Firefox fails wpt test css-grid/subgrid/grid-template-computed-nogrid.html and grid-template-valid.html due to serialization differences
Categories
(Core :: Layout: Grid, defect)
Tracking
()
People
(Reporter: dholbert, Unassigned)
References
(Blocks 1 open bug)
Details
wpt.fyi:
https://wpt.fyi/results/css/css-grid/subgrid/grid-template-computed-nogrid.html
https://wpt.fyi/results/css/css-grid/subgrid/grid-template-valid.html
In both tests, the failure has to do with cases where we're expanding repeat(...)
expressions in the computed style, where the test expects us to not-expand them.
e.g.
!EQ("subgrid [a] repeat(2, [c] [d e])", "subgrid [a] [c] [d e] [c] [d e]")
!EQ("subgrid repeat(1, [])", "subgrid []")
Here's a reduced testcase, which alert()'s its computed style:
data:text/html,<body style="grid-template-columns: subgrid repeat(1, [])"><script>alert(getComputedStyle(document.body).gridTemplateColumns)</script>
In Firefox, this alerts subgrid []
. In WebKit, this alerts subgrid repeat(1, [])
Reporter | ||
Comment 1•1 year ago
|
||
It looks like the tests' expectations do match the spec. Specifically, this section:
https://drafts.csswg.org/css-grid-2/#computed-tracks
https://drafts.csswg.org/css-grid-2/#computed-repeat-notation
That section defines "computed track list" and "computed repeat notation" in a way that sounds like it's asking for the repeat(...) notation.
Reporter | ||
Comment 2•1 year ago
|
||
(Looks like this is essentially a dupe of bug 1583429, as Emily noted via linking it up. We should probably implement the spec text as a patch for bug 1583429, and we can update this bug to be a dupe of that, if it's indeed fixed by our change over there.)
Reporter | ||
Comment 3•9 months ago
|
||
Bug 1583429 landed and removed the failure-annotations for these WPTs, and wpt.fyi does show them passing now, so we're good to dupe. Hooray!
Description
•