Remove Bitmaps Raw Pixel Support.

Remove Bitmaps Raw Pixel Support.

bitmap.setPixels(...) now creates a mutable pixelRef instead
of just setting fPixels.

[DerekS, 2012-07-18 14:01:44]

Here is a new CL to address step 1 of the changes I outlined in
http://codereview.appspot.com/6344076/

[reed1, 2012-07-18 16:13:37]

http://codereview.appspot.com/6419044/diff/3001/src/core/SkMallocPixelRef.cpp
File src/core/SkMallocPixelRef.cpp (right):

http://codereview.appspot.com/6419044/diff/3001/src/core/SkMallocPixelRef.cpp...
src/core/SkMallocPixelRef.cpp:13: SkColorTable* ctable, bool managePixels) {
As written, if I call

SkMallocPixelRef(NULL, 100, false)

We will have NULL for fStorage, and then crash.

My early comment was meant to notice that if the caller is passing NULL for
storage, then *they must be asking us to alloc/dealloc the storage*, so that
managePixels *must* be true.

[TomH, 2012-07-19 19:06:26]

http://codereview.appspot.com/6419044/diff/3001/include/core/SkBitmap.h
File include/core/SkBitmap.h (right):

http://codereview.appspot.com/6419044/diff/3001/include/core/SkBitmap.h#newco...
include/core/SkBitmap.h:363: /** Returns a non-zero, unique value corresponding
to the pixels in our
"unique" to this particular SkBitmap instance's history, right? Or "unique"
across all SkBitmaps for the entire program execution? This reads to me like the
latter, even though I suspect it's the former.

http://codereview.appspot.com/6419044/diff/3001/include/core/SkBitmap.h#newco...
include/core/SkBitmap.h:363: /** Returns a non-zero, unique value corresponding
to the pixels in our
Also, when (0 == fPixelRef), the implementation returns 0.

[DerekS, 2012-07-23 12:47:31]

http://codereview.appspot.com/6419044/diff/3001/include/core/SkBitmap.h
File include/core/SkBitmap.h (right):

http://codereview.appspot.com/6419044/diff/3001/include/core/SkBitmap.h#newco...
include/core/SkBitmap.h:363: /** Returns a non-zero, unique value corresponding
to the pixels in our
On 2012/07/19 19:06:26, TomH wrote:
> Also, when (0 == fPixelRef), the implementation returns 0. 

I updated the comment to include that we return 0 if there is no pixelRef. 
However, I think the comment as written is correct with respect to uniqueness as
the pixelRefs are globally unique with respect to one another.

http://codereview.appspot.com/6419044/diff/3001/src/core/SkMallocPixelRef.cpp
File src/core/SkMallocPixelRef.cpp (right):

http://codereview.appspot.com/6419044/diff/3001/src/core/SkMallocPixelRef.cpp...
src/core/SkMallocPixelRef.cpp:13: SkColorTable* ctable, bool managePixels) {
On 2012/07/18 16:13:37, reed1 wrote:
> As written, if I call
> 
> SkMallocPixelRef(NULL, 100, false)
> 
> We will have NULL for fStorage, and then crash.
> 
> My early comment was meant to notice that if the caller is passing NULL for
> storage, then *they must be asking us to alloc/dealloc the storage*, so that
> managePixels *must* be true.

Done.

[junov1, 2012-07-23 14:07:24]

Suggestion: managePixels -> ownPixels

IMHO: The term manage usually implies more than just being responsible for
releasing a resource.

lgtm