make compare_rendered_pictures_test.py run end-to-end test

make compare_rendered_pictures_test.py run end-to-end test
1. generate test SKPs using skpmaker
2. render those SKPs using render_pictures
3. compare results using compare_rendered_pictures.py

BUG=skia:1942, skia:2230
NOTRY=True

Committed: http://code.google.com/p/skia/source/detail?r=14624

[epoger, 2014-05-07 14:56:25]

Ready for review at patchset 1.

https://codereview.chromium.org/274623002/diff/1/gm/rebaseline_server/testdat...
File
gm/rebaseline_server/testdata/outputs/expected/compare_rendered_pictures_test.CompareRenderedPicturesTest.test_endToEnd/compare_rendered_pictures.json
(right):

https://codereview.chromium.org/274623002/diff/1/gm/rebaseline_server/testdat...
gm/rebaseline_server/testdata/outputs/expected/compare_rendered_pictures_test.CompareRenderedPicturesTest.test_endToEnd/compare_rendered_pictures.json:58:
"imageAUrl": "changed_skp/bitmap-64bitMD5_3101044995537104462.png",
I haven't validated the new checksums, but I'm not surprised they are different
(since this CL changes the input SKP files).  I don't think we need to manually
validate them--I trust the render_pictures unittest has kept this calculation in
line.

https://codereview.chromium.org/274623002/diff/1/gm/rebaseline_server/testdat...
gm/rebaseline_server/testdata/outputs/expected/compare_rendered_pictures_test.CompareRenderedPicturesTest.test_endToEnd/compare_rendered_pictures.json:80:
"imageAUrl": "only-in-before_skp/bitmap-64bitMD5_2320185040577047131.png",
Now that this test is running render_pictures (instead of using canned data),
the imageURLs now include the skpname in the directory path (as they should).

[epoger, 2014-05-07 15:28:29]

The trybot failed in ApplyPatch, but I think that's probably because of the
image files deleted within this CL.  Ready for review.

[borenet, 2014-05-07 18:47:48]

Nice!  LGTM but wondering why you didn't use the library function.

https://codereview.chromium.org/274623002/diff/1/gm/rebaseline_server/base_un...
File gm/rebaseline_server/base_unittest.py (right):

https://codereview.chromium.org/274623002/diff/1/gm/rebaseline_server/base_un...
gm/rebaseline_server/base_unittest.py:64: def run_command(self, args):
Why not just use subprocess.check_output which seems to do the same thing?

[epoger, 2014-05-07 19:27:49]

https://codereview.chromium.org/274623002/diff/1/gm/rebaseline_server/base_un...
File gm/rebaseline_server/base_unittest.py (right):

https://codereview.chromium.org/274623002/diff/1/gm/rebaseline_server/base_un...
gm/rebaseline_server/base_unittest.py:64: def run_command(self, args):
On 2014/05/07 18:47:49, borenet wrote:
> Why not just use subprocess.check_output which seems to do the same thing?

Because I'm stupid, that's why.

[epoger, 2014-05-07 19:28:02]

The CQ bit was checked by epoger@google.com

[commit-bot: I haz the power, 2014-05-07 19:28:50]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/epoger@google.com/274623002/20001

[commit-bot: I haz the power, 2014-05-07 19:29:12]

Change committed as 14624