[wpt-sync] Sync PR 45148 - [css-color] test conversion lossiness
Categories
(Core :: CSS Parsing and Computation, task, P4)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox126 | --- | fixed |
People
(Reporter: wpt-sync, Unassigned)
References
()
Details
(Whiteboard: [wptsync downstream])
Attachments
(5 files)
Sync web-platform-tests PR 45148 into mozilla-central (this bug is closed when the sync is complete).
PR: https://github.com/web-platform-tests/wpt/pull/45148
Details from upstream follow.
Romain Menke <romainmenke@gmail.com> wrote:
[css-color] test conversion lossiness
These tests are a bit over the top, so if someone has a suggestion for a more succinct test?
All these cases pass with at least 5 decimal places on my end, but most fail in Safari and Chrome. (Firefox doesn't support relative color syntax).
I would expect these to also pass in a browser, especially simple cases like this one:
color(from color(from color(rec2020 0.99 0.88 0.77) srgb r g b) rec2020 r g b)
- Chrome:
color(rec2020 0.989868 0.879845 0.769978)- Safari:
color(rec2020 0.9531038 0.88519126 0.8055929)When I was first testing with these values, I was mostly interested in how good implementations would perform at preserving the input after many conversions.
But given the results I think that even a single conversion is done incorrectly in some cases.Should there be more exhaustive tests?
From all, to all notations and back for a wider range of values?
| Assignee | ||
Updated•2 years ago
|
| Assignee | ||
Comment 1•2 years ago
|
||
| Assignee | ||
Comment 2•2 years ago
|
||
| Assignee | ||
Comment 3•2 years ago
|
||
| Assignee | ||
Comment 4•2 years ago
|
||
| Assignee | ||
Comment 5•2 years ago
|
||
| Assignee | ||
Comment 7•2 years ago
|
||
| Assignee | ||
Comment 8•2 years ago
|
||
Comment 9•2 years ago
|
||
| bugherder | ||
https://hg.mozilla.org/mozilla-central/rev/19e67a628516
https://hg.mozilla.org/mozilla-central/rev/f53791765c64
Description
•