cc: Add software renderer pixel tests

cc: Add software renderer pixel tests

Parameterize the existing gl renderer tests to use the software renderer
and reuse them when appropriate.  As the software renderer doesn't
support filters or antialiasing, these are only turned on for the gl
renderer at this point.

R=danakj@chromium.org
BUG=none

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=197121

[enne (OOO), 2013-04-28 02:56:55]

https://codereview.chromium.org/14135011/diff/1/cc/output/renderer_pixeltest.cc
File cc/output/renderer_pixeltest.cc (right):

https://codereview.chromium.org/14135011/diff/1/cc/output/renderer_pixeltest....
cc/output/renderer_pixeltest.cc:87: gfx::Rect rect(this->device_viewport_size_);
The "this->" requirement is an unfortunate requirement of derived class
templates.

https://codereview.chromium.org/14135011/diff/1/cc/test/pixel_comparator.cc
File cc/test/pixel_comparator.cc (left):

https://codereview.chromium.org/14135011/diff/1/cc/test/pixel_comparator.cc#o...
cc/test/pixel_comparator.cc:117: LOG(ERROR) << "Pixel error at x=" << x << " y="
<< y << "; "
This is spammy.  No errors should be logged unless the test is going to fail.

[danakj, 2013-04-28 14:15:15]

On Apr 27, 2013 10:56 PM, <enne@chromium.org> wrote:
>
>
>
https://codereview.chromium.org/14135011/diff/1/cc/output/renderer_pixeltest.cc
> File cc/output/renderer_pixeltest.cc (right):
>
>
https://codereview.chromium.org/14135011/diff/1/cc/output/renderer_pixeltest....
> cc/output/renderer_pixeltest.cc:87: gfx::Rect
> rect(this->device_viewport_size_);
> The "this->" requirement is an unfortunate requirement of derived class
> templates.
>
>
https://codereview.chromium.org/14135011/diff/1/cc/test/pixel_comparator.cc
> File cc/test/pixel_comparator.cc (left):
>
>
https://codereview.chromium.org/14135011/diff/1/cc/test/pixel_comparator.cc#o...
> cc/test/pixel_comparator.cc:117: LOG(ERROR) << "Pixel error at x=" << x
> << " y=" << y << "; "
> This is spammy.  No errors should be logged unless the test is going to
> fail.

Can you make it print the same thing when the test does fail then?
>
> https://codereview.chromium.org/14135011/

[enne (OOO), 2013-04-28 14:31:48]

On 2013/04/28 14:15:15, danakj wrote:

> Can you make it print the same thing when the test does fail then?

Done.

[danakj, 2013-04-29 14:15:48]

https://codereview.chromium.org/14135011/diff/5001/cc/cc_tests.gyp
File cc/cc_tests.gyp (right):

https://codereview.chromium.org/14135011/diff/5001/cc/cc_tests.gyp#newcode42
cc/cc_tests.gyp:42: 'output/renderer_pixeltest.cc',
Does e sort before _? Don't think it does.

https://codereview.chromium.org/14135011/diff/5001/cc/output/renderer_pixelte...
File cc/output/renderer_pixeltest.cc (right):

https://codereview.chromium.org/14135011/diff/5001/cc/output/renderer_pixelte...
cc/output/renderer_pixeltest.cc:78: // All pixels can be off by one, but any
more than that is an error.
Not super excited about this, we could easily cover a bug.

Can we limit this to the software renderer? Or make different baselines?

https://codereview.chromium.org/14135011/diff/5001/cc/output/renderer_pixelte...
cc/output/renderer_pixeltest.cc:495: TEST_F(GLRendererPixelTest, AntiAliasing) {
Should GLRendererPixelTests move to a separate file?

[enne (OOO), 2013-04-29 16:00:54]

https://codereview.chromium.org/14135011/diff/5001/cc/cc_tests.gyp
File cc/cc_tests.gyp (right):

https://codereview.chromium.org/14135011/diff/5001/cc/cc_tests.gyp#newcode42
cc/cc_tests.gyp:42: 'output/renderer_pixeltest.cc',
On 2013/04/29 14:15:49, danakj wrote:
> Does e sort before _? Don't think it does.

Presubmit was happy.

https://codereview.chromium.org/14135011/diff/5001/cc/output/renderer_pixelte...
File cc/output/renderer_pixeltest.cc (right):

https://codereview.chromium.org/14135011/diff/5001/cc/output/renderer_pixelte...
cc/output/renderer_pixeltest.cc:78: // All pixels can be off by one, but any
more than that is an error.
On 2013/04/29 14:15:49, danakj wrote:
> Not super excited about this, we could easily cover a bug.
> 
> Can we limit this to the software renderer? Or make different baselines?

Isn't the point of the fuzzy comparator so that we don't need different
baselines that differ by very small amounts?

At any rate, I've limited this to the software renderer.  PTAL.

https://codereview.chromium.org/14135011/diff/5001/cc/output/renderer_pixelte...
cc/output/renderer_pixeltest.cc:495: TEST_F(GLRendererPixelTest, AntiAliasing) {
On 2013/04/29 14:15:49, danakj wrote:
> Should GLRendererPixelTests move to a separate file?

Maybe if we get more and this file gets unwieldy?  Antialiasing will probably
never be implemented, but some things like filters and picture draw quads are
just not implemented yet and it'd just be needless churn to move them back and
forth before the dust settles.

[danakj, 2013-04-29 18:05:04]

Thanks, LGTM

[commit-bot: I haz the power, 2013-04-29 18:08:16]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/enne@chromium.org/14135011/9001

[commit-bot: I haz the power, 2013-04-29 20:35:08]

Change committed as 197121