Re: [Boost-testing] False failures?

From: David Abrahams (dave_at_[hidden])
Date: 2007-02-20 11:43:44

Misha Bergal <mbergal_at_[hidden]> writes:

> David Abrahams wrote:
>> "Misha Berga;" <mbergal_at_[hidden]> writes:
>> 
>>
>> Are you not updating regression.py on the branch?  
>
> No we are. My bad, I looked into one I had on one of my test machines, 
> not realizing that it has been recently updated to change 
> --send-bjam-log to --dont-send-bjam-log.
>
>  >
> That might explain
>> why I'm not ending up with the latest.  The one I have only has a
>> --dont-send-logs option.  IIRC regression.py tries to update itself
>> from the downloaded source during the update-self phase.  I think
>> maybe either you need to separately update it from the HEAD or make
>> sure you keep the RC branch in synch.
>> 
>> Hmm, even on HEAD I don't see --send-bjam-log.  What's up?
>> 
>>> and rerun the full cycle
>>>    (sorry) 
>> 
>> What does that mean?  A clean run from the beginning?
>
> Full (not an incremental) run.

Misha,

as far as I can tell, we're still having this problem.  What do you
need from me to help fix it?

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com