Do the ReloadLocaleResources() call on the IO thread to avoid synchronization issues.

Do the ReloadLocaleResources() call on the IO thread to avoid synchronization issues.

Followup to https://chromiumcodereview.appspot.com/10543015/
Forced the Reload on the same thread (IO) that was racing with the call.  Since Tony doesn't want us to add synchronization to ResourceBundle from a case that's only exercised in test, we'll go this route to stop the race.

R=tony@chromium.org, estade@chromium.org
BUG=95425
TEST=browser_tests --gtest_filter=WebUIBidiCheckerBrowserTestRTL.*TestSettingsFrameStartup --gtest_also_run_disabled_tests

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=141964

[tony, 2012-06-13 16:38:15]

Clever, this seems like it should work since the UI thread won't be accessing
ResourceBundle.

LGTM

https://chromiumcodereview.appspot.com/10546139/diff/1/chrome/browser/ui/webu...
File chrome/browser/ui/webui/bidi_checker_web_ui_test.cc (right):

https://chromiumcodereview.appspot.com/10546139/diff/1/chrome/browser/ui/webu...
chrome/browser/ui/webui/bidi_checker_web_ui_test.cc:80:
base::ThreadRestrictions::SetIOAllowed(false);
Nit: Can you use ScopedAllowIO here?  You can put it in a block so
event->Signal() doesn't get included.

[Sheridan Rawlins, 2012-06-13 20:10:34]

https://chromiumcodereview.appspot.com/10546139/diff/1/chrome/browser/ui/webu...
File chrome/browser/ui/webui/bidi_checker_web_ui_test.cc (right):

https://chromiumcodereview.appspot.com/10546139/diff/1/chrome/browser/ui/webu...
chrome/browser/ui/webui/bidi_checker_web_ui_test.cc:80:
base::ThreadRestrictions::SetIOAllowed(false);
On 2012/06/13 16:38:15, tony wrote:
> Nit: Can you use ScopedAllowIO here?  You can put it in a block so
> event->Signal() doesn't get included.

When I do, I hit the pre-submit:

"New code should not use ScopedAllowIO. Post a task to the blocking pool or the
FILE thread instead."

As discussed on chat, it should be ok in new test code.

[tony, 2012-06-13 20:14:34]

New patch LGTM.

[Evan Stade, 2012-06-13 20:47:21]

lgtm

[commit-bot: I haz the power, 2012-06-13 20:51:58]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/scr@chromium.org/10546139/12001

[commit-bot: I haz the power, 2012-06-13 20:52:00]

Presubmit check for 10546139-12001 failed and returned exit status 1.

Running presubmit commit checks ...

** Presubmit Warnings **
New code should not use ScopedAllowIO. Post a task to the blocking pool or the
FILE thread instead.
    chrome/browser/ui/webui/bidi_checker_web_ui_test.cc:79

[Joao da Silva, 2012-06-13 22:09:16]

http://codereview.chromium.org/10546139/diff/12001/chrome/browser/ui/webui/bi...
File chrome/browser/ui/webui/bidi_checker_web_ui_test.cc (right):

http://codereview.chromium.org/10546139/diff/12001/chrome/browser/ui/webui/bi...
chrome/browser/ui/webui/bidi_checker_web_ui_test.cc:105:
ui_test_utils::WaitEventSignaled(&event);
Why not PostTaskAndReply to IO? The posted task does the SetUpOnIOThread, and
the reply is MessageLoop::QuitClosure. Then call Run() on the current loop (or
maybe ui_test_utils::RunMessageLoop()). The reply task will make that return.

[Sheridan Rawlins, 2012-06-15 00:36:23]

http://codereview.chromium.org/10546139/diff/12001/chrome/browser/ui/webui/bi...
File chrome/browser/ui/webui/bidi_checker_web_ui_test.cc (right):

http://codereview.chromium.org/10546139/diff/12001/chrome/browser/ui/webui/bi...
chrome/browser/ui/webui/bidi_checker_web_ui_test.cc:105:
ui_test_utils::WaitEventSignaled(&event);
This was already committed, but I'll try this on a followup.

On 2012/06/13 22:09:16, Joao da Silva wrote:
> Why not PostTaskAndReply to IO? The posted task does the SetUpOnIOThread, and
> the reply is MessageLoop::QuitClosure. Then call Run() on the current loop (or
> maybe ui_test_utils::RunMessageLoop()). The reply task will make that return.