I just sent a proposal to www-style about having css3-transitions describe a clearer model for how transitions and animations interact and how transitions are started: http://lists.w3.org/Archives/Public/www-style/2013Mar/0297.html This would hopefully clear up a bunch of areas where there's significant lack of interoperability between implementations of transitions and animations, and where the spec is very unclear.
I wanted to write a bit more here about how I see this being implemented in Gecko. I think the big pieces would be: (1) Removing the concepts of "style with animation" and "style without animation" that we have, which I think has proven not to meet the goals I originally intended them to meet (providing sane behavior for requirements (3) and (5) in the above message). This would be replaced with what I propose in item (A) in that message, which is essentially the miniflush concept that we have implemented for suppressing the main thread style updates for animations that are running on the compositor, but which we don't currently use beyond that code. I think this removal will end up being a pretty substantial code simplification, and the performance cost of the miniflush will be roughly the same as the performance cost of the current dual-style setup. (2) Changing how animations and transitions work in the cascade. Note that currently compositor-driven animations have different cascading behavior than animations not running on the compositor, which is a relatively serious bug (847287). I admit I don't have a completely concrete idea about how to do this, but I think it's not going to be all that hard; at a concrete level it might come down to having nsIStyleRules in the cascade at two different places (both at the top of the cascade and at the official location) in order to let the rule safely record which properties the animation or transition is the winning rule for, and then recording that information so that we can use it for the compositor-driven animations. I don't think this is too hard, though; I just haven't implemented it yet. I wanted to call this out both because it implies some somewhat significant changes coming in code that we're depending on pretty heavily, and to encourage others to comment either here about the Gecko-specific parts or on www-style about the standard bits. -David -- 𝄞 L. David Baron http://dbaron.org/ 𝄂 𝄢 Mozilla http://www.mozilla.org/ 𝄂 _______________________________________________ dev-tech-layout mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-layout
