Closed nairnandu closed 9 months ago
@nt1m / @gsnedders could you please review for WebKit? @jgraham / @zcorpan could you please review for Gecko?
As mentioned in the meeting, these tests were invalid in the first place, the commits fix them.
They both test core functionality, so I would be against removing them.
adding @mysteryDate as well for review
For https://github.com/web-platform-tests/wpt/commit/892583e4adb98da9b63c6ceaf1e0896f2cfb71e7 :
I know that hwb/hsl serialize as legacy srgb, which never serializes "none" and simply swaps it for zero at parse time. According to the spec "During serialization, any missing values are converted to 0." and the tests agree. But in this instance, the colors are being serialized as non-legacy color(srgb ... ) format, so the "nones" should remain in my opinion. Regardless, I think we should have had a test change proposal.
Could you also explain your reasoning here: https://github.com/web-platform-tests/wpt/commit/79c26e1b48072705e936333a5443aa2aaf829809
I don't understand what this change is meant to be about.
I would ideally like this to just go through a normal test change proposal so we can have an obvious documentation of the reasoning for the test changes.
Ah, I now see this
https://github.com/web-platform-tests/wpt/issues/42677
I'll respond there. I have no further objections here.
The commit message from https://github.com/web-platform-tests/wpt/commit/79c26e1b48072705e936333a5443aa2aaf829809 says it all fwiw, computed value serialization should resolve the positions. Which is something neither Blink or WebKit are doing correctly in this case atm.
Gotcha, I didn't see the spec link in the commit message, my apologies.
Test List
Rationale
There were changes made to these tests, without an associated test-change-proposal: