Svg filter feTurbulence with some specific parameters does not display.

Svg filter feTurbulence with some specific parameters does not display.

Ignore the location part of absolutePaintRegion as the ImageBuffer is (0, 0) based.
I found this to only be a problem in the accelerated path.

BUG=162545
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=162582

[rwlbuis, 2013-11-22 14:49:58]

Simple filter rendering fix for the accelerated path.

[Stephen Chennney, 2013-11-22 21:32:45]

On 2013/11/22 14:49:58, rwlbuis wrote:
> Simple filter rendering fix for the accelerated path.

How did you find the problem int he first place? Does that not lead to a test
suggestion?

If the issue is how to force the hardware path, I'm looping in Stephen White
(senorblanco) who did all that work.

[rwlbuis, 2013-11-22 21:38:08]

On 2013/11/22 21:32:45, Stephen Chenney wrote:
> On 2013/11/22 14:49:58, rwlbuis wrote:
> > Simple filter rendering fix for the accelerated path.
> 
> How did you find the problem int he first place? Does that not lead to a test
> suggestion?

It can be seen by loading the mentioned feTile test file, but also the regular
ancient W3C-SVG-1.1/filters-turb-01-f.svg test. Unfortunately it is a fix
specific to the runtime flag "GPU Accelerated
SVG Filters" being enabled. If it is possible to enable that in a test then I
could add one, I just don't know how to do it atm.

> If the issue is how to force the hardware path, I'm looping in Stephen White
> (senorblanco) who did all that work.

Thanks.

[Stephen White, 2013-11-22 21:52:36]

On 2013/11/22 21:38:08, rwlbuis wrote:
> On 2013/11/22 21:32:45, Stephen Chenney wrote:
> > On 2013/11/22 14:49:58, rwlbuis wrote:
> > > Simple filter rendering fix for the accelerated path.
> > 
> > How did you find the problem int he first place? Does that not lead to a
test
> > suggestion?
> 
> It can be seen by loading the mentioned feTile test file, but also the regular
> ancient W3C-SVG-1.1/filters-turb-01-f.svg test. Unfortunately it is a fix
> specific to the runtime flag "GPU Accelerated
> SVG Filters" being enabled. If it is possible to enable that in a test then I
> could add one, I just don't know how to do it atm.
> 
> > If the issue is how to force the hardware path, I'm looping in Stephen White
> > (senorblanco) who did all that work.
> 
> Thanks.

Sadly, there's no easy way to trigger this from a test. This accelerated path is
likely going to away in the not-too-distant future, and the current flag won't
be flipped on; I'm just keeping it around for testing.

[Stephen White, 2013-11-22 21:52:51]

On 2013/11/22 21:52:36, Stephen White wrote:
> On 2013/11/22 21:38:08, rwlbuis wrote:
> > On 2013/11/22 21:32:45, Stephen Chenney wrote:
> > > On 2013/11/22 14:49:58, rwlbuis wrote:
> > > > Simple filter rendering fix for the accelerated path.
> > > 
> > > How did you find the problem int he first place? Does that not lead to a
> test
> > > suggestion?
> > 
> > It can be seen by loading the mentioned feTile test file, but also the
regular
> > ancient W3C-SVG-1.1/filters-turb-01-f.svg test. Unfortunately it is a fix
> > specific to the runtime flag "GPU Accelerated
> > SVG Filters" being enabled. If it is possible to enable that in a test then
I
> > could add one, I just don't know how to do it atm.
> > 
> > > If the issue is how to force the hardware path, I'm looping in Stephen
White
> > > (senorblanco) who did all that work.
> > 
> > Thanks.
> 
> Sadly, there's no easy way to trigger this from a test. This accelerated path
is
> likely going to away in the not-too-distant future, and the current flag won't
> be flipped on; I'm just keeping it around for testing.

The fix LGTM, though.

[commit-bot: I haz the power, 2013-11-22 21:56:26]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rob.buis@samsung.com/83623002/1

[commit-bot: I haz the power, 2013-11-23 12:49:40]

Change committed as 162582