[boost] Re: Beta posted for new regression tests

From: Gennadiy Rozental (rogeeff_at_[hidden])
Date: 2002-07-23 09:18:29

"Beman Dawes" <bdawes_at_[hidden]> wrote in message
news:4.3.2.7.2.20020723091440.00b93508_at_mailhost.esva.net...
> As the Boost regression tests have grown in size, and more of the Boost
> libraries require a shared library be built first, a new regression test
> framework was needed.
>
> Dave, Rene, and others working on the Boost.Build system have provided
that
> new framework.
>
> But just running the tests isn't enough; if it isn't easy to see what
tests
> are failing and why, problems just don't get fixed. I've also worried that
> compiler warnings were not being acted on often enough, probably because
> the warnings were buried so deeply in log files that developers never saw
> them.
>
> Accordingly, the Compiler Status report has been upgraded.
>
> * The old Pass/Fail result reporting has been replaced by:
>
>       Pass - The test passes and there are no warning messages.
>       Warn - The test passes, but there are warning messages.
>       Fail - The test fails.
>       Missing - No test results available (probably shared library
>                 failed to build)
>
> * If warning or error messages are available, the result is hyper-linked
to
> an entry in a status file, also in HTML format.  The hope is that by
making
> it very easy (one click!) to view the warning or error message, it will be
> acted upon more often.
>
> * It is possible to vary the Status Report; tests which pass on all
> compilers can be omitted for example.  For daily reporting, we probably
> don't want to bother with tests that pass all compilers.
>
> A beta has been posted at http://www.esva.net/~beman/jam_regr.html
>
> It only covers three compilers, but should be enough to get the flavor.
>
> Comments and suggestions appreciated.
>
> --Beman
>
>
> _______________________________________________
> Unsubscribe & other changes:
http://lists.boost.org/mailman/listinfo.cgi/boost
>