Previously, some StyleValues created a large number of intermediate
strings during serialization. Passing a StringBUilder into the
serialization function allows us to avoid a large number of these
unnecessary allocations.
The color stop members were RefPtr which compares by pointer identity.
This caused false positives in style invalidation when the same gradient
color values were recomputed as different objects.
The x and y members were RefPtr which compares by pointer identity.
This caused false positives in style invalidation when the same values
were recomputed as different objects.
The member type was NonnullRefPtr which compares by pointer identity.
This caused false positives in style invalidation when the same values
were recomputed as different objects.
The member types were NonnullRefPtr which compares by pointer identity.
This caused false positives in style invalidation when the same color
values were recomputed as different objects.
`EdgeStyleValues` which consist of an offset of a `calc()`s which
resolves to 50% should be considered "centered" for
`SerializationMode::ResolvedValue` for the purpose of omitting the
position value from gradient serialization.
As well as being required to implement absolutization this also means we
now bypass a limitation with `LengthPercentage` where we would always
use `SerializationMode::Normal` for the constituent lengths which gains
us some WPT passes
The Transformation class wasn't really accomplishing anything. It still
had to store StyleValues, so it was basically the same as
TransformationStyleValue, with extra steps to convert from one to the
other. So... let's just use TransformationStyleValue instead!
Apart from moving code around, the behavior has changed a bit. We now
actually acknowledge unresolvable parameters and return an error when
we try to produce a matrix from them. Previously we just skipped over
them, which was pretty wrong. This gets us an extra pass in the
typed-om test.
We also get some slightly different results with our transform
serialization, because we're not converting to CSSPixels and back.
Some transform-functions are not reifiable as a CSSTransformComponent,
for example a matrix() with sibling-index() in it. In that case, rather
than crashing, fall back to the basic CSSStyleValue reification.
The crash doesn't happen yet, but would with the following commit's
changes. Currently, Transformation::to_matrix() completely ignores
unsupported argument types, but in the next commit that's replaced with
TransformationStyleValue::to_matrix(), which attempts to handle them
properly.
This is equivalent to the existing `PositionStyleValue::create_center`
but returns the value expected after computation has been completed
(i.e. `50% 50%` rather than `center center`)
Previously this was implemented inline within the parsing of
`{repeating}-radial-gradient()` functions but it will also be useful for
`circle()` and `ellipse()`.
We now support the CSS Images Module Level 4 additions to the
`<radial-size>` syntax, namely:
- `<length-percentage>` rather than just `<length>` for circles.
- Distinct `<radial-extent>` values for horizontal and vertical for
ellipses.
- Mixing of `<radial-extent>` and `<length-percentage>` values for
ellipses.
The regressions are due to WPT not being updated to expect the first of
these additions.
This also allows us to remove the path generation and interpolation
handling for `rect()` and `xywh()` since that occurs after computation
Regressions in clip-path-interpolation-xywh.html are due to improper
simplification of length-percentage mixes where the length is 0px.
Storing these within `LengthPercentage`, `LengthBox`, and `Variant`
over-complicated things.
We also now use the correct `SerializationMode` when serializing `xywh`
and `rect`
This lets us not care about non-absolute Length units when resolving
gradient data, as they'll already have been converted to px.
We can also use Angle::from_style_value() safely on absolutized angles,
which reduces some boilerplate code.
