fix: correctly scope CSS selectors with descendant combinators

[Rich-Harris]

Ok, I've been on a journey.

It turns out our current logic for scoping CSS selectors with descendant combinators is flawed. We're also adding scoping classes in places where it shouldn't be necessary (though that's still true, to an extent, in this PR — might be good follow-up work).

Consider a case like this, where the <Y> component consists of <y><slot /></y>:

<script>
  import Y from './Y.svelte';
</script>

<x>
  <y>
    <z>this should be red</z>
  </y>
</x>

<x>
  <Y>
    <z>this should not be red</z>
  </Y>
</x>

<style>
  x y z {
    display: block;
    color: red;
  }
</style>

Currently, we turn that into x.blah y z.blah {...} and apply the corresponding classes to <x> and <z>. But that means we incorrectly target the second <z> element.

Fixing that without breaking anything else turned out to be hellaciously complicated. All the tests are passing now, but the code could probably use a bit of tidying up.

Before submitting the PR, please make sure you do the following

• It's really useful if your PR references an issue where it is discussed ahead of time. In many cases, features are absent for a reason. For large changes, please create an RFC: https://github.com/sveltejs/rfcs
• Prefix your PR title with feat:, fix:, chore:, or docs:.
• This message body should clearly illustrate what problems it solves.
• Ideally, include a test that fails without this PR but passes with it.

Tests and linting

• Run the tests with pnpm test and lint the project with pnpm lint

[changeset-bot]

🦋 Changeset detected

Latest commit: 7a79279

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

This PR includes changesets to release 1 package

Name	Type
svelte	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