Fix blink_perf.canvas failure on chromium.perf

Fix blink_perf.canvas failure on chromium.perf

Some platform don't support "experimental-webgl" for Canvas.getContext API anymore. Use "webgl".

BUG=488059, 455796
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=195401

[dshwang, 2015-05-15 08:34:31]

dongseong.hwang@intel.com changed reviewers:
+ kbr@chromium.org, simonhatch@chromium.org, skyostil@chromium.org

[dshwang, 2015-05-15 08:34:32]

kbr or other owners, could you review it urgently? now perf bots fails for 1.5
days

[dshwang, 2015-05-15 08:36:59]

dongseong.hwang@intel.com changed reviewers:
+ junov@chromium.org

[Sami, 2015-05-15 09:48:19]

Thank you, lgtm.

[Sami, 2015-05-15 09:48:46]

The CQ bit was checked by skyostil@chromium.org

[commit-bot: I haz the power, 2015-05-15 09:48:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1142603002/1

[commit-bot: I haz the power, 2015-05-15 09:51:46]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=195401

[Ken Russell (switch to Gerrit), 2015-05-15 21:02:56]

That's surprising. None of the code for accepting the "experimental-webgl"
context type has been removed.

[Ken Russell (switch to Gerrit), 2015-05-15 21:04:32]

On 2015/05/15 21:02:56, Ken Russell wrote:
> That's surprising. None of the code for accepting the "experimental-webgl"
> context type has been removed.

I think it's more likely that the test is running on a bot that doesn't have a
real GPU. We've encountered this problem before with the
smoothness.tough_webgl_cases benchmark. Code may be needed to avoid running
these tests on bots that don't have WebGL support.

[Sami, 2015-05-18 10:44:00]

On 2015/05/15 21:04:32, Ken Russell wrote:
> On 2015/05/15 21:02:56, Ken Russell wrote:
> > That's surprising. None of the code for accepting the "experimental-webgl"
> > context type has been removed.
> 
> I think it's more likely that the test is running on a bot that doesn't have a
> real GPU. We've encountered this problem before with the
> smoothness.tough_webgl_cases benchmark. Code may be needed to avoid running
> these tests on bots that don't have WebGL support.

Ah, right, sorry I didn't notice this during the review. Arguably using webgl
instead of experimental-webgl is still the right thing to do here.

Let me comment on the bug.