On 27/6/09 8:12 AM, L. David Baron wrote:
My initial design ideas were that: 1) We'd only store one style or the other at any given time. 2) We'd decide whether to include transitions by knowing whether we were currently advancing a transition.I think I've come to the conclusion that (2) is a mistake because of all the flushing that it would require (i.e., it would require flushing style both before and after advancing transitions) and the fragility and weird flush-timing-dependency of the process of starting a transition. So I think I'm leaning towards a model where a restyle has an explicit parameter that says whether it's with or without transitions. (When we get a without-transitions restyle, and the style changes, we then start a transition and post a with-transitions restyle to get the data from it.) But I think I'm also inclined to stick with (1) rather than adding the complexity of storing two sets of style data. Does this sound reasonable?
To me, yes. Rob _______________________________________________ dev-tech-layout mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-layout
