatchers + m_gconfig->stop(); + + // Also this one + m_event_cb.dispatcher.Deactivate(); +} + +/** + * @section + * C++ Exported prototype functions + */ + +NAN_METHOD(AdminClient::NodeConnect) { + Nan::HandleScope scope; + + AdminClient* client = ObjectWrap::Unwrap(info.This()); + + Baton b = client->Connect(); + // Let the JS library throw if we need to so the error can be more rich + int error_code = static_cast(b.err()); + return info.GetReturnValue().Set(Nan::New(error_code)); +} + +NAN_METHOD(AdminClient::NodeDisconnect) { + Nan::HandleScope scope; + + AdminClient* client = ObjectWrap::Unwrap(info.This()); + + Baton b = client->Disconnect(); + // Let the JS library throw if we need to so the error can be more rich + int error_code = static_cast(b.err()); + return info.GetReturnValue().Set(Nan::New(error_code)); +} + +/** + * Create topic + */ +NAN_METHOD(AdminClient::NodeCreateTopic) { + Nan::HandleScope scope; + + if (info.Length() < 3 || !info[2]->IsFunction()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback"); + } + + if (!info[1]->IsNumber()) { + return Nan::ThrowError("Must provide 'timeout'"); + } + + // Create the final callback object + v8::Local cb = info[2].As(); + Nan::Callback *callback = new Nan::Callback(cb); + AdminClient* client = ObjectWrap::Unwrap(info.This()); + + // Get the timeout + int timeout = Nan::To(info[1]).FromJust(); + + std::string errstr; + // Get that topic we want to create + rd_kafka_NewTopic_t* topic = Conversion::Admin::FromV8TopicObject( + info[0].As(), errstr); + + if (topic == NULL) { + Nan::ThrowError(errstr.c_str()); + return; + } + + // Queue up dat work + Nan::AsyncQueueWorker( + new Workers::AdminClientCreateTopic(callback, client, topic, timeout)); + + return info.GetReturnValue().Set(Nan::Null()); +} + +/** + * Delete topic + */ +NAN_METHOD(AdminClient::NodeDeleteTopic) { + Nan::HandleScope scope; + + if (info.Length() < 3 || !info[2]->IsFunction()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback"); + } + + if (!info[1]->IsNumber() || !info[0]->IsString()) { + return Nan::ThrowError("Must provide 'timeout', and 'topicName'"); + } + + // Create the final callback object + v8::Local cb = info[2].As(); + Nan::Callback *callback = new Nan::Callback(cb); + AdminClient* client = ObjectWrap::Unwrap(info.This()); + + // Get the topic name from the string + std::string topic_name = Util::FromV8String( + Nan::To(info[0]).ToLocalChecked()); + + // Get the timeout + int timeout = Nan::To(info[1]).FromJust(); + + // Get that topic we want to create + rd_kafka_DeleteTopic_t* topic = rd_kafka_DeleteTopic_new( + topic_name.c_str()); + + // Queue up dat work + Nan::AsyncQueueWorker( + new Workers::AdminClientDeleteTopic(callback, client, topic, timeout)); + + return info.GetReturnValue().Set(Nan::Null()); +} + +/** + * Delete topic + */ +NAN_METHOD(AdminClient::NodeCreatePartitions) { + Nan::HandleScope scope; + + if (info.Length() < 4) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback"); + } + + if (!info[3]->IsFunction()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback 2"); + } + + if (!info[2]->IsNumber() || !info[1]->IsNumber() || !info[0]->IsString()) { + return Nan::ThrowError( + "Must provide 'totalPartitions', 'timeout', and 'topicName'"); + } + + // Create the final callback object + v8::Local cb = info[3].As(); + Nan::Callback *callback = new Nan::Callback(cb); + AdminClient* client = ObjectWrap::Unwrap(info.This()); + + // Get the timeout + int timeout = Nan::To(info[2]).FromJust(); + + // Get the total number of desired partitions + int partition_total_count = Nan::To(info[1]).FromJust(); + + // Get the topic name from the string + std::string topic_name = Util::FromV8String( + Nan::To(info[0]).ToLocalChecked()); + + // Create an error buffer we can throw + char* errbuf = reinterpret_cast(malloc(100)); + + // Create the new partitions request + rd_kafka_NewPartitions_t* new_partitions = rd_kafka_NewPartitions_new( + topic_name.c_str(), partition_total_count, errbuf, 100); + + // If we got a failure on the create new partitions request, + // fail here + if (new_partitions == NULL) { + return Nan::ThrowError(errbuf); + } + + // Queue up dat work + Nan::AsyncQueueWorker(new Workers::AdminClientCreatePartitions( + callback, client, new_partitions, timeout)); + + return info.GetReturnValue().Set(Nan::Null()); +} + +} // namespace NodeKafka diff --git a/src/admin.h b/src/admin.h new file mode 100644 index 0000000..db91238 --- /dev/null +++ b/src/admin.h @@ -0,0 +1,77 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_ADMIN_H_ +#define SRC_ADMIN_H_ + +#include +#include +#include +#include +#include + +#include "rdkafkacpp.h" +#include "rdkafka.h" // NOLINT + +#include "src/common.h" +#include "src/connection.h" +#include "src/callbacks.h" + +namespace NodeKafka { + +/** + * @brief KafkaConsumer v8 wrapped object. + * + * Specializes the connection to wrap a consumer object through compositional + * inheritence. Establishes its prototype in node through `Init` + * + * @sa RdKafka::Handle + * @sa NodeKafka::Client + */ + +class AdminClient : public Connection { + public: + static void Init(v8::Local); + static v8::Local NewInstance(v8::Local); + + void ActivateDispatchers(); + void DeactivateDispatchers(); + + Baton Connect(); + Baton Disconnect(); + + Baton CreateTopic(rd_kafka_NewTopic_t* topic, int timeout_ms); + Baton DeleteTopic(rd_kafka_DeleteTopic_t* topic, int timeout_ms); + Baton CreatePartitions(rd_kafka_NewPartitions_t* topic, int timeout_ms); + // Baton AlterConfig(rd_kafka_NewTopic_t* topic, int timeout_ms); + // Baton DescribeConfig(rd_kafka_NewTopic_t* topic, int timeout_ms); + + protected: + static Nan::Persistent constructor; + static void New(const Nan::FunctionCallbackInfo& info); + + explicit AdminClient(Conf* globalConfig); + ~AdminClient(); + + rd_kafka_queue_t* rkqu; + + private: + // Node methods + // static NAN_METHOD(NodeValidateTopic); + static NAN_METHOD(NodeCreateTopic); + static NAN_METHOD(NodeDeleteTopic); + static NAN_METHOD(NodeCreatePartitions); + + static NAN_METHOD(NodeConnect); + static NAN_METHOD(NodeDisconnect); +}; + +} // namespace NodeKafka + +#endif // SRC_ADMIN_H_ diff --git a/src/binding.cc b/src/binding.cc new file mode 100644 index 0000000..badbc68 --- /dev/null +++ b/src/binding.cc @@ -0,0 +1,94 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include "src/binding.h" + +using NodeKafka::Producer; +using NodeKafka::KafkaConsumer; +using NodeKafka::AdminClient; +using NodeKafka::Topic; + +using node::AtExit; +using RdKafka::ErrorCode; + +static void RdKafkaCleanup(void*) { // NOLINT + /* + * Wait for RdKafka to decommission. + * This is not strictly needed but + * allows RdKafka to clean up all its resources before the application + * exits so that memory profilers such as valgrind wont complain about + * memory leaks. + */ + + RdKafka::wait_destroyed(5000); +} + +NAN_METHOD(NodeRdKafkaErr2Str) { + int points = Nan::To(info[0]).FromJust(); + // Cast to error code + RdKafka::ErrorCode err = static_cast(points); + + std::string errstr = RdKafka::err2str(err); + + info.GetReturnValue().Set(Nan::New(errstr).ToLocalChecked()); +} + +NAN_METHOD(NodeRdKafkaBuildInFeatures) { + RdKafka::Conf * config = RdKafka::Conf::create(RdKafka::Conf::CONF_GLOBAL); + + std::string features; + + if (RdKafka::Conf::CONF_OK == config->get("builtin.features", features)) { + info.GetReturnValue().Set(Nan::New(features).ToLocalChecked()); + } else { + info.GetReturnValue().Set(Nan::Undefined()); + } + + delete config; +} + +void ConstantsInit(v8::Local exports) { + v8::Local topicConstants = Nan::New(); + + // RdKafka Error Code definitions + NODE_DEFINE_CONSTANT(topicConstants, RdKafka::Topic::PARTITION_UA); + NODE_DEFINE_CONSTANT(topicConstants, RdKafka::Topic::OFFSET_BEGINNING); + NODE_DEFINE_CONSTANT(topicConstants, RdKafka::Topic::OFFSET_END); + NODE_DEFINE_CONSTANT(topicConstants, RdKafka::Topic::OFFSET_STORED); + NODE_DEFINE_CONSTANT(topicConstants, RdKafka::Topic::OFFSET_INVALID); + + Nan::Set(exports, Nan::New("topic").ToLocalChecked(), topicConstants); + + Nan::Set(exports, Nan::New("err2str").ToLocalChecked(), + Nan::GetFunction(Nan::New(NodeRdKafkaErr2Str)).ToLocalChecked()); // NOLINT + + Nan::Set(exports, Nan::New("features").ToLocalChecked(), + Nan::GetFunction(Nan::New(NodeRdKafkaBuildInFeatures)).ToLocalChecked()); // NOLINT +} + +void Init(v8::Local exports, v8::Local m_, void* v_) { +#if NODE_MAJOR_VERSION <= 9 || (NODE_MAJOR_VERSION == 10 && NODE_MINOR_VERSION <= 15) + AtExit(RdKafkaCleanup); +#else + v8::Local context = Nan::GetCurrentContext(); + node::Environment* env = node::GetCurrentEnvironment(context); + AtExit(env, RdKafkaCleanup, NULL); +#endif + KafkaConsumer::Init(exports); + Producer::Init(exports); + AdminClient::Init(exports); + Topic::Init(exports); + ConstantsInit(exports); + + Nan::Set(exports, Nan::New("librdkafkaVersion").ToLocalChecked(), + Nan::New(RdKafka::version_str().c_str()).ToLocalChecked()); +} + +NODE_MODULE(kafka, Init) diff --git a/src/binding.h b/src/binding.h new file mode 100644 index 0000000..c032a4e --- /dev/null +++ b/src/binding.h @@ -0,0 +1,25 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_BINDING_H_ +#define SRC_BINDING_H_ + +#include +#include +#include "rdkafkacpp.h" +#include "src/common.h" +#include "src/errors.h" +#include "src/config.h" +#include "src/connection.h" +#include "src/kafka-consumer.h" +#include "src/producer.h" +#include "src/topic.h" +#include "src/admin.h" + +#endif // SRC_BINDING_H_ diff --git a/src/callbacks.cc b/src/callbacks.cc new file mode 100644 index 0000000..4dfe2f3 --- /dev/null +++ b/src/callbacks.cc @@ -0,0 +1,612 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include +#include + +#include "src/callbacks.h" +#include "src/kafka-consumer.h" + +using v8::Local; +using v8::Value; +using v8::Object; +using v8::String; +using v8::Array; +using v8::Number; + +namespace NodeKafka { +namespace Callbacks { + +v8::Local TopicPartitionListToV8Array( + std::vector parts) { + v8::Local tp_array = Nan::New(); + + for (size_t i = 0; i < parts.size(); i++) { + v8::Local tp_obj = Nan::New(); + event_topic_partition_t tp = parts[i]; + + Nan::Set(tp_obj, Nan::New("topic").ToLocalChecked(), + Nan::New(tp.topic.c_str()).ToLocalChecked()); + Nan::Set(tp_obj, Nan::New("partition").ToLocalChecked(), + Nan::New(tp.partition)); + + if (tp.offset >= 0) { + Nan::Set(tp_obj, Nan::New("offset").ToLocalChecked(), + Nan::New(tp.offset)); + } + + Nan::Set(tp_array, i, tp_obj); + } + + return tp_array; +} + +Dispatcher::Dispatcher() { + async = NULL; + uv_mutex_init(&async_lock); +} + +Dispatcher::~Dispatcher() { + if (callbacks.size() < 1) return; + + for (size_t i=0; i < callbacks.size(); i++) { + callbacks[i].Reset(); + } + + uv_mutex_destroy(&async_lock); +} + +// Only run this if we aren't already listening +void Dispatcher::Activate() { + if (!async) { + async = new uv_async_t; + uv_async_init(uv_default_loop(), async, AsyncMessage_); + + async->data = this; + } +} + +// Should be able to run this regardless of whether it is active or not +void Dispatcher::Deactivate() { + if (async) { + uv_close(reinterpret_cast(async), NULL); + async = NULL; + } +} + +bool Dispatcher::HasCallbacks() { + return callbacks.size() > 0; +} + +void Dispatcher::Execute() { + if (async) { + uv_async_send(async); + } +} + +void Dispatcher::Dispatch(const int _argc, Local _argv[]) { + // This should probably be an array of v8 values + if (!HasCallbacks()) { + return; + } + + for (size_t i=0; i < callbacks.size(); i++) { + v8::Local f = Nan::New(callbacks[i]); + Nan::Callback cb(f); + cb.Call(_argc, _argv); + } +} + +void Dispatcher::AddCallback(const v8::Local &cb) { + Nan::Persistent > value(cb); + // PersistentCopyableFunction value(func); + callbacks.push_back(value); +} + +void Dispatcher::RemoveCallback(const v8::Local &cb) { + for (size_t i=0; i < callbacks.size(); i++) { + if (callbacks[i] == cb) { + callbacks[i].Reset(); + callbacks.erase(callbacks.begin() + i); + break; + } + } +} + +event_t::event_t(const RdKafka::Event &event) { + message = ""; + fac = ""; + + type = event.type(); + + switch (type = event.type()) { + case RdKafka::Event::EVENT_ERROR: + message = RdKafka::err2str(event.err()); + break; + case RdKafka::Event::EVENT_STATS: + message = event.str(); + break; + case RdKafka::Event::EVENT_LOG: + severity = event.severity(); + fac = event.fac(); + message = event.str(); + break; + case RdKafka::Event::EVENT_THROTTLE: + message = RdKafka::err2str(event.err()); + throttle_time = event.throttle_time(); + broker_name = event.broker_name(); + broker_id = static_cast(event.broker_id()); + break; + default: + message = event.str(); + break; + } +} +event_t::~event_t() {} + +// Event callback +Event::Event(): + dispatcher() {} + +Event::~Event() {} + +void Event::event_cb(RdKafka::Event &event) { + // Second parameter is going to be an object with properties to + // represent the others. + + if (!dispatcher.HasCallbacks()) { + return; + } + + event_t e(event); + + dispatcher.Add(e); + dispatcher.Execute(); +} + +EventDispatcher::EventDispatcher() {} +EventDispatcher::~EventDispatcher() {} + +void EventDispatcher::Add(const event_t &e) { + scoped_mutex_lock lock(async_lock); + events.push_back(e); +} + +void EventDispatcher::Flush() { + Nan::HandleScope scope; + // Iterate through each of the currently stored events + // generate a callback object for each, setting to the members + // then + if (events.size() < 1) return; + + const unsigned int argc = 2; + + std::vector _events; + { + scoped_mutex_lock lock(async_lock); + events.swap(_events); + } + + for (size_t i=0; i < _events.size(); i++) { + Local argv[argc] = {}; + Local jsobj = Nan::New(); + + switch (_events[i].type) { + case RdKafka::Event::EVENT_ERROR: + argv[0] = Nan::New("error").ToLocalChecked(); + argv[1] = Nan::Error(_events[i].message.c_str()); + + // if (event->err() == RdKafka::ERR__ALL_BROKERS_DOWN). Stop running + // This may be better suited to the node side of things + break; + case RdKafka::Event::EVENT_STATS: + argv[0] = Nan::New("stats").ToLocalChecked(); + + Nan::Set(jsobj, Nan::New("message").ToLocalChecked(), + Nan::New(_events[i].message.c_str()).ToLocalChecked()); + + break; + case RdKafka::Event::EVENT_LOG: + argv[0] = Nan::New("log").ToLocalChecked(); + + Nan::Set(jsobj, Nan::New("severity").ToLocalChecked(), + Nan::New(_events[i].severity)); + Nan::Set(jsobj, Nan::New("fac").ToLocalChecked(), + Nan::New(_events[i].fac.c_str()).ToLocalChecked()); + Nan::Set(jsobj, Nan::New("message").ToLocalChecked(), + Nan::New(_events[i].message.c_str()).ToLocalChecked()); + + break; + case RdKafka::Event::EVENT_THROTTLE: + argv[0] = Nan::New("throttle").ToLocalChecked(); + + Nan::Set(jsobj, Nan::New("message").ToLocalChecked(), + Nan::New(_events[i].message.c_str()).ToLocalChecked()); + + Nan::Set(jsobj, Nan::New("throttleTime").ToLocalChecked(), + Nan::New(_events[i].throttle_time)); + Nan::Set(jsobj, Nan::New("brokerName").ToLocalChecked(), + Nan::New(_events[i].broker_name).ToLocalChecked()); + Nan::Set(jsobj, Nan::New("brokerId").ToLocalChecked(), + Nan::New(_events[i].broker_id)); + + break; + default: + argv[0] = Nan::New("event").ToLocalChecked(); + + Nan::Set(jsobj, Nan::New("message").ToLocalChecked(), + Nan::New(events[i].message.c_str()).ToLocalChecked()); + + break; + } + + if (_events[i].type != RdKafka::Event::EVENT_ERROR) { + // error would be assigned already + argv[1] = jsobj; + } + + Dispatch(argc, argv); + } +} + +DeliveryReportDispatcher::DeliveryReportDispatcher() {} +DeliveryReportDispatcher::~DeliveryReportDispatcher() {} + +size_t DeliveryReportDispatcher::Add(const DeliveryReport &e) { + scoped_mutex_lock lock(async_lock); + events.push_back(e); + return events.size(); +} + +void DeliveryReportDispatcher::Flush() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + + size_t outstanding_event_count = 0; + std::vector events_list; + { + scoped_mutex_lock lock(async_lock); + outstanding_event_count = events.size(); + const size_t flush_count = std::min(outstanding_event_count, 100UL); + events_list.reserve(flush_count); + for (size_t i = 0; i < flush_count; i++) { + events_list.emplace_back(std::move(events.front())); + events.pop_front(); + } + } + + for (size_t i = 0; i < events_list.size(); i++) { + v8::Local argv[argc] = {}; + + const DeliveryReport& event = events_list[i]; + + if (event.is_error) { + // If it is an error we need the first argument to be set + argv[0] = Nan::Error(event.error_string.c_str()); + } else { + argv[0] = Nan::Null(); + } + Local jsobj(Nan::New()); + + Nan::Set(jsobj, Nan::New("topic").ToLocalChecked(), + Nan::New(event.topic_name).ToLocalChecked()); + Nan::Set(jsobj, Nan::New("partition").ToLocalChecked(), + Nan::New(event.partition)); + Nan::Set(jsobj, Nan::New("offset").ToLocalChecked(), + Nan::New(event.offset)); + + if (event.key) { + Nan::MaybeLocal buff = Nan::NewBuffer( + static_cast(event.key), + static_cast(event.key_len)); + + Nan::Set(jsobj, Nan::New("key").ToLocalChecked(), + buff.ToLocalChecked()); + } else { + Nan::Set(jsobj, Nan::New("key").ToLocalChecked(), Nan::Null()); + } + + if (event.opaque) { + Nan::Persistent * persistent = + static_cast *>(event.opaque); + v8::Local object = Nan::New(*persistent); + Nan::Set(jsobj, Nan::New("opaque").ToLocalChecked(), object); + + // Okay... now reset and destroy the persistent handle + persistent->Reset(); + + // Get rid of the persistent since we are making it local + delete persistent; + } + + if (event.timestamp > -1) { + Nan::Set(jsobj, Nan::New("timestamp").ToLocalChecked(), + Nan::New(event.timestamp)); + } + + if (event.m_include_payload) { + if (event.payload) { + Nan::MaybeLocal buff = Nan::NewBuffer( + static_cast(event.payload), + static_cast(event.len)); + + Nan::Set(jsobj, Nan::New("value").ToLocalChecked(), + buff.ToLocalChecked()); + } else { + Nan::Set(jsobj, Nan::New("value").ToLocalChecked(), + Nan::Null()); + } + } + + Nan::Set(jsobj, Nan::New("size").ToLocalChecked(), + Nan::New(event.len)); + + argv[1] = jsobj; + + Dispatch(argc, argv); + } + if (outstanding_event_count > events_list.size()) { + Execute(); + } +} + +// This only exists to circumvent the problem with not being able to execute JS +// on any thread other than the main thread. + +// I still think there may be better alternatives, because there is a lot of +// duplication here +DeliveryReport::DeliveryReport(RdKafka::Message &message, bool include_payload) : // NOLINT + m_include_payload(include_payload) { + if (message.err() == RdKafka::ERR_NO_ERROR) { + is_error = false; + } else { + is_error = true; + error_code = message.err(); + error_string = message.errstr(); + } + + topic_name = message.topic_name(); + partition = message.partition(); + offset = message.offset(); + + if (message.timestamp().type != + RdKafka::MessageTimestamp::MSG_TIMESTAMP_NOT_AVAILABLE) { + timestamp = message.timestamp().timestamp; + } else { + timestamp = -1; + } + + + // Key length. + key_len = message.key_len(); + + // It is okay if this is null + if (message.key_pointer()) { + key = malloc(message.key_len()); + memcpy(key, message.key_pointer(), message.key_len()); + } else { + key = NULL; + } + + if (message.msg_opaque()) { + opaque = message.msg_opaque(); + } else { + opaque = NULL; + } + + len = message.len(); + + if (m_include_payload && message.payload()) { + // this pointer will be owned and freed by the Nan::NewBuffer + // created in DeliveryReportDispatcher::Flush() + payload = malloc(len); + memcpy(payload, message.payload(), len); + } else { + payload = NULL; + } +} + +DeliveryReport::~DeliveryReport() {} + +// Delivery Report + +Delivery::Delivery(): + dispatcher() { + m_dr_msg_cb = false; + } +Delivery::~Delivery() {} + + +void Delivery::SendMessageBuffer(bool send_dr_msg) { + m_dr_msg_cb = true; +} + +void Delivery::dr_cb(RdKafka::Message &message) { + if (!dispatcher.HasCallbacks()) { + return; + } + + DeliveryReport msg(message, m_dr_msg_cb); + if (dispatcher.Add(msg) == 1) { + dispatcher.Execute(); + } +} + +// Rebalance CB + +RebalanceDispatcher::RebalanceDispatcher() {} +RebalanceDispatcher::~RebalanceDispatcher() {} + +void RebalanceDispatcher::Add(const rebalance_event_t &e) { + scoped_mutex_lock lock(async_lock); + m_events.push_back(e); +} + +void RebalanceDispatcher::Flush() { + Nan::HandleScope scope; + // Iterate through each of the currently stored events + // generate a callback object for each, setting to the members + // then + + if (m_events.size() < 1) return; + + const unsigned int argc = 2; + + std::vector events; + { + scoped_mutex_lock lock(async_lock); + m_events.swap(events); + } + + for (size_t i=0; i < events.size(); i++) { + v8::Local argv[argc] = {}; + + if (events[i].err == RdKafka::ERR_NO_ERROR) { + argv[0] = Nan::Undefined(); + } else { + // ERR__ASSIGN_PARTITIONS? Special case? Nah + argv[0] = Nan::New(events[i].err); + } + + std::vector parts = events[i].partitions; + + // Now convert the TopicPartition list to a JS array + argv[1] = TopicPartitionListToV8Array(events[i].partitions); + + Dispatch(argc, argv); + } +} + +void Rebalance::rebalance_cb(RdKafka::KafkaConsumer *consumer, + RdKafka::ErrorCode err, std::vector &partitions) { + dispatcher.Add(rebalance_event_t(err, partitions)); + dispatcher.Execute(); +} + +// Offset Commit CB + +OffsetCommitDispatcher::OffsetCommitDispatcher() {} +OffsetCommitDispatcher::~OffsetCommitDispatcher() {} + +void OffsetCommitDispatcher::Add(const offset_commit_event_t &e) { + scoped_mutex_lock lock(async_lock); + m_events.push_back(e); +} + +void OffsetCommitDispatcher::Flush() { + Nan::HandleScope scope; + // Iterate through each of the currently stored events + // generate a callback object for each, setting to the members + // then + + if (m_events.size() < 1) return; + + const unsigned int argc = 2; + + std::vector events; + { + scoped_mutex_lock lock(async_lock); + m_events.swap(events); + } + + for (size_t i = 0; i < events.size(); i++) { + v8::Local argv[argc] = {}; + + if (events[i].err == RdKafka::ERR_NO_ERROR) { + argv[0] = Nan::Undefined(); + } else { + argv[0] = Nan::New(events[i].err); + } + + // Now convert the TopicPartition list to a JS array + argv[1] = TopicPartitionListToV8Array(events[i].partitions); + + Dispatch(argc, argv); + } +} + +void OffsetCommit::offset_commit_cb(RdKafka::ErrorCode err, + std::vector &offsets) { + dispatcher.Add(offset_commit_event_t(err, offsets)); + dispatcher.Execute(); +} + +// Partitioner callback + +Partitioner::Partitioner() {} +Partitioner::~Partitioner() {} + +int32_t Partitioner::partitioner_cb(const RdKafka::Topic *topic, + const std::string *key, + int32_t partition_cnt, + void *msg_opaque) { + // Send this and get the callback and parse the int + if (callback.IsEmpty()) { + // default behavior + return random(topic, partition_cnt); + } + + Local argv[3] = {}; + + argv[0] = Nan::New(topic->name().c_str()).ToLocalChecked(); + if (key->empty()) { + argv[1] = Nan::Null(); + } else { + argv[1] = Nan::New(key->c_str()).ToLocalChecked(); + } + + argv[2] = Nan::New(partition_cnt); + + v8::Local return_value = callback.Call(3, argv); + + Nan::Maybe partition_return = Nan::To(return_value); + + int32_t chosen_partition; + + if (partition_return.IsNothing()) { + chosen_partition = RdKafka::Topic::PARTITION_UA; + } else { + chosen_partition = partition_return.FromJust(); + } + + if (!topic->partition_available(chosen_partition)) { + return RdKafka::Topic::PARTITION_UA; + } + + return chosen_partition; +} + +unsigned int Partitioner::djb_hash(const char *str, size_t len) { + unsigned int hash = 5381; + for (size_t i = 0 ; i < len ; i++) + hash = ((hash << 5) + hash) + str[i]; + return hash; +} + +unsigned int Partitioner::random(const RdKafka::Topic *topic, int32_t max) { + int32_t random_partition = rand() % max; // NOLINT + + if (topic->partition_available(random_partition)) { + return random_partition; + } else { + return RdKafka::Topic::PARTITION_UA; + } +} + +void Partitioner::SetCallback(v8::Local cb) { + callback(cb); +} + + +} // end namespace Callbacks + +} // End namespace NodeKafka diff --git a/src/callbacks.h b/src/callbacks.h new file mode 100644 index 0000000..1271fbe --- /dev/null +++ b/src/callbacks.h @@ -0,0 +1,265 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_CALLBACKS_H_ +#define SRC_CALLBACKS_H_ + +#include +#include + +#include +#include + +#include "rdkafkacpp.h" +#include "src/common.h" + +typedef Nan::Persistent > PersistentCopyableFunction; +typedef std::vector CopyableFunctionList; + +namespace NodeKafka { + +class KafkaConsumer; + +namespace Callbacks { + +class Dispatcher { + public: + Dispatcher(); + ~Dispatcher(); + void Dispatch(const int, v8::Local []); + void AddCallback(const v8::Local&); + void RemoveCallback(const v8::Local&); + bool HasCallbacks(); + virtual void Flush() = 0; + void Execute(); + void Activate(); + void Deactivate(); + + protected: + std::vector > > callbacks; // NOLINT + + uv_mutex_t async_lock; + + private: + NAN_INLINE static NAUV_WORK_CB(AsyncMessage_) { + Dispatcher *dispatcher = + static_cast(async->data); + dispatcher->Flush(); + } + + uv_async_t *async; +}; + +struct event_t { + RdKafka::Event::Type type; + std::string message; + + RdKafka::Event::Severity severity; + std::string fac; + + std::string broker_name; + int throttle_time; + int broker_id; + + explicit event_t(const RdKafka::Event &); + ~event_t(); +}; + +class EventDispatcher : public Dispatcher { + public: + EventDispatcher(); + ~EventDispatcher(); + void Add(const event_t &); + void Flush(); + protected: + std::vector events; +}; + +class Event : public RdKafka::EventCb { + public: + Event(); + ~Event(); + void event_cb(RdKafka::Event&); + EventDispatcher dispatcher; +}; + +/** + * Delivery report class + * + * Class exists because the callback needs to be able to give information + * to the v8 main thread that it can use to formulate its object. + */ +class DeliveryReport { + public: + DeliveryReport(RdKafka::Message &, bool); + ~DeliveryReport(); + + // Whether we include the payload. Is the second parameter to the constructor + bool m_include_payload; + + // If it is an error these will be set + bool is_error; + std::string error_string; + RdKafka::ErrorCode error_code; + + // If it is not + std::string topic_name; + int32_t partition; + int64_t offset; + int64_t timestamp; + + // Opaque token used. Local value + void* opaque; + + // Key. It is a pointer to avoid corrupted values + // https://github.com/Blizzard/node-rdkafka/issues/208 + void* key; + size_t key_len; + + size_t len; + void* payload; +}; + +class DeliveryReportDispatcher : public Dispatcher { + public: + DeliveryReportDispatcher(); + ~DeliveryReportDispatcher(); + void Flush(); + size_t Add(const DeliveryReport &); + protected: + std::deque events; +}; + +class Delivery : public RdKafka::DeliveryReportCb { + public: + Delivery(); + ~Delivery(); + void dr_cb(RdKafka::Message&); + DeliveryReportDispatcher dispatcher; + void SendMessageBuffer(bool dr_copy_payload); + protected: + bool m_dr_msg_cb; +}; + +// Rebalance dispatcher + +struct event_topic_partition_t { + std::string topic; + int partition; + int64_t offset; + + event_topic_partition_t(std::string p_topic, int p_partition, int64_t p_offset): // NOLINT + topic(p_topic), + partition(p_partition), + offset(p_offset) {} +}; + +struct rebalance_event_t { + RdKafka::ErrorCode err; + std::vector partitions; + + rebalance_event_t(RdKafka::ErrorCode p_err, + std::vector p_partitions): + err(p_err) { + // Iterate over the topic partitions because we won't have them later + for (size_t topic_partition_i = 0; + topic_partition_i < p_partitions.size(); topic_partition_i++) { + RdKafka::TopicPartition* topic_partition = + p_partitions[topic_partition_i]; + + event_topic_partition_t tp( + topic_partition->topic(), + topic_partition->partition(), + topic_partition->offset()); + + partitions.push_back(tp); + } + } +}; + +struct offset_commit_event_t { + RdKafka::ErrorCode err; + std::vector partitions; + + offset_commit_event_t(RdKafka::ErrorCode p_err, + const std::vector &p_partitions): + err(p_err) { + // Iterate over the topic partitions because we won't have them later + for (size_t topic_partition_i = 0; + topic_partition_i < p_partitions.size(); topic_partition_i++) { + RdKafka::TopicPartition* topic_partition = + p_partitions[topic_partition_i]; + + // Just reuse this thing because it's the same exact thing we need + event_topic_partition_t tp( + topic_partition->topic(), + topic_partition->partition(), + topic_partition->offset()); + + partitions.push_back(tp); + } + } +}; + +class RebalanceDispatcher : public Dispatcher { + public: + RebalanceDispatcher(); + ~RebalanceDispatcher(); + void Add(const rebalance_event_t &); + void Flush(); + protected: + std::vector m_events; +}; + +class Rebalance : public RdKafka::RebalanceCb { + public: + void rebalance_cb(RdKafka::KafkaConsumer *, RdKafka::ErrorCode, + std::vector &); + + RebalanceDispatcher dispatcher; + private: + v8::Persistent m_cb; +}; + +class OffsetCommitDispatcher : public Dispatcher { + public: + OffsetCommitDispatcher(); + ~OffsetCommitDispatcher(); + void Add(const offset_commit_event_t &); + void Flush(); + protected: + std::vector m_events; +}; + +class OffsetCommit : public RdKafka::OffsetCommitCb { + public: + void offset_commit_cb(RdKafka::ErrorCode, std::vector &); // NOLINT + + OffsetCommitDispatcher dispatcher; + private: + v8::Persistent m_cb; +}; + +class Partitioner : public RdKafka::PartitionerCb { + public: + Partitioner(); + ~Partitioner(); + int32_t partitioner_cb( const RdKafka::Topic*, const std::string*, int32_t, void*); // NOLINT + Nan::Callback callback; // NOLINT + void SetCallback(v8::Local); + private: + static unsigned int djb_hash(const char*, size_t); + static unsigned int random(const RdKafka::Topic*, int32_t); +}; + +} // namespace Callbacks + +} // namespace NodeKafka + +#endif // SRC_CALLBACKS_H_ diff --git a/src/common.cc b/src/common.cc new file mode 100644 index 0000000..c14b9d8 --- /dev/null +++ b/src/common.cc @@ -0,0 +1,582 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include + +#include "src/common.h" + +namespace NodeKafka { + +void Log(std::string str) { + std::cerr << "% " << str.c_str() << std::endl; +} + +template +T GetParameter(v8::Local object, std::string field_name, T def) { + v8::Local field = Nan::New(field_name.c_str()).ToLocalChecked(); + if (Nan::Has(object, field).FromMaybe(false)) { + Nan::Maybe maybeT = Nan::To(Nan::Get(object, field).ToLocalChecked()); + if (maybeT.IsNothing()) { + return def; + } else { + return maybeT.FromJust(); + } + } + return def; +} + +template<> +int64_t GetParameter(v8::Local object, + std::string field_name, int64_t def) { + v8::Local field = Nan::New(field_name.c_str()).ToLocalChecked(); + if (Nan::Has(object, field).FromMaybe(false)) { + v8::Local v = Nan::Get(object, field).ToLocalChecked(); + + if (!v->IsNumber()) { + return def; + } + + Nan::Maybe maybeInt = Nan::To(v); + if (maybeInt.IsNothing()) { + return def; + } else { + return maybeInt.FromJust(); + } + } + return def; +} + +template<> +bool GetParameter(v8::Local object, + std::string field_name, bool def) { + v8::Local field = Nan::New(field_name.c_str()).ToLocalChecked(); + if (Nan::Has(object, field).FromMaybe(false)) { + v8::Local v = Nan::Get(object, field).ToLocalChecked(); + + if (!v->IsBoolean()) { + return def; + } + + Nan::Maybe maybeInt = Nan::To(v); + if (maybeInt.IsNothing()) { + return def; + } else { + return maybeInt.FromJust(); + } + } + return def; +} + +template<> +int GetParameter(v8::Local object, + std::string field_name, int def) { + return static_cast(GetParameter(object, field_name, def)); +} + +template<> +std::string GetParameter(v8::Local object, + std::string field_name, + std::string def) { + v8::Local field = Nan::New(field_name.c_str()).ToLocalChecked(); + if (Nan::Has(object, field).FromMaybe(false)) { + v8::Local parameter = + Nan::Get(object, field).ToLocalChecked(); + // Nan::To(); + + if (!parameter->IsUndefined() && !parameter->IsNull()) { + v8::Local val = Nan::To(parameter) + .ToLocalChecked(); + + if (!val->IsUndefined() && !val->IsNull()) { + Nan::Utf8String parameterValue(val); + std::string parameterString(*parameterValue); + + return parameterString; + } + } + } + return def; +} + +template<> +std::vector GetParameter >( + v8::Local object, std::string field_name, + std::vector def) { + v8::Local field = Nan::New(field_name.c_str()).ToLocalChecked(); + + if (Nan::Has(object, field).FromMaybe(false)) { + v8::Local maybeArray = Nan::Get(object, field).ToLocalChecked(); + if (maybeArray->IsArray()) { + v8::Local parameter = maybeArray.As(); + return v8ArrayToStringVector(parameter); + } + } + return def; +} + +std::vector v8ArrayToStringVector(v8::Local parameter) { + std::vector newItem; + + if (parameter->Length() >= 1) { + for (unsigned int i = 0; i < parameter->Length(); i++) { + v8::Local v; + if (!Nan::Get(parameter, i).ToLocal(&v)) { + continue; + } + Nan::MaybeLocal p = Nan::To(v); + if (p.IsEmpty()) { + continue; + } + Nan::Utf8String pVal(p.ToLocalChecked()); + std::string pString(*pVal); + newItem.push_back(pString); + } + } + return newItem; +} + +namespace Conversion { +namespace Topic { + +std::vector ToStringVector(v8::Local parameter) { + std::vector newItem; + + if (parameter->Length() >= 1) { + for (unsigned int i = 0; i < parameter->Length(); i++) { + v8::Local element; + if (!Nan::Get(parameter, i).ToLocal(&element)) { + continue; + } + + if (!element->IsRegExp()) { + Nan::MaybeLocal p = Nan::To(element); + + if (p.IsEmpty()) { + continue; + } + + Nan::Utf8String pVal(p.ToLocalChecked()); + std::string pString(*pVal); + + newItem.push_back(pString); + } else { + Nan::Utf8String pVal(element.As()->GetSource()); + std::string pString(*pVal); + + Log(pString); + + newItem.push_back(pString); + } + } + } + + return newItem; +} + +v8::Local ToV8Array(std::vector parameter) { + v8::Local newItem = Nan::New(); + + for (size_t i = 0; i < parameter.size(); i++) { + std::string topic = parameter[i]; + + Nan::Set(newItem, i, Nan::New(topic).ToLocalChecked()); + } + + return newItem; +} + +} // namespace Topic + +namespace TopicPartition { + +/** + * @brief RdKafka::TopicPartition vector to a v8 Array + * + * @see v8ArrayToTopicPartitionVector + */ +v8::Local ToV8Array( + std::vector & topic_partition_list) { // NOLINT + v8::Local array = Nan::New(); + for (size_t topic_partition_i = 0; + topic_partition_i < topic_partition_list.size(); topic_partition_i++) { + RdKafka::TopicPartition* topic_partition = + topic_partition_list[topic_partition_i]; + + if (topic_partition->err() != RdKafka::ErrorCode::ERR_NO_ERROR) { + Nan::Set(array, topic_partition_i, + Nan::Error(Nan::New(RdKafka::err2str(topic_partition->err())) + .ToLocalChecked())); + } else { + // We have the list now let's get the properties from it + v8::Local obj = Nan::New(); + + if (topic_partition->offset() != RdKafka::Topic::OFFSET_INVALID) { + Nan::Set(obj, Nan::New("offset").ToLocalChecked(), + Nan::New(topic_partition->offset())); + } + Nan::Set(obj, Nan::New("partition").ToLocalChecked(), + Nan::New(topic_partition->partition())); + Nan::Set(obj, Nan::New("topic").ToLocalChecked(), + Nan::New(topic_partition->topic().c_str()) + .ToLocalChecked()); + + Nan::Set(array, topic_partition_i, obj); + } + } + + return array; +} + + + +/** + * @brief v8 Array of topic partitions to RdKafka::TopicPartition vector + * + * @see v8ArrayToTopicPartitionVector + * + * @note You must delete all the pointers inside here when you are done!! + */ +std::vector FromV8Array( + const v8::Local & topic_partition_list) { + // NOTE: ARRAY OF POINTERS! DELETE THEM WHEN YOU ARE FINISHED + std::vector array; + + for (size_t topic_partition_i = 0; + topic_partition_i < topic_partition_list->Length(); topic_partition_i++) { + v8::Local topic_partition_value; + if (!Nan::Get(topic_partition_list, topic_partition_i) + .ToLocal(&topic_partition_value)) { + continue; + } + + if (topic_partition_value->IsObject()) { + array.push_back(FromV8Object( + Nan::To(topic_partition_value).ToLocalChecked())); + } + } + + return array; +} + +/** + * @brief v8::Object to RdKafka::TopicPartition + * + */ +RdKafka::TopicPartition * FromV8Object(v8::Local topic_partition) { + std::string topic = GetParameter(topic_partition, "topic", ""); + int partition = GetParameter(topic_partition, "partition", -1); + int64_t offset = GetParameter(topic_partition, "offset", 0); + + if (partition == -1) { + return NULL; + } + + if (topic.empty()) { + return NULL; + } + + return RdKafka::TopicPartition::create(topic, partition, offset); +} + +} // namespace TopicPartition + +namespace Metadata { + +/** + * @brief RdKafka::Metadata to v8::Object + * + */ +v8::Local ToV8Object(RdKafka::Metadata* metadata) { + v8::Local obj = Nan::New(); + + v8::Local broker_data = Nan::New(); + v8::Local topic_data = Nan::New(); + + const BrokerMetadataList* brokers = metadata->brokers(); // NOLINT + + unsigned int broker_i = 0; + + for (BrokerMetadataList::const_iterator it = brokers->begin(); + it != brokers->end(); ++it, broker_i++) { + // Start iterating over brokers and set the object up + + const RdKafka::BrokerMetadata* x = *it; + + v8::Local current_broker = Nan::New(); + + Nan::Set(current_broker, Nan::New("id").ToLocalChecked(), + Nan::New(x->id())); + Nan::Set(current_broker, Nan::New("host").ToLocalChecked(), + Nan::New(x->host().c_str()).ToLocalChecked()); + Nan::Set(current_broker, Nan::New("port").ToLocalChecked(), + Nan::New(x->port())); + + Nan::Set(broker_data, broker_i, current_broker); + } + + unsigned int topic_i = 0; + + const TopicMetadataList* topics = metadata->topics(); + + for (TopicMetadataList::const_iterator it = topics->begin(); + it != topics->end(); ++it, topic_i++) { + // Start iterating over topics + + const RdKafka::TopicMetadata* x = *it; + + v8::Local current_topic = Nan::New(); + + Nan::Set(current_topic, Nan::New("name").ToLocalChecked(), + Nan::New(x->topic().c_str()).ToLocalChecked()); + + v8::Local current_topic_partitions = Nan::New(); + + const PartitionMetadataList* current_partition_data = x->partitions(); + + unsigned int partition_i = 0; + PartitionMetadataList::const_iterator itt; + + for (itt = current_partition_data->begin(); + itt != current_partition_data->end(); ++itt, partition_i++) { + // partition iterate + const RdKafka::PartitionMetadata* xx = *itt; + + v8::Local current_partition = Nan::New(); + + Nan::Set(current_partition, Nan::New("id").ToLocalChecked(), + Nan::New(xx->id())); + Nan::Set(current_partition, Nan::New("leader").ToLocalChecked(), + Nan::New(xx->leader())); + + const std::vector * replicas = xx->replicas(); + const std::vector * isrs = xx->isrs(); + + std::vector::const_iterator r_it; + std::vector::const_iterator i_it; + + unsigned int r_i = 0; + unsigned int i_i = 0; + + v8::Local current_replicas = Nan::New(); + + for (r_it = replicas->begin(); r_it != replicas->end(); ++r_it, r_i++) { + Nan::Set(current_replicas, r_i, Nan::New(*r_it)); + } + + v8::Local current_isrs = Nan::New(); + + for (i_it = isrs->begin(); i_it != isrs->end(); ++i_it, i_i++) { + Nan::Set(current_isrs, i_i, Nan::New(*i_it)); + } + + Nan::Set(current_partition, Nan::New("replicas").ToLocalChecked(), + current_replicas); + Nan::Set(current_partition, Nan::New("isrs").ToLocalChecked(), + current_isrs); + + Nan::Set(current_topic_partitions, partition_i, current_partition); + } // iterate over partitions + + Nan::Set(current_topic, Nan::New("partitions").ToLocalChecked(), + current_topic_partitions); + + Nan::Set(topic_data, topic_i, current_topic); + } // End iterating over topics + + Nan::Set(obj, Nan::New("orig_broker_id").ToLocalChecked(), + Nan::New(metadata->orig_broker_id())); + + Nan::Set(obj, Nan::New("orig_broker_name").ToLocalChecked(), + Nan::New(metadata->orig_broker_name()).ToLocalChecked()); + + Nan::Set(obj, Nan::New("topics").ToLocalChecked(), topic_data); + Nan::Set(obj, Nan::New("brokers").ToLocalChecked(), broker_data); + + return obj; +} + +} // namespace Metadata + +namespace Message { + +// Overload for all use cases except delivery reports +v8::Local ToV8Object(RdKafka::Message *message) { + return ToV8Object(message, true, true); +} + +v8::Local ToV8Object(RdKafka::Message *message, + bool include_payload, + bool include_headers) { + if (message->err() == RdKafka::ERR_NO_ERROR) { + v8::Local pack = Nan::New(); + + const void* message_payload = message->payload(); + + if (!include_payload) { + Nan::Set(pack, Nan::New("value").ToLocalChecked(), + Nan::Undefined()); + } else if (message_payload) { + Nan::Set(pack, Nan::New("value").ToLocalChecked(), + Nan::Encode(message_payload, message->len(), Nan::Encoding::BUFFER)); + } else { + Nan::Set(pack, Nan::New("value").ToLocalChecked(), + Nan::Null()); + } + + RdKafka::Headers* headers; + if (((headers = message->headers()) != 0) && include_headers) { + v8::Local v8headers = Nan::New(); + int index = 0; + std::vector all = headers->get_all(); + for (std::vector::iterator it = all.begin(); + it != all.end(); it++) { + v8::Local v8header = Nan::New(); + Nan::Set(v8header, Nan::New(it->key()).ToLocalChecked(), + Nan::Encode(it->value_string(), + it->value_size(), Nan::Encoding::BUFFER)); + Nan::Set(v8headers, index, v8header); + index++; + } + Nan::Set(pack, + Nan::New("headers").ToLocalChecked(), v8headers); + } + + Nan::Set(pack, Nan::New("size").ToLocalChecked(), + Nan::New(message->len())); + + const void* key_payload = message->key_pointer(); + + if (key_payload) { + // We want this to also be a buffer to avoid corruption + // https://github.com/Blizzard/node-rdkafka/issues/208 + Nan::Set(pack, Nan::New("key").ToLocalChecked(), + Nan::Encode(key_payload, message->key_len(), Nan::Encoding::BUFFER)); + } else { + Nan::Set(pack, Nan::New("key").ToLocalChecked(), + Nan::Null()); + } + + Nan::Set(pack, Nan::New("topic").ToLocalChecked(), + Nan::New(message->topic_name()).ToLocalChecked()); + Nan::Set(pack, Nan::New("offset").ToLocalChecked(), + Nan::New(message->offset())); + Nan::Set(pack, Nan::New("partition").ToLocalChecked(), + Nan::New(message->partition())); + Nan::Set(pack, Nan::New("timestamp").ToLocalChecked(), + Nan::New(message->timestamp().timestamp)); + + return pack; + } else { + return RdKafkaError(message->err()); + } +} + +} // namespace Message + +/** + * @section Admin API models + */ + +namespace Admin { + +/** + * Create a low level rdkafka handle to represent a topic + * + * + */ +rd_kafka_NewTopic_t* FromV8TopicObject( + v8::Local object, std::string &errstr) { // NOLINT + std::string topic_name = GetParameter(object, "topic", ""); + int num_partitions = GetParameter(object, "num_partitions", 0); + int replication_factor = GetParameter(object, "replication_factor", 0); + + // Too slow to allocate this every call but admin api + // shouldn't be called that often + char* errbuf = reinterpret_cast(malloc(100)); + size_t errstr_size = 100; + + rd_kafka_NewTopic_t* new_topic = rd_kafka_NewTopic_new( + topic_name.c_str(), + num_partitions, + replication_factor, + errbuf, + errstr_size); + + if (new_topic == NULL) { + errstr = std::string(errbuf, errstr_size); + free(errbuf); + return NULL; + } + + rd_kafka_resp_err_t err; + + if (Nan::Has(object, Nan::New("config").ToLocalChecked()).FromMaybe(false)) { + // Get the config v8::Object that we can get parameters on + v8::Local config = + Nan::Get(object, Nan::New("config").ToLocalChecked()) + .ToLocalChecked().As(); + + // Get all of the keys of the object + v8::MaybeLocal config_keys = Nan::GetOwnPropertyNames(config); + + if (!config_keys.IsEmpty()) { + v8::Local field_array = config_keys.ToLocalChecked(); + for (size_t i = 0; i < field_array->Length(); i++) { + v8::Local config_key = Nan::Get(field_array, i) + .ToLocalChecked().As(); + v8::Local config_value = Nan::Get(config, config_key) + .ToLocalChecked(); + + // If the config value is a string... + if (config_value->IsString()) { + Nan::Utf8String pKeyVal(config_key); + std::string pKeyString(*pKeyVal); + + Nan::Utf8String pValueVal(config_value.As()); + std::string pValString(*pValueVal); + + err = rd_kafka_NewTopic_set_config( + new_topic, pKeyString.c_str(), pValString.c_str()); + + if (err != RD_KAFKA_RESP_ERR_NO_ERROR) { + errstr = rd_kafka_err2str(err); + rd_kafka_NewTopic_destroy(new_topic); + return NULL; + } + } else { + errstr = "Config values must all be provided as strings."; + rd_kafka_NewTopic_destroy(new_topic); + return NULL; + } + } + } + } + + // Free it again cuz we malloc'd it. + // free(errbuf); + return new_topic; +} + +rd_kafka_NewTopic_t** FromV8TopicObjectArray(v8::Local) { + return NULL; +} + +} // namespace Admin + +} // namespace Conversion + +namespace Util { + std::string FromV8String(v8::Local val) { + Nan::Utf8String keyUTF8(val); + return std::string(*keyUTF8); + } +} // Namespace Util + +} // namespace NodeKafka diff --git a/src/common.h b/src/common.h new file mode 100644 index 0000000..1509aee --- /dev/null +++ b/src/common.h @@ -0,0 +1,134 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_COMMON_H_ +#define SRC_COMMON_H_ + +#include + +#include +#include +#include + +#include "rdkafkacpp.h" +#include "rdkafka.h" // NOLINT + +#include "src/errors.h" + +typedef std::vector BrokerMetadataList; +typedef std::vector PartitionMetadataList; +typedef std::vector TopicMetadataList; + +namespace NodeKafka { + +void Log(std::string); + +template T GetParameter(v8::Local, std::string, T); +template<> std::string GetParameter( + v8::Local, std::string, std::string); +template<> std::vector GetParameter >( + v8::Local, std::string, std::vector); +// template int GetParameter(v8::Local v8ArrayToStringVector(v8::Local); + +class scoped_mutex_lock { + public: + explicit scoped_mutex_lock(uv_mutex_t& lock_) : // NOLINT + async_lock(lock_) { + uv_mutex_lock(&async_lock); + } + + ~scoped_mutex_lock() { + uv_mutex_unlock(&async_lock); + } + + private: + uv_mutex_t &async_lock; +}; + +/* +int uv_rwlock_tryrdlock(uv_rwlock_t* rwlock) + +int uv_rwlock_trywrlock(uv_rwlock_t* rwlock) + */ + +class scoped_shared_write_lock { + public: + explicit scoped_shared_write_lock(uv_rwlock_t& lock_) : // NOLINT + async_lock(lock_) { + uv_rwlock_wrlock(&async_lock); + } + + ~scoped_shared_write_lock() { + uv_rwlock_wrunlock(&async_lock); + } + + private: + uv_rwlock_t &async_lock; +}; + +class scoped_shared_read_lock { + public: + explicit scoped_shared_read_lock(uv_rwlock_t& lock_) : // NOLINT + async_lock(lock_) { + uv_rwlock_rdlock(&async_lock); + } + + ~scoped_shared_read_lock() { + uv_rwlock_rdunlock(&async_lock); + } + + private: + uv_rwlock_t &async_lock; +}; + +namespace Conversion { + +namespace Admin { + // Topics from topic object, or topic object array + rd_kafka_NewTopic_t* FromV8TopicObject( + v8::Local, std::string &errstr); // NOLINT + rd_kafka_NewTopic_t** FromV8TopicObjectArray(v8::Local); +} + +namespace Topic { + std::vector ToStringVector(v8::Local); + v8::Local ToV8Array(std::vector); +} // namespace Topic + +namespace TopicPartition { + +v8::Local ToV8Array(std::vector &); +RdKafka::TopicPartition * FromV8Object(v8::Local); +std::vector FromV8Array(const v8::Local &); // NOLINT + +} // namespace TopicPartition + +namespace Metadata { + +v8::Local ToV8Object(RdKafka::Metadata*); + +} // namespace Metadata + +namespace Message { + +v8::Local ToV8Object(RdKafka::Message*); +v8::Local ToV8Object(RdKafka::Message*, bool, bool); + +} + +} // namespace Conversion + +namespace Util { + std::string FromV8String(v8::Local); +} + +} // namespace NodeKafka + +#endif // SRC_COMMON_H_ diff --git a/src/config.cc b/src/config.cc new file mode 100644 index 0000000..aa296b8 --- /dev/null +++ b/src/config.cc @@ -0,0 +1,155 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include +#include + +#include "src/config.h" + +using Nan::MaybeLocal; +using Nan::Maybe; +using v8::Local; +using v8::String; +using v8::Object; +using std::cout; +using std::endl; + +namespace NodeKafka { + +void Conf::DumpConfig(std::list *dump) { + for (std::list::iterator it = dump->begin(); + it != dump->end(); ) { + std::cout << *it << " = "; + it++; + std::cout << *it << std::endl; + it++; + } + std::cout << std::endl; +} + +Conf * Conf::create(RdKafka::Conf::ConfType type, v8::Local object, std::string &errstr) { // NOLINT + v8::Local context = Nan::GetCurrentContext(); + Conf* rdconf = static_cast(RdKafka::Conf::create(type)); + + v8::MaybeLocal _property_names = object->GetOwnPropertyNames( + Nan::GetCurrentContext()); + v8::Local property_names = _property_names.ToLocalChecked(); + + for (unsigned int i = 0; i < property_names->Length(); ++i) { + std::string string_value; + std::string string_key; + + v8::Local key = Nan::Get(property_names, i).ToLocalChecked(); + v8::Local value = Nan::Get(object, key).ToLocalChecked(); + + if (key->IsString()) { + Nan::Utf8String utf8_key(key); + string_key = std::string(*utf8_key); + } else { + continue; + } + + if (!value->IsFunction()) { +#if NODE_MAJOR_VERSION > 6 + if (value->IsInt32()) { + string_value = std::to_string( + value->Int32Value(context).ToChecked()); + } else if (value->IsUint32()) { + string_value = std::to_string( + value->Uint32Value(context).ToChecked()); + } else if (value->IsBoolean()) { + const bool v = Nan::To(value).ToChecked(); + string_value = v ? "true" : "false"; + } else { + Nan::Utf8String utf8_value(value.As()); + string_value = std::string(*utf8_value); + } +#else + Nan::Utf8String utf8_value(value.As()); + string_value = std::string(*utf8_value); +#endif + if (rdconf->set(string_key, string_value, errstr) + != Conf::CONF_OK) { + delete rdconf; + return NULL; + } + } else { + v8::Local cb = value.As(); + rdconf->ConfigureCallback(string_key, cb, true, errstr); + if (!errstr.empty()) { + delete rdconf; + return NULL; + } + rdconf->ConfigureCallback(string_key, cb, false, errstr); + if (!errstr.empty()) { + delete rdconf; + return NULL; + } + } + } + + return rdconf; +} + +void Conf::ConfigureCallback(const std::string &string_key, const v8::Local &cb, bool add, std::string &errstr) { + if (string_key.compare("rebalance_cb") == 0) { + if (add) { + if (this->m_rebalance_cb == NULL) { + this->m_rebalance_cb = new NodeKafka::Callbacks::Rebalance(); + } + this->m_rebalance_cb->dispatcher.AddCallback(cb); + this->set(string_key, this->m_rebalance_cb, errstr); + } else { + if (this->m_rebalance_cb != NULL) { + this->m_rebalance_cb->dispatcher.RemoveCallback(cb); + } + } + } else if (string_key.compare("offset_commit_cb") == 0) { + if (add) { + if (this->m_offset_commit_cb == NULL) { + this->m_offset_commit_cb = new NodeKafka::Callbacks::OffsetCommit(); + } + this->m_offset_commit_cb->dispatcher.AddCallback(cb); + this->set(string_key, this->m_offset_commit_cb, errstr); + } else { + if (this->m_offset_commit_cb != NULL) { + this->m_offset_commit_cb->dispatcher.RemoveCallback(cb); + } + } + } +} + +void Conf::listen() { + if (m_rebalance_cb) { + m_rebalance_cb->dispatcher.Activate(); + } + + if (m_offset_commit_cb) { + m_offset_commit_cb->dispatcher.Activate(); + } +} + +void Conf::stop() { + if (m_rebalance_cb) { + m_rebalance_cb->dispatcher.Deactivate(); + } + + if (m_offset_commit_cb) { + m_offset_commit_cb->dispatcher.Deactivate(); + } +} + +Conf::~Conf() { + if (m_rebalance_cb) { + delete m_rebalance_cb; + } +} + +} // namespace NodeKafka diff --git a/src/config.h b/src/config.h new file mode 100644 index 0000000..112bed9 --- /dev/null +++ b/src/config.h @@ -0,0 +1,43 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_CONFIG_H_ +#define SRC_CONFIG_H_ + +#include +#include +#include +#include +#include + +#include "rdkafkacpp.h" +#include "src/common.h" +#include "src/callbacks.h" + +namespace NodeKafka { + +class Conf : public RdKafka::Conf { + public: + ~Conf(); + + static Conf* create(RdKafka::Conf::ConfType, v8::Local, std::string &); // NOLINT + static void DumpConfig(std::list *); + + void listen(); + void stop(); + + void ConfigureCallback(const std::string &string_key, const v8::Local &cb, bool add, std::string &errstr); + protected: + NodeKafka::Callbacks::Rebalance * m_rebalance_cb = NULL; + NodeKafka::Callbacks::OffsetCommit * m_offset_commit_cb = NULL; +}; + +} // namespace NodeKafka + +#endif // SRC_CONFIG_H_ diff --git a/src/connection.cc b/src/connection.cc new file mode 100644 index 0000000..9de3b3c --- /dev/null +++ b/src/connection.cc @@ -0,0 +1,423 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include + +#include "src/connection.h" +#include "src/workers.h" + +using RdKafka::Conf; + +namespace NodeKafka { + +/** + * @brief Connection v8 wrapped object. + * + * Wraps the RdKafka::Handle object with compositional inheritence and + * provides sensible defaults for exposing callbacks to node + * + * This object can't itself expose methods to the prototype directly, as far + * as I can tell. But it can provide the NAN_METHODS that just need to be added + * to the prototype. Since connections, etc. are managed differently based on + * whether it is a producer or consumer, they manage that. This base class + * handles some of the wrapping functionality and more importantly, the + * configuration of callbacks + * + * Any callback available to both consumers and producers, like logging or + * events will be handled in here. + * + * @sa RdKafka::Handle + * @sa NodeKafka::Client + */ + +Connection::Connection(Conf* gconfig, Conf* tconfig): + m_event_cb(), + m_gconfig(gconfig), + m_tconfig(tconfig) { + std::string errstr; + + m_client = NULL; + m_is_closing = false; + uv_rwlock_init(&m_connection_lock); + + // Try to set the event cb. Shouldn't be an error here, but if there + // is, it doesn't get reported. + // + // Perhaps node new methods should report this as an error? But there + // isn't anything the user can do about it. + m_gconfig->set("event_cb", &m_event_cb, errstr); + } + +Connection::~Connection() { + uv_rwlock_destroy(&m_connection_lock); + + if (m_tconfig) { + delete m_tconfig; + } + + if (m_gconfig) { + delete m_gconfig; + } +} + +RdKafka::TopicPartition* Connection::GetPartition(std::string &topic) { + return RdKafka::TopicPartition::create(topic, RdKafka::Topic::PARTITION_UA); +} + +RdKafka::TopicPartition* Connection::GetPartition(std::string &topic, int partition) { // NOLINT + return RdKafka::TopicPartition::create(topic, partition); +} + +bool Connection::IsConnected() { + return !m_is_closing && m_client != NULL; +} + +bool Connection::IsClosing() { + return m_client != NULL && m_is_closing; +} + +RdKafka::Handle* Connection::GetClient() { + return m_client; +} + +Baton Connection::CreateTopic(std::string topic_name) { + return CreateTopic(topic_name, NULL); +} + +Baton Connection::CreateTopic(std::string topic_name, RdKafka::Conf* conf) { + std::string errstr; + + RdKafka::Topic* topic = NULL; + + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (IsConnected()) { + topic = RdKafka::Topic::create(m_client, topic_name, conf, errstr); + } else { + return Baton(RdKafka::ErrorCode::ERR__STATE); + } + } else { + return Baton(RdKafka::ErrorCode::ERR__STATE); + } + + if (!errstr.empty()) { + return Baton(RdKafka::ErrorCode::ERR_TOPIC_EXCEPTION, errstr); + } + + // Maybe do it this way later? Then we don't need to do static_cast + // + return Baton(topic); +} + +Baton Connection::QueryWatermarkOffsets( + std::string topic_name, int32_t partition, + int64_t* low_offset, int64_t* high_offset, + int timeout_ms) { + // Check if we are connected first + + RdKafka::ErrorCode err; + + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (IsConnected()) { + // Always send true - we + err = m_client->query_watermark_offsets(topic_name, partition, + low_offset, high_offset, timeout_ms); + + } else { + err = RdKafka::ERR__STATE; + } + } else { + err = RdKafka::ERR__STATE; + } + + return Baton(err); +} + +/** + * Look up the offsets for the given partitions by timestamp. + * + * The returned offset for each partition is the earliest offset whose + * timestamp is greater than or equal to the given timestamp in the + * corresponding partition. + * + * @returns A baton specifying the error state. If there was no error, + * there still may be an error on a topic partition basis. + */ +Baton Connection::OffsetsForTimes( + std::vector &toppars, + int timeout_ms) { + // Check if we are connected first + + RdKafka::ErrorCode err; + + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (IsConnected()) { + // Always send true - we + err = m_client->offsetsForTimes(toppars, timeout_ms); + + } else { + err = RdKafka::ERR__STATE; + } + } else { + err = RdKafka::ERR__STATE; + } + + return Baton(err); +} + +Baton Connection::GetMetadata( + bool all_topics, std::string topic_name, int timeout_ms) { + RdKafka::Topic* topic = NULL; + RdKafka::ErrorCode err; + + std::string errstr; + + if (!topic_name.empty()) { + Baton b = CreateTopic(topic_name); + if (b.err() == RdKafka::ErrorCode::ERR_NO_ERROR) { + topic = b.data(); + } + } + + RdKafka::Metadata* metadata = NULL; + + if (!errstr.empty()) { + return Baton(RdKafka::ERR_TOPIC_EXCEPTION); + } + + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (IsConnected()) { + // Always send true - we + err = m_client->metadata(all_topics, topic, &metadata, timeout_ms); + } else { + err = RdKafka::ERR__STATE; + } + } else { + err = RdKafka::ERR__STATE; + } + + if (err == RdKafka::ERR_NO_ERROR) { + return Baton(metadata); + } else { + // metadata is not set here + // @see https://github.com/edenhill/librdkafka/blob/master/src-cpp/rdkafkacpp.h#L860 + return Baton(err); + } +} + +void Connection::ConfigureCallback(const std::string &string_key, const v8::Local &cb, bool add) { + if (string_key.compare("event_cb") == 0) { + if (add) { + this->m_event_cb.dispatcher.AddCallback(cb); + } else { + this->m_event_cb.dispatcher.RemoveCallback(cb); + } + } +} + +// NAN METHODS + +NAN_METHOD(Connection::NodeGetMetadata) { + Nan::HandleScope scope; + + Connection* obj = ObjectWrap::Unwrap(info.This()); + + v8::Local config; + if (info[0]->IsObject()) { + config = info[0].As(); + } else { + config = Nan::New(); + } + + if (!info[1]->IsFunction()) { + Nan::ThrowError("Second parameter must be a callback"); + return; + } + + v8::Local cb = info[1].As(); + + std::string topic = GetParameter(config, "topic", ""); + bool allTopics = GetParameter(config, "allTopics", true); + int timeout_ms = GetParameter(config, "timeout", 30000); + + Nan::Callback *callback = new Nan::Callback(cb); + + Nan::AsyncQueueWorker(new Workers::ConnectionMetadata( + callback, obj, topic, timeout_ms, allTopics)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Connection::NodeOffsetsForTimes) { + Nan::HandleScope scope; + + if (info.Length() < 3 || !info[0]->IsArray()) { + // Just throw an exception + return Nan::ThrowError("Need to specify an array of topic partitions"); + } + + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + + int timeout_ms; + Nan::Maybe maybeTimeout = + Nan::To(info[1].As()); + + if (maybeTimeout.IsNothing()) { + timeout_ms = 1000; + } else { + timeout_ms = static_cast(maybeTimeout.FromJust()); + } + + v8::Local cb = info[2].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Connection* handle = ObjectWrap::Unwrap(info.This()); + + Nan::AsyncQueueWorker( + new Workers::Handle::OffsetsForTimes(callback, handle, + toppars, timeout_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Connection::NodeQueryWatermarkOffsets) { + Nan::HandleScope scope; + + Connection* obj = ObjectWrap::Unwrap(info.This()); + + if (!info[0]->IsString()) { + Nan::ThrowError("1st parameter must be a topic string");; + return; + } + + if (!info[1]->IsNumber()) { + Nan::ThrowError("2nd parameter must be a partition number"); + return; + } + + if (!info[2]->IsNumber()) { + Nan::ThrowError("3rd parameter must be a number of milliseconds"); + return; + } + + if (!info[3]->IsFunction()) { + Nan::ThrowError("4th parameter must be a callback"); + return; + } + + // Get string pointer for the topic name + Nan::Utf8String topicUTF8(Nan::To(info[0]).ToLocalChecked()); + // The first parameter is the topic + std::string topic_name(*topicUTF8); + + // Second parameter is the partition + int32_t partition = Nan::To(info[1]).FromJust(); + + // Third parameter is the timeout + int timeout_ms = Nan::To(info[2]).FromJust(); + + // Fourth parameter is the callback + v8::Local cb = info[3].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Nan::AsyncQueueWorker(new Workers::ConnectionQueryWatermarkOffsets( + callback, obj, topic_name, partition, timeout_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +// Node methods +NAN_METHOD(Connection::NodeConfigureCallbacks) { + Nan::HandleScope scope; + + if (info.Length() < 2 || + !info[0]->IsBoolean() || + !info[1]->IsObject()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callbacks object"); + } + v8::Local context = Nan::GetCurrentContext(); + Connection* obj = ObjectWrap::Unwrap(info.This()); + + const bool add = Nan::To(info[0]).ToChecked(); + v8::Local configs_object = info[1]->ToObject(context).ToLocalChecked(); + v8::Local configs_property_names = configs_object->GetOwnPropertyNames(context).ToLocalChecked(); + + for (unsigned int j = 0; j < configs_property_names->Length(); ++j) { + std::string configs_string_key; + + v8::Local configs_key = Nan::Get(configs_property_names, j).ToLocalChecked(); + v8::Local configs_value = Nan::Get(configs_object, configs_key).ToLocalChecked(); + + int config_type = 0; + if (configs_value->IsObject() && configs_key->IsString()) { + Nan::Utf8String configs_utf8_key(configs_key); + configs_string_key = std::string(*configs_utf8_key); + if (configs_string_key.compare("global") == 0) { + config_type = 1; + } else if (configs_string_key.compare("topic") == 0) { + config_type = 2; + } else if (configs_string_key.compare("event") == 0) { + config_type = 3; + } else { + continue; + } + } else { + continue; + } + + v8::Local object = configs_value->ToObject(context).ToLocalChecked(); + v8::Local property_names = object->GetOwnPropertyNames(context).ToLocalChecked(); + + for (unsigned int i = 0; i < property_names->Length(); ++i) { + std::string errstr; + std::string string_key; + + v8::Local key = Nan::Get(property_names, i).ToLocalChecked(); + v8::Local value = Nan::Get(object, key).ToLocalChecked(); + + if (key->IsString()) { + Nan::Utf8String utf8_key(key); + string_key = std::string(*utf8_key); + } else { + continue; + } + + if (value->IsFunction()) { + v8::Local cb = value.As(); + switch (config_type) { + case 1: + obj->m_gconfig->ConfigureCallback(string_key, cb, add, errstr); + if (!errstr.empty()) { + return Nan::ThrowError(errstr.c_str()); + } + break; + case 2: + obj->m_tconfig->ConfigureCallback(string_key, cb, add, errstr); + if (!errstr.empty()) { + return Nan::ThrowError(errstr.c_str()); + } + break; + case 3: + obj->ConfigureCallback(string_key, cb, add); + break; + } + } + } + } + + info.GetReturnValue().Set(Nan::True()); +} + +} // namespace NodeKafka diff --git a/src/connection.h b/src/connection.h new file mode 100644 index 0000000..8c4ac73 --- /dev/null +++ b/src/connection.h @@ -0,0 +1,97 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_CONNECTION_H_ +#define SRC_CONNECTION_H_ + +#include +#include +#include +#include + +#include "rdkafkacpp.h" + +#include "src/common.h" +#include "src/errors.h" +#include "src/config.h" +#include "src/callbacks.h" + +namespace NodeKafka { + +/** + * @brief Connection v8 wrapped object. + * + * Wraps the RdKafka::Handle object with compositional inheritence and + * provides sensible defaults for exposing callbacks to node + * + * This object can't itself expose methods to the prototype directly, as far + * as I can tell. But it can provide the NAN_METHODS that just need to be added + * to the prototype. Since connections, etc. are managed differently based on + * whether it is a producer or consumer, they manage that. This base class + * handles some of the wrapping functionality and more importantly, the + * configuration of callbacks + * + * Any callback available to both consumers and producers, like logging or + * events will be handled in here. + * + * @sa RdKafka::Handle + * @sa NodeKafka::Client + */ + +class Connection : public Nan::ObjectWrap { + public: + bool IsConnected(); + bool IsClosing(); + + // Baton + Baton CreateTopic(std::string); + Baton CreateTopic(std::string, RdKafka::Conf*); + Baton GetMetadata(bool, std::string, int); + Baton QueryWatermarkOffsets(std::string, int32_t, int64_t*, int64_t*, int); + Baton OffsetsForTimes(std::vector &, int); + + RdKafka::Handle* GetClient(); + + static RdKafka::TopicPartition* GetPartition(std::string &); + static RdKafka::TopicPartition* GetPartition(std::string &, int); + + Callbacks::Event m_event_cb; + + virtual void ActivateDispatchers() = 0; + virtual void DeactivateDispatchers() = 0; + + virtual void ConfigureCallback(const std::string &string_key, const v8::Local &cb, bool add); + + protected: + Connection(Conf*, Conf*); + ~Connection(); + + static Nan::Persistent constructor; + static void New(const Nan::FunctionCallbackInfo& info); + + bool m_has_been_disconnected; + bool m_is_closing; + + Conf* m_gconfig; + Conf* m_tconfig; + std::string m_errstr; + + uv_rwlock_t m_connection_lock; + + RdKafka::Handle* m_client; + + static NAN_METHOD(NodeConfigureCallbacks); + static NAN_METHOD(NodeGetMetadata); + static NAN_METHOD(NodeQueryWatermarkOffsets); + static NAN_METHOD(NodeOffsetsForTimes); +}; + +} // namespace NodeKafka + +#endif // SRC_CONNECTION_H_ diff --git a/src/errors.cc b/src/errors.cc new file mode 100644 index 0000000..220773f --- /dev/null +++ b/src/errors.cc @@ -0,0 +1,96 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include + +#include "src/errors.h" + +namespace NodeKafka { + +v8::Local RdKafkaError(const RdKafka::ErrorCode &err, std::string errstr) { // NOLINT + // + int code = static_cast(err); + + v8::Local ret = Nan::New(); + + Nan::Set(ret, Nan::New("message").ToLocalChecked(), + Nan::New(errstr).ToLocalChecked()); + Nan::Set(ret, Nan::New("code").ToLocalChecked(), + Nan::New(code)); + + return ret; +} + +v8::Local RdKafkaError(const RdKafka::ErrorCode &err) { + return RdKafkaError(err, RdKafka::err2str(err)); +} + +v8::Local RdKafkaError(const RdKafka::ErrorCode &err, std::string errstr, + bool isFatal, bool isRetriable, bool isTxnRequiresAbort) { + v8::Local ret = RdKafkaError(err, errstr); + + Nan::Set(ret, Nan::New("isFatal").ToLocalChecked(), + Nan::New(isFatal)); + Nan::Set(ret, Nan::New("isRetriable").ToLocalChecked(), + Nan::New(isRetriable)); + Nan::Set(ret, Nan::New("isTxnRequiresAbort").ToLocalChecked(), + Nan::New(isTxnRequiresAbort)); + + return ret; +} + +Baton::Baton(const RdKafka::ErrorCode &code) { + m_err = code; +} + +Baton::Baton(const RdKafka::ErrorCode &code, std::string errstr) { + m_err = code; + m_errstr = errstr; +} + +Baton::Baton(void* data) { + m_err = RdKafka::ERR_NO_ERROR; + m_data = data; +} + +Baton::Baton(const RdKafka::ErrorCode &code, std::string errstr, bool isFatal, + bool isRetriable, bool isTxnRequiresAbort) { + m_err = code; + m_errstr = errstr; + m_isFatal = isFatal; + m_isRetriable = isRetriable; + m_isTxnRequiresAbort = isTxnRequiresAbort; +} + + +v8::Local Baton::ToObject() { + if (m_errstr.empty()) { + return RdKafkaError(m_err); + } else { + return RdKafkaError(m_err, m_errstr); + } +} + +v8::Local Baton::ToTxnObject() { + return RdKafkaError(m_err, m_errstr, m_isFatal, m_isRetriable, m_isTxnRequiresAbort); +} + +RdKafka::ErrorCode Baton::err() { + return m_err; +} + +std::string Baton::errstr() { + if (m_errstr.empty()) { + return RdKafka::err2str(m_err); + } else { + return m_errstr; + } +} + +} // namespace NodeKafka diff --git a/src/errors.h b/src/errors.h new file mode 100644 index 0000000..bec0beb --- /dev/null +++ b/src/errors.h @@ -0,0 +1,54 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_ERRORS_H_ +#define SRC_ERRORS_H_ + +#include +#include +#include + +#include "rdkafkacpp.h" + +#include "src/common.h" + +namespace NodeKafka { + +class Baton { + public: + explicit Baton(const RdKafka::ErrorCode &); + explicit Baton(void* data); + explicit Baton(const RdKafka::ErrorCode &, std::string); + explicit Baton(const RdKafka::ErrorCode &, std::string, bool isFatal, + bool isRetriable, bool isTxnRequiresAbort); + + template T data() { + return static_cast(m_data); + } + + RdKafka::ErrorCode err(); + std::string errstr(); + + v8::Local ToObject(); + v8::Local ToTxnObject(); + + private: + void* m_data; + std::string m_errstr; + RdKafka::ErrorCode m_err; + bool m_isFatal; + bool m_isRetriable; + bool m_isTxnRequiresAbort; +}; + +v8::Local RdKafkaError(const RdKafka::ErrorCode &); + +} // namespace NodeKafka + +#endif // SRC_ERRORS_H_ diff --git a/src/kafka-consumer.cc b/src/kafka-consumer.cc new file mode 100644 index 0000000..2991140 --- /dev/null +++ b/src/kafka-consumer.cc @@ -0,0 +1,1229 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include + +#include "src/kafka-consumer.h" +#include "src/workers.h" + +using Nan::FunctionCallbackInfo; + +namespace NodeKafka { + +/** + * @brief KafkaConsumer v8 wrapped object. + * + * Specializes the connection to wrap a consumer object through compositional + * inheritence. Establishes its prototype in node through `Init` + * + * @sa RdKafka::Handle + * @sa NodeKafka::Client + */ + +KafkaConsumer::KafkaConsumer(Conf* gconfig, Conf* tconfig): + Connection(gconfig, tconfig) { + std::string errstr; + + m_gconfig->set("default_topic_conf", m_tconfig, errstr); + } + +KafkaConsumer::~KafkaConsumer() { + // We only want to run this if it hasn't been run already + Disconnect(); +} + +Baton KafkaConsumer::Connect() { + if (IsConnected()) { + return Baton(RdKafka::ERR_NO_ERROR); + } + + std::string errstr; + { + scoped_shared_write_lock lock(m_connection_lock); + m_client = RdKafka::KafkaConsumer::create(m_gconfig, errstr); + } + + if (!m_client || !errstr.empty()) { + return Baton(RdKafka::ERR__STATE, errstr); + } + + if (m_partitions.size() > 0) { + m_client->resume(m_partitions); + } + + return Baton(RdKafka::ERR_NO_ERROR); +} + +void KafkaConsumer::ActivateDispatchers() { + // Listen to global config + m_gconfig->listen(); + + // Listen to non global config + // tconfig->listen(); + + // This should be refactored to config based management + m_event_cb.dispatcher.Activate(); +} + +Baton KafkaConsumer::Disconnect() { + // Only close client if it is connected + RdKafka::ErrorCode err = RdKafka::ERR_NO_ERROR; + + if (IsConnected()) { + m_is_closing = true; + { + scoped_shared_write_lock lock(m_connection_lock); + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + err = consumer->close(); + + delete m_client; + m_client = NULL; + } + } + + m_is_closing = false; + + return Baton(err); +} + +void KafkaConsumer::DeactivateDispatchers() { + // Stop listening to the config dispatchers + m_gconfig->stop(); + + // Also this one + m_event_cb.dispatcher.Deactivate(); +} + +bool KafkaConsumer::IsSubscribed() { + if (!IsConnected()) { + return false; + } + + if (!m_is_subscribed) { + return false; + } + + return true; +} + + +bool KafkaConsumer::HasAssignedPartitions() { + return !m_partitions.empty(); +} + +int KafkaConsumer::AssignedPartitionCount() { + return m_partition_cnt; +} + +Baton KafkaConsumer::GetWatermarkOffsets( + std::string topic_name, int32_t partition, + int64_t* low_offset, int64_t* high_offset) { + // Check if we are connected first + + RdKafka::ErrorCode err; + + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (IsConnected()) { + // Always send true - we + err = m_client->get_watermark_offsets(topic_name, partition, + low_offset, high_offset); + } else { + err = RdKafka::ERR__STATE; + } + } else { + err = RdKafka::ERR__STATE; + } + + return Baton(err); +} + +void KafkaConsumer::part_list_print(const std::vector &partitions) { // NOLINT + for (unsigned int i = 0 ; i < partitions.size() ; i++) + std::cerr << partitions[i]->topic() << + "[" << partitions[i]->partition() << "], "; + std::cerr << std::endl; +} + +Baton KafkaConsumer::Assign(std::vector partitions) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is disconnected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode errcode = consumer->assign(partitions); + + if (errcode == RdKafka::ERR_NO_ERROR) { + m_partition_cnt = partitions.size(); + m_partitions.swap(partitions); + } + + // Destroy the partitions: Either we're using them (and partitions + // is now our old vector), or we're not using it as there was an + // error. + RdKafka::TopicPartition::destroy(partitions); + + return Baton(errcode); +} + +Baton KafkaConsumer::Unassign() { + if (!IsClosing() && !IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode errcode = consumer->unassign(); + + if (errcode != RdKafka::ERR_NO_ERROR) { + return Baton(errcode); + } + + // Destroy the old list of partitions since we are no longer using it + RdKafka::TopicPartition::destroy(m_partitions); + + m_partition_cnt = 0; + + return Baton(RdKafka::ERR_NO_ERROR); +} + +Baton KafkaConsumer::Commit(std::vector toppars) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode err = consumer->commitAsync(toppars); + + return Baton(err); +} + +Baton KafkaConsumer::Commit(RdKafka::TopicPartition * toppar) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + // Need to put topic in a vector for it to work + std::vector offsets = {toppar}; + RdKafka::ErrorCode err = consumer->commitAsync(offsets); + + return Baton(err); +} + +Baton KafkaConsumer::Commit() { + // sets an error message + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode err = consumer->commitAsync(); + + return Baton(err); +} + +// Synchronous commit events +Baton KafkaConsumer::CommitSync(std::vector toppars) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode err = consumer->commitSync(toppars); + // RdKafka::TopicPartition::destroy(toppars); + + return Baton(err); +} + +Baton KafkaConsumer::CommitSync(RdKafka::TopicPartition * toppar) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + // Need to put topic in a vector for it to work + std::vector offsets = {toppar}; + RdKafka::ErrorCode err = consumer->commitSync(offsets); + + return Baton(err); +} + +Baton KafkaConsumer::CommitSync() { + // sets an error message + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode err = consumer->commitSync(); + + return Baton(err); +} + +Baton KafkaConsumer::Seek(const RdKafka::TopicPartition &partition, int timeout_ms) { // NOLINT + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode err = consumer->seek(partition, timeout_ms); + + return Baton(err); +} + +Baton KafkaConsumer::Committed(std::vector &toppars, + int timeout_ms) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode err = consumer->committed(toppars, timeout_ms); + + return Baton(err); +} + +Baton KafkaConsumer::Position(std::vector &toppars) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode err = consumer->position(toppars); + + return Baton(err); +} + +Baton KafkaConsumer::Subscription() { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "Consumer is not connected"); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + // Needs to be a pointer since we're returning it through the baton + std::vector * topics = new std::vector; + + RdKafka::ErrorCode err = consumer->subscription(*topics); + + if (err == RdKafka::ErrorCode::ERR_NO_ERROR) { + // Good to go + return Baton(topics); + } + + return Baton(err); +} + +Baton KafkaConsumer::Unsubscribe() { + if (IsConnected() && IsSubscribed()) { + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + consumer->unsubscribe(); + m_is_subscribed = false; + } + + return Baton(RdKafka::ERR_NO_ERROR); +} + +Baton KafkaConsumer::Pause(std::vector & toppars) { + if (IsConnected()) { + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + RdKafka::ErrorCode err = consumer->pause(toppars); + + return Baton(err); + } + + return Baton(RdKafka::ERR__STATE); +} + +Baton KafkaConsumer::Resume(std::vector & toppars) { + if (IsConnected()) { + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + RdKafka::ErrorCode err = consumer->resume(toppars); + + return Baton(err); + } + + return Baton(RdKafka::ERR__STATE); +} + +Baton KafkaConsumer::OffsetsStore(std::vector & toppars) { // NOLINT + if (IsConnected() && IsSubscribed()) { + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + RdKafka::ErrorCode err = consumer->offsets_store(toppars); + + return Baton(err); + } + + return Baton(RdKafka::ERR__STATE); +} + +Baton KafkaConsumer::Subscribe(std::vector topics) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::ErrorCode errcode = consumer->subscribe(topics); + if (errcode != RdKafka::ERR_NO_ERROR) { + return Baton(errcode); + } + + m_is_subscribed = true; + + return Baton(RdKafka::ERR_NO_ERROR); +} + +Baton KafkaConsumer::Consume(int timeout_ms) { + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } else { + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + RdKafka::Message * message = consumer->consume(timeout_ms); + RdKafka::ErrorCode response_code = message->err(); + // we want to handle these errors at the call site + if (response_code != RdKafka::ERR_NO_ERROR && + response_code != RdKafka::ERR__PARTITION_EOF && + response_code != RdKafka::ERR__TIMED_OUT && + response_code != RdKafka::ERR__TIMED_OUT_QUEUE + ) { + delete message; + return Baton(response_code); + } + + return Baton(message); + } + } else { + return Baton(RdKafka::ERR__STATE, "KafkaConsumer is not connected"); + } +} + +Baton KafkaConsumer::RefreshAssignments() { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::KafkaConsumer* consumer = + dynamic_cast(m_client); + + std::vector partition_list; + RdKafka::ErrorCode err = consumer->assignment(partition_list); + + switch (err) { + case RdKafka::ERR_NO_ERROR: + m_partition_cnt = partition_list.size(); + m_partitions.swap(partition_list); + + // These are pointers so we need to delete them somewhere. + // Do it here because we're only going to convert when we're ready + // to return to v8. + RdKafka::TopicPartition::destroy(partition_list); + return Baton(RdKafka::ERR_NO_ERROR); + break; + default: + return Baton(err); + break; + } +} + +std::string KafkaConsumer::Name() { + if (!IsConnected()) { + return std::string(""); + } + return std::string(m_client->name()); +} + +Nan::Persistent KafkaConsumer::constructor; + +void KafkaConsumer::Init(v8::Local exports) { + Nan::HandleScope scope; + + v8::Local tpl = Nan::New(New); + tpl->SetClassName(Nan::New("KafkaConsumer").ToLocalChecked()); + tpl->InstanceTemplate()->SetInternalFieldCount(1); + + /* + * Lifecycle events inherited from NodeKafka::Connection + * + * @sa NodeKafka::Connection + */ + + Nan::SetPrototypeMethod(tpl, "configureCallbacks", NodeConfigureCallbacks); + + /* + * @brief Methods to do with establishing state + */ + + Nan::SetPrototypeMethod(tpl, "connect", NodeConnect); + Nan::SetPrototypeMethod(tpl, "disconnect", NodeDisconnect); + Nan::SetPrototypeMethod(tpl, "getMetadata", NodeGetMetadata); + Nan::SetPrototypeMethod(tpl, "queryWatermarkOffsets", NodeQueryWatermarkOffsets); // NOLINT + Nan::SetPrototypeMethod(tpl, "offsetsForTimes", NodeOffsetsForTimes); + Nan::SetPrototypeMethod(tpl, "getWatermarkOffsets", NodeGetWatermarkOffsets); + + /* + * @brief Methods exposed to do with message retrieval + */ + Nan::SetPrototypeMethod(tpl, "subscription", NodeSubscription); + Nan::SetPrototypeMethod(tpl, "subscribe", NodeSubscribe); + Nan::SetPrototypeMethod(tpl, "unsubscribe", NodeUnsubscribe); + Nan::SetPrototypeMethod(tpl, "consumeLoop", NodeConsumeLoop); + Nan::SetPrototypeMethod(tpl, "consume", NodeConsume); + Nan::SetPrototypeMethod(tpl, "seek", NodeSeek); + + /** + * @brief Pausing and resuming + */ + Nan::SetPrototypeMethod(tpl, "pause", NodePause); + Nan::SetPrototypeMethod(tpl, "resume", NodeResume); + + /* + * @brief Methods to do with partition assignment / rebalancing + */ + + Nan::SetPrototypeMethod(tpl, "committed", NodeCommitted); + Nan::SetPrototypeMethod(tpl, "position", NodePosition); + Nan::SetPrototypeMethod(tpl, "assign", NodeAssign); + Nan::SetPrototypeMethod(tpl, "unassign", NodeUnassign); + Nan::SetPrototypeMethod(tpl, "assignments", NodeAssignments); + + Nan::SetPrototypeMethod(tpl, "commit", NodeCommit); + Nan::SetPrototypeMethod(tpl, "commitSync", NodeCommitSync); + Nan::SetPrototypeMethod(tpl, "offsetsStore", NodeOffsetsStore); + + constructor.Reset((tpl->GetFunction(Nan::GetCurrentContext())) + .ToLocalChecked()); + Nan::Set(exports, Nan::New("KafkaConsumer").ToLocalChecked(), + (tpl->GetFunction(Nan::GetCurrentContext())).ToLocalChecked()); +} + +void KafkaConsumer::New(const Nan::FunctionCallbackInfo& info) { + if (!info.IsConstructCall()) { + return Nan::ThrowError("non-constructor invocation not supported"); + } + + if (info.Length() < 2) { + return Nan::ThrowError("You must supply global and topic configuration"); + } + + if (!info[0]->IsObject()) { + return Nan::ThrowError("Global configuration data must be specified"); + } + + if (!info[1]->IsObject()) { + return Nan::ThrowError("Topic configuration must be specified"); + } + + std::string errstr; + + Conf* gconfig = + Conf::create(RdKafka::Conf::CONF_GLOBAL, + (info[0]->ToObject(Nan::GetCurrentContext())).ToLocalChecked(), errstr); + + if (!gconfig) { + return Nan::ThrowError(errstr.c_str()); + } + + Conf* tconfig = + Conf::create(RdKafka::Conf::CONF_TOPIC, + (info[1]->ToObject(Nan::GetCurrentContext())).ToLocalChecked(), errstr); + + if (!tconfig) { + delete gconfig; + return Nan::ThrowError(errstr.c_str()); + } + + KafkaConsumer* consumer = new KafkaConsumer(gconfig, tconfig); + + // Wrap it + consumer->Wrap(info.This()); + + // Then there is some weird initialization that happens + // basically it sets the configuration data + // we don't need to do that because we lazy load it + + info.GetReturnValue().Set(info.This()); +} + +v8::Local KafkaConsumer::NewInstance(v8::Local arg) { + Nan::EscapableHandleScope scope; + + const unsigned argc = 1; + + v8::Local argv[argc] = { arg }; + v8::Local cons = Nan::New(constructor); + v8::Local instance = + Nan::NewInstance(cons, argc, argv).ToLocalChecked(); + + return scope.Escape(instance); +} + +/* Node exposed methods */ + +NAN_METHOD(KafkaConsumer::NodeCommitted) { + Nan::HandleScope scope; + + if (info.Length() < 3 || !info[0]->IsArray()) { + // Just throw an exception + return Nan::ThrowError("Need to specify an array of topic partitions"); + } + + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + + int timeout_ms; + Nan::Maybe maybeTimeout = + Nan::To(info[1].As()); + + if (maybeTimeout.IsNothing()) { + timeout_ms = 1000; + } else { + timeout_ms = static_cast(maybeTimeout.FromJust()); + } + + v8::Local cb = info[2].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + Nan::AsyncQueueWorker( + new Workers::KafkaConsumerCommitted(callback, consumer, + toppars, timeout_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(KafkaConsumer::NodeSubscription) { + Nan::HandleScope scope; + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + Baton b = consumer->Subscription(); + + if (b.err() != RdKafka::ErrorCode::ERR_NO_ERROR) { + // Let the JS library throw if we need to so the error can be more rich + int error_code = static_cast(b.err()); + return info.GetReturnValue().Set(Nan::New(error_code)); + } + + std::vector * topics = b.data*>(); + + info.GetReturnValue().Set(Conversion::Topic::ToV8Array(*topics)); + + delete topics; +} + +NAN_METHOD(KafkaConsumer::NodePosition) { + Nan::HandleScope scope; + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + if (info.Length() < 1 || !info[0]->IsArray()) { + // Just throw an exception + return Nan::ThrowError("Need to specify an array of topic partitions"); + } + + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + + Baton b = consumer->Position(toppars); + + if (b.err() != RdKafka::ErrorCode::ERR_NO_ERROR) { + // Let the JS library throw if we need to so the error can be more rich + int error_code = static_cast(b.err()); + return info.GetReturnValue().Set(Nan::New(error_code)); + } + + info.GetReturnValue().Set( + Conversion::TopicPartition::ToV8Array(toppars)); + + // Delete the underlying topic partitions + RdKafka::TopicPartition::destroy(toppars); +} + +NAN_METHOD(KafkaConsumer::NodeAssignments) { + Nan::HandleScope scope; + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + Baton b = consumer->RefreshAssignments(); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + // Let the JS library throw if we need to so the error can be more rich + int error_code = static_cast(b.err()); + return info.GetReturnValue().Set(Nan::New(error_code)); + } + + info.GetReturnValue().Set( + Conversion::TopicPartition::ToV8Array(consumer->m_partitions)); +} + +NAN_METHOD(KafkaConsumer::NodeAssign) { + Nan::HandleScope scope; + + if (info.Length() < 1 || !info[0]->IsArray()) { + // Just throw an exception + return Nan::ThrowError("Need to specify an array of partitions"); + } + + v8::Local partitions = info[0].As(); + std::vector topic_partitions; + + for (unsigned int i = 0; i < partitions->Length(); ++i) { + v8::Local partition_obj_value; + if (!( + Nan::Get(partitions, i).ToLocal(&partition_obj_value) && + partition_obj_value->IsObject())) { + Nan::ThrowError("Must pass topic-partition objects"); + } + + v8::Local partition_obj = partition_obj_value.As(); + + // Got the object + int64_t partition = GetParameter(partition_obj, "partition", -1); + std::string topic = GetParameter(partition_obj, "topic", ""); + + if (!topic.empty()) { + RdKafka::TopicPartition* part; + + if (partition < 0) { + part = Connection::GetPartition(topic); + } else { + part = Connection::GetPartition(topic, partition); + } + + // Set the default value to offset invalid. If provided, we will not set + // the offset. + int64_t offset = GetParameter( + partition_obj, "offset", RdKafka::Topic::OFFSET_INVALID); + if (offset != RdKafka::Topic::OFFSET_INVALID) { + part->set_offset(offset); + } + + topic_partitions.push_back(part); + } + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + // Hand over the partitions to the consumer. + Baton b = consumer->Assign(topic_partitions); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + Nan::ThrowError(RdKafka::err2str(b.err()).c_str()); + } + + info.GetReturnValue().Set(Nan::True()); +} + +NAN_METHOD(KafkaConsumer::NodeUnassign) { + Nan::HandleScope scope; + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + + if (!consumer->IsClosing() && !consumer->IsConnected()) { + Nan::ThrowError("KafkaConsumer is disconnected"); + return; + } + + Baton b = consumer->Unassign(); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + Nan::ThrowError(RdKafka::err2str(b.err()).c_str()); + } + + info.GetReturnValue().Set(Nan::True()); +} + +NAN_METHOD(KafkaConsumer::NodeUnsubscribe) { + Nan::HandleScope scope; + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + Baton b = consumer->Unsubscribe(); + + info.GetReturnValue().Set(Nan::New(static_cast(b.err()))); +} + +NAN_METHOD(KafkaConsumer::NodeCommit) { + Nan::HandleScope scope; + int error_code; + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + if (!consumer->IsConnected()) { + Nan::ThrowError("KafkaConsumer is disconnected"); + return; + } + + if (info[0]->IsNull() || info[0]->IsUndefined()) { + Baton b = consumer->Commit(); + error_code = static_cast(b.err()); + } else if (info[0]->IsArray()) { + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + + Baton b = consumer->Commit(toppars); + error_code = static_cast(b.err()); + + RdKafka::TopicPartition::destroy(toppars); + } else if (info[0]->IsObject()) { + RdKafka::TopicPartition * toppar = + Conversion::TopicPartition::FromV8Object(info[0].As()); + + if (toppar == NULL) { + Nan::ThrowError("Invalid topic partition provided"); + return; + } + + Baton b = consumer->Commit(toppar); + error_code = static_cast(b.err()); + + delete toppar; + } else { + Nan::ThrowError("First parameter must be an object or an array"); + return; + } + + info.GetReturnValue().Set(Nan::New(error_code)); +} + +NAN_METHOD(KafkaConsumer::NodeCommitSync) { + Nan::HandleScope scope; + int error_code; + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + if (!consumer->IsConnected()) { + Nan::ThrowError("KafkaConsumer is disconnected"); + return; + } + + if (info[0]->IsNull() || info[0]->IsUndefined()) { + Baton b = consumer->CommitSync(); + error_code = static_cast(b.err()); + } else if (info[0]->IsArray()) { + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + + Baton b = consumer->CommitSync(toppars); + error_code = static_cast(b.err()); + + RdKafka::TopicPartition::destroy(toppars); + } else if (info[0]->IsObject()) { + RdKafka::TopicPartition * toppar = + Conversion::TopicPartition::FromV8Object(info[0].As()); + + if (toppar == NULL) { + Nan::ThrowError("Invalid topic partition provided"); + return; + } + + Baton b = consumer->CommitSync(toppar); + error_code = static_cast(b.err()); + + delete toppar; + } else { + Nan::ThrowError("First parameter must be an object or an array"); + return; + } + + info.GetReturnValue().Set(Nan::New(error_code)); +} + +NAN_METHOD(KafkaConsumer::NodeSubscribe) { + Nan::HandleScope scope; + + if (info.Length() < 1 || !info[0]->IsArray()) { + // Just throw an exception + return Nan::ThrowError("First parameter must be an array"); + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + v8::Local topicsArray = info[0].As(); + std::vector topics = Conversion::Topic::ToStringVector(topicsArray); // NOLINT + + Baton b = consumer->Subscribe(topics); + + int error_code = static_cast(b.err()); + info.GetReturnValue().Set(Nan::New(error_code)); +} + +NAN_METHOD(KafkaConsumer::NodeSeek) { + Nan::HandleScope scope; + + // If number of parameters is less than 3 (need topic partition, timeout, + // and callback), we can't call this thing + if (info.Length() < 3) { + return Nan::ThrowError("Must provide a topic partition, timeout, and callback"); // NOLINT + } + + if (!info[0]->IsObject()) { + return Nan::ThrowError("Topic partition must be an object"); + } + + if (!info[1]->IsNumber() && !info[1]->IsNull()) { + return Nan::ThrowError("Timeout must be a number."); + } + + if (!info[2]->IsFunction()) { + return Nan::ThrowError("Callback must be a function"); + } + + int timeout_ms; + Nan::Maybe maybeTimeout = + Nan::To(info[1].As()); + + if (maybeTimeout.IsNothing()) { + timeout_ms = 1000; + } else { + timeout_ms = static_cast(maybeTimeout.FromJust()); + // Do not allow timeouts of less than 10. Providing 0 causes segfaults + // because it makes it asynchronous. + if (timeout_ms < 10) { + timeout_ms = 10; + } + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + const RdKafka::TopicPartition * toppar = + Conversion::TopicPartition::FromV8Object(info[0].As()); + + if (!toppar) { + return Nan::ThrowError("Invalid topic partition provided"); + } + + Nan::Callback *callback = new Nan::Callback(info[2].As()); + Nan::AsyncQueueWorker( + new Workers::KafkaConsumerSeek(callback, consumer, toppar, timeout_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(KafkaConsumer::NodeOffsetsStore) { + Nan::HandleScope scope; + + // If number of parameters is less than 3 (need topic partition, timeout, + // and callback), we can't call this thing + if (info.Length() < 1) { + return Nan::ThrowError("Must provide a list of topic partitions"); + } + + if (!info[0]->IsArray()) { + return Nan::ThrowError("Topic partition must be an array of objects"); + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + + Baton b = consumer->OffsetsStore(toppars); + RdKafka::TopicPartition::destroy(toppars); + + int error_code = static_cast(b.err()); + info.GetReturnValue().Set(Nan::New(error_code)); +} + +NAN_METHOD(KafkaConsumer::NodePause) { + Nan::HandleScope scope; + + // If number of parameters is less than 3 (need topic partition, timeout, + // and callback), we can't call this thing + if (info.Length() < 1) { + return Nan::ThrowError("Must provide a list of topic partitions"); + } + + if (!info[0]->IsArray()) { + return Nan::ThrowError("Topic partition must be an array of objects"); + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + + Baton b = consumer->Pause(toppars); + RdKafka::TopicPartition::destroy(toppars); + + #if 0 + // Now iterate through and delete these toppars + for (std::vector::const_iterator it = toppars.begin(); // NOLINT + it != toppars.end(); it++) { + RdKafka::TopicPartition* toppar = *it; + if (toppar->err() != RdKafka::ERR_NO_ERROR) { + // Need to somehow transmit this information. + // @TODO(webmakersteve) + } + delete toppar; + } + #endif + + int error_code = static_cast(b.err()); + info.GetReturnValue().Set(Nan::New(error_code)); +} + +NAN_METHOD(KafkaConsumer::NodeResume) { + Nan::HandleScope scope; + + // If number of parameters is less than 3 (need topic partition, timeout, + // and callback), we can't call this thing + if (info.Length() < 1) { + return Nan::ThrowError("Must provide a list of topic partitions"); // NOLINT + } + + if (!info[0]->IsArray()) { + return Nan::ThrowError("Topic partition must be an array of objects"); + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + + Baton b = consumer->Resume(toppars); + + // Now iterate through and delete these toppars + for (std::vector::const_iterator it = toppars.begin(); // NOLINT + it != toppars.end(); it++) { + RdKafka::TopicPartition* toppar = *it; + if (toppar->err() != RdKafka::ERR_NO_ERROR) { + // Need to somehow transmit this information. + // @TODO(webmakersteve) + } + delete toppar; + } + + int error_code = static_cast(b.err()); + info.GetReturnValue().Set(Nan::New(error_code)); +} + +NAN_METHOD(KafkaConsumer::NodeConsumeLoop) { + Nan::HandleScope scope; + + if (info.Length() < 3) { + // Just throw an exception + return Nan::ThrowError("Invalid number of parameters"); + } + + if (!info[0]->IsNumber()) { + return Nan::ThrowError("Need to specify a timeout"); + } + + if (!info[1]->IsNumber()) { + return Nan::ThrowError("Need to specify a sleep delay"); + } + + if (!info[2]->IsFunction()) { + return Nan::ThrowError("Need to specify a callback"); + } + + int timeout_ms; + Nan::Maybe maybeTimeout = + Nan::To(info[0].As()); + + if (maybeTimeout.IsNothing()) { + timeout_ms = 1000; + } else { + timeout_ms = static_cast(maybeTimeout.FromJust()); + } + int retry_read_ms; + Nan::Maybe maybeSleep = + Nan::To(info[1].As()); + + if (maybeSleep.IsNothing()) { + retry_read_ms = 500; + } else { + retry_read_ms = static_cast(maybeSleep.FromJust()); + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + v8::Local cb = info[2].As(); + + Nan::Callback *callback = new Nan::Callback(cb); + Nan::AsyncQueueWorker( + new Workers::KafkaConsumerConsumeLoop(callback, consumer, timeout_ms, retry_read_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(KafkaConsumer::NodeConsume) { + Nan::HandleScope scope; + + if (info.Length() < 2) { + // Just throw an exception + return Nan::ThrowError("Invalid number of parameters"); + } + + int timeout_ms; + Nan::Maybe maybeTimeout = + Nan::To(info[0].As()); + + if (maybeTimeout.IsNothing()) { + timeout_ms = 1000; + } else { + timeout_ms = static_cast(maybeTimeout.FromJust()); + } + + if (info[1]->IsNumber()) { + if (!info[2]->IsFunction()) { + return Nan::ThrowError("Need to specify a callback"); + } + + v8::Local numMessagesNumber = info[1].As(); + Nan::Maybe numMessagesMaybe = Nan::To(numMessagesNumber); // NOLINT + + uint32_t numMessages; + if (numMessagesMaybe.IsNothing()) { + return Nan::ThrowError("Parameter must be a number over 0"); + } else { + numMessages = numMessagesMaybe.FromJust(); + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + v8::Local cb = info[2].As(); + Nan::Callback *callback = new Nan::Callback(cb); + Nan::AsyncQueueWorker( + new Workers::KafkaConsumerConsumeNum(callback, consumer, numMessages, timeout_ms)); // NOLINT + + } else { + if (!info[1]->IsFunction()) { + return Nan::ThrowError("Need to specify a callback"); + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + v8::Local cb = info[1].As(); + Nan::Callback *callback = new Nan::Callback(cb); + Nan::AsyncQueueWorker( + new Workers::KafkaConsumerConsume(callback, consumer, timeout_ms)); + } + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(KafkaConsumer::NodeConnect) { + Nan::HandleScope scope; + + if (info.Length() < 1 || !info[0]->IsFunction()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback"); + } + + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + Nan::Callback *callback = new Nan::Callback(info[0].As()); + Nan::AsyncQueueWorker(new Workers::KafkaConsumerConnect(callback, consumer)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(KafkaConsumer::NodeDisconnect) { + Nan::HandleScope scope; + + if (info.Length() < 1 || !info[0]->IsFunction()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback"); + } + + v8::Local cb = info[0].As(); + Nan::Callback *callback = new Nan::Callback(cb); + KafkaConsumer* consumer = ObjectWrap::Unwrap(info.This()); + + Nan::AsyncQueueWorker( + new Workers::KafkaConsumerDisconnect(callback, consumer)); + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(KafkaConsumer::NodeGetWatermarkOffsets) { + Nan::HandleScope scope; + + KafkaConsumer* obj = ObjectWrap::Unwrap(info.This()); + + if (!info[0]->IsString()) { + Nan::ThrowError("1st parameter must be a topic string");; + return; + } + + if (!info[1]->IsNumber()) { + Nan::ThrowError("2nd parameter must be a partition number"); + return; + } + + // Get string pointer for the topic name + Nan::Utf8String topicUTF8(Nan::To(info[0]).ToLocalChecked()); + // The first parameter is the topic + std::string topic_name(*topicUTF8); + + // Second parameter is the partition + int32_t partition = Nan::To(info[1]).FromJust(); + + // Set these ints which will store the return data + int64_t low_offset; + int64_t high_offset; + + Baton b = obj->GetWatermarkOffsets( + topic_name, partition, &low_offset, &high_offset); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + // Let the JS library throw if we need to so the error can be more rich + int error_code = static_cast(b.err()); + return info.GetReturnValue().Set(Nan::New(error_code)); + } else { + v8::Local offsetsObj = Nan::New(); + Nan::Set(offsetsObj, Nan::New("lowOffset").ToLocalChecked(), + Nan::New(low_offset)); + Nan::Set(offsetsObj, Nan::New("highOffset").ToLocalChecked(), + Nan::New(high_offset)); + + return info.GetReturnValue().Set(offsetsObj); + } +} + +} // namespace NodeKafka diff --git a/src/kafka-consumer.h b/src/kafka-consumer.h new file mode 100644 index 0000000..57b9290 --- /dev/null +++ b/src/kafka-consumer.h @@ -0,0 +1,126 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_KAFKA_CONSUMER_H_ +#define SRC_KAFKA_CONSUMER_H_ + +#include +#include +#include +#include +#include + +#include "rdkafkacpp.h" + +#include "src/common.h" +#include "src/connection.h" +#include "src/callbacks.h" + +namespace NodeKafka { + +/** + * @brief KafkaConsumer v8 wrapped object. + * + * Specializes the connection to wrap a consumer object through compositional + * inheritence. Establishes its prototype in node through `Init` + * + * @sa RdKafka::Handle + * @sa NodeKafka::Client + */ + +class KafkaConsumer : public Connection { + friend class Producer; + public: + static void Init(v8::Local); + static v8::Local NewInstance(v8::Local); + + Baton Connect(); + Baton Disconnect(); + + Baton Subscription(); + Baton Unsubscribe(); + bool IsSubscribed(); + + Baton Pause(std::vector &); + Baton Resume(std::vector &); + + // Asynchronous commit events + Baton Commit(std::vector); + Baton Commit(RdKafka::TopicPartition*); + Baton Commit(); + + Baton OffsetsStore(std::vector &); + Baton GetWatermarkOffsets(std::string, int32_t, int64_t*, int64_t*); + + // Synchronous commit events + Baton CommitSync(std::vector); + Baton CommitSync(RdKafka::TopicPartition*); + Baton CommitSync(); + + Baton Committed(std::vector &, int timeout_ms); + Baton Position(std::vector &); + + Baton RefreshAssignments(); + + bool HasAssignedPartitions(); + int AssignedPartitionCount(); + + Baton Assign(std::vector); + Baton Unassign(); + + Baton Seek(const RdKafka::TopicPartition &partition, int timeout_ms); + + std::string Name(); + + Baton Subscribe(std::vector); + Baton Consume(int timeout_ms); + + void ActivateDispatchers(); + void DeactivateDispatchers(); + + protected: + static Nan::Persistent constructor; + static void New(const Nan::FunctionCallbackInfo& info); + + KafkaConsumer(Conf *, Conf *); + ~KafkaConsumer(); + + private: + static void part_list_print(const std::vector&); + + std::vector m_partitions; + int m_partition_cnt; + bool m_is_subscribed = false; + + // Node methods + static NAN_METHOD(NodeConnect); + static NAN_METHOD(NodeSubscribe); + static NAN_METHOD(NodeDisconnect); + static NAN_METHOD(NodeAssign); + static NAN_METHOD(NodeUnassign); + static NAN_METHOD(NodeAssignments); + static NAN_METHOD(NodeUnsubscribe); + static NAN_METHOD(NodeCommit); + static NAN_METHOD(NodeCommitSync); + static NAN_METHOD(NodeOffsetsStore); + static NAN_METHOD(NodeCommitted); + static NAN_METHOD(NodePosition); + static NAN_METHOD(NodeSubscription); + static NAN_METHOD(NodeSeek); + static NAN_METHOD(NodeGetWatermarkOffsets); + static NAN_METHOD(NodeConsumeLoop); + static NAN_METHOD(NodeConsume); + + static NAN_METHOD(NodePause); + static NAN_METHOD(NodeResume); +}; + +} // namespace NodeKafka + +#endif // SRC_KAFKA_CONSUMER_H_ diff --git a/src/producer.cc b/src/producer.cc new file mode 100644 index 0000000..c3b9824 --- /dev/null +++ b/src/producer.cc @@ -0,0 +1,849 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include + +#include "src/producer.h" +#include "src/kafka-consumer.h" +#include "src/workers.h" + +namespace NodeKafka { + +/** + * @brief Producer v8 wrapped object. + * + * Wraps the RdKafka::Producer object with compositional inheritence and + * provides methods for interacting with it exposed to node. + * + * The base wrappable RdKafka::Handle deals with most of the wrapping but + * we still need to declare its prototype. + * + * @sa RdKafka::Producer + * @sa NodeKafka::Connection + */ + +Producer::Producer(Conf* gconfig, Conf* tconfig): + Connection(gconfig, tconfig), + m_dr_cb(), + m_partitioner_cb() { + std::string errstr; + + m_gconfig->set("default_topic_conf", m_tconfig, errstr); + m_gconfig->set("dr_cb", &m_dr_cb, errstr); + } + +Producer::~Producer() { + Disconnect(); +} + +Nan::Persistent Producer::constructor; + +void Producer::Init(v8::Local exports) { + Nan::HandleScope scope; + + v8::Local tpl = Nan::New(New); + tpl->SetClassName(Nan::New("Producer").ToLocalChecked()); + tpl->InstanceTemplate()->SetInternalFieldCount(1); + + /* + * Lifecycle events inherited from NodeKafka::Connection + * + * @sa NodeKafka::Connection + */ + + Nan::SetPrototypeMethod(tpl, "configureCallbacks", NodeConfigureCallbacks); + + /* + * @brief Methods to do with establishing state + */ + + Nan::SetPrototypeMethod(tpl, "connect", NodeConnect); + Nan::SetPrototypeMethod(tpl, "disconnect", NodeDisconnect); + Nan::SetPrototypeMethod(tpl, "getMetadata", NodeGetMetadata); + Nan::SetPrototypeMethod(tpl, "queryWatermarkOffsets", NodeQueryWatermarkOffsets); // NOLINT + Nan::SetPrototypeMethod(tpl, "poll", NodePoll); + + /* + * @brief Methods exposed to do with message production + */ + + Nan::SetPrototypeMethod(tpl, "setPartitioner", NodeSetPartitioner); + Nan::SetPrototypeMethod(tpl, "produce", NodeProduce); + + Nan::SetPrototypeMethod(tpl, "flush", NodeFlush); + + /* + * @brief Methods exposed to do with transactions + */ + + Nan::SetPrototypeMethod(tpl, "initTransactions", NodeInitTransactions); + Nan::SetPrototypeMethod(tpl, "beginTransaction", NodeBeginTransaction); + Nan::SetPrototypeMethod(tpl, "commitTransaction", NodeCommitTransaction); + Nan::SetPrototypeMethod(tpl, "abortTransaction", NodeAbortTransaction); + Nan::SetPrototypeMethod(tpl, "sendOffsetsToTransaction", NodeSendOffsetsToTransaction); + + // connect. disconnect. resume. pause. get meta data + constructor.Reset((tpl->GetFunction(Nan::GetCurrentContext())) + .ToLocalChecked()); + + Nan::Set(exports, Nan::New("Producer").ToLocalChecked(), + tpl->GetFunction(Nan::GetCurrentContext()).ToLocalChecked()); +} + +void Producer::New(const Nan::FunctionCallbackInfo& info) { + if (!info.IsConstructCall()) { + return Nan::ThrowError("non-constructor invocation not supported"); + } + + if (info.Length() < 2) { + return Nan::ThrowError("You must supply global and topic configuration"); + } + + if (!info[0]->IsObject()) { + return Nan::ThrowError("Global configuration data must be specified"); + } + + if (!info[1]->IsObject()) { + return Nan::ThrowError("Topic configuration must be specified"); + } + + std::string errstr; + + Conf* gconfig = + Conf::create(RdKafka::Conf::CONF_GLOBAL, + (info[0]->ToObject(Nan::GetCurrentContext())).ToLocalChecked(), errstr); + + if (!gconfig) { + return Nan::ThrowError(errstr.c_str()); + } + + Conf* tconfig = + Conf::create(RdKafka::Conf::CONF_TOPIC, + (info[1]->ToObject(Nan::GetCurrentContext())).ToLocalChecked(), errstr); + + if (!tconfig) { + // No longer need this since we aren't instantiating anything + delete gconfig; + return Nan::ThrowError(errstr.c_str()); + } + + Producer* producer = new Producer(gconfig, tconfig); + + // Wrap it + producer->Wrap(info.This()); + + // Then there is some weird initialization that happens + // basically it sets the configuration data + // we don't need to do that because we lazy load it + + info.GetReturnValue().Set(info.This()); +} + +v8::Local Producer::NewInstance(v8::Local arg) { + Nan::EscapableHandleScope scope; + + const unsigned argc = 1; + + v8::Local argv[argc] = { arg }; + v8::Local cons = Nan::New(constructor); + v8::Local instance = + Nan::NewInstance(cons, argc, argv).ToLocalChecked(); + + return scope.Escape(instance); +} + + +std::string Producer::Name() { + if (!IsConnected()) { + return std::string(""); + } + return std::string(m_client->name()); +} + +Baton Producer::Connect() { + if (IsConnected()) { + return Baton(RdKafka::ERR_NO_ERROR); + } + + std::string errstr; + { + scoped_shared_read_lock lock(m_connection_lock); + m_client = RdKafka::Producer::create(m_gconfig, errstr); + } + + if (!m_client) { + // @todo implement errstr into this somehow + return Baton(RdKafka::ERR__STATE, errstr); + } + + return Baton(RdKafka::ERR_NO_ERROR); +} + +void Producer::ActivateDispatchers() { + m_event_cb.dispatcher.Activate(); // From connection + m_dr_cb.dispatcher.Activate(); +} + +void Producer::DeactivateDispatchers() { + m_event_cb.dispatcher.Deactivate(); // From connection + m_dr_cb.dispatcher.Deactivate(); +} + +void Producer::Disconnect() { + if (IsConnected()) { + scoped_shared_write_lock lock(m_connection_lock); + delete m_client; + m_client = NULL; + } +} + +/** + * [Producer::Produce description] + * @param message - pointer to the message we are sending. This method will + * create a copy of it, so you are still required to free it when done. + * @param size - size of the message. We are copying the memory so we need + * the size + * @param topic - RdKafka::Topic* object to send the message to. Generally + * created by NodeKafka::Topic::toRDKafkaTopic + * @param partition - partition to send it to. Send in + * RdKafka::Topic::PARTITION_UA to send to an unassigned topic + * @param key - a string pointer for the key, or null if there is none. + * @return - A baton object with error code set if it failed. + */ +Baton Producer::Produce(void* message, size_t size, RdKafka::Topic* topic, + int32_t partition, const void *key, size_t key_len, void* opaque) { + RdKafka::ErrorCode response_code; + + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (IsConnected()) { + RdKafka::Producer* producer = dynamic_cast(m_client); + response_code = producer->produce(topic, partition, + RdKafka::Producer::RK_MSG_COPY, + message, size, key, key_len, opaque); + } else { + response_code = RdKafka::ERR__STATE; + } + } else { + response_code = RdKafka::ERR__STATE; + } + + // These topics actually link to the configuration + // they are made from. It's so we can reuse topic configurations + // That means if we delete it here and librd thinks its still linked, + // producing to the same topic will try to reuse it and it will die. + // + // Honestly, we may need to make configuration a first class object + // @todo(Conf needs to be a first class object that is passed around) + // delete topic; + + if (response_code != RdKafka::ERR_NO_ERROR) { + return Baton(response_code); + } + + return Baton(RdKafka::ERR_NO_ERROR); +} + +/** + * [Producer::Produce description] + * @param message - pointer to the message we are sending. This method will + * create a copy of it, so you are still required to free it when done. + * @param size - size of the message. We are copying the memory so we need + * the size + * @param topic - String topic to use so we do not need to create + * an RdKafka::Topic* + * @param partition - partition to send it to. Send in + * RdKafka::Topic::PARTITION_UA to send to an unassigned topic + * @param key - a string pointer for the key, or null if there is none. + * @return - A baton object with error code set if it failed. + */ +Baton Producer::Produce(void* message, size_t size, std::string topic, + int32_t partition, std::string *key, int64_t timestamp, void* opaque, + RdKafka::Headers* headers) { + return Produce(message, size, topic, partition, + key ? key->data() : NULL, key ? key->size() : 0, + timestamp, opaque, headers); +} + +/** + * [Producer::Produce description] + * @param message - pointer to the message we are sending. This method will + * create a copy of it, so you are still required to free it when done. + * @param size - size of the message. We are copying the memory so we need + * the size + * @param topic - String topic to use so we do not need to create + * an RdKafka::Topic* + * @param partition - partition to send it to. Send in + * RdKafka::Topic::PARTITION_UA to send to an unassigned topic + * @param key - a string pointer for the key, or null if there is none. + * @return - A baton object with error code set if it failed. + */ +Baton Producer::Produce(void* message, size_t size, std::string topic, + int32_t partition, const void *key, size_t key_len, + int64_t timestamp, void* opaque, RdKafka::Headers* headers) { + RdKafka::ErrorCode response_code; + + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (IsConnected()) { + RdKafka::Producer* producer = dynamic_cast(m_client); + // This one is a bit different + response_code = producer->produce(topic, partition, + RdKafka::Producer::RK_MSG_COPY, + message, size, + key, key_len, + timestamp, headers, opaque); + } else { + response_code = RdKafka::ERR__STATE; + } + } else { + response_code = RdKafka::ERR__STATE; + } + + // These topics actually link to the configuration + // they are made from. It's so we can reuse topic configurations + // That means if we delete it here and librd thinks its still linked, + // producing to the same topic will try to reuse it and it will die. + // + // Honestly, we may need to make configuration a first class object + // @todo(Conf needs to be a first class object that is passed around) + // delete topic; + + if (response_code != RdKafka::ERR_NO_ERROR) { + return Baton(response_code); + } + + return Baton(RdKafka::ERR_NO_ERROR); +} + +void Producer::Poll() { + m_client->poll(0); +} + +void Producer::ConfigureCallback(const std::string &string_key, const v8::Local &cb, bool add) { + if (string_key.compare("delivery_cb") == 0) { + if (add) { + bool dr_msg_cb = false; + v8::Local dr_msg_cb_key = Nan::New("dr_msg_cb").ToLocalChecked(); + if (Nan::Has(cb, dr_msg_cb_key).FromMaybe(false)) { + v8::Local v = Nan::Get(cb, dr_msg_cb_key).ToLocalChecked(); + if (v->IsBoolean()) { + dr_msg_cb = Nan::To(v).ToChecked(); + } + } + if (dr_msg_cb) { + this->m_dr_cb.SendMessageBuffer(true); + } + this->m_dr_cb.dispatcher.AddCallback(cb); + } else { + this->m_dr_cb.dispatcher.RemoveCallback(cb); + } + } else { + Connection::ConfigureCallback(string_key, cb, add); + } +} + +Baton rdkafkaErrorToBaton(RdKafka::Error* error) { + if ( NULL == error) { + return Baton(RdKafka::ERR_NO_ERROR); + } + else { + Baton result(error->code(), error->str(), error->is_fatal(), + error->is_retriable(), error->txn_requires_abort()); + delete error; + return result; + } +} + +Baton Producer::InitTransactions(int32_t timeout_ms) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::Producer* producer = dynamic_cast(m_client); + RdKafka::Error* error = producer->init_transactions(timeout_ms); + + return rdkafkaErrorToBaton( error); +} + +Baton Producer::BeginTransaction() { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::Producer* producer = dynamic_cast(m_client); + RdKafka::Error* error = producer->begin_transaction(); + + return rdkafkaErrorToBaton( error); +} + +Baton Producer::CommitTransaction(int32_t timeout_ms) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::Producer* producer = dynamic_cast(m_client); + RdKafka::Error* error = producer->commit_transaction(timeout_ms); + + return rdkafkaErrorToBaton( error); +} + +Baton Producer::AbortTransaction(int32_t timeout_ms) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::Producer* producer = dynamic_cast(m_client); + RdKafka::Error* error = producer->abort_transaction(timeout_ms); + + return rdkafkaErrorToBaton( error); +} + +Baton Producer::SendOffsetsToTransaction( + std::vector &offsets, + NodeKafka::KafkaConsumer* consumer, + int timeout_ms) { + if (!IsConnected()) { + return Baton(RdKafka::ERR__STATE); + } + + RdKafka::ConsumerGroupMetadata* group_metadata = dynamic_cast(consumer->m_client)->groupMetadata(); + + RdKafka::Producer* producer = dynamic_cast(m_client); + RdKafka::Error* error = producer->send_offsets_to_transaction(offsets, group_metadata, timeout_ms); + delete group_metadata; + + return rdkafkaErrorToBaton( error); +} + +/* Node exposed methods */ + +/** + * @brief Producer::NodeProduce - produce a message through a producer + * + * This is a synchronous method. You may ask, "why?". The answer is because + * there is no true value doing this asynchronously. All it does is degrade + * performance. This method does not block - all it does is add a message + * to a queue. In the case where the queue is full, it will return an error + * immediately. The only way this method blocks is when you provide it a + * flag to do so, which we never do. + * + * Doing it asynchronously eats up the libuv threadpool for no reason and + * increases execution time by a very small amount. It will take two ticks of + * the event loop to execute at minimum - 1 for executing it and another for + * calling back the callback. + * + * @sa RdKafka::Producer::produce + */ +NAN_METHOD(Producer::NodeProduce) { + Nan::HandleScope scope; + + // Need to extract the message data here. + if (info.Length() < 3) { + // Just throw an exception + return Nan::ThrowError("Need to specify a topic, partition, and message"); + } + + // Second parameter is the partition + int32_t partition; + + if (info[1]->IsNull() || info[1]->IsUndefined()) { + partition = RdKafka::Topic::PARTITION_UA; + } else { + partition = Nan::To(info[1]).FromJust(); + } + + if (partition < 0) { + partition = RdKafka::Topic::PARTITION_UA; + } + + size_t message_buffer_length; + void* message_buffer_data; + + if (info[2]->IsNull()) { + // This is okay for whatever reason + message_buffer_length = 0; + message_buffer_data = NULL; + } else if (!node::Buffer::HasInstance(info[2])) { + return Nan::ThrowError("Message must be a buffer or null"); + } else { + v8::Local message_buffer_object = + (info[2]->ToObject(Nan::GetCurrentContext())).ToLocalChecked(); + + // v8 handles the garbage collection here so we need to make a copy of + // the buffer or assign the buffer to a persistent handle. + + // I'm not sure which would be the more performant option. I assume + // the persistent handle would be but for now we'll try this one + // which should be more memory-efficient and allow v8 to dispose of the + // buffer sooner + + message_buffer_length = node::Buffer::Length(message_buffer_object); + message_buffer_data = node::Buffer::Data(message_buffer_object); + if (message_buffer_data == NULL) { + // empty string message buffer should not end up as null message + v8::Local message_buffer_object_emptystring = Nan::NewBuffer(new char[0], 0).ToLocalChecked(); + message_buffer_length = node::Buffer::Length(message_buffer_object_emptystring); + message_buffer_data = node::Buffer::Data(message_buffer_object_emptystring); + } + } + + size_t key_buffer_length; + const void* key_buffer_data; + std::string * key = NULL; + + if (info[3]->IsNull() || info[3]->IsUndefined()) { + // This is okay for whatever reason + key_buffer_length = 0; + key_buffer_data = NULL; + } else if (node::Buffer::HasInstance(info[3])) { + v8::Local key_buffer_object = + (info[3]->ToObject(Nan::GetCurrentContext())).ToLocalChecked(); + + // v8 handles the garbage collection here so we need to make a copy of + // the buffer or assign the buffer to a persistent handle. + + // I'm not sure which would be the more performant option. I assume + // the persistent handle would be but for now we'll try this one + // which should be more memory-efficient and allow v8 to dispose of the + // buffer sooner + + key_buffer_length = node::Buffer::Length(key_buffer_object); + key_buffer_data = node::Buffer::Data(key_buffer_object); + if (key_buffer_data == NULL) { + // empty string key buffer should not end up as null key + v8::Local key_buffer_object_emptystring = Nan::NewBuffer(new char[0], 0).ToLocalChecked(); + key_buffer_length = node::Buffer::Length(key_buffer_object_emptystring); + key_buffer_data = node::Buffer::Data(key_buffer_object_emptystring); + } + } else { + // If it was a string just use the utf8 value. + v8::Local val = Nan::To(info[3]).ToLocalChecked(); + // Get string pointer for this thing + Nan::Utf8String keyUTF8(val); + key = new std::string(*keyUTF8); + + key_buffer_data = key->data(); + key_buffer_length = key->length(); + } + + int64_t timestamp; + + if (info.Length() > 4 && !info[4]->IsUndefined() && !info[4]->IsNull()) { + if (!info[4]->IsNumber()) { + return Nan::ThrowError("Timestamp must be a number"); + } + + timestamp = Nan::To(info[4]).FromJust(); + } else { + timestamp = 0; + } + + void* opaque = NULL; + // Opaque handling + if (info.Length() > 5 && !info[5]->IsUndefined()) { + // We need to create a persistent handle + opaque = new Nan::Persistent(info[5]); + // To get the local from this later, + // v8::Local object = Nan::New(persistent); + } + + std::vector headers; + if (info.Length() > 6 && !info[6]->IsUndefined()) { + v8::Local v8Headers = v8::Local::Cast(info[6]); + + if (v8Headers->Length() >= 1) { + for (unsigned int i = 0; i < v8Headers->Length(); i++) { + v8::Local header = Nan::Get(v8Headers, i).ToLocalChecked() + ->ToObject(Nan::GetCurrentContext()).ToLocalChecked(); + if (header.IsEmpty()) { + continue; + } + + v8::Local props = header->GetOwnPropertyNames( + Nan::GetCurrentContext()).ToLocalChecked(); + Nan::MaybeLocal v8Key = Nan::To( + Nan::Get(props, 0).ToLocalChecked()); + Nan::MaybeLocal v8Value = Nan::To( + Nan::Get(header, v8Key.ToLocalChecked()).ToLocalChecked()); + + Nan::Utf8String uKey(v8Key.ToLocalChecked()); + std::string key(*uKey); + + Nan::Utf8String uValue(v8Value.ToLocalChecked()); + std::string value(*uValue); + headers.push_back( + RdKafka::Headers::Header(key, value.c_str(), value.size())); + } + } + } + + Producer* producer = ObjectWrap::Unwrap(info.This()); + + // Let the JS library throw if we need to so the error can be more rich + int error_code; + + if (info[0]->IsString()) { + // Get string pointer for this thing + Nan::Utf8String topicUTF8(Nan::To(info[0]).ToLocalChecked()); + std::string topic_name(*topicUTF8); + RdKafka::Headers *rd_headers = RdKafka::Headers::create(headers); + + Baton b = producer->Produce(message_buffer_data, message_buffer_length, + topic_name, partition, key_buffer_data, key_buffer_length, + timestamp, opaque, rd_headers); + + error_code = static_cast(b.err()); + if (error_code != 0 && rd_headers) { + delete rd_headers; + } + } else { + // First parameter is a topic OBJECT + Topic* topic = ObjectWrap::Unwrap(info[0].As()); + + // Unwrap it and turn it into an RdKafka::Topic* + Baton topic_baton = topic->toRDKafkaTopic(producer); + + if (topic_baton.err() != RdKafka::ERR_NO_ERROR) { + // Let the JS library throw if we need to so the error can be more rich + error_code = static_cast(topic_baton.err()); + + return info.GetReturnValue().Set(Nan::New(error_code)); + } + + RdKafka::Topic* rd_topic = topic_baton.data(); + + Baton b = producer->Produce(message_buffer_data, message_buffer_length, + rd_topic, partition, key_buffer_data, key_buffer_length, opaque); + + // Delete the topic when we are done. + delete rd_topic; + + error_code = static_cast(b.err()); + } + + if (key != NULL) { + delete key; + } + + info.GetReturnValue().Set(Nan::New(error_code)); +} + +NAN_METHOD(Producer::NodeSetPartitioner) { + Nan::HandleScope scope; + + if (info.Length() < 1 || !info[0]->IsFunction()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback"); + } + + Producer* producer = ObjectWrap::Unwrap(info.This()); + v8::Local cb = info[0].As(); + producer->m_partitioner_cb.SetCallback(cb); + info.GetReturnValue().Set(Nan::True()); +} + +NAN_METHOD(Producer::NodeConnect) { + Nan::HandleScope scope; + + if (info.Length() < 1 || !info[0]->IsFunction()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback"); + } + + // This needs to be offloaded to libuv + v8::Local cb = info[0].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Producer* producer = ObjectWrap::Unwrap(info.This()); + Nan::AsyncQueueWorker(new Workers::ProducerConnect(callback, producer)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Producer::NodePoll) { + Nan::HandleScope scope; + + Producer* producer = ObjectWrap::Unwrap(info.This()); + + if (!producer->IsConnected()) { + Nan::ThrowError("Producer is disconnected"); + } else { + producer->Poll(); + info.GetReturnValue().Set(Nan::True()); + } +} + +Baton Producer::Flush(int timeout_ms) { + RdKafka::ErrorCode response_code; + if (IsConnected()) { + scoped_shared_read_lock lock(m_connection_lock); + if (IsConnected()) { + RdKafka::Producer* producer = dynamic_cast(m_client); + response_code = producer->flush(timeout_ms); + } else { + response_code = RdKafka::ERR__STATE; + } + } else { + response_code = RdKafka::ERR__STATE; + } + + return Baton(response_code); +} + +NAN_METHOD(Producer::NodeFlush) { + Nan::HandleScope scope; + + if (info.Length() < 2 || !info[1]->IsFunction() || !info[0]->IsNumber()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a timeout and a callback"); + } + + int timeout_ms = Nan::To(info[0]).FromJust(); + + v8::Local cb = info[1].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Producer* producer = ObjectWrap::Unwrap(info.This()); + + Nan::AsyncQueueWorker( + new Workers::ProducerFlush(callback, producer, timeout_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Producer::NodeDisconnect) { + Nan::HandleScope scope; + + if (info.Length() < 1 || !info[0]->IsFunction()) { + // Just throw an exception + return Nan::ThrowError("Need to specify a callback"); + } + + + v8::Local cb = info[0].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Producer* producer = ObjectWrap::Unwrap(info.This()); + Nan::AsyncQueueWorker(new Workers::ProducerDisconnect(callback, producer)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Producer::NodeInitTransactions) { + Nan::HandleScope scope; + + if (info.Length() < 2 || !info[1]->IsFunction() || !info[0]->IsNumber()) { + return Nan::ThrowError("Need to specify a timeout and a callback"); + } + + int timeout_ms = Nan::To(info[0]).FromJust(); + + v8::Local cb = info[1].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Producer* producer = ObjectWrap::Unwrap(info.This()); + Nan::AsyncQueueWorker(new Workers::ProducerInitTransactions(callback, producer, timeout_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Producer::NodeBeginTransaction) { + Nan::HandleScope scope; + + if (info.Length() < 1 || !info[0]->IsFunction()) { + return Nan::ThrowError("Need to specify a callback"); + } + + v8::Local cb = info[0].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Producer* producer = ObjectWrap::Unwrap(info.This()); + Nan::AsyncQueueWorker(new Workers::ProducerBeginTransaction(callback, producer)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Producer::NodeCommitTransaction) { + Nan::HandleScope scope; + + if (info.Length() < 2 || !info[1]->IsFunction() || !info[0]->IsNumber()) { + return Nan::ThrowError("Need to specify a timeout and a callback"); + } + + int timeout_ms = Nan::To(info[0]).FromJust(); + + v8::Local cb = info[1].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Producer* producer = ObjectWrap::Unwrap(info.This()); + Nan::AsyncQueueWorker(new Workers::ProducerCommitTransaction(callback, producer, timeout_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Producer::NodeAbortTransaction) { + Nan::HandleScope scope; + + if (info.Length() < 2 || !info[1]->IsFunction() || !info[0]->IsNumber()) { + return Nan::ThrowError("Need to specify a timeout and a callback"); + } + + int timeout_ms = Nan::To(info[0]).FromJust(); + + v8::Local cb = info[1].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Producer* producer = ObjectWrap::Unwrap(info.This()); + Nan::AsyncQueueWorker(new Workers::ProducerAbortTransaction(callback, producer, timeout_ms)); + + info.GetReturnValue().Set(Nan::Null()); +} + +NAN_METHOD(Producer::NodeSendOffsetsToTransaction) { + Nan::HandleScope scope; + + if (info.Length() < 4) { + return Nan::ThrowError("Need to specify offsets, consumer, timeout for 'send offsets to transaction', and callback"); + } + if (!info[0]->IsArray()) { + return Nan::ThrowError("First argument to 'send offsets to transaction' has to be a consumer object"); + } + if (!info[1]->IsObject()) { + Nan::ThrowError("Kafka consumer must be provided"); + } + if (!info[2]->IsNumber()) { + Nan::ThrowError("Timeout must be provided"); + } + if (!info[3]->IsFunction()) { + return Nan::ThrowError("Need to specify a callback"); + } + + std::vector toppars = + Conversion::TopicPartition::FromV8Array(info[0].As()); + NodeKafka::KafkaConsumer* consumer = + ObjectWrap::Unwrap(info[1].As()); + int timeout_ms = Nan::To(info[2]).FromJust(); + v8::Local cb = info[3].As(); + Nan::Callback *callback = new Nan::Callback(cb); + + Producer* producer = ObjectWrap::Unwrap(info.This()); + Nan::AsyncQueueWorker(new Workers::ProducerSendOffsetsToTransaction( + callback, + producer, + toppars, + consumer, + timeout_ms + )); + + info.GetReturnValue().Set(Nan::Null()); +} + +} // namespace NodeKafka diff --git a/src/producer.h b/src/producer.h new file mode 100644 index 0000000..5674aa4 --- /dev/null +++ b/src/producer.h @@ -0,0 +1,123 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_PRODUCER_H_ +#define SRC_PRODUCER_H_ + +#include +#include +#include +#include + +#include "rdkafkacpp.h" + +#include "src/common.h" +#include "src/connection.h" +#include "src/callbacks.h" +#include "src/topic.h" + +namespace NodeKafka { + +class ProducerMessage { + public: + explicit ProducerMessage(v8::Local, NodeKafka::Topic*); + ~ProducerMessage(); + + void* Payload(); + size_t Size(); + bool IsEmpty(); + RdKafka::Topic * GetTopic(); + + std::string m_errstr; + + Topic * m_topic; + int32_t m_partition; + std::string m_key; + + void* m_buffer_data; + size_t m_buffer_length; + + bool m_is_empty; +}; + +class Producer : public Connection { + public: + static void Init(v8::Local); + static v8::Local NewInstance(v8::Local); + + Baton Connect(); + void Disconnect(); + void Poll(); + #if RD_KAFKA_VERSION > 0x00090200 + Baton Flush(int timeout_ms); + #endif + + Baton Produce(void* message, size_t message_size, + RdKafka::Topic* topic, int32_t partition, + const void* key, size_t key_len, + void* opaque); + + Baton Produce(void* message, size_t message_size, + std::string topic, int32_t partition, + std::string* key, + int64_t timestamp, void* opaque, + RdKafka::Headers* headers); + + Baton Produce(void* message, size_t message_size, + std::string topic, int32_t partition, + const void* key, size_t key_len, + int64_t timestamp, void* opaque, + RdKafka::Headers* headers); + + std::string Name(); + + void ActivateDispatchers(); + void DeactivateDispatchers(); + + void ConfigureCallback(const std::string &string_key, const v8::Local &cb, bool add) override; + + Baton InitTransactions(int32_t timeout_ms); + Baton BeginTransaction(); + Baton CommitTransaction(int32_t timeout_ms); + Baton AbortTransaction(int32_t timeout_ms); + Baton SendOffsetsToTransaction( + std::vector &offsets, + NodeKafka::KafkaConsumer* consumer, + int timeout_ms + ); + + protected: + static Nan::Persistent constructor; + static void New(const Nan::FunctionCallbackInfo&); + + Producer(Conf*, Conf*); + ~Producer(); + + private: + static NAN_METHOD(NodeProduce); + static NAN_METHOD(NodeSetPartitioner); + static NAN_METHOD(NodeConnect); + static NAN_METHOD(NodeDisconnect); + static NAN_METHOD(NodePoll); + #if RD_KAFKA_VERSION > 0x00090200 + static NAN_METHOD(NodeFlush); + #endif + static NAN_METHOD(NodeInitTransactions); + static NAN_METHOD(NodeBeginTransaction); + static NAN_METHOD(NodeCommitTransaction); + static NAN_METHOD(NodeAbortTransaction); + static NAN_METHOD(NodeSendOffsetsToTransaction); + + Callbacks::Delivery m_dr_cb; + Callbacks::Partitioner m_partitioner_cb; +}; + +} // namespace NodeKafka + +#endif // SRC_PRODUCER_H_ diff --git a/src/topic.cc b/src/topic.cc new file mode 100644 index 0000000..4c9d663 --- /dev/null +++ b/src/topic.cc @@ -0,0 +1,173 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include + +#include "src/common.h" +#include "src/connection.h" +#include "src/topic.h" + +namespace NodeKafka { + +/** + * @brief Producer v8 wrapped object. + * + * Wraps the RdKafka::Producer object with compositional inheritence and + * provides methods for interacting with it exposed to node. + * + * The base wrappable RdKafka::Handle deals with most of the wrapping but + * we still need to declare its prototype. + * + * @sa RdKafka::Producer + * @sa NodeKafka::Connection + */ + +Topic::Topic(std::string topic_name, RdKafka::Conf* config): + m_topic_name(topic_name), + m_config(config) { + // We probably want to copy the config. May require refactoring if we do not +} + +Topic::~Topic() { + if (m_config) { + delete m_config; + } +} + +std::string Topic::name() { + return m_topic_name; +} + +Baton Topic::toRDKafkaTopic(Connection* handle) { + if (m_config) { + return handle->CreateTopic(m_topic_name, m_config); + } else { + return handle->CreateTopic(m_topic_name); + } +} + +/* + +bool partition_available(int32_t partition) { + return topic_->partition_available(partition); +} + +Baton offset_store (int32_t partition, int64_t offset) { + RdKafka::ErrorCode err = topic_->offset_store(partition, offset); + + switch (err) { + case RdKafka::ERR_NO_ERROR: + + break; + default: + + break; + } +} + +*/ + +Nan::Persistent Topic::constructor; + +void Topic::Init(v8::Local exports) { + Nan::HandleScope scope; + + v8::Local tpl = Nan::New(New); + tpl->SetClassName(Nan::New("Topic").ToLocalChecked()); + tpl->InstanceTemplate()->SetInternalFieldCount(1); + + Nan::SetPrototypeMethod(tpl, "name", NodeGetName); + + // connect. disconnect. resume. pause. get meta data + constructor.Reset((tpl->GetFunction(Nan::GetCurrentContext())) + .ToLocalChecked()); + + Nan::Set(exports, Nan::New("Topic").ToLocalChecked(), + tpl->GetFunction(Nan::GetCurrentContext()).ToLocalChecked()); +} + +void Topic::New(const Nan::FunctionCallbackInfo& info) { + if (!info.IsConstructCall()) { + return Nan::ThrowError("non-constructor invocation not supported"); + } + + if (info.Length() < 1) { + return Nan::ThrowError("topic name is required"); + } + + if (!info[0]->IsString()) { + return Nan::ThrowError("Topic name must be a string"); + } + + RdKafka::Conf* config = NULL; + + if (info.Length() >= 2 && !info[1]->IsUndefined() && !info[1]->IsNull()) { + // If they gave us two parameters, or the 3rd parameter is null or + // undefined, we want to pass null in for the config + + std::string errstr; + if (!info[1]->IsObject()) { + return Nan::ThrowError("Configuration data must be specified"); + } + + config = Conf::create(RdKafka::Conf::CONF_TOPIC, (info[1]->ToObject(Nan::GetCurrentContext())).ToLocalChecked(), errstr); // NOLINT + + if (!config) { + return Nan::ThrowError(errstr.c_str()); + } + } + + Nan::Utf8String parameterValue(Nan::To(info[0]).ToLocalChecked()); + std::string topic_name(*parameterValue); + + Topic* topic = new Topic(topic_name, config); + + // Wrap it + topic->Wrap(info.This()); + + // Then there is some weird initialization that happens + // basically it sets the configuration data + // we don't need to do that because we lazy load it + + info.GetReturnValue().Set(info.This()); +} + +// handle + +v8::Local Topic::NewInstance(v8::Local arg) { + Nan::EscapableHandleScope scope; + + const unsigned argc = 1; + + v8::Local argv[argc] = { arg }; + v8::Local cons = Nan::New(constructor); + v8::Local instance = + Nan::NewInstance(cons, argc, argv).ToLocalChecked(); + + return scope.Escape(instance); +} + +NAN_METHOD(Topic::NodeGetName) { + Nan::HandleScope scope; + + Topic* topic = ObjectWrap::Unwrap(info.This()); + + info.GetReturnValue().Set(Nan::New(topic->name()).ToLocalChecked()); +} + +NAN_METHOD(Topic::NodePartitionAvailable) { + // @TODO(sparente) +} + +NAN_METHOD(Topic::NodeOffsetStore) { + // @TODO(sparente) +} + +} // namespace NodeKafka diff --git a/src/topic.h b/src/topic.h new file mode 100644 index 0000000..425bf7f --- /dev/null +++ b/src/topic.h @@ -0,0 +1,54 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_TOPIC_H_ +#define SRC_TOPIC_H_ + +#include +#include + +#include "rdkafkacpp.h" + +#include "src/config.h" + +namespace NodeKafka { + +class Topic : public Nan::ObjectWrap { + public: + static void Init(v8::Local); + static v8::Local NewInstance(v8::Local arg); + + Baton toRDKafkaTopic(Connection *handle); + + protected: + static Nan::Persistent constructor; + static void New(const Nan::FunctionCallbackInfo& info); + + static NAN_METHOD(NodeGetMetadata); + + // TopicConfig * config_; + + std::string errstr; + std::string name(); + + private: + Topic(std::string, RdKafka::Conf *); + ~Topic(); + + std::string m_topic_name; + RdKafka::Conf * m_config; + + static NAN_METHOD(NodeGetName); + static NAN_METHOD(NodePartitionAvailable); + static NAN_METHOD(NodeOffsetStore); +}; + +} // namespace NodeKafka + +#endif // SRC_TOPIC_H_ diff --git a/src/workers.cc b/src/workers.cc new file mode 100644 index 0000000..8930636 --- /dev/null +++ b/src/workers.cc @@ -0,0 +1,1238 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#include +#include + +#include "src/workers.h" + +#ifndef _WIN32 +#include +#else +// Windows specific +#include +#endif + +using NodeKafka::Producer; +using NodeKafka::Connection; + +namespace NodeKafka { +namespace Workers { + +namespace Handle { +/** + * @brief Handle: get offsets for times. + * + * This callback will take a topic partition list with timestamps and + * for each topic partition, will fill in the offsets. It is done async + * because it has a timeout and I don't want node to block + * + * @see RdKafka::KafkaConsumer::Committed + */ + +OffsetsForTimes::OffsetsForTimes(Nan::Callback *callback, + Connection* handle, + std::vector & t, + const int & timeout_ms) : + ErrorAwareWorker(callback), + m_handle(handle), + m_topic_partitions(t), + m_timeout_ms(timeout_ms) {} + +OffsetsForTimes::~OffsetsForTimes() { + // Delete the underlying topic partitions as they are ephemeral or cloned + RdKafka::TopicPartition::destroy(m_topic_partitions); +} + +void OffsetsForTimes::Execute() { + Baton b = m_handle->OffsetsForTimes(m_topic_partitions, m_timeout_ms); + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void OffsetsForTimes::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + v8::Local argv[argc]; + + argv[0] = Nan::Null(); + argv[1] = Conversion::TopicPartition::ToV8Array(m_topic_partitions); + + callback->Call(argc, argv); +} + +void OffsetsForTimes::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} +} // namespace Handle + +ConnectionMetadata::ConnectionMetadata( + Nan::Callback *callback, Connection* connection, + std::string topic, int timeout_ms, bool all_topics) : + ErrorAwareWorker(callback), + m_connection(connection), + m_topic(topic), + m_timeout_ms(timeout_ms), + m_all_topics(all_topics), + m_metadata(NULL) {} + +ConnectionMetadata::~ConnectionMetadata() {} + +void ConnectionMetadata::Execute() { + Baton b = m_connection->GetMetadata(m_all_topics, m_topic, m_timeout_ms); + + if (b.err() == RdKafka::ERR_NO_ERROR) { + // No good way to do this except some stupid string delimiting. + // maybe we'll delimit it by a | or something and just split + // the string to create the object + m_metadata = b.data(); + } else { + SetErrorBaton(b); + } +} + +void ConnectionMetadata::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + + // This is a big one! + v8::Local argv[argc] = { Nan::Null(), + Conversion::Metadata::ToV8Object(m_metadata)}; + + callback->Call(argc, argv); + + delete m_metadata; +} + +void ConnectionMetadata::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Client query watermark offsets worker + * + * Easy Nan::AsyncWorker for getting watermark offsets from a broker + * + * @sa RdKafka::Handle::query_watermark_offsets + * @sa NodeKafka::Connection::QueryWatermarkOffsets + */ + +ConnectionQueryWatermarkOffsets::ConnectionQueryWatermarkOffsets( + Nan::Callback *callback, Connection* connection, + std::string topic, int32_t partition, int timeout_ms) : + ErrorAwareWorker(callback), + m_connection(connection), + m_topic(topic), + m_partition(partition), + m_timeout_ms(timeout_ms) {} + +ConnectionQueryWatermarkOffsets::~ConnectionQueryWatermarkOffsets() {} + +void ConnectionQueryWatermarkOffsets::Execute() { + Baton b = m_connection->QueryWatermarkOffsets( + m_topic, m_partition, &m_low_offset, &m_high_offset, m_timeout_ms); + + // If we got any error here we need to bail out + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void ConnectionQueryWatermarkOffsets::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + + v8::Local offsetsObj = Nan::New(); + Nan::Set(offsetsObj, Nan::New("lowOffset").ToLocalChecked(), + Nan::New(m_low_offset)); + Nan::Set(offsetsObj, Nan::New("highOffset").ToLocalChecked(), + Nan::New(m_high_offset)); + + // This is a big one! + v8::Local argv[argc] = { Nan::Null(), offsetsObj}; + + callback->Call(argc, argv); +} + +void ConnectionQueryWatermarkOffsets::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Producer connect worker. + * + * Easy Nan::AsyncWorker for setting up client connections + * + * @sa RdKafka::Producer::connect + * @sa NodeKafka::Producer::Connect + */ + +ProducerConnect::ProducerConnect(Nan::Callback *callback, Producer* producer): + ErrorAwareWorker(callback), + producer(producer) {} + +ProducerConnect::~ProducerConnect() {} + +void ProducerConnect::Execute() { + Baton b = producer->Connect(); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void ProducerConnect::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + + v8::Local obj = Nan::New(); + Nan::Set(obj, Nan::New("name").ToLocalChecked(), + Nan::New(producer->Name()).ToLocalChecked()); + + v8::Local argv[argc] = { Nan::Null(), obj}; + + // Activate the dispatchers + producer->ActivateDispatchers(); + + callback->Call(argc, argv); +} + +void ProducerConnect::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Producer disconnect worker + * + * Easy Nan::AsyncWorker for disconnecting from clients + */ + +ProducerDisconnect::ProducerDisconnect(Nan::Callback *callback, + Producer* producer): + ErrorAwareWorker(callback), + producer(producer) {} + +ProducerDisconnect::~ProducerDisconnect() {} + +void ProducerDisconnect::Execute() { + producer->Disconnect(); +} + +void ProducerDisconnect::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + v8::Local argv[argc] = { Nan::Null(), Nan::True()}; + + // Deactivate the dispatchers + producer->DeactivateDispatchers(); + + callback->Call(argc, argv); +} + +void ProducerDisconnect::HandleErrorCallback() { + // This should never run + assert(0); +} + +/** + * @brief Producer flush worker + * + * Easy Nan::AsyncWorker for flushing a producer. + */ + +ProducerFlush::ProducerFlush(Nan::Callback *callback, + Producer* producer, int timeout_ms): + ErrorAwareWorker(callback), + producer(producer), + timeout_ms(timeout_ms) {} + +ProducerFlush::~ProducerFlush() {} + +void ProducerFlush::Execute() { + if (!producer->IsConnected()) { + SetErrorMessage("Producer is disconnected"); + return; + } + + Baton b = producer->Flush(timeout_ms); + if (b.err() != RdKafka::ErrorCode::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void ProducerFlush::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { Nan::Null() }; + + callback->Call(argc, argv); +} + +/** + * @brief Producer init transactions worker. + * + * Easy Nan::AsyncWorker for initiating transactions + * + * @sa RdKafka::Producer::init_transactions + * @sa NodeKafka::Producer::InitTransactions + */ + +ProducerInitTransactions::ProducerInitTransactions(Nan::Callback *callback, + Producer* producer, const int & timeout_ms): + ErrorAwareWorker(callback), + producer(producer), + m_timeout_ms(timeout_ms) {} + +ProducerInitTransactions::~ProducerInitTransactions() {} + +void ProducerInitTransactions::Execute() { + Baton b = producer->InitTransactions(m_timeout_ms); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void ProducerInitTransactions::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { Nan::Null() }; + + // Activate the dispatchers + producer->ActivateDispatchers(); + + callback->Call(argc, argv); +} + +void ProducerInitTransactions::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { m_baton.ToTxnObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Producer begin transaction worker. + * + * Easy Nan::AsyncWorker for begin transaction + * + * @sa RdKafka::Producer::begin_transaction + * @sa NodeKafka::Producer::BeginTransaction + */ + +ProducerBeginTransaction::ProducerBeginTransaction(Nan::Callback *callback, Producer* producer): + ErrorAwareWorker(callback), + producer(producer) {} + +ProducerBeginTransaction::~ProducerBeginTransaction() {} + +void ProducerBeginTransaction::Execute() { + Baton b = producer->BeginTransaction(); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void ProducerBeginTransaction::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + + v8::Local argv[argc] = { Nan::Null() }; + + // Activate the dispatchers + producer->ActivateDispatchers(); + + callback->Call(argc, argv); +} + +void ProducerBeginTransaction::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Producer commit transaction worker. + * + * Easy Nan::AsyncWorker for committing transaction + * + * @sa RdKafka::Producer::commit_transaction + * @sa NodeKafka::Producer::CommitTransaction + */ + +ProducerCommitTransaction::ProducerCommitTransaction(Nan::Callback *callback, + Producer* producer, const int & timeout_ms): + ErrorAwareWorker(callback), + producer(producer), + m_timeout_ms(timeout_ms) {} + +ProducerCommitTransaction::~ProducerCommitTransaction() {} + +void ProducerCommitTransaction::Execute() { + Baton b = producer->CommitTransaction(m_timeout_ms); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void ProducerCommitTransaction::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { Nan::Null() }; + + // Activate the dispatchers + producer->ActivateDispatchers(); + + callback->Call(argc, argv); +} + +void ProducerCommitTransaction::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { m_baton.ToTxnObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Producer abort transaction worker. + * + * Easy Nan::AsyncWorker for aborting transaction + * + * @sa RdKafka::Producer::abort_transaction + * @sa NodeKafka::Producer::AbortTransaction + */ + +ProducerAbortTransaction::ProducerAbortTransaction(Nan::Callback *callback, + Producer* producer, const int & timeout_ms): + ErrorAwareWorker(callback), + producer(producer), + m_timeout_ms(timeout_ms) {} + +ProducerAbortTransaction::~ProducerAbortTransaction() {} + +void ProducerAbortTransaction::Execute() { + Baton b = producer->AbortTransaction(m_timeout_ms); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void ProducerAbortTransaction::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { Nan::Null() }; + + // Activate the dispatchers + producer->ActivateDispatchers(); + + callback->Call(argc, argv); +} + +void ProducerAbortTransaction::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { m_baton.ToTxnObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Producer SendOffsetsToTransaction worker. + * + * Easy Nan::AsyncWorker for SendOffsetsToTransaction + * + * @sa RdKafka::Producer::send_offsets_to_transaction + * @sa NodeKafka::Producer::SendOffsetsToTransaction + */ + +ProducerSendOffsetsToTransaction::ProducerSendOffsetsToTransaction( + Nan::Callback *callback, + Producer* producer, + std::vector & t, + KafkaConsumer* consumer, + const int & timeout_ms): + ErrorAwareWorker(callback), + producer(producer), + m_topic_partitions(t), + consumer(consumer), + m_timeout_ms(timeout_ms) {} + +ProducerSendOffsetsToTransaction::~ProducerSendOffsetsToTransaction() {} + +void ProducerSendOffsetsToTransaction::Execute() { + Baton b = producer->SendOffsetsToTransaction( + m_topic_partitions, + consumer, + m_timeout_ms + ); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void ProducerSendOffsetsToTransaction::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { Nan::Null() }; + + // Activate the dispatchers + producer->ActivateDispatchers(); + + callback->Call(argc, argv); +} + +void ProducerSendOffsetsToTransaction::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { m_baton.ToTxnObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief KafkaConsumer connect worker. + * + * Easy Nan::AsyncWorker for setting up client connections + * + * @sa RdKafka::KafkaConsumer::connect + * @sa NodeKafka::KafkaConsumer::Connect + */ + +KafkaConsumerConnect::KafkaConsumerConnect(Nan::Callback *callback, + KafkaConsumer* consumer): + ErrorAwareWorker(callback), + consumer(consumer) {} + +KafkaConsumerConnect::~KafkaConsumerConnect() {} + +void KafkaConsumerConnect::Execute() { + Baton b = consumer->Connect(); + // consumer->Wait(); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void KafkaConsumerConnect::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + + // Create the object + v8::Local obj = Nan::New(); + Nan::Set(obj, Nan::New("name").ToLocalChecked(), + Nan::New(consumer->Name()).ToLocalChecked()); + + v8::Local argv[argc] = { Nan::Null(), obj }; + consumer->ActivateDispatchers(); + + callback->Call(argc, argv); +} + +void KafkaConsumerConnect::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { Nan::Error(ErrorMessage()) }; + + callback->Call(argc, argv); +} + +/** + * @brief KafkaConsumer disconnect worker. + * + * Easy Nan::AsyncWorker for disconnecting and cleaning up librdkafka artifacts + * + * @sa RdKafka::KafkaConsumer::disconnect + * @sa NodeKafka::KafkaConsumer::Disconnect + */ + +KafkaConsumerDisconnect::KafkaConsumerDisconnect(Nan::Callback *callback, + KafkaConsumer* consumer): + ErrorAwareWorker(callback), + consumer(consumer) {} + +KafkaConsumerDisconnect::~KafkaConsumerDisconnect() {} + +void KafkaConsumerDisconnect::Execute() { + Baton b = consumer->Disconnect(); + + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void KafkaConsumerDisconnect::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + v8::Local argv[argc] = { Nan::Null(), Nan::True() }; + + consumer->DeactivateDispatchers(); + + callback->Call(argc, argv); +} + +void KafkaConsumerDisconnect::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + consumer->DeactivateDispatchers(); + + callback->Call(argc, argv); +} + +/** + * @brief KafkaConsumer get messages worker. + * + * A more complex Nan::AsyncProgressWorker. I made a custom superclass to deal + * with more real-time progress points. Instead of using ProgressWorker, which + * is not time sensitive, this custom worker will poll using libuv and send + * data back to v8 as it comes available without missing any + * + * The actual event runs through a continuous while loop. It stops when the + * consumer is flagged as disconnected or as unsubscribed. + * + * @todo thread-safe isConnected checking + * @note Chances are, when the connection is broken with the way librdkafka works, + * we are shutting down. But we want it to shut down properly so we probably + * need the consumer to have a thread lock that can be used when + * we are dealing with manipulating the `client` + * + * @sa RdKafka::KafkaConsumer::Consume + * @sa NodeKafka::KafkaConsumer::GetMessage + */ + +KafkaConsumerConsumeLoop::KafkaConsumerConsumeLoop(Nan::Callback *callback, + KafkaConsumer* consumer, + const int & timeout_ms, + const int & timeout_sleep_delay_ms) : + MessageWorker(callback), + consumer(consumer), + m_timeout_ms(timeout_ms), + m_timeout_sleep_delay_ms(timeout_sleep_delay_ms), + m_rand_seed(time(NULL)) {} + +KafkaConsumerConsumeLoop::~KafkaConsumerConsumeLoop() {} + +void KafkaConsumerConsumeLoop::Execute(const ExecutionMessageBus& bus) { + // Do one check here before we move forward + bool looping = true; + while (consumer->IsConnected() && looping) { + Baton b = consumer->Consume(m_timeout_ms); + RdKafka::ErrorCode ec = b.err(); + if (ec == RdKafka::ERR_NO_ERROR) { + RdKafka::Message *message = b.data(); + switch (message->err()) { + case RdKafka::ERR__PARTITION_EOF: + bus.Send(message); + // EOF means there are no more messages to read. + // We should wait a little bit for more messages to come in + // when in consume loop mode + // Randomise the wait time to avoid contention on different + // slow topics + #ifndef _WIN32 + usleep(static_cast(rand_r(&m_rand_seed) * 1000 * 1000 / RAND_MAX)); + #else + _sleep(1000); + #endif + break; + case RdKafka::ERR__TIMED_OUT: + case RdKafka::ERR__TIMED_OUT_QUEUE: + delete message; + // If it is timed out this could just mean there were no + // new messages fetched quickly enough. This isn't really + // an error that should kill us. + #ifndef _WIN32 + usleep(m_timeout_sleep_delay_ms*1000); + #else + _sleep(m_timeout_sleep_delay_ms); + #endif + break; + case RdKafka::ERR_NO_ERROR: + bus.Send(message); + break; + default: + // Unknown error. We need to break out of this + SetErrorBaton(b); + looping = false; + break; + } + } else if (ec == RdKafka::ERR_UNKNOWN_TOPIC_OR_PART || ec == RdKafka::ERR_TOPIC_AUTHORIZATION_FAILED) { + bus.SendWarning(ec); + } else { + // Unknown error. We need to break out of this + SetErrorBaton(b); + looping = false; + } + } +} + +void KafkaConsumerConsumeLoop::HandleMessageCallback(RdKafka::Message* msg, RdKafka::ErrorCode ec) { + Nan::HandleScope scope; + + const unsigned int argc = 4; + v8::Local argv[argc]; + + argv[0] = Nan::Null(); + if (msg == NULL) { + argv[1] = Nan::Null(); + argv[2] = Nan::Null(); + argv[3] = Nan::New(ec); + } else { + argv[3] = Nan::Null(); + switch (msg->err()) { + case RdKafka::ERR__PARTITION_EOF: { + argv[1] = Nan::Null(); + v8::Local eofEvent = Nan::New(); + + Nan::Set(eofEvent, Nan::New("topic").ToLocalChecked(), + Nan::New(msg->topic_name()).ToLocalChecked()); + Nan::Set(eofEvent, Nan::New("offset").ToLocalChecked(), + Nan::New(msg->offset())); + Nan::Set(eofEvent, Nan::New("partition").ToLocalChecked(), + Nan::New(msg->partition())); + + argv[2] = eofEvent; + break; + } + default: + argv[1] = Conversion::Message::ToV8Object(msg); + argv[2] = Nan::Null(); + break; + } + + // We can delete msg now + delete msg; + } + + callback->Call(argc, argv); +} + +void KafkaConsumerConsumeLoop::HandleOKCallback() { + Nan::HandleScope scope; +} + +void KafkaConsumerConsumeLoop::HandleErrorCallback() { + Nan::HandleScope scope; + + + const unsigned int argc = 1; + v8::Local argv[argc] = { Nan::Error(ErrorMessage()) }; + + callback->Call(argc, argv); +} + +/** + * @brief KafkaConsumer get messages worker. + * + * This callback will get a number of messages. Can be of use in streams or + * places where you don't want an infinite loop managed in C++land and would + * rather manage it in Node. + * + * @see RdKafka::KafkaConsumer::Consume + * @see NodeKafka::KafkaConsumer::GetMessage + */ + +KafkaConsumerConsumeNum::KafkaConsumerConsumeNum(Nan::Callback *callback, + KafkaConsumer* consumer, + const uint32_t & num_messages, + const int & timeout_ms) : + ErrorAwareWorker(callback), + m_consumer(consumer), + m_num_messages(num_messages), + m_timeout_ms(timeout_ms) {} + +KafkaConsumerConsumeNum::~KafkaConsumerConsumeNum() {} + +void KafkaConsumerConsumeNum::Execute() { + std::size_t max = static_cast(m_num_messages); + bool looping = true; + int timeout_ms = m_timeout_ms; + std::size_t eof_event_count = 0; + + while (m_messages.size() - eof_event_count < max && looping) { + // Get a message + Baton b = m_consumer->Consume(timeout_ms); + if (b.err() == RdKafka::ERR_NO_ERROR) { + RdKafka::Message *message = b.data(); + RdKafka::ErrorCode errorCode = message->err(); + switch (errorCode) { + case RdKafka::ERR__PARTITION_EOF: + // If partition EOF and have consumed messages, retry with timeout 1 + // This allows getting ready messages, while not waiting for new ones + if (m_messages.size() > eof_event_count) { + timeout_ms = 1; + } + + // We will only go into this code path when `enable.partition.eof` is set to true + // In this case, consumer is also interested in EOF messages, so we return an EOF message + m_messages.push_back(message); + eof_event_count += 1; + break; + case RdKafka::ERR__TIMED_OUT: + case RdKafka::ERR__TIMED_OUT_QUEUE: + // Break of the loop if we timed out + delete message; + looping = false; + break; + case RdKafka::ERR_NO_ERROR: + m_messages.push_back(b.data()); + break; + default: + // Set the error for any other errors and break + delete message; + if (m_messages.size() == 0) { + SetErrorBaton(Baton(errorCode)); + } + looping = false; + break; + } + } else { + if (m_messages.size() == 0) { + SetErrorBaton(b); + } + looping = false; + } + } +} + +void KafkaConsumerConsumeNum::HandleOKCallback() { + Nan::HandleScope scope; + const unsigned int argc = 3; + v8::Local argv[argc]; + argv[0] = Nan::Null(); + + v8::Local returnArray = Nan::New(); + v8::Local eofEventsArray = Nan::New(); + + if (m_messages.size() > 0) { + int returnArrayIndex = -1; + int eofEventsArrayIndex = -1; + for (std::vector::iterator it = m_messages.begin(); + it != m_messages.end(); ++it) { + RdKafka::Message* message = *it; + + switch (message->err()) { + case RdKafka::ERR_NO_ERROR: + ++returnArrayIndex; + Nan::Set(returnArray, returnArrayIndex, Conversion::Message::ToV8Object(message)); + break; + case RdKafka::ERR__PARTITION_EOF: + ++eofEventsArrayIndex; + + // create EOF event + v8::Local eofEvent = Nan::New(); + + Nan::Set(eofEvent, Nan::New("topic").ToLocalChecked(), + Nan::New(message->topic_name()).ToLocalChecked()); + Nan::Set(eofEvent, Nan::New("offset").ToLocalChecked(), + Nan::New(message->offset())); + Nan::Set(eofEvent, Nan::New("partition").ToLocalChecked(), + Nan::New(message->partition())); + + // also store index at which position in the message array this event was emitted + // this way, we can later emit it at the right point in time + Nan::Set(eofEvent, Nan::New("messageIndex").ToLocalChecked(), + Nan::New(returnArrayIndex)); + + Nan::Set(eofEventsArray, eofEventsArrayIndex, eofEvent); + } + + delete message; + } + } + + argv[1] = returnArray; + argv[2] = eofEventsArray; + + callback->Call(argc, argv); +} + +void KafkaConsumerConsumeNum::HandleErrorCallback() { + Nan::HandleScope scope; + + if (m_messages.size() > 0) { + for (std::vector::iterator it = m_messages.begin(); + it != m_messages.end(); ++it) { + RdKafka::Message* message = *it; + delete message; + } + } + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief KafkaConsumer get message worker. + * + * This callback will get a single message. Can be of use in streams or places + * where you don't want an infinite loop managed in C++land and would rather + * manage it in Node. + * + * @see RdKafka::KafkaConsumer::Consume + * @see NodeKafka::KafkaConsumer::GetMessage + */ + +KafkaConsumerConsume::KafkaConsumerConsume(Nan::Callback *callback, + KafkaConsumer* consumer, + const int & timeout_ms) : + ErrorAwareWorker(callback), + consumer(consumer), + m_timeout_ms(timeout_ms) {} + +KafkaConsumerConsume::~KafkaConsumerConsume() {} + +void KafkaConsumerConsume::Execute() { + Baton b = consumer->Consume(m_timeout_ms); + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } else { + RdKafka::Message *message = b.data(); + RdKafka::ErrorCode errorCode = message->err(); + if (errorCode == RdKafka::ERR_NO_ERROR) { + m_message = message; + } else { + delete message; + } + } +} + +void KafkaConsumerConsume::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + v8::Local argv[argc]; + + argv[0] = Nan::Null(); + argv[1] = Conversion::Message::ToV8Object(m_message); + + delete m_message; + + callback->Call(argc, argv); +} + +void KafkaConsumerConsume::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief KafkaConsumer get committed topic partitions worker. + * + * This callback will get a topic partition list of committed offsets + * for each topic partition. It is done async because it has a timeout + * and I don't want node to block + * + * @see RdKafka::KafkaConsumer::Committed + */ + +KafkaConsumerCommitted::KafkaConsumerCommitted(Nan::Callback *callback, + KafkaConsumer* consumer, + std::vector & t, + const int & timeout_ms) : + ErrorAwareWorker(callback), + m_consumer(consumer), + m_topic_partitions(t), + m_timeout_ms(timeout_ms) {} + +KafkaConsumerCommitted::~KafkaConsumerCommitted() { + // Delete the underlying topic partitions as they are ephemeral or cloned + RdKafka::TopicPartition::destroy(m_topic_partitions); +} + +void KafkaConsumerCommitted::Execute() { + Baton b = m_consumer->Committed(m_topic_partitions, m_timeout_ms); + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void KafkaConsumerCommitted::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 2; + v8::Local argv[argc]; + + argv[0] = Nan::Null(); + argv[1] = Conversion::TopicPartition::ToV8Array(m_topic_partitions); + + callback->Call(argc, argv); +} + +void KafkaConsumerCommitted::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief KafkaConsumer seek + * + * This callback will take a topic partition list with offsets and + * seek messages from there + * + * @see RdKafka::KafkaConsumer::seek + * + * @remark Consumption for the given partition must have started for the + * seek to work. Use assign() to set the starting offset. + */ + +KafkaConsumerSeek::KafkaConsumerSeek(Nan::Callback *callback, + KafkaConsumer* consumer, + const RdKafka::TopicPartition * toppar, + const int & timeout_ms) : + ErrorAwareWorker(callback), + m_consumer(consumer), + m_toppar(toppar), + m_timeout_ms(timeout_ms) {} + +KafkaConsumerSeek::~KafkaConsumerSeek() { + if (m_timeout_ms > 0) { + // Delete it when we are done with it. + // However, if the timeout was less than 1, that means librdkafka is going + // to queue the request up asynchronously, which apparently looks like if + // we delete the memory here, since it was a pointer, librdkafka segfaults + // when it actually does the operation (since it no longer blocks). + + // Well, that means we will be leaking memory when people do a timeout of 0 + // so... we should never get to this block. But just in case... + delete m_toppar; + } +} + +void KafkaConsumerSeek::Execute() { + Baton b = m_consumer->Seek(*m_toppar, m_timeout_ms); + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void KafkaConsumerSeek::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc]; + + argv[0] = Nan::Null(); + + callback->Call(argc, argv); +} + +void KafkaConsumerSeek::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief createTopic + * + * This callback will create a topic + * + */ +AdminClientCreateTopic::AdminClientCreateTopic(Nan::Callback *callback, + AdminClient* client, + rd_kafka_NewTopic_t* topic, + const int & timeout_ms) : + ErrorAwareWorker(callback), + m_client(client), + m_topic(topic), + m_timeout_ms(timeout_ms) {} + +AdminClientCreateTopic::~AdminClientCreateTopic() { + // Destroy the topic creation request when we are done + rd_kafka_NewTopic_destroy(m_topic); +} + +void AdminClientCreateTopic::Execute() { + Baton b = m_client->CreateTopic(m_topic, m_timeout_ms); + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void AdminClientCreateTopic::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc]; + + argv[0] = Nan::Null(); + + callback->Call(argc, argv); +} + +void AdminClientCreateTopic::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Delete a topic in an asynchronous worker. + * + * This callback will delete a topic + * + */ +AdminClientDeleteTopic::AdminClientDeleteTopic(Nan::Callback *callback, + AdminClient* client, + rd_kafka_DeleteTopic_t* topic, + const int & timeout_ms) : + ErrorAwareWorker(callback), + m_client(client), + m_topic(topic), + m_timeout_ms(timeout_ms) {} + +AdminClientDeleteTopic::~AdminClientDeleteTopic() { + // Destroy the topic creation request when we are done + rd_kafka_DeleteTopic_destroy(m_topic); +} + +void AdminClientDeleteTopic::Execute() { + Baton b = m_client->DeleteTopic(m_topic, m_timeout_ms); + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void AdminClientDeleteTopic::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc]; + + argv[0] = Nan::Null(); + + callback->Call(argc, argv); +} + +void AdminClientDeleteTopic::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +/** + * @brief Delete a topic in an asynchronous worker. + * + * This callback will delete a topic + * + */ +AdminClientCreatePartitions::AdminClientCreatePartitions( + Nan::Callback *callback, + AdminClient* client, + rd_kafka_NewPartitions_t* partitions, + const int & timeout_ms) : + ErrorAwareWorker(callback), + m_client(client), + m_partitions(partitions), + m_timeout_ms(timeout_ms) {} + +AdminClientCreatePartitions::~AdminClientCreatePartitions() { + // Destroy the topic creation request when we are done + rd_kafka_NewPartitions_destroy(m_partitions); +} + +void AdminClientCreatePartitions::Execute() { + Baton b = m_client->CreatePartitions(m_partitions, m_timeout_ms); + if (b.err() != RdKafka::ERR_NO_ERROR) { + SetErrorBaton(b); + } +} + +void AdminClientCreatePartitions::HandleOKCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc]; + + argv[0] = Nan::Null(); + + callback->Call(argc, argv); +} + +void AdminClientCreatePartitions::HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { GetErrorObject() }; + + callback->Call(argc, argv); +} + +} // namespace Workers +} // namespace NodeKafka diff --git a/src/workers.h b/src/workers.h new file mode 100644 index 0000000..62a9cca --- /dev/null +++ b/src/workers.h @@ -0,0 +1,505 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +#ifndef SRC_WORKERS_H_ +#define SRC_WORKERS_H_ + +#include +#include +#include +#include + +#include "src/common.h" +#include "src/producer.h" +#include "src/kafka-consumer.h" +#include "src/admin.h" +#include "rdkafka.h" // NOLINT + +namespace NodeKafka { +namespace Workers { + +class ErrorAwareWorker : public Nan::AsyncWorker { + public: + explicit ErrorAwareWorker(Nan::Callback* callback_) : + Nan::AsyncWorker(callback_), + m_baton(RdKafka::ERR_NO_ERROR) {} + virtual ~ErrorAwareWorker() {} + + virtual void Execute() = 0; + virtual void HandleOKCallback() = 0; + void HandleErrorCallback() { + Nan::HandleScope scope; + + const unsigned int argc = 1; + v8::Local argv[argc] = { Nan::Error(ErrorMessage()) }; + + callback->Call(argc, argv); + } + + protected: + void SetErrorCode(const int & code) { + RdKafka::ErrorCode rd_err = static_cast(code); + SetErrorCode(rd_err); + } + void SetErrorCode(const RdKafka::ErrorCode & err) { + SetErrorBaton(Baton(err)); + } + void SetErrorBaton(const NodeKafka::Baton & baton) { + m_baton = baton; + SetErrorMessage(m_baton.errstr().c_str()); + } + + int GetErrorCode() { + return m_baton.err(); + } + + v8::Local GetErrorObject() { + return m_baton.ToObject(); + } + + Baton m_baton; +}; + +class MessageWorker : public ErrorAwareWorker { + public: + explicit MessageWorker(Nan::Callback* callback_) + : ErrorAwareWorker(callback_), m_asyncdata() { + m_async = new uv_async_t; + uv_async_init( + uv_default_loop(), + m_async, + m_async_message); + m_async->data = this; + + uv_mutex_init(&m_async_lock); + } + + virtual ~MessageWorker() { + uv_mutex_destroy(&m_async_lock); + } + + void WorkMessage() { + if (!callback) { + return; + } + + std::vector message_queue; + std::vector warning_queue; + + { + scoped_mutex_lock lock(m_async_lock); + // Copy the vector and empty it + m_asyncdata.swap(message_queue); + m_asyncwarning.swap(warning_queue); + } + + for (unsigned int i = 0; i < message_queue.size(); i++) { + HandleMessageCallback(message_queue[i], RdKafka::ERR_NO_ERROR); + + // we are done with it. it is about to go out of scope + // for the last time so let's just free it up here. can't rely + // on the destructor + } + + for (unsigned int i = 0; i < warning_queue.size(); i++) { + HandleMessageCallback(NULL, warning_queue[i]); + } + } + + class ExecutionMessageBus { + friend class MessageWorker; + public: + void Send(RdKafka::Message* m) const { + that_->Produce_(m); + } + void SendWarning(RdKafka::ErrorCode c) const { + that_->ProduceWarning_(c); + } + private: + explicit ExecutionMessageBus(MessageWorker* that) : that_(that) {} + MessageWorker* const that_; + }; + + virtual void Execute(const ExecutionMessageBus&) = 0; + virtual void HandleMessageCallback(RdKafka::Message*, RdKafka::ErrorCode) = 0; + + virtual void Destroy() { + uv_close(reinterpret_cast(m_async), AsyncClose_); + } + + private: + void Execute() { + ExecutionMessageBus message_bus(this); + Execute(message_bus); + } + + void Produce_(RdKafka::Message* m) { + scoped_mutex_lock lock(m_async_lock); + m_asyncdata.push_back(m); + uv_async_send(m_async); + } + + void ProduceWarning_(RdKafka::ErrorCode c) { + scoped_mutex_lock lock(m_async_lock); + m_asyncwarning.push_back(c); + uv_async_send(m_async); + } + + NAN_INLINE static NAUV_WORK_CB(m_async_message) { + MessageWorker *worker = static_cast(async->data); + worker->WorkMessage(); + } + + NAN_INLINE static void AsyncClose_(uv_handle_t* handle) { + MessageWorker *worker = static_cast(handle->data); + delete reinterpret_cast(handle); + delete worker; + } + + uv_async_t *m_async; + uv_mutex_t m_async_lock; + std::vector m_asyncdata; + std::vector m_asyncwarning; +}; + +namespace Handle { +class OffsetsForTimes : public ErrorAwareWorker { + public: + OffsetsForTimes(Nan::Callback*, NodeKafka::Connection*, + std::vector &, + const int &); + ~OffsetsForTimes(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Connection * m_handle; + std::vector m_topic_partitions; + const int m_timeout_ms; +}; +} // namespace Handle + +class ConnectionMetadata : public ErrorAwareWorker { + public: + ConnectionMetadata(Nan::Callback*, NodeKafka::Connection*, + std::string, int, bool); + ~ConnectionMetadata(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Connection * m_connection; + std::string m_topic; + int m_timeout_ms; + bool m_all_topics; + + RdKafka::Metadata* m_metadata; +}; + +class ConnectionQueryWatermarkOffsets : public ErrorAwareWorker { + public: + ConnectionQueryWatermarkOffsets(Nan::Callback*, NodeKafka::Connection*, + std::string, int32_t, int); + ~ConnectionQueryWatermarkOffsets(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Connection * m_connection; + std::string m_topic; + int32_t m_partition; + int m_timeout_ms; + + int64_t m_high_offset; + int64_t m_low_offset; +}; + +class ProducerConnect : public ErrorAwareWorker { + public: + ProducerConnect(Nan::Callback*, NodeKafka::Producer*); + ~ProducerConnect(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Producer * producer; +}; + +class ProducerDisconnect : public ErrorAwareWorker { + public: + ProducerDisconnect(Nan::Callback*, NodeKafka::Producer*); + ~ProducerDisconnect(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Producer * producer; +}; + +class ProducerFlush : public ErrorAwareWorker { + public: + ProducerFlush(Nan::Callback*, NodeKafka::Producer*, int); + ~ProducerFlush(); + + void Execute(); + void HandleOKCallback(); + + private: + NodeKafka::Producer * producer; + int timeout_ms; +}; + +class ProducerInitTransactions : public ErrorAwareWorker { + public: + ProducerInitTransactions(Nan::Callback*, NodeKafka::Producer*, const int &); + ~ProducerInitTransactions(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Producer * producer; + const int m_timeout_ms; +}; + +class ProducerBeginTransaction : public ErrorAwareWorker { + public: + ProducerBeginTransaction(Nan::Callback*, NodeKafka::Producer*); + ~ProducerBeginTransaction(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Producer * producer; +}; + +class ProducerCommitTransaction : public ErrorAwareWorker { + public: + ProducerCommitTransaction(Nan::Callback*, NodeKafka::Producer*, const int &); + ~ProducerCommitTransaction(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Producer * producer; + const int m_timeout_ms; +}; + +class ProducerAbortTransaction : public ErrorAwareWorker { + public: + ProducerAbortTransaction(Nan::Callback*, NodeKafka::Producer*, const int &); + ~ProducerAbortTransaction(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Producer * producer; + const int m_timeout_ms; +}; + +class ProducerSendOffsetsToTransaction : public ErrorAwareWorker { + public: + ProducerSendOffsetsToTransaction( + Nan::Callback*, NodeKafka::Producer*, + std::vector &, + KafkaConsumer*, + const int &); + ~ProducerSendOffsetsToTransaction(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::Producer * producer; + std::vector m_topic_partitions; + NodeKafka::KafkaConsumer* consumer; + const int m_timeout_ms; +}; + +class KafkaConsumerConnect : public ErrorAwareWorker { + public: + KafkaConsumerConnect(Nan::Callback*, NodeKafka::KafkaConsumer*); + ~KafkaConsumerConnect(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::KafkaConsumer * consumer; +}; + +class KafkaConsumerDisconnect : public ErrorAwareWorker { + public: + KafkaConsumerDisconnect(Nan::Callback*, NodeKafka::KafkaConsumer*); + ~KafkaConsumerDisconnect(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + + private: + NodeKafka::KafkaConsumer * consumer; +}; + +class KafkaConsumerConsumeLoop : public MessageWorker { + public: + KafkaConsumerConsumeLoop(Nan::Callback*, + NodeKafka::KafkaConsumer*, const int &, const int &); + ~KafkaConsumerConsumeLoop(); + + void Execute(const ExecutionMessageBus&); + void HandleOKCallback(); + void HandleErrorCallback(); + void HandleMessageCallback(RdKafka::Message*, RdKafka::ErrorCode); + private: + NodeKafka::KafkaConsumer * consumer; + const int m_timeout_ms; + unsigned int m_rand_seed; + const int m_timeout_sleep_delay_ms; +}; + +class KafkaConsumerConsume : public ErrorAwareWorker { + public: + KafkaConsumerConsume(Nan::Callback*, NodeKafka::KafkaConsumer*, const int &); + ~KafkaConsumerConsume(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + private: + NodeKafka::KafkaConsumer * consumer; + const int m_timeout_ms; + RdKafka::Message* m_message; +}; + +class KafkaConsumerCommitted : public ErrorAwareWorker { + public: + KafkaConsumerCommitted(Nan::Callback*, + NodeKafka::KafkaConsumer*, std::vector &, + const int &); + ~KafkaConsumerCommitted(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + private: + NodeKafka::KafkaConsumer * m_consumer; + std::vector m_topic_partitions; + const int m_timeout_ms; +}; + +class KafkaConsumerSeek : public ErrorAwareWorker { + public: + KafkaConsumerSeek(Nan::Callback*, NodeKafka::KafkaConsumer*, + const RdKafka::TopicPartition *, const int &); + ~KafkaConsumerSeek(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + private: + NodeKafka::KafkaConsumer * m_consumer; + const RdKafka::TopicPartition * m_toppar; + const int m_timeout_ms; +}; + +class KafkaConsumerConsumeNum : public ErrorAwareWorker { + public: + KafkaConsumerConsumeNum(Nan::Callback*, NodeKafka::KafkaConsumer*, + const uint32_t &, const int &); + ~KafkaConsumerConsumeNum(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + private: + NodeKafka::KafkaConsumer * m_consumer; + const uint32_t m_num_messages; + const int m_timeout_ms; + std::vector m_messages; +}; + +/** + * @brief Create a kafka topic on a remote broker cluster + */ +class AdminClientCreateTopic : public ErrorAwareWorker { + public: + AdminClientCreateTopic(Nan::Callback*, NodeKafka::AdminClient*, + rd_kafka_NewTopic_t*, const int &); + ~AdminClientCreateTopic(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + private: + NodeKafka::AdminClient * m_client; + rd_kafka_NewTopic_t* m_topic; + const int m_timeout_ms; +}; + +/** + * @brief Delete a kafka topic on a remote broker cluster + */ +class AdminClientDeleteTopic : public ErrorAwareWorker { + public: + AdminClientDeleteTopic(Nan::Callback*, NodeKafka::AdminClient*, + rd_kafka_DeleteTopic_t*, const int &); + ~AdminClientDeleteTopic(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + private: + NodeKafka::AdminClient * m_client; + rd_kafka_DeleteTopic_t* m_topic; + const int m_timeout_ms; +}; + +/** + * @brief Delete a kafka topic on a remote broker cluster + */ +class AdminClientCreatePartitions : public ErrorAwareWorker { + public: + AdminClientCreatePartitions(Nan::Callback*, NodeKafka::AdminClient*, + rd_kafka_NewPartitions_t*, const int &); + ~AdminClientCreatePartitions(); + + void Execute(); + void HandleOKCallback(); + void HandleErrorCallback(); + private: + NodeKafka::AdminClient * m_client; + rd_kafka_NewPartitions_t* m_partitions; + const int m_timeout_ms; +}; + +} // namespace Workers + +} // namespace NodeKafka + +#endif // SRC_WORKERS_H_ diff --git a/test/binding.spec.js b/test/binding.spec.js new file mode 100644 index 0000000..b82c7bc --- /dev/null +++ b/test/binding.spec.js @@ -0,0 +1,66 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var addon = require('bindings')('node-librdkafka'); +var t = require('assert'); + +var consumerConfig = { + 'group.id': 'awesome' +}; + +var producerConfig = { + 'client.id': 'kafka-mocha', + 'metadata.broker.list': 'localhost:9092', + 'socket.timeout.ms': 250 +}; + +var client; + +module.exports = { + 'native addon': { + 'exports something': function() { + t.equal(typeof(addon), 'object'); + }, + 'exports valid producer': function() { + t.equal(typeof(addon.Producer), 'function'); + t.throws(addon.Producer); // Requires constructor + t.equal(typeof(new addon.Producer({}, {})), 'object'); + }, + 'exports valid consumer': function() { + t.equal(typeof(addon.KafkaConsumer), 'function'); + t.throws(addon.KafkaConsumer); // Requires constructor + t.equal(typeof(new addon.KafkaConsumer(consumerConfig, {})), 'object'); + }, + 'exports version': function() { + t.ok(addon.librdkafkaVersion); + }, + 'Producer client': { + 'beforeEach': function() { + client = new addon.Producer(producerConfig, {}); + }, + 'afterEach': function() { + client = null; + }, + 'is an object': function() { + t.equal(typeof(client), 'object'); + }, + 'requires configuration': function() { + t.throws(function() { + return new addon.Producer(); + }); + }, + 'has necessary methods from superclass': function() { + var methods = ['connect', 'disconnect', 'configureCallbacks', 'getMetadata']; + methods.forEach(function(m) { + t.equal(typeof(client[m]), 'function', 'Client is missing ' + m + ' method'); + }); + } + } + }, +}; diff --git a/test/consumer.spec.js b/test/consumer.spec.js new file mode 100644 index 0000000..40b52ee --- /dev/null +++ b/test/consumer.spec.js @@ -0,0 +1,86 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var addon = require('bindings')('node-librdkafka'); +var t = require('assert'); + +var client; +var defaultConfig = { + 'client.id': 'kafka-mocha', + 'group.id': 'kafka-mocha-grp', + 'metadata.broker.list': 'localhost:9092' +}; + +module.exports = { + 'Consumer': { + 'afterEach': function() { + client = null; + }, + 'cannot be set without a topic config': function() { + t.throws(function() { + client = new addon.KafkaConsumer(defaultConfig); + }); + }, + 'can be given a topic config': function() { + client = new addon.KafkaConsumer(defaultConfig, {}); + }, + 'throws us an error if we provide an invalid configuration value': function() { + t.throws(function() { + client = new addon.KafkaConsumer({ + 'foo': 'bar' + }); + }, 'should throw because the key is invalid1'); + }, + 'throws us an error if topic config is given something invalid': function() { + t.throws(function() { + client = new addon.KafkaConsumer(defaultConfig, { 'foo': 'bar' }); + }); + }, + 'ignores function arguments for global configuration': function() { + client = new addon.KafkaConsumer({ + 'event_cb': function() {}, + 'group.id': 'mocha-test' + }, {}); + t.ok(client); + }, + 'ignores function arguments for topic configuration': function() { + client = new addon.KafkaConsumer(defaultConfig, { + 'partitioner_cb': function() {} + }); + } + }, + 'KafkaConsumer client': { + 'beforeEach': function() { + client = new addon.KafkaConsumer(defaultConfig, {}); + }, + 'afterEach': function() { + client = null; + }, + 'is an object': function() { + t.equal(typeof(client), 'object'); + }, + 'requires configuration': function() { + t.throws(function() { + return new addon.KafkaConsumer(); + }); + }, + 'has necessary methods from superclass': function() { + var methods = ['connect', 'disconnect', 'configureCallbacks', 'getMetadata']; + methods.forEach(function(m) { + t.equal(typeof(client[m]), 'function', 'Client is missing ' + m + ' method'); + }); + }, + 'has necessary bindings for librdkafka 1:1 binding': function() { + var methods = ['assign', 'unassign', 'subscribe']; + methods.forEach(function(m) { + t.equal(typeof(client[m]), 'function', 'Client is missing ' + m + ' method'); + }); + } + }, +}; diff --git a/test/error.spec.js b/test/error.spec.js new file mode 100644 index 0000000..13587fc --- /dev/null +++ b/test/error.spec.js @@ -0,0 +1,8 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ diff --git a/test/index.spec.js b/test/index.spec.js new file mode 100644 index 0000000..13587fc --- /dev/null +++ b/test/index.spec.js @@ -0,0 +1,8 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ diff --git a/test/kafka-consumer-stream.spec.js b/test/kafka-consumer-stream.spec.js new file mode 100644 index 0000000..72e4630 --- /dev/null +++ b/test/kafka-consumer-stream.spec.js @@ -0,0 +1,298 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var KafkaConsumerStream = require('../lib/kafka-consumer-stream'); +var t = require('assert'); +var Writable = require('stream').Writable; +var Emitter = require('events'); + +var fakeClient; + +module.exports = { + 'KafkaConsumerStream stream': { + 'beforeEach': function() { + fakeClient = new Emitter(); + fakeClient._isConnecting = false; + fakeClient._isConnected = true; + fakeClient.isConnected = function() { + return true; + }; + fakeClient.unsubscribe = function() { + this.emit('unsubscribed'); + return true; + }; + fakeClient.disconnect = function(cb) { + this.emit('disconnected'); + if (cb) { + t.equal(typeof cb, 'function'); + setImmediate(cb); + } + }; + fakeClient.consume = function(size, cb) { + if (!size) { + cb = size; + } + + t.equal(typeof cb, 'function', + 'Provided callback should always be a function'); + setImmediate(function() { + cb(null, [{ + value: Buffer.from('test'), + key: 'testkey', + offset: 1 + }]); + }); + }; + fakeClient.subscribe = function(topics) { + t.equal(Array.isArray(topics), true); + return this; + }; + }, + + 'exports a stream class': function() { + t.equal(typeof(KafkaConsumerStream), 'function'); + }, + + 'can be instantiated': function() { + t.equal(typeof new KafkaConsumerStream(fakeClient, { + topics: 'topic' + }), 'object'); + }, + + 'properly reads off the fake client': function(cb) { + var stream = new KafkaConsumerStream(fakeClient, { + topics: 'topic' + }); + stream.on('error', function(err) { + t.fail(err); + }); + stream.once('readable', function() { + var message = stream.read(); + t.notEqual(message, null); + t.ok(Buffer.isBuffer(message.value)); + t.equal('test', message.value.toString()); + t.equal('testkey', message.key); + t.equal(typeof message.offset, 'number'); + stream.pause(); + cb(); + }); + }, + + 'properly reads off the fake with a topic function': function(cb) { + fakeClient._metadata = { + orig_broker_id: 1, + orig_broker_name: "broker_name", + brokers: [ + { + id: 1, + host: 'localhost', + port: 40 + } + ], + topics: [ + { + name: 'awesome-topic', + partitions: [ + { + id: 1, + leader: 20, + replicas: [1, 2], + isrs: [1, 2] + } + ] + } + ] + }; + + var stream = new KafkaConsumerStream(fakeClient, { + topics: function(metadata) { + var topics = metadata.topics.map(function(v) { + return v.name; + }); + + return topics; + } + }); + fakeClient.subscribe = function(topics) { + t.equal(Array.isArray(topics), true); + t.equal(topics[0], 'awesome-topic'); + t.equal(topics.length, 1); + return this; + }; + + stream.on('error', function(err) { + t.fail(err); + }); + stream.once('readable', function() { + var message = stream.read(); + t.notEqual(message, null); + t.ok(Buffer.isBuffer(message.value)); + t.equal('test', message.value.toString()); + t.equal('testkey', message.key); + t.equal(typeof message.offset, 'number'); + stream.pause(); + cb(); + }); + }, + + 'properly reads correct number of messages but does not stop': function(next) { + var numMessages = 10; + var numReceived = 0; + var numSent = 0; + + fakeClient.consume = function(size, cb) { + if (numSent < numMessages) { + numSent++; + setImmediate(function() { + cb(null, [{ + value: Buffer.from('test'), + offset: 1 + }]); + }); + } else { + } + }; + var stream = new KafkaConsumerStream(fakeClient, { + topics: 'topic' + }); + stream.on('error', function(err) { + // Ignore + }); + stream.on('readable', function() { + var message = stream.read(); + numReceived++; + t.notEqual(message, null); + t.ok(Buffer.isBuffer(message.value)); + t.equal(typeof message.offset, 'number'); + if (numReceived === numMessages) { + // give it a second to get an error + next(); + } + }); + }, + + 'can be piped around': function(cb) { + var stream = new KafkaConsumerStream(fakeClient, { + topics: 'topic' + }); + var writable = new Writable({ + write: function(message, encoding, next) { + t.notEqual(message, null); + t.ok(Buffer.isBuffer(message.value)); + t.equal(typeof message.offset, 'number'); + this.cork(); + cb(); + }, + objectMode: true + }); + + stream.pipe(writable); + stream.on('error', function(err) { + t.fail(err); + }); + + }, + + 'streams as batch when specified': function(next) { + var numMessages = 10; + var numReceived = 0; + var numSent = 0; + + fakeClient.consume = function(size, cb) { + if (numSent < numMessages) { + numSent++; + setImmediate(function() { + cb(null, [{ + value: Buffer.from('test'), + offset: 1 + }]); + }); + } else { + } + }; + var stream = new KafkaConsumerStream(fakeClient, { + topics: 'topic', + streamAsBatch: true + }); + stream.on('error', function(err) { + // Ignore + }); + stream.on('readable', function() { + var messages = stream.read(); + numReceived++; + t.equal(Array.isArray(messages), true); + t.equal(messages.length, 1); + var message = messages[0]; + + t.notEqual(message, null); + t.ok(Buffer.isBuffer(message.value)); + t.equal(typeof message.offset, 'number'); + if (numReceived === numMessages) { + // give it a second to get an error + next(); + } + }); + }, + + 'stops reading on unsubscribe': function(next) { + var numMessages = 10; + var numReceived = 0; + var numSent = 0; + + fakeClient.consume = function(size, cb) { + if (numSent < numMessages) { + numSent++; + setImmediate(function() { + cb(null, [{ + value: Buffer.from('test'), + offset: 1 + }]); + }); + } else { + } + }; + + var stream = new KafkaConsumerStream(fakeClient, { + topics: 'topic' + }); + stream.on('error', function(err) { + // Ignore + }); + stream.on('readable', function() { + var message = stream.read(); + numReceived++; + if (message) { + t.ok(Buffer.isBuffer(message.value)); + t.equal(typeof message.offset, 'number'); + if (numReceived === numMessages) { + // give it a second to get an error + fakeClient.emit('unsubscribed'); + } + } + }); + + stream.on('end', function() { + next(); + }); + }, + + 'calls the callback on destroy': function (next) { + + fakeClient.unsubscribe = function () {}; + var stream = new KafkaConsumerStream(fakeClient, { + topics: 'topic' + }); + stream.once('readable', function () { + stream.destroy(); + stream.once('close', next); + }); + + }, + } +}; diff --git a/test/kafka-consumer.spec.js b/test/kafka-consumer.spec.js new file mode 100644 index 0000000..0f4de52 --- /dev/null +++ b/test/kafka-consumer.spec.js @@ -0,0 +1,48 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var KafkaConsumer = require('../lib/kafka-consumer'); +var t = require('assert'); + +var client; +var defaultConfig = { + 'client.id': 'kafka-mocha', + 'group.id': 'kafka-mocha-grp', + 'metadata.broker.list': 'localhost:9092' +}; +var topicConfig = {}; + +module.exports = { + 'KafkaConsumer client': { + 'beforeEach': function() { + client = new KafkaConsumer(defaultConfig, topicConfig); + }, + 'afterEach': function() { + client = null; + }, + 'does not modify config and clones it': function () { + t.deepStrictEqual(defaultConfig, { + 'client.id': 'kafka-mocha', + 'group.id': 'kafka-mocha-grp', + 'metadata.broker.list': 'localhost:9092' + }); + t.deepStrictEqual(client.globalConfig, { + 'client.id': 'kafka-mocha', + 'group.id': 'kafka-mocha-grp', + 'metadata.broker.list': 'localhost:9092' + }); + t.notEqual(defaultConfig, client.globalConfig); + }, + 'does not modify topic config and clones it': function () { + t.deepStrictEqual(topicConfig, {}); + t.deepStrictEqual(client.topicConfig, {}); + t.notEqual(topicConfig, client.topicConfig); + }, + }, +}; diff --git a/test/mocha.opts b/test/mocha.opts new file mode 100644 index 0000000..e3ca00a --- /dev/null +++ b/test/mocha.opts @@ -0,0 +1 @@ +--ui exports diff --git a/test/mock.js b/test/mock.js new file mode 100644 index 0000000..4a06bc8 --- /dev/null +++ b/test/mock.js @@ -0,0 +1,55 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +'use strict'; + +var net = require('net'); +var util = require('util'); +var Emitter = require('events'); + +function KafkaServer(config) { + if (!(this instanceof KafkaServer)) { + return new KafkaServer(config); + } + + if (config === undefined) { + config = {}; + } else if (typeof config !== 'object') { + throw new TypeError('"config" must be an object'); + } + + Emitter.call(this); + + var self = this; + + this.socket = net.createServer(function(socket) { + socket.end(); + }); //.unref(); + + this.socket.on('error', function(err) { + console.error(err); + }); + + this.socket.listen({ + port: 9092, + host: 'localhost' + }, function() { + self.address = self.socket.address(); + self.emit('ready'); + }); + +} + +util.inherits(KafkaServer, Emitter); + +KafkaServer.prototype.close = function(cb) { + this.socket.close(cb); +}; + +module.exports = KafkaServer; diff --git a/test/producer-stream.spec.js b/test/producer-stream.spec.js new file mode 100644 index 0000000..8e38c6c --- /dev/null +++ b/test/producer-stream.spec.js @@ -0,0 +1,723 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var ProducerStream = require('../lib/producer-stream'); +var t = require('assert'); +var Readable = require('stream').Readable; +var Emitter = require('events'); + +var fakeClient; + +module.exports = { + 'ProducerStream stream': { + 'beforeEach': function() { + fakeClient = new Emitter(); + + fakeClient._isConnected = true; + fakeClient._isConnecting = false; + + fakeClient.isConnected = function() { + return true; + }; + fakeClient.connect = function(opts, cb) { + setImmediate(function() { + this.emit('ready'); + }.bind(this)); + return this; + }; + fakeClient.disconnect = function(cb) { + setImmediate(function() { + this.emit('disconnected'); + }.bind(this)); + return this; + }; + fakeClient.poll = function() { + return this; + }; + fakeClient.setPollInterval = function() { + return this; + }; + }, + + 'exports a stream class': function() { + t.equal(typeof(ProducerStream), 'function'); + }, + + 'in buffer mode': { + 'requires a topic be provided when running in buffer mode': function() { + t.throws(function() { + var x = new ProducerStream(fakeClient, {}); + }); + }, + + 'can be instantiated': function() { + t.equal(typeof new ProducerStream(fakeClient, { + topic: 'topic' + }), 'object'); + }, + + 'does not run connect if the client is already connected': function(cb) { + fakeClient.connect = function() { + t.fail('Should not run connect if the client is already connected'); + }; + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + + setTimeout(cb, 10); + }, + + 'does run connect if the client is not already connected': function(cb) { + fakeClient._isConnected = false; + fakeClient.isConnected = function() { + return false; + }; + + fakeClient.once('ready', cb); + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + }, + + 'forwards connectOptions to client options when provided': function(cb) { + var testClientOptions = { timeout: 3000 }; + + fakeClient._isConnected = false; + fakeClient.isConnected = function() { + return false; + }; + var fakeConnect = fakeClient.connect; + fakeClient.connect = function(opts, callback) { + t.deepEqual(opts, testClientOptions); + cb(); + }; + + var stream = new ProducerStream(fakeClient, { + topic: 'topic', + connectOptions: testClientOptions + }); + }, + + 'automatically disconnects when autoclose is not provided': function(cb) { + fakeClient.once('disconnected', cb); + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + + stream.end(); + }, + + 'does not automatically disconnect when autoclose is set to false': function(done) { + fakeClient.once('disconnected', function() { + t.fail('Should not run disconnect'); + }); + + var stream = new ProducerStream(fakeClient, { + topic: 'topic', + autoClose: false + }); + + stream.end(); + + setTimeout(done, 10); + }, + + 'properly reads off the fake client': function(done) { + var message; + + fakeClient.produce = function(topic, partition, message, key) { + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome'); + t.equal(Buffer.isBuffer(message), true); + done(); + }; + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + stream.on('error', function(err) { + t.fail(err); + }); + + stream.write(Buffer.from('Awesome')); + }, + + 'passes a topic string if options are not provided': function(done) { + var message; + + fakeClient.produce = function(topic, partition, message, key) { + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome'); + t.equal(Buffer.isBuffer(message), true); + done(); + }; + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + stream.on('error', function(err) { + t.fail(err); + }); + + stream.write(Buffer.from('Awesome')); + }, + + 'properly handles queue errors': function(done) { + var message; + + var first = true; + + fakeClient.produce = function(topic, partition, message, key) { + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome'); + t.equal(Buffer.isBuffer(message), true); + if (first) { + first = false; + var err = new Error('Queue full'); + err.code = -184; + throw err; + } else { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + stream.on('error', function(err) { + t.fail(err); + }); + + stream.write(Buffer.from('Awesome')); + }, + + 'errors out when a non-queue related error occurs': function(done) { + fakeClient.produce = function(topic, partition, message, key) { + var err = new Error('ERR_MSG_SIZE_TOO_LARGE '); + err.code = 10; + throw err; + }; + + fakeClient.on('disconnected', function() { + done(); + }); + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + stream.on('error', function(err) { + t.equal(err.code, 10, 'Error was unexpected'); + // This is good + }); + + stream.write(Buffer.from('Awesome')); + }, + + 'errors out when a non-queue related error occurs but does not disconnect if autoclose is false': function(done) { + fakeClient.produce = function(topic, partition, message, key) { + var err = new Error('ERR_MSG_SIZE_TOO_LARGE '); + err.code = 10; + throw err; + }; + + fakeClient.on('disconnected', function() { + t.fail('Should not try to disconnect'); + }); + + var stream = new ProducerStream(fakeClient, { + topic: 'topic', + autoClose: false + }); + stream.on('error', function(err) { + t.equal(err.code, 10, 'Error was unexpected'); + // This is good + }); + + stream.write(Buffer.from('Awesome')); + + setTimeout(done, 10); + }, + + 'properly reads more than one message in order': function(done) { + + var message; + var currentMessage = 0; + + fakeClient.produce = function(topic, partition, message, key) { + currentMessage++; + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome' + currentMessage); + t.equal(Buffer.isBuffer(message), true); + if (currentMessage === 2) { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + stream.on('error', function(err) { + t.fail(err); + }); + + stream.write(Buffer.from('Awesome1')); + stream.write(Buffer.from('Awesome2')); + }, + + 'can be piped into a readable': function(done) { + + var message; + var currentMessage = 0; + var iteration = 0; + + var readable = new Readable({ + read: function(size) { + iteration++; + if (iteration > 1) { + + } else { + this.push('Awesome1'); + this.push('Awesome2'); + } + } + }); + + fakeClient.produce = function(topic, partition, message, key) { + currentMessage++; + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome' + currentMessage); + t.equal(Buffer.isBuffer(message), true); + if (currentMessage === 2) { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + stream.on('error', function(err) { + t.fail(err); + }); + + readable.pipe(stream); + }, + 'can drain buffered chunks': function(done) { + + var message; + var currentMessage = 0; + + fakeClient.produce = function(topic, partition, message, key) { + currentMessage++; + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome' + currentMessage); + t.equal(Buffer.isBuffer(message), true); + if (currentMessage === 3) { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + topic: 'topic' + }); + stream.on('error', function(err) { + t.fail(err); + }); + + fakeClient._isConnected = false; + fakeClient._isConnecting = true; + fakeClient.isConnected = function() { + return false; + }; + + stream.write(Buffer.from('Awesome1')); + stream.write(Buffer.from('Awesome2')); + stream.write(Buffer.from('Awesome3')); + + fakeClient._isConnected = true; + fakeClient._isConnecting = false; + fakeClient.isConnected = function() { + return true; + }; + fakeClient.connect(); + }, + }, + + 'in objectMode': { + 'can be instantiated': function() { + t.equal(typeof new ProducerStream(fakeClient, { + objectMode: true + }), 'object'); + }, + + 'properly produces message objects': function(done) { + var _timestamp = Date.now(); + var _opaque = { + foo: 'bar' + }; + var _headers = { + header: 'header value' + }; + + fakeClient.produce = function(topic, partition, message, key, timestamp, opaque, headers) { + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome'); + t.equal(Buffer.isBuffer(message), true); + t.equal(partition, 10); + t.equal(key, 'key'); + t.deepEqual(_opaque, opaque); + t.deepEqual(_timestamp, timestamp); + t.deepEqual(_headers, headers); + done(); + }; + + var stream = new ProducerStream(fakeClient, { + objectMode: true + }); + stream.on('error', function(err) { + t.fail(err); + }); + + stream.write({ + topic: 'topic', + value: Buffer.from('Awesome'), + partition: 10, + key: 'key', + timestamp: _timestamp, + opaque: _opaque, + headers: _headers + }); + }, + + 'properly handles queue errors': function(done) { + var message; + + var first = true; + + fakeClient.produce = function(topic, partition, message, key) { + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome'); + t.equal(Buffer.isBuffer(message), true); + t.equal(partition, 10); + t.equal(key, 'key'); + if (first) { + first = false; + var err = new Error('Queue full'); + err.code = -184; + throw err; + } else { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + objectMode: true + }); + stream.on('error', function(err) { + t.fail(err); + }); + + stream.write({ + topic: 'topic', + value: Buffer.from('Awesome'), + partition: 10, + key: 'key' + }); + }, + + 'errors out when a non-queue related error occurs': function(done) { + fakeClient.produce = function(topic, partition, message, key) { + var err = new Error('ERR_MSG_SIZE_TOO_LARGE '); + err.code = 10; + throw err; + }; + + fakeClient.on('disconnected', function() { + done(); + }); + + var stream = new ProducerStream(fakeClient, { + objectMode: true + }); + stream.on('error', function(err) { + t.equal(err.code, 10, 'Error was unexpected'); + // This is good + }); + + stream.write(Buffer.from('Awesome')); + }, + + 'errors out when a non-queue related error occurs but does not disconnect if autoclose is false': function(done) { + fakeClient.produce = function(topic, partition, message, key) { + var err = new Error('ERR_MSG_SIZE_TOO_LARGE '); + err.code = 10; + throw err; + }; + + fakeClient.on('disconnected', function() { + t.fail('Should not try to disconnect'); + }); + + var stream = new ProducerStream(fakeClient, { + objectMode: true, + autoClose: false + }); + stream.on('error', function(err) { + t.equal(err.code, 10, 'Error was unexpected'); + // This is good + }); + + stream.write({ + value: Buffer.from('Awesome'), + topic: 'topic' + }); + + setTimeout(done, 10); + }, + + 'properly reads more than one message in order': function(done) { + + var message; + var currentMessage = 0; + + fakeClient.produce = function(topic, partition, message, key) { + currentMessage++; + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome' + currentMessage); + t.equal(Buffer.isBuffer(message), true); + if (currentMessage === 2) { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + objectMode: true + }); + stream.on('error', function(err) { + t.fail(err); + }); + + stream.write({ + value: Buffer.from('Awesome1'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome2'), + topic: 'topic' + }); + }, + + 'can be piped into a readable': function(done) { + + var message; + var currentMessage = 0; + var iteration = 0; + + var readable = new Readable({ + objectMode: true, + read: function(size) { + iteration++; + if (iteration > 1) { + + } else { + this.push({ + topic: 'topic', + value: Buffer.from('Awesome1') + }); + this.push({ + topic: 'topic', + value: Buffer.from('Awesome2') + }); + } + } + }); + + fakeClient.produce = function(topic, partition, message, key) { + currentMessage++; + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome' + currentMessage); + t.equal(Buffer.isBuffer(message), true); + if (currentMessage === 2) { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + objectMode: true + }); + stream.on('error', function(err) { + t.fail(err); + }); + + readable.pipe(stream); + }, + + 'can drain buffered messages': function(done) { + + var message; + var currentMessage = 0; + + fakeClient.produce = function(topic, partition, message, key) { + currentMessage++; + t.equal('topic', topic); + t.equal(message.toString(), 'Awesome' + currentMessage); + t.equal(Buffer.isBuffer(message), true); + if (currentMessage === 3) { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + objectMode: true + }); + stream.on('error', function(err) { + t.fail(err); + }); + + fakeClient._isConnected = false; + fakeClient._isConnecting = true; + fakeClient.isConnected = function() { + return false; + }; + + stream.write({ + value: Buffer.from('Awesome1'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome2'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome3'), + topic: 'topic' + }); + + fakeClient._isConnected = true; + fakeClient._isConnecting = false; + fakeClient.isConnected = function() { + return true; + }; + fakeClient.connect(); + }, + + 'properly handles queue errors while draining': function(done) { + var message; + var currentMessage = 0; + + fakeClient.produce = function(topic, partition, message, key) { + currentMessage++; + if (currentMessage === 3) { + var err = new Error('Queue full'); + err.code = -184; + throw err; + } else if (currentMessage === 4) { + done(); + } + }; + + var stream = new ProducerStream(fakeClient, { + objectMode: true + }); + stream.on('error', function(err) { + t.fail(err); + }); + + fakeClient._isConnected = false; + fakeClient._isConnecting = true; + fakeClient.isConnected = function() { + return false; + }; + + stream.write({ + value: Buffer.from('Awesome1'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome2'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome3'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome4'), + topic: 'topic' + }); + + fakeClient._isConnected = true; + fakeClient._isConnecting = false; + fakeClient.isConnected = function() { + return true; + }; + fakeClient.connect(); + }, + + 'errors out for non-queue related errors while draining': function (done) { + var currentMessage = 0; + + fakeClient.produce = function(topic, partition, message, key) { + currentMessage++; + if (currentMessage === 3) { + var err = new Error('ERR_MSG_SIZE_TOO_LARGE '); + err.code = 10; + throw err; + } + }; + + fakeClient.on('disconnected', function() { + done(); + }); + + var stream = new ProducerStream(fakeClient, { + objectMode: true + }); + stream.on('error', function(err) { + t.equal(err.code, 10, 'Error was unexpected'); + // This is good + }); + + fakeClient._isConnected = false; + fakeClient._isConnecting = true; + fakeClient.isConnected = function() { + return false; + }; + + stream.write({ + value: Buffer.from('Awesome1'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome2'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome3'), + topic: 'topic' + }); + stream.write({ + value: Buffer.from('Awesome4'), + topic: 'topic' + }); + + fakeClient._isConnected = true; + fakeClient._isConnecting = false; + fakeClient.isConnected = function() { + return true; + }; + fakeClient.connect(); + }, + + } + + } +}; diff --git a/test/producer.spec.js b/test/producer.spec.js new file mode 100644 index 0000000..dac2e22 --- /dev/null +++ b/test/producer.spec.js @@ -0,0 +1,100 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var Producer = require('../lib/producer'); +var t = require('assert'); +// var Mock = require('./mock'); + +var client; +var defaultConfig = { + 'client.id': 'kafka-mocha', + 'metadata.broker.list': 'localhost:9092', + 'socket.timeout.ms': 250 +}; +var topicConfig = {}; + +var server; + +module.exports = { + 'Producer client': { + 'beforeEach': function() { + client = new Producer(defaultConfig, topicConfig); + }, + 'afterEach': function() { + client = null; + }, + 'is an object': function() { + t.equal(typeof(client), 'object'); + }, + 'requires configuration': function() { + t.throws(function() { + return new Producer(); + }); + }, + 'has necessary methods from superclass': function() { + var methods = ['connect', 'disconnect', 'getMetadata']; + methods.forEach(function(m) { + t.equal(typeof(client[m]), 'function', 'Client is missing ' + m + ' method'); + }); + }, + 'has "_disconnect" override': function() { + t.equal(typeof(client._disconnect), 'function', 'Producer is missing base _disconnect method'); + }, + 'does not modify config and clones it': function () { + t.deepStrictEqual(defaultConfig, { + 'client.id': 'kafka-mocha', + 'metadata.broker.list': 'localhost:9092', + 'socket.timeout.ms': 250 + }); + t.deepStrictEqual(client.globalConfig, { + 'client.id': 'kafka-mocha', + 'metadata.broker.list': 'localhost:9092', + 'socket.timeout.ms': 250 + }); + t.notEqual(defaultConfig, client.globalConfig); + }, + 'does not modify topic config and clones it': function () { + t.deepStrictEqual(topicConfig, {}); + t.deepStrictEqual(client.topicConfig, {}); + t.notEqual(topicConfig, client.topicConfig); + }, + 'disconnect method': { + 'calls flush before it runs': function(next) { + var providedTimeout = 1; + + client.flush = function(timeout, cb) { + t.equal(providedTimeout, timeout, 'Timeouts do not match'); + t.equal(typeof(cb), 'function'); + setImmediate(cb); + }; + + client._disconnect = function(cb) { + setImmediate(cb); + }; + + client.disconnect(providedTimeout, next); + }, + 'provides a default timeout when none is provided': function(next) { + client.flush = function(timeout, cb) { + t.notEqual(timeout, undefined); + t.notEqual(timeout, null); + t.notEqual(timeout, 0); + t.equal(typeof(cb), 'function'); + setImmediate(cb); + }; + + client._disconnect = function(cb) { + setImmediate(cb); + }; + + client.disconnect(next); + } + } + }, +}; diff --git a/test/producer/high-level-producer.spec.js b/test/producer/high-level-producer.spec.js new file mode 100644 index 0000000..e40254e --- /dev/null +++ b/test/producer/high-level-producer.spec.js @@ -0,0 +1,496 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var HighLevelProducer = require('../../lib/producer/high-level-producer'); +var t = require('assert'); +var Promise = require('bluebird'); +// var Mock = require('./mock'); + +var client; +var defaultConfig = { + 'client.id': 'kafka-mocha', + 'metadata.broker.list': 'localhost:9092', + 'socket.timeout.ms': 250 +}; +var topicConfig = {}; + +var server; + +module.exports = { + 'High Level Producer client': { + 'beforeEach': function() { + client = new HighLevelProducer(defaultConfig, topicConfig); + }, + 'afterEach': function() { + client = null; + }, + 'is an object': function() { + t.equal(typeof(client), 'object'); + }, + 'requires configuration': function() { + t.throws(function() { + return new HighLevelProducer(); + }); + }, + 'has necessary methods from superclass': function() { + var methods = ['_oldProduce']; + methods.forEach(function(m) { + t.equal(typeof(client[m]), 'function', 'Client is missing ' + m + ' method'); + }); + }, + 'does not modify config and clones it': function () { + t.deepStrictEqual(defaultConfig, { + 'client.id': 'kafka-mocha', + 'metadata.broker.list': 'localhost:9092', + 'socket.timeout.ms': 250 + }); + t.deepStrictEqual(client.globalConfig, { + 'client.id': 'kafka-mocha', + 'metadata.broker.list': 'localhost:9092', + 'socket.timeout.ms': 250 + }); + t.notEqual(defaultConfig, client.globalConfig); + }, + 'does not modify topic config and clones it': function () { + t.deepStrictEqual(topicConfig, {}); + t.deepStrictEqual(client.topicConfig, {}); + t.notEqual(topicConfig, client.topicConfig); + }, + 'produce method': { + 'headers support': function(next) { + var v = 'foo'; + var k = 'key'; + var h = [ + { key1: "value1A" }, + { key1: "value1B" }, + { key2: "value2" }, + { key1: "value1C" }, + ]; + var jsonH = JSON.stringify(h); + + client._oldProduce = function(topic, partition, value, key, timestamp, opaque, headers) { + t.equal(value, 'foo'); + t.equal(key, 'key'); + t.equal(JSON.stringify(headers), jsonH); + next(); + }; + + client.produce('tawpic', 0, v, k, null, h, function() { + + }); + }, + + 'can use a custom serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + var valueSerializerCalled = false; + var keySerializerCalled = false; + + client._oldProduce = function(topic, partition, v, k, timestamp, opaque) { + t.equal(valueSerializerCalled, true); + t.equal(keySerializerCalled, true); + t.deepEqual(v, Buffer.from('foo')); + t.equal(k, 'key'); + next(); + }; + + client.setValueSerializer(function(_) { + valueSerializerCalled = true; + t.deepEqual(_, v); + return Buffer.from('foo'); + }); + + client.setKeySerializer(function(_) { + keySerializerCalled = true; + t.deepEqual(_, k); + return 'key'; + }); + + client.produce('tawpic', 0, v, k, null, function() { + + }); + }, + + 'can use a value asynchronous custom serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + var valueSerializerCalled = false; + var keySerializerCalled = false; + + client._oldProduce = function(topic, partition, v, k, timestamp, opaque) { + t.equal(valueSerializerCalled, true); + t.equal(keySerializerCalled, true); + t.deepEqual(v, Buffer.from('foo')); + t.equal(k, 'key'); + next(); + }; + + client.setValueSerializer(function(_, cb) { + valueSerializerCalled = true; + t.deepEqual(_, v); + setImmediate(function() { + cb(null, Buffer.from('foo')); + }); + }); + + client.setKeySerializer(function(_) { + keySerializerCalled = true; + t.deepEqual(_, k); + return 'key'; + }); + + client.produce('tawpic', 0, v, k, null, function() { + + }); + }, + + 'can use a key asynchronous custom serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + var valueSerializerCalled = false; + var keySerializerCalled = false; + + client._oldProduce = function(topic, partition, v, k, timestamp, opaque) { + t.equal(valueSerializerCalled, true); + t.equal(keySerializerCalled, true); + t.deepEqual(v, Buffer.from('foo')); + t.equal(k, 'key'); + next(); + }; + + client.setValueSerializer(function(_) { + valueSerializerCalled = true; + t.deepEqual(_, v); + return Buffer.from('foo'); + }); + + client.setKeySerializer(function(_, cb) { + keySerializerCalled = true; + t.deepEqual(_, k); + setImmediate(function() { + cb(null, 'key'); + }); + }); + + client.produce('tawpic', 0, v, k, null, function() { + + }); + }, + + 'can use two asynchronous custom serializers': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + var valueSerializerCalled = false; + var keySerializerCalled = false; + + client._oldProduce = function(topic, partition, v, k, timestamp, opaque) { + t.equal(valueSerializerCalled, true); + t.equal(keySerializerCalled, true); + t.deepEqual(v, Buffer.from('foo')); + t.equal(k, 'key'); + next(); + }; + + client.setValueSerializer(function(_, cb) { + valueSerializerCalled = true; + t.deepEqual(_, v); + setImmediate(function() { + cb(null, Buffer.from('foo')); + }); + }); + + client.setKeySerializer(function(_, cb) { + keySerializerCalled = true; + t.deepEqual(_, k); + setImmediate(function() { + cb(null, 'key'); + }); + }); + + client.produce('tawpic', 0, v, k, null, function() { + + }); + }, + + // Promise API + 'can use a value promise-based custom serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + var valueSerializerCalled = false; + var keySerializerCalled = false; + + client._oldProduce = function(topic, partition, v, k, timestamp, opaque) { + t.equal(valueSerializerCalled, true); + t.equal(keySerializerCalled, true); + next(); + }; + + client.setValueSerializer(function(_) { + valueSerializerCalled = true; + t.deepEqual(_, v); + return new Promise(function(resolve) { + resolve(Buffer.from('')); + }); + }); + + client.setKeySerializer(function(_) { + keySerializerCalled = true; + t.deepEqual(_, k); + return null; + }); + + client.produce('tawpic', 0, v, k, null, function() { + + }); + }, + + 'can use a key promise-based custom serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + var valueSerializerCalled = false; + var keySerializerCalled = false; + + client._oldProduce = function(topic, partition, v, k, timestamp, opaque) { + t.equal(valueSerializerCalled, true); + t.equal(keySerializerCalled, true); + t.deepEqual(v, Buffer.from('foo')); + t.equal(k, 'key'); + next(); + }; + + client.setValueSerializer(function(_) { + valueSerializerCalled = true; + t.deepEqual(_, v); + return Buffer.from('foo'); + }); + + client.setKeySerializer(function(_) { + keySerializerCalled = true; + t.deepEqual(_, k); + return new Promise(function(resolve) { + resolve('key'); + }); + }); + + client.produce('tawpic', 0, v, k, null, function() { + + }); + }, + + 'can use two promise-based custom serializers': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + var valueSerializerCalled = false; + var keySerializerCalled = false; + + client._oldProduce = function(topic, partition, v, k, timestamp, opaque) { + t.equal(valueSerializerCalled, true); + t.equal(keySerializerCalled, true); + t.deepEqual(v, Buffer.from('foo')); + t.equal(k, 'key'); + next(); + }; + + client.setValueSerializer(function(_) { + valueSerializerCalled = true; + t.deepEqual(_, v); + return new Promise(function(resolve) { + resolve(Buffer.from('foo')); + }); + }); + + client.setKeySerializer(function(_) { + keySerializerCalled = true; + t.deepEqual(_, k); + return new Promise(function(resolve) { + resolve('key'); + }); + }); + + client.produce('tawpic', 0, v, k, null, function() { + + }); + }, + + 'bubbles up serializer errors in an async value serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + client.setValueSerializer(function(_, cb) { + t.deepEqual(_, v); + setImmediate(function() { + cb(new Error('even together we failed')); + }); + }); + + client.produce('tawpic', 0, v, k, null, function(err) { + t.equal(typeof err, 'object', 'an error should be returned'); + next(); + }); + }, + + 'bubbles up serializer errors in an async key serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + client.setKeySerializer(function(_, cb) { + t.deepEqual(_, v); + setImmediate(function() { + cb(new Error('even together we failed')); + }); + }); + + client.produce('tawpic', 0, v, k, null, function(err) { + t.equal(typeof err, 'object', 'an error should be returned'); + next(); + }); + }, + + 'bubbles up serializer errors in a sync value serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + client.setValueSerializer(function(_, cb) { + t.deepEqual(_, v); + throw new Error('even together we failed'); + }); + + client.produce('tawpic', 0, v, k, null, function(err) { + t.equal(typeof err, 'object', 'an error should be returned'); + next(); + }); + }, + + 'bubbles up serializer errors in a sync key serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + client.setKeySerializer(function(_, cb) { + t.deepEqual(_, v); + throw new Error('even together we failed'); + }); + + client.produce('tawpic', 0, v, k, null, function(err) { + t.equal(typeof err, 'object', 'an error should be returned'); + next(); + }); + }, + + 'bubbles up serializer errors in a promise-based value serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + client.setValueSerializer(function(_) { + t.deepEqual(_, v); + + return new Promise(function (resolve, reject) { + reject(new Error('even together we failed')); + }); + }); + + client.produce('tawpic', 0, v, k, null, function(err) { + t.equal(typeof err, 'object', 'an error should be returned'); + next(); + }); + }, + + 'bubbles up serializer errors in a promise-based key serializer': function(next) { + var v = { + disparaging: 'hyena', + }; + + var k = { + delicious: 'cookie', + }; + + client.setKeySerializer(function(_) { + t.deepEqual(_, v); + + return new Promise(function(resolve, reject) { + return new Promise(function (resolve, reject) { + reject(new Error('even together we failed')); + }); + }); + }); + + client.produce('tawpic', 0, v, k, null, function(err) { + t.equal(typeof err, 'object', 'an error should be returned'); + next(); + }); + }, + } + }, +}; diff --git a/test/tools/ref-counter.spec.js b/test/tools/ref-counter.spec.js new file mode 100644 index 0000000..41c53f5 --- /dev/null +++ b/test/tools/ref-counter.spec.js @@ -0,0 +1,53 @@ +var t = require('assert'); +var RefCounter = require('../../lib/tools/ref-counter'); + +function noop() {} + +module.exports = { + 'RefCounter': { + 'is an object': function() { + t.equal(typeof(RefCounter), 'function'); + }, + 'should become active when incremented': function(next) { + var refCounter = new RefCounter(function() { next(); }, noop); + + refCounter.increment(); + }, + 'should become inactive when incremented and decremented': function(next) { + var refCounter = new RefCounter(noop, function() { next(); }); + + refCounter.increment(); + setImmediate(function() { + refCounter.decrement(); + }); + }, + 'should support multiple accesses': function(next) { + var refCounter = new RefCounter(noop, function() { next(); }); + + refCounter.increment(); + refCounter.increment(); + refCounter.decrement(); + setImmediate(function() { + refCounter.decrement(); + }); + }, + 'should be reusable': function(next) { + var numActives = 0; + var numPassives = 0; + var refCounter = new RefCounter(function() { + numActives += 1; + }, function() { + numPassives += 1; + + if (numActives === 2 && numPassives === 2) { + next(); + } + }); + + refCounter.increment(); + refCounter.decrement(); + refCounter.increment(); + refCounter.decrement(); + } + } +}; diff --git a/test/topic-partition.spec.js b/test/topic-partition.spec.js new file mode 100644 index 0000000..1961e8a --- /dev/null +++ b/test/topic-partition.spec.js @@ -0,0 +1,104 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var TopicPartition = require('../lib/topic-partition'); +var Topic = require('../lib/topic'); + +var t = require('assert'); + +module.exports = { + 'TopicPartition': { + 'is a function': function() { + t.equal(typeof(TopicPartition), 'function'); + }, + 'be constructable': function() { + var toppar = new TopicPartition('topic', 1, 0); + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.offset, 0); + t.equal(toppar.partition, 1); + }, + 'be creatable using 0 as the partition': function() { + var toppar = new TopicPartition('topic', 0, 0); + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.offset, 0); + t.equal(toppar.partition, 0); + }, + 'throw if partition is null or undefined': function() { + t.throws(function() { + var tp = new TopicPartition('topic', undefined, 0); + }); + + t.throws(function() { + var tp = new TopicPartition('topic', null, 0); + }); + }, + 'sets offset to stored by default': function() { + var toppar = new TopicPartition('topic', 1); + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.partition, 1); + t.equal(toppar.offset, Topic.OFFSET_STORED); + }, + 'sets offset to end if "end" is provided"': function() { + var toppar = new TopicPartition('topic', 1, 'end'); + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.partition, 1); + t.equal(toppar.offset, Topic.OFFSET_END); + }, + 'sets offset to end if "latest" is provided"': function() { + var toppar = new TopicPartition('topic', 1, 'latest'); + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.partition, 1); + t.equal(toppar.offset, Topic.OFFSET_END); + }, + 'sets offset to beginning if "beginning" is provided"': function() { + var toppar = new TopicPartition('topic', 1, 'beginning'); + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.partition, 1); + t.equal(toppar.offset, Topic.OFFSET_BEGINNING); + }, + 'sets offset to start if "beginning" is provided"': function() { + var toppar = new TopicPartition('topic', 1, 'beginning'); + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.partition, 1); + t.equal(toppar.offset, Topic.OFFSET_BEGINNING); + }, + 'sets offset to stored if "stored" is provided"': function() { + var toppar = new TopicPartition('topic', 1, 'stored'); + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.partition, 1); + t.equal(toppar.offset, Topic.OFFSET_STORED); + }, + 'throws when an invalid special offset is provided"': function() { + t.throws(function() { + var toppar = new TopicPartition('topic', 1, 'fake'); + }); + } + }, + 'TopicPartition.map': { + 'is a function': function() { + t.equal(typeof(TopicPartition.map), 'function'); + }, + 'converts offsets inside the array': function() { + var result = TopicPartition.map([{ topic: 'topic', partition: 1, offset: 'stored' }]); + var toppar = result[0]; + + t.equal(toppar.topic, 'topic'); + t.equal(toppar.partition, 1); + t.equal(toppar.offset, Topic.OFFSET_STORED); + }, + }, +}; diff --git a/test/util.spec.js b/test/util.spec.js new file mode 100644 index 0000000..b58db0e --- /dev/null +++ b/test/util.spec.js @@ -0,0 +1,46 @@ +/* + * node-rdkafka - Node.js wrapper for RdKafka C/C++ library + * + * Copyright (c) 2016 Blizzard Entertainment + * + * This software may be modified and distributed under the terms + * of the MIT license. See the LICENSE.txt file for details. + */ + +var shallowCopy = require('../lib/util').shallowCopy; +var t = require('assert'); + +module.exports = { + 'shallowCopy utility': { + 'returns value itself when it is not an object': function () { + t.strictEqual(10, shallowCopy(10)); + t.strictEqual('str', shallowCopy('str')); + t.strictEqual(null, shallowCopy(null)); + t.strictEqual(undefined, shallowCopy(undefined)); + t.strictEqual(false, shallowCopy(false)); + }, + 'returns shallow copy of the passed object': function () { + var obj = { + sub: { a: 10 }, + b: 'str', + }; + var copy = shallowCopy(obj); + + t.notEqual(obj, copy); + t.deepStrictEqual(obj, copy); + t.equal(obj.sub, copy.sub); + }, + 'does not copy non-enumerable and inherited properties': function () { + var obj = Object.create({ + a: 10, + }, { + b: { value: 'str' }, + c: { value: true, enumerable: true }, + }); + var copy = shallowCopy(obj); + + t.notEqual(obj, copy); + t.deepStrictEqual(copy, { c: true }); + }, + }, +}; diff --git a/util/configure.js b/util/configure.js new file mode 100644 index 0000000..240d944 --- /dev/null +++ b/util/configure.js @@ -0,0 +1,30 @@ +'use strict'; + +var query = process.argv[2]; + +var fs = require('fs'); +var path = require('path'); + +var baseDir = path.resolve(__dirname, '../'); +var releaseDir = path.join(baseDir, 'build', 'deps'); + +var isWin = /^win/.test(process.platform); + +// Skip running this if we are running on a windows system +if (isWin) { + process.stderr.write('Skipping run because we are on windows\n'); + process.exit(0); +} + +var childProcess = require('child_process'); + +try { + childProcess.execSync('./configure --prefix=' + releaseDir + ' --libdir=' + releaseDir, { + cwd: baseDir, + stdio: [0,1,2] + }); + process.exit(0); +} catch (e) { + process.stderr.write(e.message + '\n'); + process.exit(1); +} diff --git a/util/get-env.js b/util/get-env.js new file mode 100644 index 0000000..1ef46b0 --- /dev/null +++ b/util/get-env.js @@ -0,0 +1,6 @@ +'use strict'; + +var env = process.argv[2]; +var def = process.argv[3] || ''; + +process.stdout.write(process.env[env] || def); diff --git a/util/test-compile.js b/util/test-compile.js new file mode 100644 index 0000000..2976336 --- /dev/null +++ b/util/test-compile.js @@ -0,0 +1,11 @@ +var kafka = require('../lib'); + +var p = new kafka.Producer({ 'bootstrap.servers': 'localhost:9092' }, {}); + +p.connect({ timeout: 1000 }, function(err) { + if (!err) { + p.disconnect(); + } else { + process.exit(0); + } +}); diff --git a/util/test-producer-delivery.js b/util/test-producer-delivery.js new file mode 100644 index 0000000..cda39c4 --- /dev/null +++ b/util/test-producer-delivery.js @@ -0,0 +1,100 @@ +const Kafka = require("../lib/index.js"); + +const wait = (ms) => new Promise((resolve) => setTimeout(resolve, ms)); + +const sendData = async (producer, totalMessages) => { + const topic = "node"; + const msg = "dkfljaskldfajkldsjfklasdjfalk;dsjfkl;asjfdskl;fjda;lkfjsdklfsajlkfjdsklfajsklfjsklanklsalkjkljkasfak"; + const buffer = Buffer.from(msg); + const key = "test"; + for (let n = 0; n < totalMessages; ++n) { + let bufferIsFull = false; + do { + bufferIsFull = false; + try { + producer.produce(topic, -1, buffer, key, null, n); + } + catch (error) { + // Based on config, and messages, this will execute once + if (error.code === Kafka.CODES.ERRORS.ERR__QUEUE_FULL) { + producer.poll(); + // The wait introduces 11-12 seconds of latency when dr_cb is true + const start = process.hrtime(); + await wait(50); + const latency = process.hrtime(start); + console.info(`Wait took ${latency[0]} seconds`); + bufferIsFull = true; + } else { + throw error; + } + } + } while (bufferIsFull); + } + console.log("Finished producing"); +}; + +const verifyReports = async (reports, reportsComplete, totalMessages) => { + const reportsTimeout = new Promise((resolve, reject) => { + setTimeout(() => { + reject("Delivery report timed out"); + }, 10000); + }); + await Promise.race([reportsComplete, reportsTimeout]); + await wait(500); // wait for some more delivery reports. + if (reports.length === totalMessages) { + console.log("Reports count match"); + } else { + console.error("Reports count doesn't match"); + return; + } + for(let n = 0; n < totalMessages; ++n) { + if(reports[n].opaque !== n) { + console.error("Expect message number does not match"); + } + } +}; + +const run = async () => { + const reports = []; + const totalMessages = 1000100; + const producer = new Kafka.Producer({ + "batch.num.messages": 50000, + "compression.codec": "lz4", + "delivery.report.only.error": false, + "dr_cb": true, + "metadata.broker.list": "localhost:9092", + "message.send.max.retries": 10000000, + "queue.buffering.max.kbytes": 2000000, + "queue.buffering.max.messages": 1000000, + "queue.buffering.max.ms": 0, + "socket.keepalive.enable": true, + }, {}); + + producer.setPollInterval(100); + producer.on("event.log", (obj) => console.log(obj)); + const reportsComplete = new Promise((resolve) => { + producer.on("delivery-report", (err, report) => { + reports.push(report); + if(reports.length === totalMessages) { + resolve(); + } + }); + }); + + const readyPromise = new Promise((resolve) => { + producer.on("ready", async () => { + console.log("Producer is ready"); + resolve(); + }); + producer.connect(); + }); + await readyPromise; + + await sendData(producer, totalMessages); + await verifyReports(reports, reportsComplete, totalMessages); + process.exit(0); +}; + +run().catch((err) => { + console.error(err); +}); diff --git a/win_install.bat b/win_install.bat new file mode 100644 index 0000000..a0229c9 --- /dev/null +++ b/win_install.bat @@ -0,0 +1,9 @@ +@echo off +choco install openssl.light +REM if /i %TRAVIS_NODE_VERSION% gtr 6 ( +REM npm install --global --production windows-build-tools +REM ) else ( +REM npm install --global --production windows-build-tools@3.1.0 +REM ) + +choco install make \ No newline at end of file