This PR builds on top of the new [JS config to CSS
migration](https://github.com/tailwindlabs/tailwindcss/pull/14651) and
extends it to support migrating _static_ plugins.
What are _static_ plugins you might ask? Static plugins are plugins
where we can statically determine that these are coming from a different
file (so there is nothing inside the JS config that creates them). An
example for this is this config file:
```js
import typographyPlugin from '@tailwindcss/typography'
import { type Config } from 'tailwindcss'
export default {
content: ['./src/**/*.{js,jsx,ts,tsx}'],
darkMode: 'selector',
plugins: [typographyPlugin],
} satisfies Config
```
Here, the `plugins` array only has one element and it is a static import
from the `@tailwindcss/typography` module. In this PR we attempt to
parse the config file via Tree-sitter to extract the following
information from this file:
- What are the contents of the `plugins` array
- What are statically imported resources from the file
We then check if _all_ entries in the `plugins` array are either static
resources or _strings_ (something I saw working in some tests but I’m
not sure it still does). We migrate the JS config file to CSS if all
plugins are static and we can migrate them to CSS `@plugin` calls.
## Todo
This will need to be rebased after the updated tests in #14648
Inside the `theme(…)` function, we can use the `/` character for
applying an opacity. For example `theme(colors.red.500 / 50%)` will
apply a 50% opacity to the `colors.red.500` value.
However, if you used a variable instead of the hardcoded `50%` value,
then this was not parsed correctly. E.g.: `theme(colors.red.500 /
var(--opacity))`
_If_ we have this exact syntax (with the spaces), then it parses, but
some information is lost:
```html
<div class="bg-[theme(colors.red.500_/_var(--opacity))]"></div>
```
Results in:
```css
.bg-\[theme\(colors\.red\.500_\/_var\(--opacity\)\)\] {
background-color: color-mix(in srgb, #ef4444 calc(var * 100%), transparent);
}
```
Notice that the `var(--opacity)` is just parsed as `var`, and the
`--opacity` is lost.
Additionally, if we drop the spaces, then it doesn't parse at all:
```html
<div class="bg-[theme(colors.red.500/var(--opacity))]"></div>
```
Results in:
```css
```
This means that we have to handle 2 issues to make this work:
1. We have to properly handle the `/` character as a proper separator.
2. If we have sub-functions, we have to make sure to print them in full
(instead of only the very first node (`var` in this case)).
Quick follow-up to #14659 base don @thecrypticace's idea:
- This behavior is no longer added to the types of the Plugin API to be
consistent with v3
- When the plugin argument is used as a function, we now warn the first
time
Something we noticed while testing the codemods on one of our codebases
is that the callback passed to the `theme` function properties doesn't
only expose some properties like `colors`, but it's also a function
itself.
```ts
/** @type {import('tailwindcss').Config} */
export default {
theme: {
extend: {
colors: (theme) => {
// The `theme` is a theme function _and_ the object...
console.log(theme('spacing.2'), theme.colors.red['500'])
return {}
},
},
},
plugins: [],
}
```
E.g.: https://play.tailwindcss.com/eV7Jgv17X1?file=config
---
h/t to @RobinMalfait for the issue description
While working on some fixes for #14639 I noticed that the following v3
configuration file would not load properly in v4:
```ts
import { type Config } from 'tailwindcss'
export default {
content: ['./src/**/*.{js,jsx,ts,tsx}'],
theme: {
extend: {
colors: ({ colors }) => ({
gray: colors.neutral,
}),
},
} satisfies Config
```
The reason for this is that we did not pass the `colors` property to the
callback function. Since we have colors available now, we can easily add
it.
---------
Co-authored-by: Adam Wathan <adam.wathan@gmail.com>
This PR adds a new codemod that can migrate `data-*` and `aria-*`
variants using arbitrary values to bare values.
In Tailwind CSS v3, if you want to conditionally apply a class using
data attributes, then you can write `data-[selected]:flex`. This
requires the DOM element to have a `data-selected=""` attribute. In
Tailwind CSS v4 we can simplify this, by dropping the brackets and by
using `data-selected:flex` directly.
This migration operates on the internal AST, which means that this also
just works for compound variants such as
`group-has-data-[selected]:flex` (which turns into
`group-has-data-selected:flex`).
