add asAShadowBlur for android to drawlooper

add asAShadowBlur for android to drawlooper

BUG=skia:
R=djsollen@google.com, scroggo@google.com

Committed: https://code.google.com/p/skia/source/detail?r=14431

[reed1, 2014-04-25 21:35:08]

Things still to do:

1. write some tests
2. confirm the math in both Convert utilities
3. guard with #ifdef ANDROID (?)
4. what to do about the current flags (e.g. IgnoreTransform) that Android does
not know about...

[scroggo, 2014-04-28 13:28:05]

https://codereview.chromium.org/253633003/diff/1/include/core/SkDrawLooper.h
File include/core/SkDrawLooper.h (right):

https://codereview.chromium.org/253633003/diff/1/include/core/SkDrawLooper.h#...
include/core/SkDrawLooper.h:107: virtual bool asABlurShadow(BlurShadowRec*)
const;
Does this need to be on the base class? It seems very specific to one subclass.

[djsollen, 2014-04-28 13:55:40]

https://codereview.chromium.org/253633003/diff/1/include/core/SkDrawLooper.h
File include/core/SkDrawLooper.h (right):

https://codereview.chromium.org/253633003/diff/1/include/core/SkDrawLooper.h#...
include/core/SkDrawLooper.h:94: SkScalar    fRadius;
As we discussed offline, I think this should return sigma so we can let the
caller do the conversion if they want to.

https://codereview.chromium.org/253633003/diff/1/include/core/SkDrawLooper.h#...
include/core/SkDrawLooper.h:107: virtual bool asABlurShadow(BlurShadowRec*)
const;
On 2014/04/28 13:28:06, scroggo wrote:
> Does this need to be on the base class? It seems very specific to one
subclass.

It does need to be on the base class since it allows you to inspect any
DrawLooper and see if it is just implementing a blur.  The subclasses that do
not will just return false.

https://codereview.chromium.org/253633003/diff/1/src/effects/SkBlurDrawLooper...
File src/effects/SkBlurDrawLooper.cpp (right):

https://codereview.chromium.org/253633003/diff/1/src/effects/SkBlurDrawLooper...
src/effects/SkBlurDrawLooper.cpp:88: buffer.writeScalar(fDy);
should we serialize fSigma too?

[reed1, 2014-04-28 20:59:26]

ptal

https://codereview.chromium.org/253633003/diff/1/src/effects/SkBlurDrawLooper...
File src/effects/SkBlurDrawLooper.cpp (right):

https://codereview.chromium.org/253633003/diff/1/src/effects/SkBlurDrawLooper...
src/effects/SkBlurDrawLooper.cpp:88: buffer.writeScalar(fDy);
On 2014/04/28 13:55:40, djsollen wrote:
> should we serialize fSigma too?

Done.

[scroggo, 2014-04-28 21:19:12]

lgtm

https://codereview.chromium.org/253633003/diff/10001/src/effects/SkBlurDrawLo...
File src/effects/SkBlurDrawLooper.cpp (right):

https://codereview.chromium.org/253633003/diff/10001/src/effects/SkBlurDrawLo...
src/effects/SkBlurDrawLooper.cpp:74: fSigma = buffer.readScalar();
Did you say chrome never uses this? If not, I guess there's no need to worry
about the picture version?

[djsollen, 2014-04-29 12:53:36]

lgtm

https://codereview.chromium.org/253633003/diff/30001/src/effects/SkBlurDrawLo...
File src/effects/SkBlurDrawLooper.cpp (right):

https://codereview.chromium.org/253633003/diff/30001/src/effects/SkBlurDrawLo...
src/effects/SkBlurDrawLooper.cpp:98: if (fSigma <= 0 || (fBlurFlags & fBlurFlags
& kIgnoreTransform_BlurFlag)) {
can you add extra parans to your bit and to make it clear the order of operation
you expect?

[reed1, 2014-04-29 13:54:45]

The CQ bit was checked by reed@google.com

[commit-bot: I haz the power, 2014-04-29 13:55:02]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/reed@google.com/253633003/50001

[reed1, 2014-04-29 15:18:53]

The CQ bit was unchecked by reed@google.com

[reed1, 2014-04-29 15:20:44]

Committed patchset #4 manually as r14431 (presubmit successful).