Implement crop rect support for SkMatrixConvolutionImageFilter.

Implement crop rect support for SkMatrixConvolutionImageFilter.

R=bsalomon@google.com

Committed: https://code.google.com/p/skia/source/detail?r=11863

[Stephen White, 2013-10-18 18:45:27]

Brian: PTAL. Thanks!

[bsalomon, 2013-10-18 18:56:22]

lgtm w/ a couple small inline comments

https://codereview.chromium.org/27471002/diff/36001/gm/matrixconvolution.cpp
File gm/matrixconvolution.cpp (right):

https://codereview.chromium.org/27471002/diff/36001/gm/matrixconvolution.cpp#...
gm/matrixconvolution.cpp:80: typedef SkMatrixConvolutionImageFilter MCIF;
nice :)

https://codereview.chromium.org/27471002/diff/36001/gm/matrixconvolution.cpp#...
gm/matrixconvolution.cpp:97: draw(canvas, 310, 10, target,
MCIF::kClamp_TileMode, true, &rect);
this->

https://codereview.chromium.org/27471002/diff/36001/src/effects/SkMatrixConvo...
File src/effects/SkMatrixConvolutionImageFilter.cpp (right):

https://codereview.chromium.org/27471002/diff/36001/src/effects/SkMatrixConvo...
src/effects/SkMatrixConvolutionImageFilter.cpp:593: SkIRect bounds =
SkIRect::MakeXYWH(random->nextRangeU(0, textures[texIdx]->width()),
I guess it doesn't really matter, but it's a little weird to generate bounds
where left > right or top > bottom.

[reed1, 2013-10-18 18:57:35]

https://codereview.chromium.org/27471002/diff/36001/include/core/SkImageFilter.h
File include/core/SkImageFilter.h (right):

https://codereview.chromium.org/27471002/diff/36001/include/core/SkImageFilte...
include/core/SkImageFilter.h:190: const SkIRect& bounds) const;
shot in the dark : is this correct to be IRect, and not Rect?

[Stephen White, 2013-10-18 19:47:16]

On 2013/10/18 18:57:35, reed1 wrote:
>
https://codereview.chromium.org/27471002/diff/36001/include/core/SkImageFilter.h
> File include/core/SkImageFilter.h (right):
> 
>
https://codereview.chromium.org/27471002/diff/36001/include/core/SkImageFilte...
> include/core/SkImageFilter.h:190: const SkIRect& bounds) const;
> shot in the dark : is this correct to be IRect, and not Rect?

Yes, it's correct -- by this time, we're in device space, so we've rounded up to
the nearest pixel.

[Stephen White, 2013-10-18 20:05:44]

Thanks for your review!

https://codereview.chromium.org/27471002/diff/36001/gm/matrixconvolution.cpp
File gm/matrixconvolution.cpp (right):

https://codereview.chromium.org/27471002/diff/36001/gm/matrixconvolution.cpp#...
gm/matrixconvolution.cpp:97: draw(canvas, 310, 10, target,
MCIF::kClamp_TileMode, true, &rect);
On 2013/10/18 18:56:23, bsalomon wrote:
> this->

Done.

https://codereview.chromium.org/27471002/diff/36001/src/effects/SkMatrixConvo...
File src/effects/SkMatrixConvolutionImageFilter.cpp (right):

https://codereview.chromium.org/27471002/diff/36001/src/effects/SkMatrixConvo...
src/effects/SkMatrixConvolutionImageFilter.cpp:593: SkIRect bounds =
SkIRect::MakeXYWH(random->nextRangeU(0, textures[texIdx]->width()),
On 2013/10/18 18:56:23, bsalomon wrote:
> I guess it doesn't really matter, but it's a little weird to generate bounds
> where left > right or top > bottom.

True, but we might catch more bugs this way too. :)

[Stephen White, 2013-10-18 20:08:23]

Committed patchset #5 manually as r11863 (presubmit successful).