Removed arbitrary 4096 size limit on WebGL canvas backbuffers, Take 3

Removed arbitrary 4096 size limit on WebGL canvas backbuffers, Take 3

Worked out the issue from the previous attempt with not consuming the
out of memory error and verified with synthetic OOM generation. (See
https://codereview.chromium.org/1050043003/)

BUG=445542
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=193315

[bajones, 2015-04-06 22:06:14]

bajones@chromium.org changed reviewers:
+ kbr@chromium.org

[bajones, 2015-04-06 22:06:15]

PTAL... again. :P

[Ken Russell (switch to Gerrit), 2015-04-06 23:18:24]

Per piman's comment on https://codereview.chromium.org/1040153002 -- would you
be willing to change ScopedErrorCache to something like a
ScopedConditionalErrorCache? Then it could be conditionally run only if the new
texture's width or height are >= 4096, for example.

[bajones, 2015-04-07 22:16:06]

On 2015/04/06 23:18:24, Ken Russell wrote:
> Per piman's comment on https://codereview.chromium.org/1040153002 -- would you
> be willing to change ScopedErrorCache to something like a
> ScopedConditionalErrorCache? Then it could be conditionally run only if the
new
> texture's width or height are >= 4096, for example.

Done.

[Ken Russell (switch to Gerrit), 2015-04-07 22:18:00]

LGTM if this has been tested with your new failure injection code in the
separate CL.

https://codereview.chromium.org/1066643003/diff/20001/Source/platform/graphic...
File Source/platform/graphics/gpu/DrawingBuffer.cpp (right):

https://codereview.chromium.org/1066643003/diff/20001/Source/platform/graphic...
Source/platform/graphics/gpu/DrawingBuffer.cpp:94: // Avoid infinite looping if
glGetError is behaving badly.
This won't affect the behavior of the product, only debug builds -- it would be
safer to explicitly check a loop variable.

[bajones, 2015-04-07 22:37:01]

The CQ bit was checked by bajones@chromium.org

[bajones, 2015-04-07 22:37:01]

The patchset sent to the CQ was uploaded after l-g-t-m from kbr@chromium.org
Link to the patchset: https://codereview.chromium.org/1066643003/#ps40001
(title: "Addressing kbr@")

[commit-bot: I haz the power, 2015-04-07 22:37:15]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1066643003/40001

[commit-bot: I haz the power, 2015-04-08 01:21:48]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=193315

[Michael Achenbach, 2015-04-08 10:03:00]

A revert of this CL (patchset #3 id:40001) has been created in
https://codereview.chromium.org/1069363002/ by machenbach@chromium.org.

The reason for reverting is: [Sheriff] Breaks
http://build.chromium.org/p/chromium.webkit/builders/GPU%20Win7%20%28NVIDIA%2...

With OOM.

[bajones, 2015-04-08 17:05:26]

On 2015/04/08 10:03:00, Michael Achenbach wrote:
> A revert of this CL (patchset #3 id:40001) has been created in
> https://codereview.chromium.org/1069363002/ by mailto:machenbach@chromium.org.
> 
> The reason for reverting is: [Sheriff] Breaks
>
http://build.chromium.org/p/chromium.webkit/builders/GPU%20Win7%20%28NVIDIA%2...
> 
> With OOM.

*sigh* Thanks for the revert, sorry for the breakage. (Sends a stern look of
disapproval at driver behavior... ಠ_ಠ)