webpack-merge v5

[bpinto]

• Upgrades webpack-merge to latest.
• Add a test covering plugins merge.

Disclaimer: I did not reimplement merge.smart because we may not need it. I thought it would be better to keep it simple and in case someone has a special need then we work on the code to handle that instead of potentially add unnecessary code.

[bpinto]

@erezrokah I noticed I hadn't run npm run format before pushing, now I've done it. Could you allow the workflow to run again? Thank you!

[bpinto]

@erezrokah have you had the time to review this PR? Is there any change you would like to see me doing?

[erezrokah]

Hi @bpinto, sorry for the delay here - can you detail the changes between merge and merge.smart?
This might be a breaking change (as well as #492) so we'll need to put it under a major version bump

[bpinto]

I'll share some links that explain how merge.smart used to work, as it's useful to see the examples.

This is the commit that removed merge.smart method from webpack-merge library.

This is the docs for the removed method:

webpack-merge tries to be smart about merging loaders when merge.smart is used. Loaders with matching rule fields, including test/enforce/include/exclude, will be merged into a single loader value.

The standard webpack configuration that is used is very simple:

netlify-lambda/lib/build.js

Lines 137 to 177 in 2493319

	var webpackConfig = {
	mode: webpackMode,
	resolve: {
	extensions: ['.wasm', '.mjs', '.js', '.json', '.ts'],
	mainFields: ['module', 'main'],
	},
	module: {
	rules: [
	{
	test: /\.(m?js|ts)?$/,
	exclude: new RegExp(
	`(node_modules|bower_components|${testFilePattern})`,
	),
	use: {
	loader: require.resolve('babel-loader'),
	options: { ...babelOpts, babelrc: useBabelrc },
	},
	},
	],
	},
	context: dirPath,
	entry: {},
	target: 'node',
	plugins: [
	new webpack.IgnorePlugin(/vertx/),
	new webpack.DefinePlugin(defineEnv),
	],
	output: {
	path: functionsPath,
	filename: '[name].js',
	libraryTarget: 'commonjs',
	},
	optimization: {
	nodeEnv,
	},
	bail: true,
	devtool: false,
	stats: {
	colors: true,
	},
	};

The only option I thought we could have an issue with the basic merge was the module.rules however since we support a custom babelrc file I think that's not a problem. As far as I could see, we did not need the smart merge. That said, it's possible I've missed something and in that case we would need to tweak the merge strategy to handle any incompatibility issue that may arise.

The only change I can think of is that it's now not possible to disable the babel-loader, but then not sure why someone would be using netlify-lambda if they don't want webpack + babel.

By bumping the major version, I think that possible backwards incompatibility problem is reduced as most projects using this won't have issues after running a regular npm update.

[bpinto]

P.S.: I've rebased code on top of latest main.

[bpinto]

@erezrokah Sorry to ping you, but I'd love to see this PR merged :)

[bpinto]

Any news? Thank you!

[bpinto]

Any updates?

[laurentlaporte]

Why is this not merged?

[danez]

We are deprecating netlify-lambda and therefore won't need this PR anymore. Thanks for your contributions and sorry it wasn't used.