Additionally, this codemod is also applicable to `aria-*` variants. The
biggest difference is that in v4 `aria-selected` maps to an attribute of
`aria-selected="true"`. This means that we can only migrate
`aria=[selected="true"]:flex` to `aria-selected:flex`.
Last but not least, we also migrate `supports-[gap]` to `supports-gap`
if the passed in value looks like a property. If not, e.g.:
`supports-[display:grid]` then it stays as-is.
A few theme keys have changed in v4 relative to v3:
- `screens` -> `--breakpoint-*`
- `colors` -> `--color-*`
- `animation` -> `--animate-*`
- `borderRadius` -> `--radius-*`
- `boxShadow` -> `--shadow-*`
When using the `theme()` function we wouldn't pick up values from the
CSS for some of these. Likewise, when loading a v3 config not all of
these would be pushed back into the CSS theme and they should've been.
This PR addresses both of these problems.
This PR implements the first version of JS config file migration to CSS.
It is based on the most simple config setups we are using in the
Tailwind UI templates Commit, Primer, Radiant, and Studio.
The example we use in the integration test is a config that looks like
this:
```js
import { type Config } from 'tailwindcss'
import defaultTheme from 'tailwindcss/defaultTheme'
module.exports = {
darkMode: 'selector',
content: ['./src/**/*.{html,js}'],
theme: {
boxShadow: {
sm: '0 2px 6px rgb(15 23 42 / 0.08)',
},
colors: {
red: {
500: '#ef4444',
},
},
fontSize: {
xs: ['0.75rem', { lineHeight: '1rem' }],
sm: ['0.875rem', { lineHeight: '1.5rem' }],
base: ['1rem', { lineHeight: '2rem' }],
},
extend: {
colors: {
red: {
600: '#dc2626',
},
},
fontFamily: {
sans: 'Inter, system-ui, sans-serif',
display: ['Cabinet Grotesk', ...defaultTheme.fontFamily.sans],
},
borderRadius: {
'4xl': '2rem',
},
},
},
plugins: [],
} satisfies Config
```
As you can see, this file only has a `darkMode` selector, custom
`content` globs, a `theme` (with some theme keys being overwriting the
default theme and some others extending the defaults). Note that it does
not support `plugins` and/or `presets` yet.
In the case above, we will find the CSS file containing the existing
`@tailwind` directives and are migrating it to the following:
```css
@import 'tailwindcss';
@source './**/*.{html,js}';
@variant dark (&:where(.dark, .dark *));
@theme {
--box-shadow-*: initial;
--box-shadow-sm: 0 2px 6px rgb(15 23 42 / 0.08);
--color-*: initial;
--color-red-500: #ef4444;
--font-size-*: initial;
--font-size-xs: 0.75rem;
--font-size-xs--line-height: 1rem;
--font-size-sm: 0.875rem;
--font-size-sm--line-height: 1.5rem;
--font-size-base: 1rem;
--font-size-base--line-height: 2rem;
--color-red-600: #dc2626;
--font-family-sans: Inter, system-ui, sans-serif;
--font-family-display: Cabinet Grotesk, ui-sans-serif, system-ui, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji";
--border-radius-4xl: 2rem;
}
```
This replicates all features of the JS config so we can even delete the
existing JS config in this case.
---------
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
This PR adds a mapping from legacy classes to new classes. For example,
the `flex-shrink-0` is still used in our projects, but is deprecated in
v3.
The migration does a tiny bit of parsing because we can't rely on
`designSystem.parseCandidate(…)` because this requires the utility to be
defined which is not the case for legacy classes.
This migration runs _after_ the migration where we handle prefixes, so
we don't have to worry about that. We do have to worry about the `!`
location, because the `important` migration also relies on the
`designSystem`.
| Old | New |
| ------------------- | ---------------------- |
| `overflow-clip` | `text-clip` |
| `overflow-ellipsis` | `text-ellipsis` |
| `flex-grow-0` | `grow-0` |
| `flex-shrink-0` | `shrink-0` |
| `decoration-clone` | `box-decoration-clone` |
| `decoration-slice` | `box-decoration-slice` |
Here is everything you need to know about this update. Please take a
good look at what changed and the test results before merging this pull
request.
