* optimize handling of RegEx parser results
Previous:
- Copy `results`, for every subsequent result of other `patterns`
- Loop over results to filter out `undefined` values
- Loop over results to map to `clipAtBalancedParens`
Current:
- For each candidate, push the `clipAtBalancedParens(candidate)` into
the `results`
This way we are not copying existing results, and we are also avoiding
additional loops over the entire array to filter out `undefined` values
and map to `clipAtBalancedParens`.
* do not allow `]` in the first part of arbitrary properties
```
[foo:bar]
─┬─
└── This part cannot contain `]`
```
This is also a very targeted fix for when the arbitrary properties seem
to match a large piece of text, but shouldn't
* add real world tests for parsing candidate strings
* sync package-lock.json
* update changelog
* Skip over classes inside `:not(…)` when nested in an at-rule
When defining a utility we skip over classes inside `:not(…)` but we missed doing this when classes were contained within an at-rule. This fixes that.
* Update changelog
* Cleanup test
Strings are going to actually do something now
* Refactor plugin loading in `resolveConfig`
This should keep the same instance of a plugin in use too. Maybe even reduce memory usage?! idk we’ll see
* Support simple plugin configs
* Update types
* Update changelog
* Refactor
* Refactor
* Batch content file reads in Node into groups of 500
We shouldn’t need to do this for our Rust code because it utilizes Rayon’s default thread pool for parallelism. This threadpool has roughly the number of cores as the number of available threads except when overridden. This generally is much, much lower than 500 and can be explicitly overridden via an env var to work around potential issues with open file descriptors if anyone ever runs into that.
* Fix sequential/parallel flip
* Update changelog
* Extend current preflight `html` styles to support other root/host scopes
Supports SVG root scope, ShadowDOM, and presumably other root scopes
* Replace `:root` with `html`
* Update tests
* Update changelog
---------
Co-authored-by: Jonathan Reinink <jonathan@reinink.ca>
* Correct formatting for inconsistent comments
Updated the formatting for several comments to be consistent with the rest of the document.
I was incorporating the preflight in a non-Tailwind project and noticed the difference in comment styles.
* add missing period
---------
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
* feat(preflight): simplify sans-serif font stack
`-apple-system` and `BlinkMacSystemFont` were historically needed for
IE11 and chakra-based Edge (The one that wasn't chromium-based).
https://caniuse.com/font-family-system-ui has more details around it.
* further simplify `font-family`
* update tests
* update changelog
---------
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
* parse the `calc()`-like expressions and format them
* update changelog
* Add test case for double negatives
wanted to be sure this worked
---------
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
When you have a custom config for the prefix/separator values, then we
want to disable the `oxideParser` for now. But we also have to ensure
that if you use the value `"all"` for the experimental feature flag that
we enable all feature flags.
We were using the keys of the array instead of the values.
* make main plugin async
This way we can improve the `fs.readFileSync` to a bunch of
`fs.promises.readFile` in a `Promise.all` instead.
* make CLI plugin async
* update CHANGELOG
* Refactor
* Sort based on first occurence of a candidate
This primarily affects components and utilities which contain multiple matched classes
* Simplify
* Update changelog
* Update
* Refactor
* Don’t prefix classes in arbitrary values for group and peer
* use `foo` instead of `lol`
* handle the prefix inside the group/peer variants
Then add the `NoPrefix` feature to the variant itself, which will skip
prefixing any other class in the generated selector (because we already
took care of prefixing `.group` and `.peer`).
We are using an internal symbol such that:
- We can keep it as a private API
- We don't introduce a breaking change
* refactor to simple object instead
We will still use a symbol as an internal/private marker, but the data
itself will be a simple object for now.
If we want to refactor this (and more) in the future using bitflags then
we can refactor that in a separate PR.
---------
Co-authored-by: Robin Malfait <malfait.robin@gmail.com>
* always prepend the Tailwind License
Even when you are not using preflight.
* drop the Tailwind License from the preflight plugin
* drop the license in the tests
This allows us to focus on the actual generated CSS without the license.
The license also includes the version number which we would have to
update every time we release a new version.
* update the source maps snapshots
* always prepend the license when using the CLI
* update Tailwind CLI integration test
* use `defaults` for `browserslist`
+ some features that we require for Tailwind to work.
* simplify Lightning CSS features
Always transpile `Nesting`, never transpile `LogicalProperties`.
* bump caniuse-lite (`npx browserslist@latest --update-db`)
* reflect changes in the tests
* update changelog
* exclude `Features.LogicalProperties` lightningcss feature
We don't want this to automatically apply since we have dedicated
logical property based utilities already.
We also have a feature flag to toggle some utilities to make use of
logical property based ones.
This feature should be opt-in to prevent backwards compatibility (e.g.:
RTL sites that _did_ use a mr-4 explicitly).
* update default browserslist
* explicitly include/exclude Lightning CSS features
* update tests based on new browserslist and include/exclude features from Lightning CSS
* update integration tests
* prefer user browserslist, fallback to built-in browserslist
* always include Nesting
If a custom browserslist config is used, then we don't explicitly set
the include/exclude features from Lightning CSS (except for nesting, we
always want to process nesting)
* ensure to fallback to the current working directory
* update changelog
* add failling test using array syntax without space delimiter
* add Rust test with candidates in array without spaces
* Fix JS arrays without spaces
* make the `oxideParser` the default
* sync feature flags with reality
* use better example in tests to be more real
* skip failing tests in the RegEx parser
* update changelog
* make `clippy` happy
---------
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
* WIP
* use `parse` instead of `defaultExtractor`
* skip `Vue` describe block
* add a few more dedicated arbitrary values/properties tests
* use parallel parsing
* splitup Vue tests
* add some Rust specific tests
* setup parse candidate strings test system
These tests will run against the `Regex` and `Rust` based parsers. We
have groups of classes of various shapes and forms + variants and
rendered in various template situation (plain, html, Vue, ...)
+ enable all skipped tests
* ensure we also validate the classes with variants
The classes with variants are built in the `templateTable` function, so
we get them out again by using the potional arguments of the `test.each`
cb function.
* cleanup test suite
* add "anti-test" tests
To make sure that we are _not_ parsing out certain values given a
certain input.
* Add ParseAction enum
* Restart parsing following an arbitrary parse failure
* Split variants off before validating the uility part
* Collapse candidate from the end when validation fails
* Support `<`, and `>` in variant position
* fix error
* format parser.rs
* Refactor
* Update editorconfig
* wip
* wip
* Refactor
* Refactor
* Simplify
* wip
* wip
* wip
* wip
* wip
* wip
* wip
* run `cargo clippy --fix`
* run `cargo fmt`
* implement `cargo clippy` suggestions
These were not applied using `cargo clippy --fix`
* only allow `.` in the candidate part when surrounded by 0-9
This is only in the candidate part, not the arbitrary part.
* % characters can only appear at the end after digits
* > and < should only be part of variants (start OR end)
It can technically be inside the candidate when we have stacked
variants:
```
dark:<sm:underline
dark:md>:underline
```
* handle parsing utilities within quotes, parans or brackets
* mark `pt-1.5` as an expected value sliced out from `["pt-1.5"]`
* Add cursor abstraction
* wip
* disable the oxideParser if using a custom `prefix` or `separator`
* update tests
* Use cursor abstraction
* Refactor more code toward use of global cursor
* wip
* simplify
* Simplify
* Simplify
* Simplify
* Cleanup
* wip
* Simplify
* wip
* Simplify
* Handle candidates ending with % sign
* Tweak code a bit
* fmt
* Simplify
* Add cursor details to trace
* cargo fmt
* use preferred `zoom-0.5` name instead of `zoom-.5`
* drop over-extracted utilities in oxide parser
The RegEx parser does extract `underline` from
```html
<div class="peer-aria-[labelledby='a_b']:underline"></div>
```
... but that's not needed and is not happening in the oxide parser
This means that we have to make the output check a little bit different
but they are explicit based on the feature flag.
