Fix assert when interpolating using unset

Fix assert when interpolating using unset

In Debug mode interpolating using unset hit an ASSERT because
StyleBuilder::applyProperty did not take it into account and assumed the
CSSValue was a primitive value if not initial or inherit. So add the code
to handle unset and determine if we should fallback to inherit or initial
for the given property.

Add a test for interpolating unset with an inherited property (font-size)
and one which falls back to initial (background-color).

BUG=445332
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=188009

[rwlbuis, 2015-01-06 21:28:42]

rob.buis@samsung.com changed reviewers:
+ timloh@chromium.org

[rwlbuis, 2015-01-06 21:28:43]

PTAL. I think the mac bot problem is unrelated.

[Timothy Loh, 2015-01-06 23:34:51]

timloh@chromium.org changed reviewers:
+ dstockwell@chromium.org

[Timothy Loh, 2015-01-06 23:34:52]

Could you CC me to the bug? :)

This duplicates the unset handling in StyleResolver::applyProperties, but for
the animation interpolation code, right? Do we want to remove the handling
there?

[rwlbuis, 2015-01-06 23:49:43]

On 2015/01/06 23:34:52, Timothy Loh wrote:
> Could you CC me to the bug? :)

Done!

> This duplicates the unset handling in StyleResolver::applyProperties, but for
> the animation interpolation code, right? Do we want to remove the handling
> there?

I can check after fitness...

[rwlbuis, 2015-01-07 01:43:54]

On 2015/01/06 23:34:52, Timothy Loh wrote:
> Could you CC me to the bug? :)

Oops I added dstockwell by mistake, added you too now.

> This duplicates the unset handling in StyleResolver::applyProperties, but for
> the animation interpolation code, right? Do we want to remove the handling
> there?

It seems like it, well spotted. I'll try to update the patch for that tomorrow.

[rwlbuis, 2015-01-07 16:55:04]

On 2015/01/07 01:43:54, rwlbuis wrote:
> > This duplicates the unset handling in StyleResolver::applyProperties, but
for
> > the animation interpolation code, right? Do we want to remove the handling
> > there?
> 
> It seems like it, well spotted. I'll try to update the patch for that
tomorrow.

Indeed removing the code seems fine, patch set #3 shows up green.

[Timothy Loh, 2015-01-07 23:07:31]

On 2015/01/07 16:55:04, rwlbuis wrote:
> On 2015/01/07 01:43:54, rwlbuis wrote:
> > > This duplicates the unset handling in StyleResolver::applyProperties, but
> for
> > > the animation interpolation code, right? Do we want to remove the handling
> > > there?
> > 
> > It seems like it, well spotted. I'll try to update the patch for that
> tomorrow.
> 
> Indeed removing the code seems fine, patch set #3 shows up green.

lgtm

[rwlbuis, 2015-01-07 23:17:08]

The CQ bit was checked by rob.buis@samsung.com

[commit-bot: I haz the power, 2015-01-07 23:17:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/840463002/40001

[commit-bot: I haz the power, 2015-01-08 00:25:11]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=188009