### What changed?
#### ✳️ turbo (2.1.2 → 2.1.3) ·
[Repo](https://github.com/turborepo/turbo)
Sorry, we couldn't find anything useful about this release.
---
[Depfu](https://depfu.com) will automatically keep this PR
conflict-free, as long as you don't add any commits to this branch
yourself. You can also trigger a rebase manually by commenting with
`@depfu rebase`.
<details><summary>All Depfu comment commands</summary>
<blockquote><dl>
<dt>@depfu rebase</dt><dd>Rebases against your default branch and
redoes this update</dd>
<dt>@depfu recreate</dt><dd>Recreates this PR, overwriting any edits
that you've made to it</dd>
<dt>@depfu merge</dt><dd>Merges this PR once your tests are passing and
conflicts are resolved</dd>
<dt>@depfu cancel merge</dt><dd>Cancels automatic merging of this
PR</dd>
<dt>@depfu close</dt><dd>Closes this PR and deletes the branch</dd>
<dt>@depfu reopen</dt><dd>Restores the branch and reopens this PR (if
it's closed)</dd>
<dt>@depfu pause</dt><dd>Ignores all future updates for this dependency
and closes this PR</dd>
<dt>@depfu pause [minor|major]</dt><dd>Ignores all future minor/major
updates for this dependency and closes this PR</dd>
<dt>@depfu resume</dt><dd>Future versions of this dependency will
create PRs again (leaves this PR as is)</dd>
</dl></blockquote>
</details>
Co-authored-by: depfu[bot] <23717796+depfu[bot]@users.noreply.github.com>
This PR injects a `@config "…"` in the CSS file if a JS based config has
been found.
We will try to inject the `@config` in a sensible place:
1. Above the very first `@theme`
2. If that doesn't work, below the last `@import`
3. If that doesn't work, at the top of the file (as a last resort)
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>
This PR attempts to detect simple postcss setups: These are setups that
do not load dynamic modules and are based off the examples we are
[recommending in our
docs](https://tailwindcss.com/docs/installation/using-postcss). We
detect wether a config is appropriate by having it use the object plugin
config and by not requiring any other modules:
```js
module.exports = {
plugins: {
tailwindcss: {},
autoprefixer: {},
},
}
```
When we find such a config file, we will go over it line-by-line and
attempt to:
- Upgrade `tailwindcss:` to `'@tailwindcss/postcss':`
- Remove `autoprefixer` if used
We then attempt to install and remove the respective npm packages based
on the package manger we detect.
And since we now have logic to upgrade packages, this also makes sure to
install `tailwindcss@next` at the end of a sucessful migration.
---------
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
This PR adds a codemod that migrates the `@media screen(…)` to the
properly expanded `@media (…)` syntax.
```css
@media screen(md) {
.foo {
color: red;
}
}
```
Will be converted to:
```css
@media (width >= 48rem) {
.foo {
color: red;
}
}
```
If you happen to have custom screens (even complex ones), the screen
will be converted to a custom media query.
```css
@media screen(foo) {
.foo {
color: red;
}
}
```
With a custom `tailwind.config.js` file with a config like this:
```js
module.exports = {
// …
theme: {
screens: {
foo: { min: '100px', max: '123px' },
},
}
}
```
Then the codemod will convert it to:
```css
@media (123px >= width >= 100px) {
.foo {
color: red;
}
}
```
We forgot to pass `options` from `addComponents` to `addUtilities` and
from `matchComponents` to `matchUtilities`.
This didn't affect anything using addComponents but anything that used
`matchComponents` wouldn't have worked 😬
Discovered `matchUtilities(…)` wasn't populating intellisense on v4 when
working on Tailwind Play earlier today. This PR fixes this so plugins
using matchUtilities also have intellisense suggestions.
We currently have three different implementations of the
`resolveThemeValue()` Theme method:
1. The _core_ version which only resolves based on the exact CSS
variable name. This is the v4-only version.
2. A _compat light_ version which attempts to translate a dot-notation
keypath to a CSS variable name.
3. A _full compat_ version which resolves the legacy theme object which
is used whenever a `@plugin` or `@config` is added.
Unfortunately, there are some issues with this approach that we've
encountered when testing the CSS `theme()` function while upgrading
Tailwind Templates to v4 using our upgrading toolkit.
1. The _compat light_ resolver was trying to convert `theme(spacing.1)`
to tuple syntax. Tuples require a nested property access, though, and
instead this should be convert to `theme(--spacing-1)`.
2. We currently only load _full compat_ version if `@plugin` or
`@config` directives are used. It is possible, though, that we need the
full compat mapping in other cases as well. One example we encountered
is `theme(fontWeight.semibold)` which maps to a dictionary in the
default theme that that we do not have an equivalent for in v4
variables.
To fix both issues, we decided to remove the _compat light_ version of
the `theme()` function in favor of adding this behavior to an upgrade
codemod. Instead, the second layer now only decides wether it can use
the _core_ version or wether it needs to upgrade to the _full compat_
version. Since typos would trigger a _hard error_, we do not think this
has unintended performance consequences if someone wants to use the
_core_ version only (they would get an error regardless which they need
to fix in order to continue).
Fixes#14613
Don't crash when trying to read the modification time of a file that
might already be deleted.
Note: This fix is purely theoretical right now as I wasn't able to
reproduce the issue yet.
---------
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
Fixes an issue reported by the React Aria Components team here:
https://github.com/adobe/react-spectrum/issues/7160
Basically `hidden="until-found"` behaves very differently than `hidden`
and doesn't actually use `display: none`, so we don't want to apply the
behavior we apply for the regular `hidden` attribute.
This PR fixes two issues we found when testing the candidate codemodes:
1. Sometimes, core would emit the same candidate twice. This would
result into rewriting a range multiple times, without realizing that
this change might already be applied, causing it to swallow or duplicate
some bytes.
2. The codemods were mutating the `Candidate` object, however since the
`Candidate` parsing is _cached_ in core, it would sometimes return the
same instance. This is an issue especially since we monkey patch the
prefix to `null` when migrating prefixed candidates. This means that a
candidate would be cached that would be _invalid relative to the real
design system_. We fixed this by making sure the mutations would only be
applied to clones of the `Candidate` and I changed the `DesignSystem`
API to return `ReadOnly<T>` versions of these candidates. A better
solution would maybe be to disable the cache at all but this requires
broader changes in Core.
We're trialing [namespace.so](https://namespace.so/) for their custom
GitHub Action workers. So far we're seeing an improvement of 3 minutes
in our pipeline (used to take 6min 30sec and now it's 3min 30sec, an
almost 50% improvement!).
We can always revert this PR, but I recommend we try it out by enabling
it for our CI for the next few days.
This PR makes the internal `@at-root` API private. Before this PR you
could use `@at-root` in your own CSS, which means that it was part of
the public API. If you (accidentally) used it in variants, you could
generate CSS that was completely broken.
This now introduces a new private `AtRoot` node (similar to the recently
introduced `Context` node) and can only be constructed within the
framework.
In order to properly migrate your Tailwind CSS v3 project to v4, we need
access to the JavaScript configuration object. This was previously only
required for template migrations, but in this PR we're making it so that
this is also a prerequisite of the CSS migrations. This is because some
migrations, like `@apply`, also need to convert candidates that to the
v4 syntax and we need the full config in order to properly validate
them.
In addition to requiring a JS config, we also now attempt to
automatically find the right configuration file inside the current
working directory. This is now matching the behavior of the Tailwind CSS
v3 CLI where it will find the config automatically if it's in the
current directory and called `tailwind.conf.js`.
---------
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
This PR is a very small improvement. We started pretty printing the
generated CSS (proper indentation) a while ago, so that we can use the
output as-is for intellisense (on hover).
The other day I noticed that when you use `!important` that we attach it
directly to the declaration. Not the end of the world, but this PR
injects a little space to make sure that the `!important` is separated
from the value which makes it a little easier to read and looks more
like what you would write by hand.
Before:
```css
.flex\! {
display: flex!important;
}
```
After:
```css
.flex\! {
display: flex !important;
}
```
Here is everything you need to know about this update. Please take a
good look at what changed and the test results before merging this pull
request.
