Bilerp + mirror + perspective

WIP: experimental bilerp pipeline.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1775963002

Committed: https://skia.googlesource.com/skia/+/6eff52afb458bf6702a715d88611fd571544ef73

[herb_g, 2016-03-08 16:51:40]

Description was changed from

==========
WIP: experimental bilerp pipeline.

BUG=skia:
==========

to

==========
WIP: experimental bilerp pipeline.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[herb_g, 2016-03-09 22:50:11]

WIP: stack mostly working.

[herb_g, 2016-03-10 15:07:55]

WIP: working again after sync.

[herb_g, 2016-03-10 20:10:14]

WIP: tilers split. Tests passing.

[herb_g, 2016-03-10 22:44:30]

WIP: bilerp fast rate

[herb_g, 2016-03-11 19:30:36]

WIP: turn on pipeline to test.

[herb_g, 2016-03-11 20:36:27]

WIP: fix two y span expander fast rate.

[herb_g, 2016-03-12 19:22:32]

WIP: boundary asserts pass.

[herb_g, 2016-03-12 19:24:12]

WIP: rename variable.

[herb_g, 2016-03-13 15:32:41]

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

[commit-bot: I haz the power, 2016-03-13 15:32:48]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1775963002/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1775963002/160001

[commit-bot: I haz the power, 2016-03-13 15:42:18]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-03-13 15:42:19]

Dry run: This issue passed the CQ dry run.

[herb_g, 2016-03-14 20:02:46]

Bilerp working.

[herb_g, 2016-03-14 22:28:05]

Inverses working for repeat.

[herb_g, 2016-03-15 19:38:23]

Mirror working.

[herb_g, 2016-03-15 20:33:59]

Perspective working.

[herb_g, 2016-03-15 21:34:14]

Add post alpha.

[herb_g, 2016-03-16 17:37:13]

Fix right edge clamp sampling.

[herb_g, 2016-03-18 20:57:31]

Working

[herb_g, 2016-03-19 18:29:41]

sub-pixel working

[herb_g, 2016-03-19 19:17:12]

Add spans back in

[herb_g, 2016-03-19 19:20:35]

Fix negative repeat.

[herb_g, 2016-03-21 21:51:00]

Synced an unpdated to sk_sp.

[herb_g, 2016-03-22 14:46:27]

Description was changed from

==========
WIP: experimental bilerp pipeline.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
WIP: experimental bilerp pipeline.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[herb_g, 2016-03-22 14:46:28]

herb@google.com changed reviewers:
+ fmalita@chromium.org, mtklein@google.com, reed@google.com

[herb_g, 2016-03-22 17:23:50]

Fix mirror.

[herb_g, 2016-03-22 18:33:09]

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

[commit-bot: I haz the power, 2016-03-22 18:33:17]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1775963002/400001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1775963002/400001

[commit-bot: I haz the power, 2016-03-22 18:40:47]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-03-22 18:40:48]

Dry run: Try jobs failed on following builders:
  Build-Win-MSVC-x86_64-Debug-Trybot on client.skia.compile (JOB_FAILED,
http://build.chromium.org/p/client.skia.compile/builders/Build-Win-MSVC-x86_6...)

[herb_g, 2016-03-22 19:10:25]

Fix msvc compiler problem.

[reed1, 2016-03-22 19:17:45]

lgtm

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
File src/core/SkBitmapProcState.cpp (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
src/core/SkBitmapProcState.cpp:140: //SkASSERT(fPixmap.colorType() ==
fProvider.info().colorType());
can we restore this assert? do we want to?

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkLinearBitma...
File src/core/SkLinearBitmapPipeline.h (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkLinearBitma...
src/core/SkLinearBitmapPipeline.h:66: using MatrixStage =
PolymorphicUnion<PointProcessorInterface, 160>;
literal sizes are odd to look at.

[herb_g, 2016-03-22 19:17:51]

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

[commit-bot: I haz the power, 2016-03-22 19:18:00]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1775963002/420001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1775963002/420001

[commit-bot: I haz the power, 2016-03-22 19:33:58]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-03-22 19:33:59]

Dry run: This issue passed the CQ dry run.

[herb_g, 2016-03-22 20:26:57]

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
File src/core/SkBitmapProcState.cpp (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
src/core/SkBitmapProcState.cpp:140: //SkASSERT(fPixmap.colorType() ==
fProvider.info().colorType());
On 2016/03/22 19:17:45, reed1 wrote:
> can we restore this assert? do we want to?

I should remove this assert. I think we proved to ourselves that in general this
could not be true.

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkLinearBitma...
File src/core/SkLinearBitmapPipeline.h (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkLinearBitma...
src/core/SkLinearBitmapPipeline.h:66: using MatrixStage =
PolymorphicUnion<PointProcessorInterface, 160>;
On 2016/03/22 19:17:45, reed1 wrote:
> literal sizes are odd to look at.

Done.