* allow extracting variants+utilities inside `{}` for the oxide parser
* characters in candidates such as `group-${id}` should not be allowed
* do not extract any of the following candidate `w-[foo-bar]w-[bar-baz]`
* ensure we can consume the full candidate and discard it
* Add fast skipping of whitespace
* Use fast skipping whenever possible
* Add fast skipping to benchmark
* Hand-tune to generate more optimized assembly
* Move code around a bit
This makes sure all the fancy SIMD stuff is as early as possible. This results in an extremely minor perf increase.
* Undo tweak
no meaningful perf difference in real world scenarios
* Disable fast skipping for now
It needs to be done in a different spot so it doesn’t affect how things are returned
* Change test names
* Fix normalize config error
* cleanup a bit
* Cleanup
* Extract validation result enum
* Cleanup comments
* Simplify
* Fix formatting
* Run clippy
* wip
* add `md>` under the special characters test set
---------
Co-authored-by: Adam Wathan <4323180+adamwathan@users.noreply.github.com>
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
* bump `lightningcss`
* allow for non-standard selector combinators such as `::deep`
* just use Prettier when comparing CSS results
We were using 2 different engines where the stable one was not using
Lightning CSS and the Oxide one was using Lightning CSS.
To ensure that we didn't have to rewrite every single test expectation,
the `toMatchFormattedCss` parsed both the actual and expected value
using Lightning CSS (to make the result similar), then it used Prettier
to make it... pretty.
Right now we _only_ use Lightning CSS, which means that we can drop the
additional lightningcss format step and just use Prettier on both the
actual and expected values.
Pretty also only prettifies the CSS, it doesn't rewrite it. E.g.:
`@media (min-width: 768px)` will not be optimized to `@media (width >=
768px)`, that's something that Lightning CSS does for us.
This will require some changes in our test output, but it will be
consistent afterwards because there won't be hidden transformation steps
anymore. Because up until now it could be that the actual result was
`color: black` but the tests showed `color: #000` (because it is
shorter).
This change will reflect reality.
* update tests based on previous commit
* only use `toMatchFormattedCss` instead of `toMatchCss`
They both do the exact same thing right now. While `toMatchCss` is
shorter, `toMatchFormattedCss` makes a bit more sense since we are
comparing the Prettier results.
* update integration tests
* handle custom `transformer` and `extractor` implementations
Right now the Rust based parser can't work with custom `transfomers` or
`extractors`.
In a perfect world we can implement as many custom parsers/extractors in
Rust such that we don't need this at all. In an almost perfect world we
can pass the transformer and extractor to the Rust based parser and call
the callback functions to handle all of this. This is probably what we
are going to do in the future but this requires more work to make sure
that:
1. It works as expected
2. Doesn't result in (major) performance issues
Since it currently doesn't work with the Rust based parser, we can
implement a fix for this in the meantime before we reach the "perfect"
solution.
One solution to this problem is to check if we do have a custom
transformer or a custom extractor and if we do, then we can bail on the
Rust parser completely and just use the current regex based parser.
An alternative solution, the solution implemented here, is that we group
the `changedContent` into 2 buckets. The bucket where we rely on the
default transformer and extractor and a bucket where a custom
transformer or extractor is used.
Then, the bucket where we use the default transformer and extractor can
still rely on the way faster Rust based parser. For the other bucket we
fallback to the regex based parser.
The nice part about this is that we can use both parsers at the same
time, and the majority of the use cases should use the faster Rust based
parser.
* update changelog
* only try to pretty print `err.source` if it exists
* fix small typo
* drop unnecessary path resolving
* add `from` and `to` to `run` test helper
* enable all source map tests
* ensure source maps are generated properly
* update annotation values in tests
* handle `map: true`
* add test for `map: { inline: true }`
* force Lightning CSS to handle an inline source map
* base64 encode the test name
This will get rid of special characters that can cause issues /
confusion when debugging tests.
* ensure utilities and components with variants are pointing to the correct source
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
* update changelog
---------
Co-authored-by: Jordan Pittman <jordan@cryptica.me>
* Add `has-*` variants for `:has(...)` pseudo-class
* Update changelog
* Fix mistake in test
---------
Co-authored-by: Adam Wathan <4323180+adamwathan@users.noreply.github.com>
