Fix odd v4 behavior with custom @-rules

[tizu69]

Fixes #1211

I'm not sure if the .replace I removed actually had a purpose - it seemed useless to me.
I can (probably) easily bring it back if it does have a purpose, gonna have to see how to fix it in that case hten.

Also adds a QOL feature of now autosuggesting [*] in --value(.

See the issue for way more details on why some things are the way they are.

[thecrypticace]

The .replace is necessary but it's matching too much. I've fixed it up. Will get this merged in the am after some more testing.

[tizu69]

The .replace is necessary but it's matching too much. I've fixed it up. Will get this merged in the am after some more testing.

out of curiosity, what it is used for? I couldn't find anything that that would catch (or is supposed to catch).

[philipp-spiess]

@tizu69 One is for the shorthand syntax, e.g.:

@custom-variant dark (&:where(.dark, .dark *));

and one is for the syntax the block syntax:

@custom-variant dark {
  &:where(.dark, .dark *) {
    @slot;
  }
}

You can see that one regex contains the semicolon and one an opening-brace 👍

[tizu69]

ah, I see. I wasn't familiar with that first syntax, fair enough.

[thecrypticace]

Also adds a QOL feature of now autosuggesting [*] in --value(.

I'm going to land this and the syntax error fix in a separate PR. I wanna get a fix out for the @custom-variant specific issue today.