Re: [css3-transitions] Can/should a change of currentColor's computed value trigger a transition ?

On Fri, Feb 11, 2011 at 6:01 PM, Boris Zbarsky <bzbarsky@mit.edu> wrote:
> On 2/11/11 6:56 PM, Tab Atkins Jr. wrote:
>>
>> Do transitions fire based on the specified or computed value?  I can't
>> seem to find language in the spec specifying one way or the other
>
> These paragraphs:
>
>  Implementations must not start a transition when the computed
>  value of a property changes as a result of declarative animation
>  (as opposed to scripted animation).
>
>  Implementations also must not start a transition when the
>  computed value changes because it is inherited (directly or
>  indirectly) from another element that is transitioning the same
>  property.
>
> sure sound like the computed value is what matters.  However the start of
> section 3 still talks about "the value"; it should probably explicitly say
> "computed value".

Well, maybe.  It's definition-through-negation, in that you can
*infer* you should be transitioning on computed values, since the spec
makes such a big deal about not paying attention to the computed value
in certain situations.

Inference is not equivalent to definition.  At the moment, transitions
are completely untestable, because there aren't any normative
statements surrounding what *should* trigger a transition.


>> some simple testing seems to show that webkit uses specified value:
>
> Yeah, I think we decided a while back that that's not what we want here,
> since it makes transitions on things like percentage widths when ancestors
> are resized not really work that well.

I agree; that was precisely what I was using to test.

~TJ

Received on Saturday, 12 February 2011 02:05:59 UTC