|
Boost Testing : |
From: Vladimir Prus (ghost_at_[hidden])
Date: 2006-12-08 06:35:58
Markus Schöpflin wrote:
> Vladimir Prus wrote:
>
>> Markus Schöpflin wrote:
>>
>>> Starting 20061208-034259 UTC, regression testing on the RC branch has
>>> been broken for me. The following error appears in the log:
>>
>> Strange. The only recent change is some doxygen changes from Rene, which
>> should not have effect on anything.
>>
>> Can you either:
>>
>> 1. Try to "cvs up" to older versions of tools/build/v2
>> to pinpoint the change that broke things.
>>
>> 2. Send me your user-config.jam, so that I can try to
>> reproduce this myself.
>
> I think I found the reason for the problem. It's not in bbv2, but caused
> by the change to regression.py.
>
> if not toolsets is None:
> if v2:
> - result += ' ' + string.join(string.split( toolsets, ',' ), '
> ' )
> + result += ' "--toolset=%s"' % string.join( string.split(
> toolsets, ',' ), ' ' )
> else:
> result += ' "-sTOOLS=%s"' % string.join( string.split(
> toolsets, ',' ), ' ' )
>
> I run the regression tests with
> --toolsets="hp_cxx-65_042_tru64,hp_cxx-71_006_tru64" as an argument to
> regression.py.
>
> Seems like bjam doesn't like the resulting "--toolset=hp_cxx-65_042_tru64
> hp_cxx-71_006_tru64".
>
> Does that make sense?
Yes. I've undid the problematic checkin and posted a separate mail to
discuss what's wrong. Thanks for the heads-up.
- Volodya