### What changed?
#### ✳️ bun (1.1.22 → 1.1.29) · [Repo](https://github.com/oven-sh/bun)
Sorry, we couldn't find anything useful about this release.
---
[Depfu](https://depfu.com) will automatically keep this PR
conflict-free, as long as you don't add any commits to this branch
yourself. You can also trigger a rebase manually by commenting with
`@depfu rebase`.
<details><summary>All Depfu comment commands</summary>
<blockquote><dl>
<dt>@depfu rebase</dt><dd>Rebases against your default branch and
redoes this update</dd>
<dt>@depfu recreate</dt><dd>Recreates this PR, overwriting any edits
that you've made to it</dd>
<dt>@depfu merge</dt><dd>Merges this PR once your tests are passing and
conflicts are resolved</dd>
<dt>@depfu cancel merge</dt><dd>Cancels automatic merging of this
PR</dd>
<dt>@depfu close</dt><dd>Closes this PR and deletes the branch</dd>
<dt>@depfu reopen</dt><dd>Restores the branch and reopens this PR (if
it's closed)</dd>
<dt>@depfu pause</dt><dd>Ignores all future updates for this dependency
and closes this PR</dd>
<dt>@depfu pause [minor|major]</dt><dd>Ignores all future minor/major
updates for this dependency and closes this PR</dd>
<dt>@depfu resume</dt><dd>Future versions of this dependency will
create PRs again (leaves this PR as is)</dd>
</dl></blockquote>
</details>
Co-authored-by: depfu[bot] <23717796+depfu[bot]@users.noreply.github.com>
In v3 it was possible to import `*.js` variants of compat files, e.g.
`tailwindcss/plugin.js` in addition to `tailwindcss/plugin`. We need to
properly map the exports to support this.
This PR fixes an issue where CSS that existed before a layer:
```css
.foo {
color: red;
}
@layer components {
.bar {
color: blue;
}
}
```
Was turned into an `@layer` without a name:
```css
@layer {
.foo {
color: red;
}
}
@utility bar {
color: blue;
}
```
But in this case, it should stay as-is:
```css
.foo {
color: red;
}
@utility bar {
color: blue;
}
```
When a prefix is set in a stylesheet and we found a candidate that is
equal to the prefix we would crash:
```css
@import "tailwindcss" prefix(tomato);
```
```js
console.log("tomato")
```
This PR fixes this case by ensuring that we have something that looks
like a variant before considering a prefix.
This PR allows modifying utility output by wrapping all utility
declarations in a custom selector or marking all utility declarations as
`!important`. This is the v4 equivalent to the `important` option in the
`tailwind.config.js` file.
## Mark all utility declarations as `!important`
To add `!important` to all utility declarations, you add an `important`
flag after the `@import` statement for `tailwindcss` or
`tailwindcss/utilities`:
```css
/** Importing `tailwindcss` */
@import "tailwindcss" important;
/** Importing the utilities directly */
@import "tailwindcss/utilities" important;
```
Example utility output:
```css
.mx-auto {
margin-left: auto !important;
margin-right: auto !important;
}
.text-center {
text-align: center !important;
}
```
This is equivalent to adding `important: true` to the
`tailwind.config.js` file — which is still supported for backwards
compatibility.
## Wrap all utility declarations in a custom selector
To nest all utilities in an `#app` selector you add `selector(#app)`
flag after the `@import` statement for `tailwindcss` or
`tailwindcss/utilities`:
```css
/** Importing `tailwindcss` */
@import "tailwindcss" selector(#app);
/** Importing the utilities directly */
@import "tailwindcss/utilities" selector(#app);
```
Example utility output:
```css
.mx-auto {
#app & {
margin-left: auto;
margin-right: auto;
}
}
.text-center {
#app & {
text-align: center;
}
}
```
This is equivalent to adding `important: "#app"` to the
`tailwind.config.js` file — which is still supported for backwards
compatibility.
**This _does not_ bring back support for the `respectImportant` flag in
`addUtilities` / `matchUtilities`.**
This PR improves the performance of the `@tailwindcss/postcss` plugin.
