Implement support for non-scale/translate CTM in image filters.

Implement support for non-scale/translate CTM in image filters.

Clone with tweaks of https://codereview.chromium.org/986623003/

pre-cl to guard chrome here: https://codereview.chromium.org/1133403003/

BUG=skia:3288
Committed: https://skia.googlesource.com/skia/+/35786b176c159d2e7a816e9da9b051ef3129d5cb

[reed1, 2015-05-13 21:02:17]

reed@google.com changed reviewers:
+ robertphillips@google.com, senorblanco@google.com

[reed1, 2015-05-13 21:02:18]

need to not reference layer after its deleted/destroyed

[reed1, 2015-05-13 21:15:17]

The CQ bit was checked by reed@google.com to run a CQ dry run

[commit-bot: I haz the power, 2015-05-13 21:15:39]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1140943004/20001

[reed1, 2015-05-13 21:15:43]

This may change layouttests, so I will try adding a guard before I land this.

[reed1, 2015-05-13 21:20:33]

The CQ bit was unchecked by reed@google.com

[Stephen White, 2015-05-13 22:18:44]

senorblanco@chromium.org changed reviewers:
+ senorblanco@chromium.org

[Stephen White, 2015-05-13 22:18:45]

Could you remove the "This reverts commit ..." line from the CL description?

Also feel free to reference bug 3288.

https://codereview.chromium.org/1140943004/diff/20001/src/core/SkCanvas.cpp
File src/core/SkCanvas.cpp (right):

https://codereview.chromium.org/1140943004/diff/20001/src/core/SkCanvas.cpp#n...
src/core/SkCanvas.cpp:1120: fMCRec->fMatrix = layer->fStashedMatrix;
I don't know if it's still relevant, but in my original patch I had to
null-check fMCRec before unstashing here.

[Stephen White, 2015-05-13 22:18:58]

oh and LGTM

[reed2, 2015-05-14 10:53:23]

reed@chromium.org changed reviewers:
+ reed@chromium.org

[reed2, 2015-05-14 10:53:25]

https://codereview.chromium.org/1140943004/diff/20001/src/core/SkCanvas.cpp
File src/core/SkCanvas.cpp (right):

https://codereview.chromium.org/1140943004/diff/20001/src/core/SkCanvas.cpp#n...
src/core/SkCanvas.cpp:1120: fMCRec->fMatrix = layer->fStashedMatrix;
On 2015/05/13 22:18:44, Stephen White wrote:
> I don't know if it's still relevant, but in my original patch I had to
> null-check fMCRec before unstashing here.

Not needed, as we only restore the matrix when we're not shutting down now

[reed1, 2015-05-14 12:54:25]

now with guard and comment

[robertphillips, 2015-05-14 13:01:22]

lgtm - do we test this locally?

[reed1, 2015-05-14 13:22:52]

The CQ bit was checked by reed@google.com

[reed1, 2015-05-14 13:22:52]

The patchset sent to the CQ was uploaded after l-g-t-m from
senorblanco@chromium.org
Link to the patchset: https://codereview.chromium.org/1140943004/#ps60001
(title: "add comment")

[commit-bot: I haz the power, 2015-05-14 13:30:55]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1140943004/60001

[commit-bot: I haz the power, 2015-05-14 13:37:00]

Committed patchset #4 (id:60001) as
https://skia.googlesource.com/skia/+/35786b176c159d2e7a816e9da9b051ef3129d5cb

[Stephen White, 2015-05-14 14:47:58]

On 2015/05/14 12:54:25, reed1 wrote:
> now with guard and comment

No harm in a guard, but it shouldn't be necessary -- Chrome shouldn't be passing
a matrix with anything other than scale & translate.

[reed1, 2015-05-14 18:24:33]

On 2015/05/14 14:47:58, Stephen White wrote:
> On 2015/05/14 12:54:25, reed1 wrote:
> > now with guard and comment
> 
> No harm in a guard, but it shouldn't be necessary -- Chrome shouldn't be
passing
> a matrix with anything other than scale & translate.

layouttests do, so the guard was necessary.

[reed1, 2015-05-14 19:24:45]

A revert of this CL (patchset #4 id:60001) has been created in
https://codereview.chromium.org/1139173002/ by reed@google.com.

The reason for reverting is: matriximagefilter needs to antialias its draw, so
we get smooth diagonals...
.