1039799 - off-main-thread animations (OMTA, OMT Animations) incorectly handle inheritance of an animating value

Status: NEW

(Core :: CSS Parsing and Computation, defect)

Description

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Off main thread animations (OMTA, OMT Animations) incorrectly handle inheritance of a value that's animating, because:
 (1) they don't ship the animation to the compositor thread for the child element, but
 (2) they suppress the updating on the main thread, which means that only the parent will animate

This is probably a relatively rare case since people don't inherit these properties often, but we should probably handle it in some way, perhaps by avoiding the suppression of main thread updates.

(I've only actually verified that part (1) is a problem; I haven't yet verified that part (2) is a problem; there's a chance it isn't.)

Comment 1

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

One other thought:  fixing this might in some way relate to fixing the cascading levels of transitions and animations (which should probably be separated from bug 960465), since both might depend on having the style rules fill in, in MapRuleInfoInto, a dummy "get from animation" value that gets resolved later instead of filling in the current animating value.

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I think this doesn't block shipping OMTA since the properties that are animated off the main thread are so rarely inherited.

Comment 4

[Xidorn Quan [:xidorn] UTC+10]

FWIW, Blink doesn't do it right either, but Edge handles it correctly.