ThreadWatcher - give ShutdownWatchDogThread::Alarm a more unique function body.

ThreadWatcher - give ShutdownWatchDogThread::Alarm a more 
unique function body.

Changed ShutdownWatchDogThread::Alarm's to call
ShutdownCrash(), so it does not get merged by the link time 
optimization.

R=jar@chromium.org, eroman@chromium.org
BUG=232080
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=194524

[eroman, 2013-04-16 22:53:12]

https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread...
File chrome/browser/metrics/thread_watcher.cc (right):

https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread...
chrome/browser/metrics/thread_watcher.cc:857: // Disarm(). When Alarm goes off,
in release mode we get the crash dump
This comment no longer matches the code.

https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread...
chrome/browser/metrics/thread_watcher.cc:860: StartupCrash();
Is this behavior change intentional? before it would DumpWithoutCrashing()
whereas now it outright crashes.

[ramant (doing other things), 2013-04-16 23:00:39]

https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread...
File chrome/browser/metrics/thread_watcher.cc (right):

https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread...
chrome/browser/metrics/thread_watcher.cc:857: // Disarm(). When Alarm goes off,
in release mode we get the crash dump
On 2013/04/16 22:53:12, eroman wrote:
> This comment no longer matches the code.

Reverted this change. Shouldn't have changed the behavior.

https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread...
chrome/browser/metrics/thread_watcher.cc:860: StartupCrash();
On 2013/04/16 22:53:12, eroman wrote:
> Is this behavior change intentional? before it would DumpWithoutCrashing()
> whereas now it outright crashes.

Reverted the change.

[eroman, 2013-04-16 23:12:37]

LGTM, however I am not convinced that the NullPointerCrash() method will
work as promised -- in release mode the line number argument is unused, and
hence will be stripped, meaning the generated function bodies will probably
be identical and ICF will merge them all...


On Tue, Apr 16, 2013 at 4:00 PM, <rtenneti@chromium.org> wrote:

>
> https://codereview.chromium.**org/14250003/diff/1/chrome/**
>
browser/metrics/thread_**watcher.cc<https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread_watcher.cc>
> File chrome/browser/metrics/thread_**watcher.cc (right):
>
> https://codereview.chromium.**org/14250003/diff/1/chrome/**
>
browser/metrics/thread_**watcher.cc#newcode857<https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread_watcher.cc#newcode857>
> chrome/browser/metrics/thread_**watcher.cc:857: // Disarm(). When Alarm
> goes off, in release mode we get the crash dump
> On 2013/04/16 22:53:12, eroman wrote:
>
>> This comment no longer matches the code.
>>
>
> Reverted this change. Shouldn't have changed the behavior.
>
>
> https://codereview.chromium.**org/14250003/diff/1/chrome/**
>
browser/metrics/thread_**watcher.cc#newcode860<https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread_watcher.cc#newcode860>
> chrome/browser/metrics/thread_**watcher.cc:860: StartupCrash();
> On 2013/04/16 22:53:12, eroman wrote:
>
>> Is this behavior change intentional? before it would
>>
> DumpWithoutCrashing()
>
>> whereas now it outright crashes.
>>
>
> Reverted the change.
>
>
https://codereview.chromium.**org/14250003/<https://codereview.chromium.org/1...
>

[eroman, 2013-04-16 23:13:31]

Oh sorry nevermind, they are all wrapped in a  MSVC_DISABLE_OPTIMIZE()
block :)

LGTM


On Tue, Apr 16, 2013 at 4:12 PM, Eric Roman <eroman@chromium.org> wrote:

> LGTM, however I am not convinced that the NullPointerCrash() method will
> work as promised -- in release mode the line number argument is unused, and
> hence will be stripped, meaning the generated function bodies will probably
> be identical and ICF will merge them all...
>
>
> On Tue, Apr 16, 2013 at 4:00 PM, <rtenneti@chromium.org> wrote:
>
>>
>> https://codereview.chromium.**org/14250003/diff/1/chrome/**
>>
browser/metrics/thread_**watcher.cc<https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread_watcher.cc>
>> File chrome/browser/metrics/thread_**watcher.cc (right):
>>
>> https://codereview.chromium.**org/14250003/diff/1/chrome/**
>>
browser/metrics/thread_**watcher.cc#newcode857<https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread_watcher.cc#newcode857>
>> chrome/browser/metrics/thread_**watcher.cc:857: // Disarm(). When Alarm
>> goes off, in release mode we get the crash dump
>> On 2013/04/16 22:53:12, eroman wrote:
>>
>>> This comment no longer matches the code.
>>>
>>
>> Reverted this change. Shouldn't have changed the behavior.
>>
>>
>> https://codereview.chromium.**org/14250003/diff/1/chrome/**
>>
browser/metrics/thread_**watcher.cc#newcode860<https://codereview.chromium.org/14250003/diff/1/chrome/browser/metrics/thread_watcher.cc#newcode860>
>> chrome/browser/metrics/thread_**watcher.cc:860: StartupCrash();
>> On 2013/04/16 22:53:12, eroman wrote:
>>
>>> Is this behavior change intentional? before it would
>>>
>> DumpWithoutCrashing()
>>
>>> whereas now it outright crashes.
>>>
>>
>> Reverted the change.
>>
>>
https://codereview.chromium.**org/14250003/<https://codereview.chromium.org/1...
>>
>
>

[jar (doing other things), 2013-04-16 23:43:54]

lgtm