Plumb in "bleed" flag

This CL plumbs in the drawBitmapRectToRect "bleed" flag and makes it live on the simplest GPU path.

Committed: http://code.google.com/p/skia/source/detail?r=10765

[jvanverth1, 2013-07-29 14:35:37]

https://codereview.chromium.org/20806003/diff/51001/include/core/SkPicture.h
File include/core/SkPicture.h (right):

https://codereview.chromium.org/20806003/diff/51001/include/core/SkPicture.h#...
include/core/SkPicture.h:212: static const uint32_t PRIOR_PICTURE_VERSION = 12; 
// TODO: remove when .skps regenerated
Might be good to add a #define to bracket all of the code related to this, and
make it easier to find when it's time to remove it.

[edisonn, 2013-07-29 14:51:13]

https://codereview.chromium.org/20806003/diff/51001/src/core/SkCanvas.cpp
File src/core/SkCanvas.cpp (right):

https://codereview.chromium.org/20806003/diff/51001/src/core/SkCanvas.cpp#new...
src/core/SkCanvas.cpp:1865: const SkRect& dst, const SkPaint* paint) {
I am wondering why drawBitmapNine does not take the extra flag.

https://codereview.chromium.org/20806003/diff/51001/src/core/SkPicture.cpp
File src/core/SkPicture.cpp (right):

https://codereview.chromium.org/20806003/diff/51001/src/core/SkPicture.cpp#ne...
src/core/SkPicture.cpp:277: PRIOR_PICTURE_VERSION != info.fVersion) { // TODO:
remove when .skps regenerated
since there are multiple places that need to be updated, when the .skps are
regenerated, could we add an issue, and mark all the comments with the issue
number? That would make it simple to be found. then no need to explain.

If we do this on regular basis, at one point someone will forget to remove some
code.

Also I understand that introducing a process like this would annoy devs (it
would surely annoy me) so I let it up to you to make a cost/benefit judgment.

https://codereview.chromium.org/20806003/diff/51001/src/pdf/SkPDFDevice.cpp
File src/pdf/SkPDFDevice.cpp (right):

https://codereview.chromium.org/20806003/diff/51001/src/pdf/SkPDFDevice.cpp#n...
src/pdf/SkPDFDevice.cpp:862: SkCanvas::DrawBitmapRectFlags flags) {
add a todo here to implement the flag

[robertphillips, 2013-07-29 15:27:31]

Addressed initial code review comments

https://codereview.chromium.org/20806003/diff/51001/include/core/SkPicture.h
File include/core/SkPicture.h (right):

https://codereview.chromium.org/20806003/diff/51001/include/core/SkPicture.h#...
include/core/SkPicture.h:212: static const uint32_t PRIOR_PICTURE_VERSION = 12; 
// TODO: remove when .skps regenerated
On 2013/07/29 14:35:37, JimVV wrote:
> Might be good to add a #define to bracket all of the code related to this, and
> make it easier to find when it's time to remove it.

Done.

https://codereview.chromium.org/20806003/diff/51001/src/core/SkCanvas.cpp
File src/core/SkCanvas.cpp (right):

https://codereview.chromium.org/20806003/diff/51001/src/core/SkCanvas.cpp#new...
src/core/SkCanvas.cpp:1865: const SkRect& dst, const SkPaint* paint) {
drawBitmapNine operates on the entire bitmap rather then a sub-window.
Additionally the semantics of drawBitmapNine appears to be that no blurring
occurs between the nine panels.

https://codereview.chromium.org/20806003/diff/51001/src/core/SkPicture.cpp
File src/core/SkPicture.cpp (right):

https://codereview.chromium.org/20806003/diff/51001/src/core/SkPicture.cpp#ne...
src/core/SkPicture.cpp:277: PRIOR_PICTURE_VERSION != info.fVersion) { // TODO:
remove when .skps regenerated
I have used a huge #define as suggested by Jim.

https://codereview.chromium.org/20806003/diff/51001/src/pdf/SkPDFDevice.cpp
File src/pdf/SkPDFDevice.cpp (right):

https://codereview.chromium.org/20806003/diff/51001/src/pdf/SkPDFDevice.cpp#n...
src/pdf/SkPDFDevice.cpp:862: SkCanvas::DrawBitmapRectFlags flags) {
On 2013/07/29 14:51:14, edisonn wrote:
> add a todo here to implement the flag

Done.

