107609 2013-01-22 18:05:38 -0800 REGRESSION(r139800): animations with start delay may not animate if page becomes hidden 2013-04-06 01:39:53 -0700 1 1 1 Unclassified WebKit New Bugs 528+ (Nightly build) Unspecified Unspecified NEW P2 Normal --- 107521 1 dstockwell webkit-unassigned bdakin dino jamesr simon.fraser syoichi oldest_to_newest 813637 0 dstockwell 2013-01-22 18:05:38 -0800 http://trac.webkit.org/changeset/139800 It's not clear from the Changelog which of these changes are intentional: Animation events do not fire while page is hidden. Animations may not animate if page becomes hidden then visible again during animation-delay. Animations do not make progress while page is hidden. 813638 1 jamesr 2013-01-22 18:09:05 -0800 My guess would be all 3 are intentional. 813647 2 simon.fraser 2013-01-22 18:18:35 -0800 Animation suspend/resume are totally broken; they just stop existing animations, they don't prevent new ones. Not deliberate, just broken. 813661 3 dstockwell 2013-01-22 18:40:16 -0800 (In reply to comment #2) > Animation suspend/resume are totally broken; they just stop existing animations, they don't prevent new ones. Not deliberate, just broken. Can we roll it out, or is work happening on suspend/resume? If the goal was to reduce CPU usage I think we could still achieve that without suspending by only servicing animations at event points when the page is hidden. 813732 4 bdakin 2013-01-22 20:36:18 -0800 First we should agree on desired behavior. (In reply to comment #0) > http://trac.webkit.org/changeset/139800 > > It's not clear from the Changelog which of these changes are intentional: > > Animation events do not fire while page is hidden. I tend to think this is correct behavior, but I am open to a discussion. > Animations may not animate if page becomes hidden then visible again during animation-delay. Definitely a bug. > Animations do not make progress while page is hidden. Personally I like this behavior as well, but I do think it's open to discussion. And Simon pointed out that suspend does not prevent new animations, which is also definitely a bug. 813816 5 dstockwell 2013-01-22 22:04:04 -0800 (In reply to comment #4) > > Animation events do not fire while page is hidden. > > I tend to think this is correct behavior, but I am open to a discussion. > > > Animations do not make progress while page is hidden. > > Personally I like this behavior as well, but I do think it's open to discussion. I think these are effectively the same issue. I am concerned that this change diverges from IE and Firefox. I think there are many use cases where animations align with real time and will be impacted/ made more difficult. eg. http://www.paulrhayes.com/experiments/clock/#clock 813826 6 bdakin 2013-01-22 22:11:08 -0800 (In reply to comment #5) > (In reply to comment #4) > > > Animation events do not fire while page is hidden. > > > > I tend to think this is correct behavior, but I am open to a discussion. > > > > > Animations do not make progress while page is hidden. > > > > Personally I like this behavior as well, but I do think it's open to discussion. > > I think these are effectively the same issue. I am concerned that this change diverges from IE and Firefox. I think there are many use cases where animations align with real time and will be impacted/ made more difficult. > > eg. http://www.paulrhayes.com/experiments/clock/#clock Hmm! Thank you for that example; it is a compelling demo for the alternate behavior. Thinking this over… 820844 7 dstockwell 2013-01-30 17:12:39 -0800 (In reply to comment #6) > Hmm! Thank you for that example; it is a compelling demo for the alternate behavior. Thinking this over… Any thoughts? I have a patch (https://bugs.webkit.org/attachment.cgi?id=185628) on https://bugs.webkit.org/show_bug.cgi?id=107521 that implements the idea above (only service animations at events for hidden pages) which you could experiment with.