Rolling back change 22245.

Rolling back change 22245.
Too much red :-(

BUG=None
TEST=None
TBR=erikkay


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=22247

[bradn, 2009-08-02 00:12:55]

I think I'm gonna need to roll this back out Erik, sorry.
After a complete svn dir clobber, and reboot, the chromeos build is still
failing (in addition to that testlifetimeofdom test). I'd leave it in, but every
time a build goes off, the tree closes.

[Erik does not do reviews, 2009-08-02 02:28:18]

LGTM :-(

Do we know who owns this bot?  I have no way of trying to repro / fix this bug.

[bradn, 2009-08-02 02:37:31]

I think its rspangler.
According to dave moore the only difference between this and a generic
chrome builder is that OS_CHROMEOS gets set.
This affects a range of #defines, but I can't see any obvious connection
with the failures.
If I catch someone from their team I'll ask them to take a look.

-BradN

On Sat, Aug 1, 2009 at 7:28 PM, <erikkay@chromium.org> wrote:

> LGTM :-(
>
> Do we know who owns this bot?  I have no way of trying to repro / fix
> this bug.
>
>
>
> http://codereview.chromium.org/159763
>

[bradn, 2009-08-02 17:44:46]

So apparently the way you can reproduce it is to set the GYP_DEFINES
environment variable to: chromeos=1 branding=Chrome

-BradN


On Sat, Aug 1, 2009 at 7:36 PM, Bradley Nelson <bradnelson@google.com>wrote:

> I think its rspangler.
> According to dave moore the only difference between this and a generic
> chrome builder is that OS_CHROMEOS gets set.
> This affects a range of #defines, but I can't see any obvious connection
> with the failures.
> If I catch someone from their team I'll ask them to take a look.
>
> -BradN
>
>
> On Sat, Aug 1, 2009 at 7:28 PM, <erikkay@chromium.org> wrote:
>
>> LGTM :-(
>>
>> Do we know who owns this bot?  I have no way of trying to repro / fix
>> this bug.
>>
>>
>>
>> http://codereview.chromium.org/159763
>>
>
>

[bradn, 2009-08-02 17:46:15]

Oh and be sure to run gclient runhooks --force after that.
Followed by hammer --implicit-deps-changed chrome

-BradN

On Sun, Aug 2, 2009 at 10:44 AM, Bradley Nelson <bradnelson@google.com>wrote:

> So apparently the way you can reproduce it is to set the GYP_DEFINES
> environment variable to: chromeos=1 branding=Chrome
>
> -BradN
>
>
>
> On Sat, Aug 1, 2009 at 7:36 PM, Bradley Nelson <bradnelson@google.com>wrote:
>
>> I think its rspangler.
>> According to dave moore the only difference between this and a generic
>> chrome builder is that OS_CHROMEOS gets set.
>> This affects a range of #defines, but I can't see any obvious connection
>> with the failures.
>> If I catch someone from their team I'll ask them to take a look.
>>
>> -BradN
>>
>>
>> On Sat, Aug 1, 2009 at 7:28 PM, <erikkay@chromium.org> wrote:
>>
>>> LGTM :-(
>>>
>>> Do we know who owns this bot?  I have no way of trying to repro / fix
>>> this bug.
>>>
>>>
>>>
>>> http://codereview.chromium.org/159763
>>>
>>
>>
>