Added fNumTransformsExclChildren and fNumTexturesExclChildren; changed emitChild() to forward search

This change is in preparation for updating how processor keys and meta keys are generated for frag procs.

BUG=skia:4182
Committed: https://skia.googlesource.com/skia/+/93ab254b7e4ce82074e88e219a8e32b31086ca86

[wangyix, 2015-08-18 19:21:56]

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

[tomhudson, 2015-08-18 19:36:11]

lgtm

https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
File include/gpu/GrFragmentProcessor.h (right):

https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
include/gpu/GrFragmentProcessor.h:45: int numTexturesExclChildren() const {
return fNumTexturesExclChildren; }
nit: abbreviation in API smells bad - but I can understand wanting the name
shorter, and aren't coming up with better alternatives

[joshualitt, 2015-08-19 13:34:28]

On 2015/08/18 19:36:11, tomhudson wrote:
> lgtm
> 
>
https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
> File include/gpu/GrFragmentProcessor.h (right):
> 
>
https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
> include/gpu/GrFragmentProcessor.h:45: int numTexturesExclChildren() const {
> return fNumTexturesExclChildren; }
> nit: abbreviation in API smells bad - but I can understand wanting the name
> shorter, and aren't coming up with better alternatives

Looks fine to me, the thing about having to register child after texture /
coords is annoying but given we store these in a flat array I'm not sure what
else to do.  Brian might have better ideas about the name.

[bsalomon, 2015-08-19 14:22:17]

lgtm, but I'm liking the bubbling up less as we go along. Might rather have
complexity in the builder/descriptor code to iterate over the tree.

https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
File include/gpu/GrFragmentProcessor.h (right):

https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
include/gpu/GrFragmentProcessor.h:45: int numTexturesExclChildren() const {
return fNumTexturesExclChildren; }
On 2015/08/18 19:36:11, tomhudson wrote:
> nit: abbreviation in API smells bad - but I can understand wanting the name
> shorter, and aren't coming up with better alternatives

Nonrecursive?

I can't come up with anything I really like.

[wangyix, 2015-08-19 15:04:35]

On 2015/08/19 14:22:17, bsalomon wrote:
> lgtm, but I'm liking the bubbling up less as we go along. Might rather have
> complexity in the builder/descriptor code to iterate over the tree.
> 
>
https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
> File include/gpu/GrFragmentProcessor.h (right):
> 
>
https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
> include/gpu/GrFragmentProcessor.h:45: int numTexturesExclChildren() const {
> return fNumTexturesExclChildren; }
> On 2015/08/18 19:36:11, tomhudson wrote:
> > nit: abbreviation in API smells bad - but I can understand wanting the name
> > shorter, and aren't coming up with better alternatives
> 
> Nonrecursive?
> 
> I can't come up with anything I really like.

numIntrinsicTextures()?
I guess the name of this isn't a huge concern since we want to git rid of the
bubbling-up of textures/transforms eventually?

[joshualitt, 2015-08-19 15:04:59]

On 2015/08/19 15:04:35, wangyix wrote:
> On 2015/08/19 14:22:17, bsalomon wrote:
> > lgtm, but I'm liking the bubbling up less as we go along. Might rather have
> > complexity in the builder/descriptor code to iterate over the tree.
> > 
> >
>
https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
> > File include/gpu/GrFragmentProcessor.h (right):
> > 
> >
>
https://codereview.chromium.org/1298233002/diff/1/include/gpu/GrFragmentProce...
> > include/gpu/GrFragmentProcessor.h:45: int numTexturesExclChildren() const {
> > return fNumTexturesExclChildren; }
> > On 2015/08/18 19:36:11, tomhudson wrote:
> > > nit: abbreviation in API smells bad - but I can understand wanting the
name
> > > shorter, and aren't coming up with better alternatives
> > 
> > Nonrecursive?
> > 
> > I can't come up with anything I really like.
> 
> numIntrinsicTextures()?
> I guess the name of this isn't a huge concern since we want to git rid of the
> bubbling-up of textures/transforms eventually?

yea, lets not worry about the name for now.

[wangyix, 2015-08-19 15:06:31]

The CQ bit was checked by wangyix@google.com

[commit-bot: I haz the power, 2015-08-19 15:06:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1298233002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1298233002/1

[commit-bot: I haz the power, 2015-08-19 15:23:14]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/skia/+/93ab254b7e4ce82074e88e219a8e32b31086ca86

[wangyix, 2015-08-19 17:10:05]

Patchset #2 (id:20001) has been deleted