* improve split logic by delimiter
The original RegEx did mostly what we want, the idea is that we wanted
to split by a `,` but one that was not within `()`. This is useful when
you define multiple background colors for example:
```html
<div class="bg-[rgb(0,0,0),rgb(255,255,255)]"></div>
```
In this case splitting by the regex would result in the proper result:
```js
let result = [
'rgb(0,0,0)',
'rgb(255,255,255)'
]
```
Visually, you can think of it like:
```
┌─[./example.html]
│
∙ 1 │ <div class="bg-[rgb(0,0,0),rgb(255,255,255)]"></div>
· ──┬── ┬ ─────┬─────
· │ │ ╰─────── Guarded by parens
· │ ╰───────────────── We will split here
· ╰───────────────────── Guarded by parens
│
└─
```
We properly split by `,` not inside a `()`. However, this RegEx fails
the moment you have deeply nested RegEx values.
Visually, this is what's happening:
```
┌─[./example.html]
│
∙ 1 │ <div class="bg-[rgba(0,0,0,var(--alpha))]"></div>
· ┬ ┬ ┬
· ╰─┴─┴── We accidentally split here
│
└─
```
This is because on the right of the `,`, the first paren is an opening
paren `(` instead of a closing one `)`.
I'm not 100% sure how we can improve the RegEx to handle that case as
well, instead I wrote a small `splitBy` function that allows you to
split the string by a character (just like you could do before) but
ignores the ones inside the given exceptions. This keeps track of a
stack to know whether we are within parens or not.
Visually, the fix looks like this:
```
┌─[./example.html]
│
∙ 1 │ <div class="bg-[rgba(0,0,0,var(--alpha)),rgb(255,255,255,var(--alpha))]"></div>
· ┬ ┬ ┬ ┬ ┬ ┬ ┬
· │ │ │ │ ╰───┴───┴── Guarded by parens
· │ │ │ ╰────────────────── We will split here
· ╰─┴─┴──────────────────────────────── Guarded by parens
│
└─
```
* use already existing `splitAtTopLevelOnly` function
* add faster implemetation for `splitAtTopLevelOnly`
However, the faster version can't handle separators with multiple
characters right now. So instead of using buggy code or only using the
"slower" code, we've added a fast path where we use the faster code
wherever we can.
* use `splitAtTopLevelOnly` directly
* make split go brrrrrrr
* update changelog
* remove unncessary array.from call
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
* Remove remnants of the user layer
It hasn’t been used in a while
* Rewrite sort offset generation
* wip
* wip
wip
* Handle parasite utilities
* wip
* wip
* Make parallel variants sorting more resillient
It’s not perfect but it’s close
* fix
* remove todo
it adds a new bit so it can’t
* Simplify getClassOrder usage
* Simplify
oops
oops
* Add parasite utility for `dark`
dark mode class name
* Cleanup
* Cleanup
* Simplify
* format files
* Fix prettier plugin to use git build of Tailwind CSS
Symlink and build instead of adding a recursive dev dependency
It breaks node < 16
* Fix prettier error
* wip
* fix test
* Update changelog
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
* Fix issue with Tailwind modifying global state
When running Tailwind, it modifies the plugin defaults parameters. As a
result Tailwind using a Tailwind plugin in the same process twice yields
different results.
* Add failing test
* Undo defaults change
* wip
* Fix shared mutation problem
* Update changelog
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
* Don’t re-add files in the CLI watcher that are covered by dynamic patterns
They don’t have the same problem. As long as the parent directory is watched their add/change events will fire correctly
* Ignore raw events for files that don’t match the content files
* fixup
* Update changelog
It’s no longer necessary. If we have an entry in the `candidateRuleCache` then it’ll also be in the class cache and vice-versa. Also, we weren’t adding rules when hitting that cache like we should’ve been.
* Fix CLI not watching atomically renamed files
Chokdar should take care of this itself but sometimes it doesn’t do so OR is otherwise very sensitive to timing problems
* Force chokidar to always check for atomic writes
* Handle repeated atomic saves by retrying file reads
* Update changelog
* Honor the `hidden` attribute on elements
You’ll still be able to override this with utilities but this ensures that things like `<iframe hidden>` work as expected
* Update changelog
* Add support for configuring default font-feature-settings for a font-family
* Update changelog
Co-authored-by: Adam Wathan <4323180+adamwathan@users.noreply.github.com>
* Fix application of rules with multiple matches of differing selectors
`-foo-1` and `foo-1` are both matches for the class `-foo-1` but `@apply` only wants the first one. It would remove the second one and cause an error because it’s an entirely separate match that had it’s only rule removed.
* Update changelog
* Don't use `cursor: pointer` for buttons by default
This is a pretty common expectation but as outlined in in #8961 isn't really right. We considered this a long time ago but it felt too against the grain at the time. These days though very UI-forward applications like [Linear](https://linear.app/) are using the default cursor for buttons and I think this trend will continue as more people become aware that `cursor: pointer` is meant for links.
Let's update our defaults here to help nudge people in this direction and make it more common. If people want to change this in their own apps, it's just a line or two of CSS to add to their projects.
* Update changelog
Co-authored-by: Adam Wathan <4323180+adamwathan@users.noreply.github.com>
* Refactor
* Support variants with quotes in them
We have to have two regexes here because this is actually ambiguous in the general case. The regex that generally handles `[&[foo='bar']]` would incorrectly match `['bar':'baz']` for instance. So, instead we’ll use multiple regexes and match both!
* Update changelog
* add tests for spaced around operators in CSS math functions
* fix CSS math functons besides calc not getting the love they deserve
* improve comment
* update changelog
* update changelog
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
* Fix issue with returning postcss nodes in addVariant
It’s not a supported use case but it didn’t use to break so let’s just fail silently
* Update changelog
* Add support for postcss-import in watch mode
* Add regression test
* Extract shared logic
* restructure test a little bit
Instead of relying on a arbitrary setTimout value, let's wait for the
file to be created instead.
* update changelog
Co-authored-by: Adam Bergman <adam@fransvilhelm.com>
