Border-collapsed table's border width recalculation is stateful
Categories
(Core :: Layout: Tables, defect)
Tracking
()
People
(Reporter: dshin, Assigned: dshin)
References
Details
Attachments
(2 files)
STR:
- Open the testcase
- Note table shape
- Hover over "one"
- Note thick border between "one" and "two"
- Hover elsewhere
- Note table shape
Expected: 2) and 6) are identical
Actual: 6) has more space between "one" and "two"
Assignee | ||
Updated•2 years ago
|
Assignee | ||
Comment 1•2 years ago
|
||
Probably happened since bug 1464928. The testcase is derived from a test added from that bug.
The path responsible was added from bug 41262 - stored border widths are stateful, storing the max of the result of GetBorderWidth
and the incoming border (Which is 0 in this case). GetBorderWidth returns a halved value, hence why the final table's gap between cells is half of 20px.
The original test case doesn't suffer from this because it works with 1px borders, of which half is rounded down to 0.
Assignee | ||
Comment 2•1 year ago
|
||
Updated•1 year ago
|
Comment 5•1 year ago
|
||
Backed out for causing wr failures in /css/css-tables/*
- Backout link
- Push with failures
- Failure Log
- Failure line: TEST-UNEXPECTED-FAIL | /css/css-tables/table_grid_size_col_colspan.html | Testing http://web-platform.test:8000/css/css-tables/table_grid_size_col_colspan.html == http://web-platform.test:8000/css/css-tables/table_grid_size_col_colspan-ref.html
- Failure Log 2
- Failure line 2: TEST-UNEXPECTED-FAIL | /css/css-tables/tentative/paint/collapsed-border-large-cell.html | Testing http://web-platform.test:8000/css/css-tables/tentative/paint/collapsed-border-large-cell.html == http://web-platform.test:8000/css/css-tables/tentative/paint/collapsed-border-large-cell-ref.html
Assignee | ||
Comment 7•1 year ago
|
||
Changes made, try run looks good (Failure cases don't even use table, so it seems unrelated)
Comment 9•1 year ago
|
||
(In reply to David Shin[:dshin] from comment #7)
Changes made, try run looks good (Failure cases don't even use table, so it seems unrelated)
Yeah, the R5 run there looks like it's a version of bug 1895092, so it's likely an infra issue and not something you need to worry about. (It matches the "bad" pattern from bug 1895092 comment 16.)
Comment 10•1 year ago
|
||
bugherder |
Updated•1 year ago
|
Issue is reproducible on a 2024-05-05 Nightly build on Windows 10.
Verified as fixed on Firefox Nightly 128.0a1 and Firefox 127.0b2 on Windows 10, macOS 12, Ubuntu 22.
Description
•