New approach for GPU font atlas

New approach for GPU font atlas

In the previous code, each GrTextStrike had exclusive access to one or more GrPlots in the texture atlas. This led to poor packing when only a few glyphs were used. This change allows GrTextStrikes to share GrPlots, thereby getting much better utilization of the entire texture.

BUG=skia:2224
Committed: http://code.google.com/p/skia/source/detail?r=13636

[bsalomon, 2014-02-28 16:28:52]

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.cpp
File src/gpu/GrAtlas.cpp (right):

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.cpp#newc...
src/gpu/GrAtlas.cpp:150: if (NULL == plot || NULL == plot->fPrev) {
I wonder if we should be using SkTInternalLList

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.cpp#newc...
src/gpu/GrAtlas.cpp:179: moveToHead(plot);
this->

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.h
File src/gpu/GrAtlas.h (right):

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.h#newcode78
src/gpu/GrAtlas.h:78: GrPlot* getUnusedPlot();
This seems a little confusing to me. I read the comment several times to try to
understand what "unused" meant. 

What's the reason that we would want a plot that only contains rendered content?
Avoiding flushes?

[jvanverth1, 2014-02-28 19:55:26]

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.cpp
File src/gpu/GrAtlas.cpp (right):

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.cpp#newc...
src/gpu/GrAtlas.cpp:150: if (NULL == plot || NULL == plot->fPrev) {
On 2014/02/28 16:28:53, bsalomon wrote:
> I wonder if we should be using SkTInternalLList

Done.

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.cpp#newc...
src/gpu/GrAtlas.cpp:179: moveToHead(plot);
On 2014/02/28 16:28:53, bsalomon wrote:
> this->

Done.

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.h
File src/gpu/GrAtlas.h (right):

https://codereview.chromium.org/177463003/diff/60001/src/gpu/GrAtlas.h#newcode78
src/gpu/GrAtlas.h:78: GrPlot* getUnusedPlot();
On 2014/02/28 16:28:53, bsalomon wrote:
> This seems a little confusing to me. I read the comment several times to try
to
> understand what "unused" meant. 
> 
> What's the reason that we would want a plot that only contains rendered
content?
> Avoiding flushes?

I've revised the comment -- I can't think of a better method name than this.

[bsalomon, 2014-02-28 20:49:22]

lgtm

[jvanverth1, 2014-03-03 13:36:22]

The CQ bit was checked by jvanverth@google.com

[commit-bot: I haz the power, 2014-03-03 13:36:32]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/jvanverth@google.com/177463003/100001

[commit-bot: I haz the power, 2014-03-03 13:44:22]

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

[commit-bot: I haz the power, 2014-03-03 13:44:22]

Retried try job too often on Build-Mac10.8-Clang-x86-Release-Trybot for step(s)
BuildSkiaLib, BuildTests
http://108.170.219.164:10117/buildstatus?builder=Build-Mac10.8-Clang-x86-Rele...

[jvanverth1, 2014-03-03 14:15:46]

The CQ bit was checked by jvanverth@google.com

[commit-bot: I haz the power, 2014-03-03 14:15:52]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/jvanverth@google.com/177463003/120001

[commit-bot: I haz the power, 2014-03-03 14:30:28]

Change committed as 13636