[reed1, 2013-07-29 15:53:35]

Seems like, but a little ucky (this is SO CLEARLY) and api hack just to support
a wacky client request).

Could the client be happy with calling drawVertices() instead, since this will
already show the bleed?

https://codereview.chromium.org/20806003/diff/51001/include/core/SkCanvas.h
File include/core/SkCanvas.h (right):

https://codereview.chromium.org/20806003/diff/51001/include/core/SkCanvas.h#n...
include/core/SkCanvas.h:671: kNone_DrawBitmapRectflag            = 0x0,
kClamp instead of kNone?

[robertphillips, 2013-08-01 12:13:14]

Parked in favor of a different approach.

[robertphillips, 2013-08-13 16:53:47]

PTAL - back from the dead.

[edisonn, 2013-08-13 20:51:35]

LGTM

[scroggo, 2013-08-13 21:05:46]

Pipe changes look logically fine.

I guess the other approach was not fruitful?

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipePriv.h
File src/pipe/SkGPipePriv.h (right):

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipePriv.h#ne...
src/pipe/SkGPipePriv.h:153: kDrawBitmap_Bleed_DrawOpFlag = 1 << 2,
nit: = should line up with = for other flags.

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipeRead.cpp
File src/pipe/SkGPipeRead.cpp (right):

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipeRead.cpp#...
src/pipe/SkGPipeRead.cpp:577: dbmrFlags = (SkCanvas::DrawBitmapRectFlags)
(dbmrFlags | SkCanvas::kBleed_DrawBitmapRectFlag);
over 100 chars.

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipeRead.cpp#...
src/pipe/SkGPipeRead.cpp:582: canvas->drawBitmapRectToRect(*bitmap, src, *dst,
hasPaint ? &state->paint() : NULL, dbmrFlags);
over 100 chars.

[bsalomon, 2013-08-13 21:09:49]

lgtm

[robertphillips, 2013-08-14 15:37:41]

Addressed pipe-specific issues and updated to ToT.

Mike - you're the only one left.

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipePriv.h
File src/pipe/SkGPipePriv.h (right):

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipePriv.h#ne...
src/pipe/SkGPipePriv.h:153: kDrawBitmap_Bleed_DrawOpFlag = 1 << 2,
On 2013/08/13 21:05:46, scroggo wrote:
> nit: = should line up with = for other flags.

Done.

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipeRead.cpp
File src/pipe/SkGPipeRead.cpp (right):

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipeRead.cpp#...
src/pipe/SkGPipeRead.cpp:577: dbmrFlags = (SkCanvas::DrawBitmapRectFlags)
(dbmrFlags | SkCanvas::kBleed_DrawBitmapRectFlag);
On 2013/08/13 21:05:46, scroggo wrote:
> over 100 chars.

Done.

https://codereview.chromium.org/20806003/diff/56001/src/pipe/SkGPipeRead.cpp#...
src/pipe/SkGPipeRead.cpp:582: canvas->drawBitmapRectToRect(*bitmap, src, *dst,
hasPaint ? &state->paint() : NULL, dbmrFlags);
On 2013/08/13 21:05:46, scroggo wrote:
> over 100 chars.

Done.

[reed1, 2013-08-14 15:50:56]

(shuts eyes) lgtm

https://codereview.chromium.org/20806003/diff/70001/include/core/SkCanvas.h
File include/core/SkCanvas.h (right):

https://codereview.chromium.org/20806003/diff/70001/include/core/SkCanvas.h#n...
include/core/SkCanvas.h:677: kNone_DrawBitmapRectflag            = 0x0,
nit: flag should be Flag

[commit-bot: I haz the power, 2013-08-16 10:15:01]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/robertphillips@google.com/20806003/70001

[commit-bot: I haz the power, 2013-08-16 10:24:54]

Change committed as 10765