New GM for testing circle blur optimization

New GM for testing circle blur optimization

BUG=skia:

Committed: http://code.google.com/p/skia/source/detail?r=14621

[humper, 2014-05-07 14:35:58]

PTAL, just getting ready for testing GPU circle blur goodness.

[humper, 2014-05-07 14:38:49]

Here's a link to the output image if you care:

https://drive.google.com/a/google.com/file/d/0B4MzzMVAC7DgMFhOT3hmTGg3MjQ/edi...

[bsalomon, 2014-05-07 14:43:02]

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp
File gm/blurcircles.cpp (right):

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp#newcode2
gm/blurcircles.cpp:2: * Copyright 2013 Google Inc.
2014

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp#newcode30
gm/blurcircles.cpp:30: return kSkipTiled_Flag;
curious why this is skipped

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp#newcode48
gm/blurcircles.cpp:48: SkAutoCanvasRestore autoRestore(canvas, true);
are these lines weirdly indented or is that the rietveld chrome extension being
funky?

[humper, 2014-05-07 17:09:42]

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp
File gm/blurcircles.cpp (right):

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp#newcode30
gm/blurcircles.cpp:30: return kSkipTiled_Flag;
On 2014/05/07 14:43:02, bsalomon wrote:
> curious why this is skipped

Good question -- I copied it from another GM.  I'll remove it.

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp#newcode48
gm/blurcircles.cpp:48: SkAutoCanvasRestore autoRestore(canvas, true);
On 2014/05/07 14:43:02, bsalomon wrote:
> are these lines weirdly indented or is that the rietveld chrome extension
being
> funky?

I did this on Windows; looks like don't have my spaces vs. tabs settings quite
right yet.  Will fix.

[mtklein, 2014-05-07 17:13:59]

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp
File gm/blurcircles.cpp (right):

https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp#newcode30
gm/blurcircles.cpp:30: return kSkipTiled_Flag;
On 2014/05/07 17:09:42, humper wrote:
> On 2014/05/07 14:43:02, bsalomon wrote:
> > curious why this is skipped
> 
> Good question -- I copied it from another GM.  I'll remove it.

Might want to try out/Debug/dm --match blurcircles first.  Many GMs must
actually set kSkipTiled_Flag right now while we sort out a good way to relax the
bitmap equality checks for our new quilt mode (renders into 16x16 squares that
should stitch up into the original bitmap).  You'll see a _quilt mode test fail
if you remove kSkipTiled_Flag when it should be there.

[humper, 2014-05-07 17:26:40]

please take another look

[bsalomon, 2014-05-07 17:31:44]

https://codereview.chromium.org/276453002/diff/20001/gm/blurcircles.cpp
File gm/blurcircles.cpp (right):

https://codereview.chromium.org/276453002/diff/20001/gm/blurcircles.cpp#newco...
gm/blurcircles.cpp:13: #include "SkLayerDrawLooper.h"
I think some of these #includes can be removed (path, draw looper, path, point,
rect, rrect, xfermode .

Otherwise lgtm if it passes mtklein's quilt dm test.

[humper, 2014-05-07 17:31:49]

On 2014/05/07 17:13:59, mtklein wrote:
> https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp
> File gm/blurcircles.cpp (right):
> 
> https://codereview.chromium.org/276453002/diff/1/gm/blurcircles.cpp#newcode30
> gm/blurcircles.cpp:30: return kSkipTiled_Flag;
> On 2014/05/07 17:09:42, humper wrote:
> > On 2014/05/07 14:43:02, bsalomon wrote:
> > > curious why this is skipped
> > 
> > Good question -- I copied it from another GM.  I'll remove it.
> 
> Might want to try out/Debug/dm --match blurcircles first.  Many GMs must
> actually set kSkipTiled_Flag right now while we sort out a good way to relax
the
> bitmap equality checks for our new quilt mode (renders into 16x16 squares that
> should stitch up into the original bitmap).  You'll see a _quilt mode test
fail
> if you remove kSkipTiled_Flag when it should be there.

Good eye, thanks.  DM did in fact fail when I removed the skip tiled, so I'll
add it back in.

[mtklein, 2014-05-07 17:36:39]

> Good eye, thanks.  DM did in fact fail when I removed the skip tiled, so I'll
> add it back in.

Cool beans.  FYI, Mike and I are in the let-ideas-stew stage of how to handle
this sort of thing more gracefully in the future.

[humper, 2014-05-07 17:38:30]

The CQ bit was checked by humper@google.com

[commit-bot: I haz the power, 2014-05-07 17:39:28]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/humper@google.com/276453002/60001

[commit-bot: I haz the power, 2014-05-07 17:50:13]

Change committed as 14621