Suppress CHECK failure in CreateProfile for OS_WIN.

Suppress CHECK failure in CreateProfile for OS_WIN.

The CHECK(profile) was only intended for platforms other than Windows.
Prior to my fix for bug 196301, the intended behavior is for CreateProfile
to return NULL and allow Chrome to normally exit, because after showing
the user data dir picker dialog we would have already re-launched Chrome
with a new command line containing the correct user data dir.

BUG=230442
TEST=CQ

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=194003

[hshi1, 2013-04-12 17:15:42]

PTAL. Nico (OWNER), Mattias (FYI). Thanks!

This CHECK was always only intended for non-Windows platforms. Prior to my
bugfix for 196301 (see
https://chromiumcodereview.appspot.com/12662033/diff/117007/chrome/browser/ch...),
the CHECK was enclosed in the #else section of a #if defined(OS_WIN) ... #else
... #endif block. It is expected that CreateProfile would just return NULL for
the case of inaccessible user data dir, which allows Chrome to exit normally.

[Nico, 2013-04-12 17:26:54]

lgtm

Thanks for the good CL description!

ps: Do we have any tests for these code paths?

[hshi1, 2013-04-12 17:38:06]

Regarding testing: for the fix of bug 196301 I had added a windows-only browser
test for the GetUserDataDir (see user_data_dir_extractor_win_browsertest.cc)
that verifies GetUserDataDir() gets invoked prior to CreateLocalState(). At
least this ensures that we'll not crash due to inaccessible user data dir again.

Ideally a better test would set up a bogus (inaccessible) user data dir in
InProcessBrowserTest::SetUp, and verify that the mock ShowUserDataDialog is
actually shown. I had attempted to modify ChromeMainUserDataDirTest::SetUp but
gave up after realizing it could not be done in this fashion because
InProcessBrowserTest::SetUp() not only sets up the user data dir but also kicks
off chrome main, so I had to somehow add a private call to the base class
InProcessBrowserTest::SetUp to set the bogus dir.

On 2013/04/12 17:26:54, Nico wrote:
> lgtm
> 
> Thanks for the good CL description!
> 
> ps: Do we have any tests for these code paths?

[commit-bot: I haz the power, 2013-04-12 17:38:42]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/hshi@chromium.org/14109008/1

[commit-bot: I haz the power, 2013-04-12 20:41:00]

Change committed as 194003