Use Vitest instead of Jest

[siketyan]

Closes #263

This pull request replaces Jest with Vitest.

At the first try, I found that Vitest does not resolve await-lock module correctly.
The package is not maintained recently and has an issue with ESM interop.
I replaced it with async-mutex to resolve the issue.

[mizdra]

Thank you for doing this work! I will accept and use this Pull Request.

At the first try, I found that Vitest does not resolve await-lock module correctly.
The package is not maintained recently and has an issue with ESM interop.
I replaced it with async-mutex to resolve the issue.

I don't object to replacing it with async-mutex. However, I don't think that the package is not maintained. The maintainer replied to the issue comment in August 2023. It seems that the package is maintained.

[mizdra]

Should we also delete the following line?

happy-css-modules/eslint.config.js

Line 90 in b0cafcd

jest: true,

[siketyan]

@mizdra
Sorry for my miswording. I meant that await-lock is not shipped with native ESM support, breaking the type definition of the default import.

[mizdra]

When I run npm run test -- runner.test.ts, Vite seems to output the following warning

$ npm run test -- runner.test.ts
...
1:17:09 AM [vite] Failed to load source map for /Users/mizdra/src/github.com/mizdra/happy-css-modules/packages/happy-css-modules/src/runner.test.ts.
Error: An error occurred while trying to read the map file at ./1.css.d.ts.map
Error: ENOENT: no such file or directory, open '/Users/mizdra/src/github.com/mizdra/happy-css-modules/packages/happy-css-modules/src/1.css.d.ts.map'
    at async open (node:internal/fs/promises:638:25)
    at async Object.readFile (node:internal/fs/promises:1238:14)
    at async extractSourcemapFromFile (file:///Users/mizdra/src/github.com/mizdra/happy-css-modules/node_modules/vite/dist/node/chunks/dep-BWSbWtLw.js:21001:53)
    at async loadAndTransform (file:///Users/mizdra/src/github.com/mizdra/happy-css-modules/node_modules/vite/dist/node/chunks/dep-BWSbWtLw.js:51892:27)
    at async ViteNodeServer._transformRequest (file:///Users/mizdra/src/github.com/mizdra/happy-css-modules/node_modules/vite-node/dist/server.mjs:537:16)
    at async ViteNodeServer._fetchModule (file:///Users/mizdra/src/github.com/mizdra/happy-css-modules/node_modules/vite-node/dist/server.mjs:499:17)
    at async Proxy.fetch (file:///Users/mizdra/src/github.com/mizdra/happy-css-modules/node_modules/vitest/dist/chunks/resolveConfig.DPmbhVlP.js:6616:22)
    at async EventEmitter.onMessage (file:///Users/mizdra/src/github.com/mizdra/happy-css-modules/node_modules/vitest/dist/chunks/index.68735LiX.js:91:20)
...

Apparently, Vitest found the following line and misunderstood it to indicate the source map for runner.test.ts. I believe this is a bug in Vitest 😆.

happy-css-modules/packages/happy-css-modules/src/runner.test.ts

Line 371 in 41ce48c

//# sourceMappingURL=./1.css.d.ts.map

The following line seems to be causing the bug. It should be /^\/\/[@#] instead of /\/\/[@#].

https://github.com/vitest-dev/vitest/blob/85c64e35b120f240e66d5fb9ca67fa55ed8eb74d/packages/vite-node/src/source-map-handler.ts#L116

[mizdra]

Let's ignore the Vitest bug here. If you are interested in reporting this bug to Vitest team, please do so. If you are not interested, I will do it for you.

[mizdra]

LGTM! Thank you for contributing!

[mizdra]

The following line seems to be causing the bug. It should be /^//[@#] instead of ///[@#].

This may be my misunderstanding. We need to check https://tc39.es/source-map/ to determine if this is a bug.

[siketyan]

@mizdra I checked the spec and found that whitespaces before the sourceMappingURL comment are allowed, as defined in https://tc39.es/source-map/#extraction-javascript. Thus, I think it's an expected behaviour and we cannot declare this as a Vitest's bug.

[siketyan]

And, this example is similar to our situation: https://tc39.es/source-map/#example-5b7659a8. This type of code can be parsed as unintentional even if the regexp starts with ^, unless Vitest parses it using AST.