Before this change we created 2 compiler instances instead of a single
one. On a project where a `tailwindcss.config.ts` file is used, this
means that the timings look like this:
```
[@tailwindcss/postcss] Setup compiler: 137.525ms
⋮
[@tailwindcss/postcss] Setup compiler: 43.95ms
```
This means that with this small change, we can easily shave of ~50ms for
initial PostCSS builds.
---------
Co-authored-by: Philipp Spiess <hello@philippspiess.com>
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
This PR extracts all _candidate migrations_ from the existing _template
migrations_ and reuses these in the `@apply` CSS migration. Seems like
this _JustWorks✨_.
---------
Co-authored-by: Adam Wathan <adam.wathan@gmail.com>
Part-of #14558
After handling `@import` of stylesheets in core, we can no longer gate
out features based on finding a specific sequence in the input CSS, as
this string will not contain contents from other stylesheets.
So, consider the following input CSS:
```css
@import "tailwindcsss";
@import "./styles.css";
```
We can't opt-out of the traversal to search for `@apply` based on it
because it, of course, does not contain the contents of `./styles.css`
yet.
---------
Co-authored-by: Adam Wathan <adam.wathan@gmail.com>
One of the breaking changes of v4 is the [inversion of variant order
application](https://github.com/tailwindlabs/tailwindcss/pull/13478). In
v3, variants are applied "inside-out".
For example a candidate like `*:first:underline` would produce the
following CSS in v3:
```css
.\*\:first\:underline:first-child > * {
text-decoration-line: underline;
}
```
To get the same behavior in v4, you would need to invert the candidate
order to `first:*:underline`. This would generate the following CSS in
v4:
```css
:where(.first\:\*\:underline:first-child > *) {
text-decoration-line: underline;
}
```
## The Migration
The most naive approach would be to invert the variants for every
candidate with at least two variants. This, however, runs into one issue
and some unexpected inconsistencies. I have identified the following
areas:
1. Some pseudo class variants _must appear at the end of the selector_.
v3 was patching over this by doing some manual reordering in for these
variants. For example, in v3, both of these variants create the same
output CSS: `hover:before:underline` and `before:hover:underline`. In v4
we simplified this system though and no longer generate the same output
in both cases. Instead, you'd always want to write
`hover:before:underline`, ensuring that these variants will appear at
the end.
For an exact list of which variants these affect, take a look [at this
diff](https://github.com/tailwindlabs/tailwindcss/pull/13478/files#diff-7779a0eebf6b980dd3abd63b39729b3023cf9a31c91594f5a25ea020b066e1c0L228-L246).
2. The `dark` variant and other at-rule variants are usually written
before other variants. This is more of a recommendation to make it
easier to read candidates rather than a difference in behavior as
`@media` queries are hoisted by the engine. For this reason, both of
these variants are _correct_ yet in real applications we prefer the
first one: `lg:hover:underline`, `hover:lg:underline`.
To avoid shuffling these rules across all candidates during the
migration, we bucket `dark` and other at-rule variants into a special
bucket that will not have their order changed (since people wrote stacks
like `sm:max-lg:` before and we want to keep them as-is) and appear
before all other variants.
3. For some variant stacks, the order does not matter. E.g.:
`focus:hover:underline` and `hover:focus:underline` will be the same. We
don't want to needlessly shuffle their order if we have to.
With these considerations, the migration now works as follows:
- If there is less then two variants, we do not need to migrate the
candidate
- If _every_ variant in the stack is an order-independent variant, we do
not need to migrate the candidate
- _Note that this is currently hardcoded to only support `&:hover` and
`&:focus`._
- Otherwise, we loop over the candidates and put them into three
buckets:
- `mediaVariants` hold variants that only contribute `@media` rules
_and_ the `dark` variant.
- `pseudoElementVariants` hold variants that _must appear at the end of
the selector_. This is based on the allow list from v3/early v4.
- `regularVariants` contains the rest.
- We now compute if any of the variants inside `regularVariants` is
order dependent.
- With this list of variants, we now construct the new order of variants
as:
```ts
[
...atRuleVariants,
...(anyRegularVariantOrderDependent ? regularVariants.reverse() :
regularVariants),
...pseudoElementVariants,
]
```
---------
Co-authored-by: Adam Wathan <adam.wathan@gmail.com>
