Fixed WebGL printing crash due to failed buffer allocation

Re-land : Fixed WebGL printing crash due to failed buffer allocation

Change since previous revert: reduced canvas size in layout
test to prevent OOM from crbug.com/538697

BUG=537711
R=kbr@chromium.org

Committed: https://crrev.com/596b53b8c357eeb8453fc613fd5834c292ae5b96
Cr-Commit-Position: refs/heads/master@{#352170}

[Justin Novosad, 2015-10-01 01:50:36]

PTAL

[Ken Russell (switch to Gerrit), 2015-10-01 02:33:30]

Thanks for working on this.

https://codereview.chromium.org/1380633003/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/printing/webgl-oversized-printing.html
(right):

https://codereview.chromium.org/1380633003/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/printing/webgl-oversized-printing.html:13:
ctx.clearColor(0, 0, 0, 0);
Can you call ctx.clear(ctx.COLOR_BUFFER_BIT) here to get the markContextChanged
method to be called?

https://codereview.chromium.org/1380633003/diff/1/third_party/WebKit/Source/m...
File third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp
(right):

https://codereview.chromium.org/1380633003/diff/1/third_party/WebKit/Source/m...
third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp:1724:
markContextChanged(CanvasChanged);
This looks wrong. markContextChanged is only supposed to be called when the
canvas's contents are changed.

[Justin Novosad, 2015-10-01 02:40:57]

https://codereview.chromium.org/1380633003/diff/1/third_party/WebKit/Source/m...
File third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp
(right):

https://codereview.chromium.org/1380633003/diff/1/third_party/WebKit/Source/m...
third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp:1724:
markContextChanged(CanvasChanged);
On 2015/10/01 02:33:30, Ken Russell wrote:
> This looks wrong. markContextChanged is only supposed to be called when the
> canvas's contents are changed.

You are right. I am missing a call to clear() in the test. I was confused by the
fact that the clusterfuzz reduced test case did not have a call to clear(). The
way the clusterfuzz bot performs a print, the canvas does not need to be marked
dirty for the content update to happen.

[Justin Novosad, 2015-10-01 04:26:03]

Figured it out. PTAL.

[Ken Russell (switch to Gerrit), 2015-10-01 04:37:21]

Thanks. LGTM

[Justin Novosad, 2015-10-01 13:34:07]

The CQ bit was checked by junov@chromium.org

[commit-bot: I haz the power, 2015-10-01 13:34:16]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1380633003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1380633003/20001

[commit-bot: I haz the power, 2015-10-01 13:40:11]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-10-01 13:40:57]

Patchset 2 (id:??) landed as
https://crrev.com/dda8a88830593f605501500c4de9c870e70c94eb
Cr-Commit-Position: refs/heads/master@{#351790}

[dmurph, 2015-10-02 18:31:09]

A revert of this CL (patchset #2 id:20001) has been created in
https://codereview.chromium.org/1387563004/ by dmurph@chromium.org.

The reason for reverting is: BUG=538697

Failing OOM on Windows 7 bots.

[Justin Novosad, 2015-10-02 21:17:30]

The CQ bit was checked by junov@chromium.org

[Justin Novosad, 2015-10-02 21:17:30]

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/1380633003/#ps40001
(title: "reduced canvas size in test to prevent OOM on win7")

[commit-bot: I haz the power, 2015-10-02 21:19:20]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1380633003/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1380633003/40001

[commit-bot: I haz the power, 2015-10-02 23:19:14]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-10-02 23:20:23]

Patchset 3 (id:??) landed as
https://crrev.com/596b53b8c357eeb8453fc613fd5834c292ae5b96
Cr-Commit-Position: refs/heads/master@{#352170}