Use MD5 hash for LQIP class name generation

[jaroel]

Fixes #2441

[changeset-bot]

🦋 Changeset detected

Latest commit: 625eff5

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@responsive-image/build-utils	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[simonihmig]

Thanks for working on this!

Code looks good, but there are some CI failures to fix:

• a weird eslint error, but seems related to import order
• some test expectations here
• also found some vitest snapshots in the vite-plugin and webpack packages' tests referring to ri-dyn-* classes which likely need to be updated

You can run the test suites of a package by calling pnpm turbo test.vitest inside the package folder. Let me know if you need help fixing those!

[jaroel]

@simonihmig Things should work better now.

That getTokensOrCommentsAfter traceback is tracked here import-js/eslint-plugin-import#3227

[simonihmig]

@jaroel thanks for updating the PR! After running CI again, some of the class names in the vitest snapshots are not matching the expected ones. I wonder, are the tests passing locally for you? Or do we have an "environment issue" maybe, that the hashes generated on one machine (your local one) don't match the ones on another machine (GH CI runners)?

[jaroel]

@simonihmig Using hex digest should be more stable across machines :)

[jaroel]

@simonihmig It would be great if you could try to run them tests locally on your machine.

There's something really weird going on where I can't run some webpack tests related to

responsive-image/packages/webpack/tests/index.test.ts

Line 83 in e6c40a5

expect(stats.modules![0]?.modules).toHaveLength(2);

failing with AssertionError: Target cannot be null or undefined..

[simonihmig]

There's something really weird going on where I can't run some webpack tests

I was able to reproduce, but also on main branch! So those are not related to your PR. Seems something is wrong with the (turborepo) setup, likely those tests didn't even run in CI since the regression was introduced. 🙈

I'll need to investigate this!

For your PR, I still see some snapshot mismatches in the vite tests, both in CI and locally. Do the test pass for you locally?

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[jaroel]

@simonihmig Yep, locally they work just fine.
The resource we're hashing is on my machine /Users/roel/Develop/responsive-image/packages/vite-plugin/tests/fixtures/image.jpg, so that's something.