esbuild-loader VS esprima

i see. there are a loaders like https://github.com/privatenumber/esbuild-loader that works with webpack but i haven't given it a try yet.

you can use it with webpack didn't try it myself

If you are a webpack user and have heard about esbuild speed, you may start questioning your js-bundler choices. Luckily, you don't have to drop your hand-crafted webpack config just now. Thanks to esbuild-loader, you can get part of the speed improvement without doing a whole migration.

esbuild-loader (secret sauce!) fork-ts-checker-webpack-plugin (helps us with typechecking) nodemon-webpack-plugin (We can also use webpack-dev-server)

For people already on webpack, there's esbuild-loader (https://github.com/privatenumber/esbuild-loader)

esbuild is a JS and TS bundler that promises ultra-fast build times. We use webpack, and there is support to leverage esbuild with esbuild-loader.

The first thing I did was replace babel-loader and Terser (minification tool) with esbuild-loader. This made our JS compile around 12 times faster, it went down to 1.4 seconds. It was a good start.

In fact, the create-react-app-esbuild is only an encapsulation of esbuild-loader.

I just started using this: privatenumber/esbuild-loader: ⚡️ Speed up your Webpack build with esbuild (github.com)

Both would be good to file. I’m happy to investigate myself if you don’t have the time.

One thing to be aware of is that some other tools that integrate esbuild do so incorrectly. For example, the 3rd-party integration of esbuild into Webpack mis-configures esbuild in a way that causes issues with JSX: https://github.com/privatenumber/esbuild-loader/pull/107. This isn’t a problem with esbuild itself.

Focusing again on ESLint, the parser used by the linter is called Espree. This is an in-house parser built by the ESLint folks to fully support ECMAScript 6 and JSX on top of the already existing Esprima. The Espree module provide APIs for both tokenization and parsing that you can easily test out.

For TypeScript we have used AST transforms from their compiler API, and for plain JavaScript we did a similar thing using ESPrima. This helped us implement some simple optimizations like stream fusion (combining .filter and .map into a single operation) or avoiding extra object allocations in vector math, which led to nice performance improvements in code that does heavy computation (we process large amounts of data on the server and store results of physics simulations).

I used ESPrima, but any parser would do in this case. I then wrote a simple function to extract all "atomic" non-boolean expressions from it.

AST is an acronym for Abstract Syntax Tree. It's the way to represent code in a format that tools can understand. Usually in form of tree data structure. We will use AST in the format of an Esprima, which is a JavaScript parser that outputs AST.

esprima

Create an AST: Esprima

The thread was originally about CRA vs Vite size on disk (or implicitly, if we're applying it to real world applications, network cost in CI job startup times). And like I said, surrogate pairs don't apply to ASCII.

See this[0] for reference. Note how the first byte must fall within a certain range in order to signal being a surrogate pair. This fact is taken advantage of by JS parsers to make parsing of ASCII code faster by special casing that range, since checking for a valid character in the entire unicode range is quite a bit more expensive[1].

[0] https://github.com/jquery/esprima/blob/0911ad869928fd218371b...

[1] https://github.com/jquery/esprima/blob/0911ad869928fd218371b...

If you want to learn more about parsing, reading the code of an actual recursive parser might be a better idea. Esprima is a decent place to start if you're interested in JS grammar. Then you can look at the babel handbook to learn more about AST transformations. From there, the literature gets quite a bit more heavy. If you get this far and are willing to push further, you'll probably want to grab yourself a copy of the dragon book at a minimum.