|
|
Boost Testing : |
Subject: Re: [Boost-testing] Lost AIX tester?
From: K. Noel Belcourt (kbelco_at_[hidden])
Date: 2010-01-21 11:54:37
Hi Chris,
On Jan 21, 2010, at 9:41 AM, Christopher Cambly wrote:
> > This is taking longer than we expected.
> >
> > Walking though the run.py output we can see that run.py successfully
> > downloads the tarballs, builds bjam, configures python then no
> > further output exist in the log file:
> > ...
> > bjam appears fine, I can use the bjam with no issues trying to rule
> > out a bad build of bjam. It would be helpful if we can get some
> > additional tracing from regression.py. Is this possible? Any
> ideas?
>
> debug-level=x worked.
>
> I believe we have the runs going again. For some odd reason bjam
> appears to want walk the entire filesystem on the machine. Some of
> our nfs mounts had gone stale causing bjam to hang.
>
I reported this issue a while ago. I think Bjam is scanning way too
much of the disk for dependent files. I think we should limit where
it looks for files to not much more than the project files.
Glad to hear you got that resolved.
-- Noel