Disable filtering in draw Bitmap operation when pixels are aligned

Disable filtering in draw Bitmap operation when pixels are aligned

The check is now performed in the callers of internalDrawBitmap() to
avoid retrieve the texture with wrong filter mode when the pixels were
aligned.

This bug was spotted in Chrome for Android: Chrome for Androids's
Canvas drawImage loses fidelity in canvases larger than 64k pixels.

BUG=chromium:231916
R=bsalomon@google.com, robertphillips@google.com

Signed-off-by: Samuel Iglesias Gonsalvez <siglesias@igalia.com>;

Committed: http://code.google.com/p/skia/source/detail?r=14665

[bsalomon, 2014-05-08 14:13:23]

https://codereview.chromium.org/264303008/diff/1/src/gpu/SkGpuDevice.cpp
File src/gpu/SkGpuDevice.cpp (right):

https://codereview.chromium.org/264303008/diff/1/src/gpu/SkGpuDevice.cpp#newc...
src/gpu/SkGpuDevice.cpp:1192: const SkMatrix& matrix = fContext->getMatrix();
Is accessing the matrix the only reason why this is a member rather than a
static function? If so maybe we could pass it in as a param? It's really minor,
but I just don't love having a function that is so specific to one narrow code
path in the header. Otherwise this lgtm.

[siglesias, 2014-05-08 14:14:59]

On 2014/05/08 14:13:23, bsalomon wrote:
> https://codereview.chromium.org/264303008/diff/1/src/gpu/SkGpuDevice.cpp
> File src/gpu/SkGpuDevice.cpp (right):
> 
>
https://codereview.chromium.org/264303008/diff/1/src/gpu/SkGpuDevice.cpp#newc...
> src/gpu/SkGpuDevice.cpp:1192: const SkMatrix& matrix = fContext->getMatrix();
> Is accessing the matrix the only reason why this is a member rather than a
> static function? If so maybe we could pass it in as a param? It's really
minor,
> but I just don't love having a function that is so specific to one narrow code
> path in the header. Otherwise this lgtm.

OK, I will change it.

Thanks for your feedback!

[bsalomon, 2014-05-08 15:19:49]

lgtm. I submitted some try bots.

[bsalomon, 2014-05-08 17:20:29]

I was able to repro the "failure" of the ninepatch-stretch GM in msaa16. The
difference is imperceptible.

Can you add this test to the file in expectations/gm/ignored-tests.txt with a
comment that includes something about "bsalomon@ will rebaseline this test" and
a link to this codereview url?

[bsalomon, 2014-05-09 13:33:18]

Thanks for updating! BTW If you just upload a patchset and don't publish a new
msg there is no notification to the reviewers and they may not realize that it
has been updated.

[bsalomon, 2014-05-09 13:33:25]

The CQ bit was checked by bsalomon@google.com

[commit-bot: I haz the power, 2014-05-09 13:34:42]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/siglesias@igalia.com/264303008/40001

[commit-bot: I haz the power, 2014-05-09 13:34:51]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-05-09 13:34:52]

Failed to apply patch for expectations/gm/ignored-tests.txt:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file expectations/gm/ignored-tests.txt
  Hunk #1 FAILED at 41.
  1 out of 1 hunk FAILED -- saving rejects to file
expectations/gm/ignored-tests.txt.rej

Patch:       expectations/gm/ignored-tests.txt
Index: expectations/gm/ignored-tests.txt
diff --git a/expectations/gm/ignored-tests.txt
b/expectations/gm/ignored-tests.txt
index
c57a0cc596738894fb56f5505f6933cb396871a3..699c08afdcf3c2142cb7daf8aa15d9f41b51ae47
100644
--- a/expectations/gm/ignored-tests.txt
+++ b/expectations/gm/ignored-tests.txt
@@ -41,3 +41,7 @@ canvas-layer-state
 # humper: https://codereview.chromium.org/248613004/
 # Changed the test in a few ways, will need rebaselining.
 simpleblurroundrect
+
+# bsalomon: https://codereview.chromium.org/264303008/
+# bsalomon@ will rebaseline this test
+ninepatch-stretch

[siglesias, 2014-05-09 13:41:11]

On 2014/05/09 13:33:18, bsalomon wrote:
> Thanks for updating! BTW If you just upload a patchset and don't publish a new
> msg there is no notification to the reviewers and they may not realize that it
> has been updated.

I used --send-mail option to "git cl upload" thinking that it would notify the
reviewers... but I see that it's not the case.

I fixed the merge error and submitted a new version of the patch.

Sam

[bsalomon, 2014-05-09 13:42:44]

The CQ bit was checked by bsalomon@google.com

[commit-bot: I haz the power, 2014-05-09 13:44:07]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/siglesias@igalia.com/264303008/60001

[bsalomon, 2014-05-09 13:44:16]

On 2014/05/09 13:41:11, siglesias wrote:
> On 2014/05/09 13:33:18, bsalomon wrote:
> > Thanks for updating! BTW If you just upload a patchset and don't publish a
new
> > msg there is no notification to the reviewers and they may not realize that
it
> > has been updated.
> 
> I used --send-mail option to "git cl upload" thinking that it would notify the
> reviewers... but I see that it's not the case.
> 
> I fixed the merge error and submitted a new version of the patch.
> 
> Sam

Weird. I wan't aware of that flag. I just searched my mail and didn't see
anything. Maybe it only works in conjunction with the -m flag? Anyway, I
submitted the new patch to the CQ.

[commit-bot: I haz the power, 2014-05-09 13:53:42]

Change committed as 14665