Re: [Boost-testing] [EXTERNAL] URGENT: Regression testing repo migration..

Subject: Re: [Boost-testing] [EXTERNAL] URGENT: Regression testing repo migration..
From: Belcourt, Kenneth (kbelco_at_[hidden])
Date: 2015-01-19 23:56:40

On Jan 19, 2015, at 9:40 PM, Rene Rivera <grafikrobot_at_[hidden]> wrote:

> On Mon, Jan 19, 2015 at 11:29 AM, Belcourt, Kenneth <kbelco_at_[hidden]> wrote:
> 
> On Jan 19, 2015, at 10:25 AM, Rene Rivera <grafikrobot_at_[hidden]> wrote:
> 
> > On Mon, Jan 19, 2015 at 10:52 AM, Belcourt, Kenneth <kbelco_at_[hidden]> wrote:
> >
> > On Jan 18, 2015, at 9:11 PM, Rene Rivera <grafikrobot_at_[hidden]> wrote:
> >
> > > On Sun, Jan 18, 2015 at 6:11 PM, Belcourt, Kenneth <kbelco_at_[hidden]> wrote:
> > > Hi Rene
> > >
> > > On Jan 16, 2015, at 10:00 AM, Rene Rivera <grafikrobot_at_[hidden]> wrote:
> > >
> > > > I'm going to remove the outdated regression testing code in the boostorg/boost super-repo the weekend of January 24th. Everyone MUST stop using the old scripts immediately as they will disappear in approximately 7 days.
> > > >
> > > > ***
> > > > All testers please migrate to using this run.py NOW
> > > >
> > > > <https://raw.githubusercontent.com/boostorg/regression/develop/src/run.py>.
> > >
> > > I switched over our testers to the new run script.  And while clang on Darwin and gcc on Linux seem to be cycling okay, the Intel :Linux testers are having problems.  Is this a known issue or any ideas why this this happening (it’s the same issue in both develop and master).
> > >
> > > Well.. I don't expect errors. So I guess it's unexpected. But I can't say anything more since I don't know what issue you refer to.
> >
> > I’ve reverted to the old run.py script.  The new one seems to cause the Intel testers to hang part way through the test process, no idea why.  This didn’t happen with the old run.py.
> >
> > There isn't much different from the new one vs. the old one. The key differences are the repointing of git paths to the new regression repo.
> 
> Ah. that must be it.  The regression repo you’re using is out of date with the one in tools.  I think there’s been a couple of changes that probably haven’t made it into your regresssion repo.  Perhaps you can diff them and see it anything jumps out?
> 
> The only CPP used during test runs is process_jam_log.cpp and it's the same on both. Are there other code changes you are thinking of?

No, nothing in particular.  BTW, I completely removed my testing directory, used the new run.py and Intel tested okay (though I didn’t allow it to post the results).  So let’s see what happens with tonight’s tests, perhaps clearing out the old cruft will be all it needs.

— Noel