Previously, GridFormattingContext handled its own min-height constraints
internally, which caused infinite recursion when min-height was an
intrinsic sizing keyword (e.g., min-height: max-content).
This change moves the responsibility to the parent formatting context
(BFC). When any box establishing an independent formatting context has
auto height but non-auto min-height:
1. BFC measures content height using a throwaway LayoutState
2. If content height < min-height, BFC passes min-height as definite
available height to the child formatting context
3. Child FC runs once with correct constraints, unaware of min-height
Fixes https://github.com/LadybirdBrowser/ladybird/issues/4261 which
is corresponding to stack overflow on https://claude.ai/
When an item is placed with an unpositioned start and a positioned end
(e.g., `grid-row: auto / 1`), allow the start position to be negative.
This correctly creates implicit tracks before the explicit grid.
This commit includes two interdependent changes that must be applied
together:
- Treat `fit-content()` tracks as having an intrinsic min sizing
function when the limit resolves to zero.
- When clamping growth limit to fit-content limit, use `max(base_size,
fit_content_limit)` instead of just `fit_content_limit` to preserve
intrinsic sizing contributions.
These changes are coupled because the first change causes
`fit-content(0)` tracks to participate in intrinsic sizing, while the
second ensures the base size from that sizing is not discarded during
clamping.
Fixes bug when we didn't use `tracks_to_grow_beyond_limits` and instead
distributed extra space to all affected tracks. Also implements missing
"when accommodating max-content" part.
This brings parsing of grid-row-* and grid-column-* properties (and
their associated shorthands) more inline with spec.
Changes:
- Only set omitted properties for combined-value shorthands (e.g.
`grid-row: a` rather than `grid-row: a / b`) if the single value is
`<custom-ident>`.
- `[ [ <integer [-∞,-1]> | <integer [1,∞]> ] && <custom-ident>? ]`:
- Properly resolve `calc`s for `<integer>` that rely on compute-time
information.
- `[ span && [ <integer [1,∞]> || <custom-ident> ] ]`
- Allow `calc`s for `<integer>`
- Allow `<custom-ident>`
There is still work to be done to properly use these parsed values.
Gains us 46 WPT tests.
Previously, we would allow calc values such as `calc(min(1 2))`, which
would be simplified to `calc(3)` because we assumed that numbers not
separated by an operator represented a sum. We now validate that the
number of operators we see is as we would expect before collecting
these values into a sum node.
`collapse_auto_fit_tracks_if_needed()` had a check that does collapsing
only if auto-fit is used like
`grid-template-columns: repeat(auto-fit, 1px);`, and it didn't work for
valid cases when `repeat(auto-fit)` is placed in the middle of
definition like `grid-template-columns: 1px repeat(auto-fit, 1px) 1px;`.
`getComputedStyle()` for grid tracks returns style value produced during
layout. This is needed to return resolved track sizes values which are
thrown away after layout is done. Now GFC produces more correct style
value by not ignoring grid line names.
Reimplements `grid`, `grid-template`, `grid-template-rows`, and
`grid-template-columns` in a way that uses a separate function for each
grammar rule defined in the specification. This change results in many
additional passing tests from the already imported WPT suite. Most of
the remaining test failures are related to incorrect serialization of
grid properties.
The spec requires us to store properties in their shorthand-expanded
form and in the "specified" order, removing duplicates prefering based
on "cascading order". We already enforced this in `set_property` but
not at creation time (e.g. in stylesheets) or in `set_css_text` (e.g.
updating style attribute).
This commit enforces the spec requirements at all the relevant points.
We no longer include logical properties in the return value of
`getComputedStyle` as they are mapped to their physical equivalents in
`StyleComputer::for_each_property_expanding_shorthands`, but resolving
that requires a relatively large rework of how we handle logical
properties, (namely moving when we map them to their physical
equivalents from parse time to style computation time).
This also exposes two false positive tests in
wpt-import/css/cssom/border-shorthand-serialization.html related to us
not yet supporting the border-image css property.
To be vendor-prefixed, an ident has to start with a '-', then have
another '-' later. If the ident simply starts with a '-' then that's
perfectly fine.
Fixes 62 in-tree WPT subtests. :^)
If a calculation was simplified down to a single numeric node, then most
of the time we can instead return a regular StyleValue, for example
`calc(2px + 3px)` would be simplified down to a `5px` LengthStyleValue.
This means that parse_calculated_value() can't return a
CalculatedStyleValue directly, and its callers all have to handle
non-calculated values as well as calculated ones.
This simplification is reflected in the new test results. Serialization
is not yet correct in all cases but we're closer than we were. :^)
This allows us to disable test output, which performs expensive assert
tracking. This was making our imported tests run significantly slower
than tests run via `WPT.sh`.
Formatting the output ourselves also allows us to remove unnecessary
information from the test output.
This commit also rebaselines all existing imported WPT tests to follow
the new format.
This wins us 65 new WPT subtest passes! It also shows up that we're
doing the wrong thing in ShorthandStyleValue in places, notably with
the grid properties. However, having one place to fix instead of two
will make it easier to correct them. :^)
In order to be fully correct, we should use the algorithm here:
https://drafts.csswg.org/cssom/#serialize-a-css-value
However, it's quite hand-wavy. What we do have in the meantime is
`ShorthandStyleValue::to_string()`, where we special-case the
serialization rules for shorthands with a generic fallback that's
equivalent to what the previous `get_property_value()` code was doing.
This causes 36 new subtests to pass locally. :^)
Unfortunately at least one of these is flaky when it's able to load the
font file, apparently because we don't wait for the font and its
stylesheet to actually load before the tests run.
