We have a migration that adds the `layer(…)` next to the `@import`
depending on the order of original values. For example:
```css
@import "tailwindcss/utilities":
@import "./foo.css":
@import "tailwindcss/components":
```
Will be turned into:
```css
@import "tailwindcss":
@import "./foo.css" layer(utilities):
```
Because it used to exist between `utilities` and `components`. Without
this it would be _after_ `components`.
This results in an issue if an import has (deeply) nested `@utility`
at-rules after migrations. This is because if this is generated:
```css
/* ./src/index.css */
@import "tailwindcss";
@import "./foo.css" layer(utilities);
/* ./src/foo.css */
@utility foo {
color: red;
}
```
Once we interpret this (and thus flatten it), the final CSS would look
like:
```css
@layer utilities {
@utility foo {
color: red;
}
}
```
This means that `@utility` is not top-level and an error would occur.
This fixes that by removing the `layer(…)` from the import if the
imported file (or any of its children) contains an `@utility`. This is
to ensure that once everything is imported and flattened, that all
`@utility` at-rules are top-level.
When a stylesheet is imported with `@import “…” layer(utilities)` that
means that all classes in that stylesheet and any of its imported
stylesheets become candidates for `@utility` conversion.
Doing this correctly requires us to place `@utility` rules into separate
stylesheets (usually) and replicate the import tree without layers as
`@utility` MUST be root-level. If a file consists of only utilities we
won't create a separate file for it and instead place the `@utility`
rules in the same stylesheet.
Been doing a LOT of pairing with @RobinMalfait on this one but I think
this is finally ready to be looked at
---------
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
