Output stacktraces on WebGL conformance failure.

Output stacktraces on WebGL conformance failure.

If a minidump stack trace is available, output it along with the
page test failure.

R=kbr@chromium.org
BUG=609252
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/0cb2844464155e533e83b10facd8523884a406be
Cr-Commit-Position: refs/heads/master@{#400282}

[vmiura, 2016-06-16 01:40:23]

Description was changed from

==========
Output stacktraces on WebGL conformance failure.

If a minidump stack trace is available, output it along with the
page test failure.

R=kbr@chromium.org
BUG=609252
==========

to

==========
Output stacktraces on WebGL conformance failure.

If a minidump stack trace is available, output it along with the
page test failure.

R=kbr@chromium.org
BUG=609252
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[vmiura, 2016-06-16 02:10:46]

vmiura@chromium.org changed reviewers:
+ nednguyen@google.com

[vmiura, 2016-06-16 02:10:47]

Let me know if this makes sense.

I think at minimum this needs something to remove, or somehow mark the minidumps
so that we only dump crashes that occurred during each page test, and don't keep
dumping prior stack traces when other tests fail.

Another question is, should we be doing something like this for all page tests
failures?

[Ken Russell (switch to Gerrit), 2016-06-16 17:48:02]

Thank you for digging into this. LGTM as a good start.

https://codereview.chromium.org/2070233002/diff/1/content/test/gpu/gpu_tests/...
File content/test/gpu/gpu_tests/webgl_conformance.py (right):

https://codereview.chromium.org/2070233002/diff/1/content/test/gpu/gpu_tests/...
content/test/gpu/gpu_tests/webgl_conformance.py:87: is_valid_dump, trace_output
= tab.browser.GetStackTrace()
I was concerned about the performance cost of doing this after every page run
but realize that this will only be executed after a failure, which should not be
the common case.

[nednguyen, 2016-06-16 17:54:09]

lgtm

Ken: does this need to be ported to the new harness too?

[Ken Russell (switch to Gerrit), 2016-06-16 20:20:44]

On 2016/06/16 17:54:09, nednguyen wrote:
> lgtm
> 
> Ken: does this need to be ported to the new harness too?

Ned: yes, it will, but things are heavily in flux there and I'd like to avoid
merge conflicts for the time being.

We should make sure a bug is filed (assign it to me) to port this over.

[vmiura, 2016-06-16 20:26:05]

Verified that the mac build is showing a stack trace after forced hang on the
GPU process in the "conformance_misc_webgl_specific" test.

https://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel...

[vmiura, 2016-06-16 20:26:17]

Patchset #2 (id:20001) has been deleted

[vmiura, 2016-06-16 20:26:36]

The CQ bit was checked by vmiura@chromium.org

[commit-bot: I haz the power, 2016-06-16 20:27:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2070233002/1

[commit-bot: I haz the power, 2016-06-16 22:35:56]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-06-16 22:37:53]

Description was changed from

==========
Output stacktraces on WebGL conformance failure.

If a minidump stack trace is available, output it along with the
page test failure.

R=kbr@chromium.org
BUG=609252
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Output stacktraces on WebGL conformance failure.

If a minidump stack trace is available, output it along with the
page test failure.

R=kbr@chromium.org
BUG=609252
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/0cb2844464155e533e83b10facd8523884a406be
Cr-Commit-Position: refs/heads/master@{#400282}
==========

[commit-bot: I haz the power, 2016-06-16 22:37:54]

Patchset 1 (id:??) landed as
https://crrev.com/0cb2844464155e533e83b10facd8523884a406be
Cr-Commit-Position: refs/heads/master@{#400282}