Merge changes from the android repo upstream to Skia

Merge changes from the android repo upstream to Skia
Committed: https://code.google.com/p/skia/source/detail?r=3199

[reed1, 2012-01-17 14:36:54]

various small comments.
can we create a separate CL for the imagedecoder changes?

http://codereview.appspot.com/5545070/diff/1/include/core/SkScalerContext.h
File include/core/SkScalerContext.h (right):

http://codereview.appspot.com/5545070/diff/1/include/core/SkScalerContext.h#n...
include/core/SkScalerContext.h:266: SkScalerContext*
charToScalerContext(SkUnichar uni);
who calls this?

http://codereview.appspot.com/5545070/diff/1/include/views/SkEvent.h
File include/views/SkEvent.h (right):

http://codereview.appspot.com/5545070/diff/1/include/views/SkEvent.h#newcode264
include/views/SkEvent.h:264: #elif defined(SK_BUILD_FOR_UNIX)
Does this change fix something on the unix build? (i.e. have we tested this?)

http://codereview.appspot.com/5545070/diff/1/src/core/SkScalerContext.cpp
File src/core/SkScalerContext.cpp (right):

http://codereview.appspot.com/5545070/diff/1/src/core/SkScalerContext.cpp#new...
src/core/SkScalerContext.cpp:193: if (NULL == ctx) {
Should we leave in this SkDebugf? Is it better to return this or null?

[DerekS, 2012-01-17 14:52:27]

I've removed the image specific changes and have put them in a separate CL.

http://codereview.appspot.com/5545070/diff/1/include/core/SkScalerContext.h
File include/core/SkScalerContext.h (right):

http://codereview.appspot.com/5545070/diff/1/include/core/SkScalerContext.h#n...
include/core/SkScalerContext.h:266: SkScalerContext*
charToScalerContext(SkUnichar uni);
On 2012/01/17 14:36:59, reed1 wrote:
> who calls this?

SkGlyphCache.cpp has another android specific function that calls this.

http://codereview.appspot.com/5545070/diff/1/include/views/SkEvent.h
File include/views/SkEvent.h (right):

http://codereview.appspot.com/5545070/diff/1/include/views/SkEvent.h#newcode264
include/views/SkEvent.h:264: #elif defined(SK_BUILD_FOR_UNIX)
On 2012/01/17 14:36:59, reed1 wrote:
> Does this change fix something on the unix build? (i.e. have we tested this?)

good question. I'll see why it went into the android tree and I'll test it on
the linux sample app.

[DerekS, 2012-02-07 16:22:49]

I've updated the patch to include the most recent changes as a result of the
latest merge of Skia to android.

On 2012/01/17 14:52:27, djsollen wrote:
> I've removed the image specific changes and have put them in a separate CL.
> 
> http://codereview.appspot.com/5545070/diff/1/include/core/SkScalerContext.h
> File include/core/SkScalerContext.h (right):
> 
>
http://codereview.appspot.com/5545070/diff/1/include/core/SkScalerContext.h#n...
> include/core/SkScalerContext.h:266: SkScalerContext*
> charToScalerContext(SkUnichar uni);
> On 2012/01/17 14:36:59, reed1 wrote:
> > who calls this?
> 
> SkGlyphCache.cpp has another android specific function that calls this.
> 
> http://codereview.appspot.com/5545070/diff/1/include/views/SkEvent.h
> File include/views/SkEvent.h (right):
> 
>
http://codereview.appspot.com/5545070/diff/1/include/views/SkEvent.h#newcode264
> include/views/SkEvent.h:264: #elif defined(SK_BUILD_FOR_UNIX)
> On 2012/01/17 14:36:59, reed1 wrote:
> > Does this change fix something on the unix build? (i.e. have we tested
this?)
> 
> good question. I'll see why it went into the android tree and I'll test it on
> the linux sample app.

[reed1, 2012-02-13 14:33:14]

lgtm

http://codereview.appspot.com/5545070/diff/8001/src/ports/SkFontHost_android.cpp
File src/ports/SkFontHost_android.cpp (left):

http://codereview.appspot.com/5545070/diff/8001/src/ports/SkFontHost_android....
src/ports/SkFontHost_android.cpp:108: // should never get here, since the faces
list should not be empty
Do you not have SkDEBUGFAIL? I think this is the preferred technique now, rather
than !"text"

[TomH, 2012-02-13 15:37:50]

SkDEBUGFAIL must be used in headers to avoid compile breakage in our clients.
SkDEBUGFAIL is nice to use in cpp files, but there hasn't been much
dissemination yet.