Summary: | Bad animations with transform-box | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Simon Fraser (smfr) <simon.fraser> | ||||||
Component: | Animations | Assignee: | Nobody <webkit-unassigned> | ||||||
Status: | RESOLVED CONFIGURATION CHANGED | ||||||||
Severity: | Normal | CC: | ahmad.saleem792, bfulgham, dino, graouts, graouts, krit, sabouhallawa, simon.fraser, webkit-bug-importer, zimmermann | ||||||
Priority: | P2 | Keywords: | InRadar | ||||||
Version: | Safari Technology Preview | ||||||||
Hardware: | Unspecified | ||||||||
OS: | Unspecified | ||||||||
See Also: | https://bugs.webkit.org/show_bug.cgi?id=211927 | ||||||||
Attachments: |
|
Description
Simon Fraser (smfr)
2020-05-11 17:17:49 PDT
These are animations on SVG elements. Does not seem to be related to Web Animations, we get the same behavior using the legacy engine. Created attachment 399300 [details]
reduced test case
Created attachment 399394 [details] More reduced test case The bug happens because of this CSS setting: #ball { ... transform-box: unset; ... } If it is deleted, the animation works fine. It looks like the initial value of "transform-box" is set to "border-box". But according to https://drafts.csswg.org/css-transforms/#transform-box it should be "view-box". It's possible that webkit behavior is correct and the other browsers are wrong; we did do some transform-box stuff not too long ago. I think "view-box" is more appropriate initial value for SVG elements. The SVG view box represents the coordinate system for the whole SVG. I think it is hard to think of the transform-origin in terms of the "element" coordinates. I logged https://bugs.webkit.org/show_bug.cgi?id=211927 to track the transform-box initial value issue. I will keep this bug for the flickering at the end of the animation. We should file bugs on Firefox and Chrome. I don't know about the past but using 'more reduced testcase', I think Safari 16.5, Chrome Canary 115 and Firefox Nightly 115 are same (at least I think). If nothing needed, can we close this now? Closing based on Ahmad's testing on the "More reduced test case". Please open a new bug if there is an additional bug. |