parallel-webpack
allows you to run multiple webpack builds in parallel,
spreading the work across your processors and thus helping to significantly speed
up your build. For us at trivago it has reduced
the build from 16 minutes to just 2 minutes - for 32 variants. That performance
improvement naturally comes at the expense of utilizing all available CPU cores.
npm install parallel-webpack --save-dev
You can choose whether to install parallel-webpack
globally or locally.
At trivago, we keep our build tools locally to the project
so that we have full control over its versions.
Given a webpack.config.js
like this:
module.exports = [{
entry: 'pageA.js',
output: {
path: './dist',
filename: 'pageA.bundle.js'
}
}, {
entry: 'pageB.js',
output: {
path: './dist',
filename: 'pageB.bundle.js'
}
}];
parallel-webpack
will run both specified builds in parallel.
Sometimes, just using different configurations like above won't be enough and what
you really want or need is the same configuration with some adjustments.
parallel-webpack
can help you with generating those configuration variants
as
well.
var createVariants = require('parallel-webpack').createVariants;
// Those options will be mixed into every variant
// and passed to the `createConfig` callback.
var baseOptions = {
preferredDevTool: process.env.DEVTOOL || 'eval'
};
// This object defines the potential option variants
// the key of the object is used as the option name, its value must be an array
// which contains all potential values of your build.
var variants = {
minified: [true, false],
debug: [true, false],
target: ['commonjs2', 'var', 'umd', 'amd']
};
function createConfig(options) {
var plugins = [
new webpack.optimize.DedupePlugin(),
new webpack.optimize.OccurenceOrderPlugin(),
new webpack.DefinePlugin({
DEBUG: JSON.stringify(JSON.parse(options.debug))
})
];
if(options.minified) {
plugins.push(new webpack.optimize.UglifyJsPlugin({
sourceMap: false,
compress: {
warnings: false
}
}));
}
return {
entry: './index.js',
devtool: options.preferredDevTool,
output: {
path: './dist/',
filename: 'MyLib.' +
options.target +
(options.minified ? '.min' : '') +
(options.debug ? '.debug' : '')
+ '.js',
libraryTarget: options.target
},
plugins: plugins
};
}
module.exports = createVariants(baseOptions, variants, createConfig);
The above configuration will create 16 variations of the build for you, which
parallel-webpack
will distribute among your processors for building.
[WEBPACK] Building 16 targets in parallel
[WEBPACK] Started building MyLib.umd.js
[WEBPACK] Started building MyLib.umd.min.js
[WEBPACK] Started building MyLib.umd.debug.js
[WEBPACK] Started building MyLib.umd.min.debug.js
[WEBPACK] Started building MyLib.amd.js
[WEBPACK] Started building MyLib.amd.min.js
[WEBPACK] Started building MyLib.amd.debug.js
[WEBPACK] Started building MyLib.amd.min.debug.js
[WEBPACK] Started building MyLib.commonjs2.js
[WEBPACK] Started building MyLib.commonjs2.min.js
[WEBPACK] Started building MyLib.commonjs2.debug.js
[WEBPACK] Started building MyLib.commonjs2.min.debug.js
[WEBPACK] Started building MyLib.var.js
[WEBPACK] Started building MyLib.var.min.js
[WEBPACK] Started building MyLib.var.debug.js
[WEBPACK] Started building MyLib.var.min.debug.js
One of the features that made webpack so popular is certainly its watcher which continously rebuilds your application.
When using parallel-webpack
, you can easily use the same feature as well by
specifying the --watch
option on the command line:
parallel-webpack --watch
As a side-effect of using parallel-webpack
, an error will no longer lead to
you having to restart webpack. Instead, parallel-webpack
will keep retrying to
build your application until you've fixed the problem.
While that is highly useful for development it can be a nightmare for
CI builds. Thus, when building with parallel-webpack
in a CI context, you should
consider to use the --max-retries
(or -m
option) to force parallel-webpack
to give
up on your build after a certain amount of retries:
parallel-webpack --max-retries=3
When you need to use a configuration file that is not webpack.config.js
, you can
specify its name using the --config
parameter:
parallel-webpack --config=myapp.webpack.config.js
While the statistics generated by Webpack are very usually very useful, they also take time to generate and print and create a lot of visual overload if you don't actually need them.
Since version 1.3.0, generating them can be turned off:
parallel-webpack --no-stats
Under certain circumstances you might not want parallel-webpack
to use all of your
available CPUs for building your assets. In those cases, you can specify the parallel
,
or p
for short, option to tell parallel-webpack
how many CPUs it may use.
parallel-webpack -p=2
Sometimes, you might want to access command line arguments within your webpack.config.js
in order to create a more specific configuration.
parallel-webpack
will forward every parameter specified after --
to the configuration
as is:
parallel-webpack -- --app=trivago
Within webpack.config.js
:
console.log(process.argv);
// => [ 'node', 'parallel-webpack', '--app=trivago' ]
parallel-webpack
adds the first two values to process.argv
to ensure that there
are no differences between various ways of invoking the webpack.config.js
.
Just like webpack, you can also use parallel-webpack
as an API from node.js
(You can specify any other option used in worker-farm):
var run = require('parallel-webpack').run,
configPath = require.resolve('./webpack.config.js');
run(configPath, {
watch: false,
maxRetries: 1,
stats: true, // defaults to false
maxConcurrentWorkers: 2 // use 2 workers
});
Alters the given baseConfig
with all possible variants
and maps the result into
a valid webpack configuration using the given configCallback
.
Creates all possible variations as specified in the variants
object and
maps the result into a valid webpack configuration using the given configCallback
.
Alters the given baseConfig
with all possible variants
and returns it.
Creates all possible variations from the given variants
and returns them as a flat array.