| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
It feels like these tests should have a stronger affinity to `css-shapes` (rather than `css-borders`) since it's testing behavior from that spec. I think it would also be better to not rely on `border-shape` for these tests.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Commit-Queue | +1 |
It feels like these tests should have a stronger affinity to `css-shapes` (rather than `css-borders`) since it's testing behavior from that spec. I think it would also be better to not rely on `border-shape` for these tests.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
Exportable changes to web-platform-tests were detected in this CL and a pull request in the upstream repo has been made: https://github.com/web-platform-tests/wpt/pull/59236.
When this CL lands, the bot will automatically merge the PR on GitHub if the required GitHub checks pass; otherwise, ecosystem-infra@ team will triage the failures and may contact you.
WPT Export docs:
https://chromium.googlesource.com/chromium/src/+/main/docs/testing/web_platform_tests.md#Automatic-export-process
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
5 is the latest approved patch-set.
No files were changed between the latest approved patch-set and the submitted one.
Fix polygon() rounding radius clamping formula
The clamping formula in ComputeRoundedPolygonVertex used tan(θ/2) where
θ is the angle between the edge direction vectors, but the spec defines
the clamp in terms of the interior angle α = π − θ, giving tan(α/2) =
1/tan(θ/2).
This caused max_radius and tangent_distance to use inverted factors:
multiplication instead of division and vice-versa. The bug was invisible
at 90° vertices (tan(45°) = 1) but produced wrong clamping at all other
angles — triangles, pentagons, and other non-rectangular polygons.
The fix computes the interior angle directly (by negating the dot
product) so the clamping and tangent-distance formulas read exactly as
the spec writes them:
max_radius = tan(α/2) × segment / 2
tangent_distance = radius / tan(α/2)
A WPT reftest under css-shapes/shape-functions exercises clamping across
five polygon shapes using clip-path. The reference uses exact clamped
radii where possible (rectangle, equilateral triangle, diamond) and
overclamped 9999px for shapes with non-uniform vertex angles (pentagon,
isosceles triangle).
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
The WPT PR for this CL has been merged upstream! https://github.com/web-platform-tests/wpt/pull/59236
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |