Re: Agenda request: Presentation attributes in animation sandwich model

On Feb 24, 2012, at 2:47 PM, Daniel Holbert wrote:

> On 02/24/2012 02:13 PM, Dirk Schulze wrote:
>>> However, after a little testing, it looks like I was wrong on that -- it
>>> appears that CSS animations override SMIL animations in Gecko right now,
>>> when both are in play.
>> Do CSS Transitions override SMIL animations as well?
> 
> Yes: at least in Gecko, the intermediate transitioned style does appear 
> to override SMIL animations' style. (side-note: we do prevent SMIL 
> animation from *triggering* CSS transitions, to keep from entering into 
> a death-spiral)
Can you give an example of a potential depth-spiral?

> 
> Here's a testcase for transitions vs. SMIL:
> 
> http://people.mozilla.org/~dholbert/tests/smil/compat_tests/transitionsVsSMIL_v1.svg
> 
> At least on that testcase, webkit disagrees with gecko -- the 
> transitioned style loses to the SMIL-animated style.  
Please don't take WebKit as reference, we animate the property for the style attribute on WebKit (style="property: newAnimatedValue"). Which is against every specification. We will fix that :).

> I tested Opera, 
> too, but it wouldn't seem to honor the transition. (its documentation 
> seems to suggest that it doesn't support CSS transitions on SVG content[1])
Could it mean that SMIL animation just overrides Transition? What happens for additive="sum" on the SMIL animation? Will it take the transition as base value for the animation?

> 
> ~Daniel
> 
> [1] 
> http://dev.opera.com/articles/view/css3-transitions-and-2d-transforms/#introduction 
> ("This article will guide you through how to use CSS3 transitions and 
> briefly touch on how the equivalent can also be achieved in SVG via SMIL.")
I want to make sure that I am open for any solution as long all UAs agree and implement it that way (all UAs with SMIL support). Even if I think that SMIL Animations should be at the end of the chain (because it supports additive="sum" which CSS animations/transitions seem not to support). If there are a reasonable concerns, I am fine with changing the resolution (because of implementers feedback)

The main goal should be to find a solution that will work for all current presentation attributes, as well as future presentation attributes.

Greetings
Dirk

Received on Friday, 24 February 2012 23:21:22 UTC