Re: [css3-animations] Change to currentColor means it's no longer animatable?

When we discussed this issue last August (
http://web.archiveorange.com/archive/v/OOYmTrbeuqHBxqIvwMab) the consensus
seemed to be that  both bumping of overlapping stops and placement of
auto-positioned stops occurs before interpolation. You can't bump
overlapping stops until layout time.

so for the following example at a width of 100px:

linear-gradient(red 100px, green 50%);
to
linear-gradient(red 20px, green 50%);

consensus was that we first resolve to
linear-gradient(red 100px, green 100px);
to
linear-gradient(red 20px, green 50px);
(i.e. at 50% interpolation we'd have linear-gradient(red 60px, green 75px))

rather than first interpolate, then resolve.
(i.e. at 50% interpolation we'd have linear-gradient(red 60px, green 50%),
which would resolve to linear-gradient(red 60px, green 60px))

Cheers,
    -Shane

On Tue, May 1, 2012 at 9:03 AM, L. David Baron <dbaron@dbaron.org> wrote:

> On Monday 2012-04-30 15:52 -0700, Tab Atkins Jr. wrote:
> > On Mon, Apr 30, 2012 at 3:31 PM, Shane Stephens <shans@google.com>
> wrote:
> > > We already make an exception for animating gradient stops - these
> actually
> > > can't be calculated until used-time due to the need to know the size
> of the
> > > container.
> >
> > That one's even worse - it's us actually edging into the "used-time
> > transitions" space.  We want to do that eventually, but have been
> > resisting it so far (unfortunately ;_;).
> >
> > currentColor is only used-time due to a technicality, so it's even
> > easier than the gradient case.
>
> I don't think so.  Gradient stop positions that are mixes of lengths
> and percents can be represented using calc(), and thus done entirely
> at the computed value stage.
>
> -David
>
> --
> 𝄞   L. David Baron                         http://dbaron.org/   𝄂
> 𝄢   Mozilla                           http://www.mozilla.org/   𝄂
>