replace SkBitmap::Config with SkColorType in gms

replace SkBitmap::Config with SkColorType in gms

add helper installMaskPixels() to SkBitmap

BUG=skia:

Committed: http://code.google.com/p/skia/source/detail?r=13482

[hal.canary, 2014-02-17 20:11:31]

https://codereview.chromium.org/169913003/diff/30001/gm/blurrect.cpp
File gm/blurrect.cpp (right):

https://codereview.chromium.org/169913003/diff/30001/gm/blurrect.cpp#newcode205
gm/blurrect.cpp:205: canvas->drawBitmap(bm, SkIntToScalar(center_x),
SkIntToScalar(center_y), NULL);
if that's a SkPicture-backed canvas, it will copty the bitmap, but the
AutoMaskFreeImage will have already freed the pixels.

https://codereview.chromium.org/169913003/diff/30001/include/core/SkBitmap.h
File include/core/SkBitmap.h (right):

https://codereview.chromium.org/169913003/diff/30001/include/core/SkBitmap.h#...
include/core/SkBitmap.h:324: *  of the created bitmap (and its pixelRef).
I'm always afraid of these assumptions that the mask pixels will remain valid. 
What if the bitmap gets copied accidentally?  Why not allocate an A8 SkBitmap
and then have a function that makes a mask out of that?

bool SkBitmap::getMask(SkMask* mask) {
  if (this->colorType() != kAlpha_8_SkColorType)
    return false;
  if (!this->getPixels()) // must lock bitmap first.
    return false;
  if (mask) {
    mask->fImage = (uint8_t*)(this->getPixels());
    mask->fBounds = SkIRect::MakeWH(this->width(), this->height());
    mask->fRowBytes = this->rowBytes;
    mask->fFormat = Skmask::kA8_Format;
  }
  return true
}

https://codereview.chromium.org/169913003/diff/30001/src/core/SkBitmap.cpp
File src/core/SkBitmap.cpp (right):

https://codereview.chromium.org/169913003/diff/30001/src/core/SkBitmap.cpp#ne...
src/core/SkBitmap.cpp:523: return
this->installPixels(SkImageInfo::MakeA8(mask.fBounds.width(),
You are assuming that (mask.fFormat == SkMask::kA8_Format)!

[reed1, 2014-02-17 20:33:25]

On 2014/02/17 20:11:31, Hal Canary wrote:
> https://codereview.chromium.org/169913003/diff/30001/gm/blurrect.cpp
> File gm/blurrect.cpp (right):
> 
>
https://codereview.chromium.org/169913003/diff/30001/gm/blurrect.cpp#newcode205
> gm/blurrect.cpp:205: canvas->drawBitmap(bm, SkIntToScalar(center_x),
> SkIntToScalar(center_y), NULL);
> if that's a SkPicture-backed canvas, it will copty the bitmap, but the
> AutoMaskFreeImage will have already freed the pixels.
> 
> https://codereview.chromium.org/169913003/diff/30001/include/core/SkBitmap.h
> File include/core/SkBitmap.h (right):
> 
>
https://codereview.chromium.org/169913003/diff/30001/include/core/SkBitmap.h#...
> include/core/SkBitmap.h:324: *  of the created bitmap (and its pixelRef).
> I'm always afraid of these assumptions that the mask pixels will remain valid.

> What if the bitmap gets copied accidentally?  Why not allocate an A8 SkBitmap
> and then have a function that makes a mask out of that?
> 
> bool SkBitmap::getMask(SkMask* mask) {
>   if (this->colorType() != kAlpha_8_SkColorType)
>     return false;
>   if (!this->getPixels()) // must lock bitmap first.
>     return false;
>   if (mask) {
>     mask->fImage = (uint8_t*)(this->getPixels());
>     mask->fBounds = SkIRect::MakeWH(this->width(), this->height());
>     mask->fRowBytes = this->rowBytes;
>     mask->fFormat = Skmask::kA8_Format;
>   }
>   return true
> }
> 
> https://codereview.chromium.org/169913003/diff/30001/src/core/SkBitmap.cpp
> File src/core/SkBitmap.cpp (right):
> 
>
https://codereview.chromium.org/169913003/diff/30001/src/core/SkBitmap.cpp#ne...
> src/core/SkBitmap.cpp:523: return
> this->installPixels(SkImageInfo::MakeA8(mask.fBounds.width(),
> You are assuming that (mask.fFormat == SkMask::kA8_Format)!

Done

[reed1, 2014-02-17 20:33:32]

https://codereview.chromium.org/169913003/diff/30001/gm/blurrect.cpp
File gm/blurrect.cpp (right):

https://codereview.chromium.org/169913003/diff/30001/gm/blurrect.cpp#newcode205
gm/blurrect.cpp:205: canvas->drawBitmap(bm, SkIntToScalar(center_x),
SkIntToScalar(center_y), NULL);
On 2014/02/17 20:11:31, Hal Canary wrote:
> if that's a SkPicture-backed canvas, it will copty the bitmap, but the
> AutoMaskFreeImage will have already freed the pixels.

Hmmm, if true, it was this way before this CL, yes?

[reed1, 2014-02-17 21:06:10]

Filed a bug to investigate what to do with any caller of installPixels that
doesn't supply a releaseproc.

ptal

[hal.canary, 2014-02-17 21:07:51]

lgtm

[reed1, 2014-02-17 21:08:19]

The CQ bit was checked by reed@google.com

[commit-bot: I haz the power, 2014-02-17 21:08:30]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/reed@google.com/169913003/40020

[commit-bot: I haz the power, 2014-02-17 21:21:53]

Change committed as 13482