Run Chrome OS accessibility histograms on main thread.

Run Chrome OS accessibility histograms on main thread.

This code was accessing preferences and it was asserting
in Debug mode because it was being run on another thread.

BUG=99504, 164000
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=170976

[oshima, 2012-12-03 22:09:22]

https://codereview.chromium.org/11299312/diff/1/content/browser/accessibility...
File content/browser/accessibility/browser_accessibility_state_impl.cc (right):

https://codereview.chromium.org/11299312/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_state_impl.cc:52: #if
defined(OS_WIN)
From the stack trace, the problem seems to be chromeos specific.
Or other platforms (mac/linux) are accessing BrowserProcessImpl from non UI
threads as well?

https://codereview.chromium.org/11299312/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_state_impl.cc:60:
BrowserThread::ID browser_thread = BrowserThread::UI;
the name |browser_thread| doesn't make much sense. Can you use something like
histogram_update_thread?

https://codereview.chromium.org/11299312/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_state_impl.cc:66: // delete
it prematurely.
update this comment too. May be just move it inside OS_WIN?

[dmazzoni, 2012-12-03 23:32:30]

https://codereview.chromium.org/11299312/diff/1/content/browser/accessibility...
File content/browser/accessibility/browser_accessibility_state_impl.cc (right):

https://codereview.chromium.org/11299312/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_state_impl.cc:52: #if
defined(OS_WIN)
On 2012/12/03 22:09:23, oshima wrote:
> From the stack trace, the problem seems to be chromeos specific.
> Or other platforms (mac/linux) are accessing BrowserProcessImpl from non UI
> threads as well?

Currently there's no Mac-specific or Linux-specific code. The desire to do some
processing on another thread was motivated by Windows-only needs, so I think it
makes sense for Windows to be the special case - the main thread seems like a
more sensible default for future changes to other platforms.

https://codereview.chromium.org/11299312/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_state_impl.cc:60:
BrowserThread::ID browser_thread = BrowserThread::UI;
On 2012/12/03 22:09:23, oshima wrote:
> the name |browser_thread| doesn't make much sense. Can you use something like
> histogram_update_thread?

Done.

https://codereview.chromium.org/11299312/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_state_impl.cc:66: // delete
it prematurely.
On 2012/12/03 22:09:23, oshima wrote:
> update this comment too. May be just move it inside OS_WIN?

Thanks, meant to do that earlier. Done.

[oshima, 2012-12-03 23:37:58]

lgtm

Please upload new patch :)

[commit-bot: I haz the power, 2012-12-03 23:40:49]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dmazzoni@chromium.org/11299312/6002

[achuithb, 2012-12-03 23:55:30]

Thanks for fixing this!

[commit-bot: I haz the power, 2012-12-04 02:44:37]

Retried try job too often on win_rel for step(s) browser_tests