Fix a stack overflow crash in the accessibility code in the renderer. The crash occurs in the

Fix a stack overflow crash in the accessibility code in the renderer. The crash occurs in the 
RendererAccessibility::PostAccessibilityNotification function while processing the WebAccessibilityNotificationLayoutComplete 
notification from webkit. This function recursively calls itself with the WebAccessibilityNotificationLayoutComplete
notification leading to infinite recursion.

Fix is to check whether the incoming notification is WebAccessibilityNotificationLayoutComplete before the recursive
call.

BUG=none
TEST=chrome frame tests should pass on the IE6 builder.
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=114573

[David Tseng, 2011-12-15 02:03:59]

lgtm. (given the comment).

Good catch! Adding sky for OWNERS approval.

http://codereview.chromium.org/8952005/diff/1/content/renderer/renderer_acces...
File content/renderer/renderer_accessibility.cc (right):

http://codereview.chromium.org/8952005/diff/1/content/renderer/renderer_acces...
content/renderer/renderer_accessibility.cc:174: PostAccessibilityNotification(
I think we should add the exclusion here as the previous line's assignment
should happen regardless.

if (notification != WebKit::WebAccessibilityNotificationLayoutComplete) { 
  Post...

[ananta, 2011-12-15 02:09:56]

http://codereview.chromium.org/8952005/diff/1/content/renderer/renderer_acces...
File content/renderer/renderer_accessibility.cc (right):

http://codereview.chromium.org/8952005/diff/1/content/renderer/renderer_acces...
content/renderer/renderer_accessibility.cc:174: PostAccessibilityNotification(
On 2011/12/15 02:03:59, David Tseng wrote:
> I think we should add the exclusion here as the previous line's assignment
> should happen regardless.
> 
> if (notification != WebKit::WebAccessibilityNotificationLayoutComplete) { 
>   Post...
> 

Done.

[piman, 2011-12-15 02:11:20]

OWNERS lgtm

[sky, 2011-12-15 17:45:01]

LGTM