emitChild() takes an outputColor instead of generating one

emitChild() used to generate a mangled outputColor based on the parent's outputColor; now it just accepts an outputColor string.  It's now up to the programmer to declare outputColors if needed before emitting child code.

BUG=skia:4182
Committed: https://skia.googlesource.com/skia/+/54a6b1a1228489161294e4fa1fc8f93797cff259

[wangyix, 2015-08-31 20:33:08]

wangyix@google.com changed reviewers:
+ bsalomon@google.com, egdaniel@google.com, joshualitt@google.com,
tomhudson@google.com

[bsalomon, 2015-08-31 20:41:38]

otherwise lgtm

https://codereview.chromium.org/1321253003/diff/1/src/core/SkComposeShader.cpp
File src/core/SkComposeShader.cpp (right):

https://codereview.chromium.org/1321253003/diff/1/src/core/SkComposeShader.cp...
src/core/SkComposeShader.cpp:288: outputColorA.append("_dst");
Your code should already be enclosed in braces so I don't think you should need
to guard your vars with "_". Also, if you want you can just put "src" and "dst"
in string literals rather than use a SkString and %s's

[wangyix, 2015-08-31 22:01:22]

https://codereview.chromium.org/1321253003/diff/1/src/core/SkComposeShader.cpp
File src/core/SkComposeShader.cpp (right):

https://codereview.chromium.org/1321253003/diff/1/src/core/SkComposeShader.cp...
src/core/SkComposeShader.cpp:288: outputColorA.append("_dst");
On 2015/08/31 20:41:38, bsalomon wrote:
> Your code should already be enclosed in braces so I don't think you should
need
> to guard your vars with "_". Also, if you want you can just put "src" and
"dst"
> in string literals rather than use a SkString and %s's

Done.

[wangyix, 2015-08-31 22:01:40]

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

[commit-bot: I haz the power, 2015-08-31 22:01:53]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1321253003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1321253003/20001

[commit-bot: I haz the power, 2015-08-31 22:09:24]

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

[commit-bot: I haz the power, 2015-08-31 22:09:25]

Dry run: This issue passed the CQ dry run.

[joshualitt, 2015-09-01 13:59:14]

https://codereview.chromium.org/1321253003/diff/20001/src/gpu/gl/GrGLFragment...
File src/gpu/gl/GrGLFragmentProcessor.cpp (left):

https://codereview.chromium.org/1321253003/diff/20001/src/gpu/gl/GrGLFragment...
src/gpu/gl/GrGLFragmentProcessor.cpp:32: 
Is outputcolor no longer being mangled?

[wangyix, 2015-09-01 14:09:12]

https://codereview.chromium.org/1321253003/diff/20001/src/gpu/gl/GrGLFragment...
File src/gpu/gl/GrGLFragmentProcessor.cpp (left):

https://codereview.chromium.org/1321253003/diff/20001/src/gpu/gl/GrGLFragment...
src/gpu/gl/GrGLFragmentProcessor.cpp:32: 
On 2015/09/01 13:59:14, joshualitt wrote:
> Is outputcolor no longer being mangled?

outputColor is now an input, and is specified by whoever calls emitChild().  Its
their responsibility to mangle it if needed.  This was done so
GrExtractAlphaFragmentProcessor and similar wrapper processors can just do
emitChild(0, args.fInputColor, args.fOutputColor, args) instead of forcing it to
write to a mangled output color and then emitting "outputColor =
mangledOutputColor;" in the shader.

[joshualitt, 2015-09-01 14:48:51]

On 2015/09/01 14:09:12, wangyix wrote:
>
https://codereview.chromium.org/1321253003/diff/20001/src/gpu/gl/GrGLFragment...
> File src/gpu/gl/GrGLFragmentProcessor.cpp (left):
> 
>
https://codereview.chromium.org/1321253003/diff/20001/src/gpu/gl/GrGLFragment...
> src/gpu/gl/GrGLFragmentProcessor.cpp:32: 
> On 2015/09/01 13:59:14, joshualitt wrote:
> > Is outputcolor no longer being mangled?
> 
> outputColor is now an input, and is specified by whoever calls emitChild(). 
Its
> their responsibility to mangle it if needed.  This was done so
> GrExtractAlphaFragmentProcessor and similar wrapper processors can just do
> emitChild(0, args.fInputColor, args.fOutputColor, args) instead of forcing it
to
> write to a mangled output color and then emitting "outputColor =
> mangledOutputColor;" in the shader.

Okay, but at least SkComposeShader should definitely mangle dst/src.

[wangyix, 2015-09-01 17:32:47]

On 2015/09/01 14:48:51, joshualitt wrote:
> Okay, but at least SkComposeShader should definitely mangle dst/src.

So disregard Brian's comment and just go back to patchset 1's code?

[joshualitt, 2015-09-01 17:51:35]

On 2015/09/01 17:32:47, wangyix wrote:
> On 2015/09/01 14:48:51, joshualitt wrote:
> > Okay, but at least SkComposeShader should definitely mangle dst/src.
> 
> So disregard Brian's comment and just go back to patchset 1's code?

oh hmm, let me talk to him.  Hes right that the braces protect against other
fragment processors, but I'm worried about collisions with the children
processors.

[bsalomon, 2015-09-08 13:28:36]

On 2015/09/01 17:51:35, joshualitt wrote:
> On 2015/09/01 17:32:47, wangyix wrote:
> > On 2015/09/01 14:48:51, joshualitt wrote:
> > > Okay, but at least SkComposeShader should definitely mangle dst/src.
> > 
> > So disregard Brian's comment and just go back to patchset 1's code?
> 
> oh hmm, let me talk to him.  Hes right that the braces protect against other
> fragment processors, but I'm worried about collisions with the children
> processors.

I hadn't thought about that.. Makes me wish we were doing this with functions
instead of braces. We can keep the _'s for now. But let's file a bug to try
doing this with functions for children instead of braces.

[wangyix, 2015-09-08 15:33:32]

Patchset #2 (id:20001) has been deleted

[wangyix, 2015-09-08 15:36:36]

The CQ bit was checked by wangyix@google.com

[wangyix, 2015-09-08 15:36:36]

The patchset sent to the CQ was uploaded after l-g-t-m from bsalomon@google.com
Link to the patchset: https://codereview.chromium.org/1321253003/#ps40001
(title: "rebase")

[commit-bot: I haz the power, 2015-09-08 15:36:44]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1321253003/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1321253003/40001

[commit-bot: I haz the power, 2015-09-08 15:41:54]

Committed patchset #2 (id:40001) as
https://skia.googlesource.com/skia/+/54a6b1a1228489161294e4fa1fc8f93797cff259