On 14/7/09 6:37 AM, L. David Baron wrote:
I think I'd like to avoid trying to run transitions in reverse, at least for now, because that specificity comparison is a good bit of work, I think. But it might be a good idea in the future. (I think this is a weakness of the spec, especially given that the default transition-timing-function is asymmetric.)One idea I had just now is that we use the following algorithm, which I think should be pretty simple: Let O be the old value of the property (including transitions) Let E be the old value of the property (not including transitions, i.e., the endpoint of the currently running transition) Let N be the new value of the property. Run the transition from O to N. (That's the part we're doing now.) However, if abs(O-N)< abs(E-N), reduce the transition-duration by their ratio. If transition-delay is negative, also reduce the transition-delay by the same ratio. (Or maybe just reduce transition-delay unconditionally?)
Sounds good. Rob _______________________________________________ dev-tech-layout mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-layout
