Breakpad about:crash crashing browser (oops)

Check the right Info.plist to determine whether the app is the browser or a
background helper.

This fixes the Breakpad about:crash problem, but not directly.  I'm not
exactly sure of the failure mechanism here.  It's likely that doing this call
properly here caches the Info.plist dictionary, and that something in Breakpad
crashes when that's not done.  Without this fix, we're seeing a double fault,
with the second exception occurring in the renderer while it is processing
the crash.  Likely because of the use of fork && execv instead of posix_spawn
to launch the renderer, the second exception is delivered to the browser
process.  The browser's Breakpad instance recognizes that the exception is not
its own, but attempts to act on it in part anyway.  There may be a Breakpad
bug here, I'll investigate more tomorrow.

BUG=20371
TEST=In a Breakpad-enabled build, about:crash should bring up a sad tab and
     not crash the browser process

r25311

[jeremy, 2009-09-03 04:27:14]

bizarre :|  I find it strange that the Renderer Process' copy of Breakpad would
crash in the exception handler since it successfully writes and uploads a crash
dump :|

Perhaps it makes sense to hold off with this change till the underlying cause is
identified, but your call.

LGTM

[Mark Mentovai, 2009-09-03 14:31:12]

I'm going to roll with this now while I continue to debug.  This
change is the right thing to do anyway, and I suspect that having this
broken may be causing other problems.

Mark

jeremy@chromium.org wrote:
> bizarre :| =A0I find it strange that the Renderer Process' copy of
> Breakpad would crash in the exception handler since it successfully
> writes and uploads a crash dump :|
>
> Perhaps it makes sense to hold off with this change till the underlying
> cause is identified, but your call.
>
> LGTM
>
> http://codereview.chromium.org/194007