-
Notifications
You must be signed in to change notification settings - Fork 3.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Revert "Reland #2 "Implement CSSTransitionDiscrete which enables transitions on discrete properties."" #38904
Conversation
…sitions on discrete properties."" This reverts commit 268250b.
We deliberately don't require no regressions, mainly because there are plenty of reasonable cases where tests do go from pass to fail (spec changes, increasing strictness of the pass condition, etc.). |
These test changes were based on this CSSWG resolution: w3c/csswg-drafts#4441 (comment) |
Also worth noting the spec edits for this happened in w3c/csswg-drafts#8520 |
I should think the tests ought to be fixed to match the spec so that there are no failures, even if the failures are now expected. |
What class of failures do you want to avoid? (That is, what do you mean by "no failures, even if the failures are now expected"?) I think it's normal that if a spec changes so that an implementation no longer conforms to the spec, that implementation should start failing tests -- it's a good way to notice that the spec has changed. |
I'm expecting that the tests related to transitions and discrete properties are modified to produce PASS results when the implementation is correct. Right now there are heaps of tests yielding FAIL results. I just want to be sure this is correct and the tests will PASS once implementations implement the spec change. An example are the two tests highlighted in my original PR message. If those tests now FAIL'ing are doing so correctly, then this PR should be landed back, by all means. |
Based on https://chromium-review.googlesource.com/c/chromium/src/+/4308075 those tests were passing in patched Chromium. (There's no entry for them in The Chromium you see on the "experimental" channel for wpt.fyi is dev channel, which can often be a week or two behind. |
OK, then I guess this is good. Thanks for providing all the extra information, this helped. |
I made a PR to re-add the test changes: #38936 |
Reverts #38812
This made at least two tests fail across all browsers:
css/css-transitions/animations/vertical-align-interpolation.html
css/css-transitions/animations/z-index-interpolation.html
This was caught by the checks on this PR but somehow they appeared green, maybe because the WPT.fyi checks are in beta. It appears there are a lot of regressions:
https://wpt.fyi/results/css?diff&filter=ADC&run_id=5094539348410368&run_id=5077352225177600
Either this change is correct and a fair few tests need to be updated to have new expectations, or this test is wrong. Either way, I think this should be reverted until this is addressed.
Cc @josepharhar and @flackr