[f(malita), 2016-03-22 20:55:26]

lgtm with some random nits.

https://codereview.chromium.org/1775963002/diff/400001/gm/SkLinearBitmapPipel...
File gm/SkLinearBitmapPipelineGM.cpp (right):

https://codereview.chromium.org/1775963002/diff/400001/gm/SkLinearBitmapPipel...
gm/SkLinearBitmapPipelineGM.cpp:72: paint.setShader(shader);
nit: not particularly important for a GM, but std::move(shader) (or inlining the
factory call) would avoid refcount churn.

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
File src/core/SkBitmapProcShader.cpp (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
src/core/SkBitmapProcShader.cpp:181: kIndex_8_SkColorType ==
srcInfo.colorType()) {
Doesn't srcInfo.bytesPerPixel() < 4 catch kIndex_8_SkColorType?

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
src/core/SkBitmapProcShader.cpp:233: && info->fPixmap.colorType() !=
kBGRA_8888_SkColorType) {
Do we really support any swizzle != kN32_SkColorType?

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
File src/core/SkBitmapProcState.h (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
src/core/SkBitmapProcState.h:39: SkMatrix            fRealInvMatrix;     // The
actual inverse matrix.
Do we actually need/use the fInvMatrix adjustment for the linear pipeline? 
Rather than adding fRealInvMatrix, could we just skip the adjustment?

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkLinearBitma...
File src/core/SkLinearBitmapPipeline.cpp (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkLinearBitma...
src/core/SkLinearBitmapPipeline.cpp:56: // resulting Y values my be off the
tile. When that happens, the second Y is used to denote
"might"?

[herb_g, 2016-03-22 21:10:40]

https://codereview.chromium.org/1775963002/diff/400001/gm/SkLinearBitmapPipel...
File gm/SkLinearBitmapPipelineGM.cpp (right):

https://codereview.chromium.org/1775963002/diff/400001/gm/SkLinearBitmapPipel...
gm/SkLinearBitmapPipelineGM.cpp:72: paint.setShader(shader);
On 2016/03/22 20:55:26, f(malita) wrote:
> nit: not particularly important for a GM, but std::move(shader) (or inlining
the
> factory call) would avoid refcount churn.

Done.

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
File src/core/SkBitmapProcShader.cpp (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
src/core/SkBitmapProcShader.cpp:181: kIndex_8_SkColorType ==
srcInfo.colorType()) {
On 2016/03/22 20:55:26, f(malita) wrote:
> Doesn't srcInfo.bytesPerPixel() < 4 catch kIndex_8_SkColorType?

The real problem is that the provider may lie about the type. So this check
doesn't matter and I should be able to handle all cases soon.

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
src/core/SkBitmapProcShader.cpp:233: && info->fPixmap.colorType() !=
kBGRA_8888_SkColorType) {
On 2016/03/22 20:55:26, f(malita) wrote:
> Do we really support any swizzle != kN32_SkColorType?

Yes we really do support both at the same time.

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
File src/core/SkBitmapProcState.h (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkBitmapProcS...
src/core/SkBitmapProcState.h:39: SkMatrix            fRealInvMatrix;     // The
actual inverse matrix.
On 2016/03/22 20:55:26, f(malita) wrote:
> Do we actually need/use the fInvMatrix adjustment for the linear pipeline? 
> Rather than adding fRealInvMatrix, could we just skip the adjustment?

I'll add a todo to clean this up.

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkLinearBitma...
File src/core/SkLinearBitmapPipeline.cpp (right):

https://codereview.chromium.org/1775963002/diff/400001/src/core/SkLinearBitma...
src/core/SkLinearBitmapPipeline.cpp:56: // resulting Y values my be off the
tile. When that happens, the second Y is used to denote
On 2016/03/22 20:55:26, f(malita) wrote:
> "might"?

Reworded comment to be clearer.

[herb_g, 2016-03-22 21:11:33]

Address comments.

[herb_g, 2016-03-22 21:12:07]

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

[commit-bot: I haz the power, 2016-03-22 21:12:17]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1775963002/440001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1775963002/440001

[commit-bot: I haz the power, 2016-03-22 21:26:29]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-03-22 21:26:30]

Dry run: This issue passed the CQ dry run.

[herb_g, 2016-03-23 15:59:32]

The CQ bit was checked by herb@google.com

[herb_g, 2016-03-23 15:59:33]

The patchset sent to the CQ was uploaded after l-g-t-m from
fmalita@chromium.org, reed@google.com
Link to the patchset: https://codereview.chromium.org/1775963002/#ps440001
(title: "Address comments.")

[commit-bot: I haz the power, 2016-03-23 15:59:36]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1775963002/440001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1775963002/440001

[commit-bot: I haz the power, 2016-03-23 16:00:36]

Description was changed from

==========
WIP: experimental bilerp pipeline.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
WIP: experimental bilerp pipeline.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...

Committed:
https://skia.googlesource.com/skia/+/6eff52afb458bf6702a715d88611fd571544ef73
==========

[commit-bot: I haz the power, 2016-03-23 16:00:37]

Committed patchset #23 (id:440001) as
https://skia.googlesource.com/skia/+/6eff52afb458bf6702a715d88611fd571544ef73