|
|
Boost Testing : |
From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2005-04-08 22:10:34
Martin Wille writes:
> Hello,
>
> I was talking with someone on IRC who, maybe, would be able to run
> regression tests on Solaris/x86. At that point we found out we're
> unable to tell whether this is already being tested, because Caleb's
> results only show "SunOS 5" in the title (Later I found out Caleb uses
> Sparc).
>
> While "SunOS 5" is correct, it also isn't very informative, since the
> real information is in the minor number which has been used as the
> Solaris version number since SunOS 5.7/Solaris 7.
>
> However, my question is whether we should require the results for some
> platforms to indicate the CPU architecture. At least for SunOS and
> Linux it would make sense in the long run. It'll also make sense for
> *BSD if we had any testers for those operating systems.
I agree.
> What do you
> think about this? (One can find out the CPU for SunOS by looking at
> the output of config_info. However, that's a bit cumbersome.)
And doesn't allow you to mark up platform-specific failures.
>
> I could add "-x86" manually for my Linux results. However, for those
> who use the regression.py script it'll probably make more sense if
> that information would be created by the script.
For now, I'd suggest providing it through "--platform" flag to
regression.py.
I've added this on the Boost.Testing page as a TODO item co-scheduled
with Boost.Build v2 transition (when we'll be revising other toolset
names' issues as well).
-- Aleksey Gurtovoy MetaCommunications Engineering