|
|
Boost Testing : |
From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2006-12-12 13:55:13
Clark Sims writes:
> I have seen two sets of documentation with instructions for running
> the regression test:
>
> http://www.boost.org/more/regression.html
FWIW, I just updated a website copy of this one to match the latest
CVS HEAD revision.
> and
>
> http://engineering.meta-comm.com/boost.aspx
>
> Neither of them mention user-config.jam.
Yep; requiring it for running regressions with standard toolsets was
an oversight (fixed now), sorry you had to struggle with it!
> I stumbled through the process of running the regression test, and
> finally got them to work after about a week. Rene spent a lot of
> time helping me with the problem. Also Valodia pointed out the
> correct syntax for specifying an alternative gcc compiler. This
> syntax isn't documented either, but is most important:
>
> # Configure gcc (default version)
> using gcc : : /usr/bin/g++-4.1.1 ;
While this _is_ documented in the Boost.Build documentation
(http://boost.org/boost-build2/doc/html/bbv2/advanced/configuration.html),
you are absolutely right in wanting [a reference to] this to be
present in the "Running Boost Regression Tests" document as well.
Will add.
> If possible the people who manage these pages should :
> 1) metion that user-config.jam is necessary to running regression.py
> 2) the line using gcc .... is very important, and give a few examples.
> 3) the line using python ; is very important and must be
> specified.
Hmm, I wonder if we can get rid of #3. Vladimir, David?
>
> If the documentation were updated
It will be.
> maybe the noob testers would take
> quite so much time, and might not flood the test reports with
> misconfigured compiler results.
Thank you for your feedback!
-- Aleksey Gurtovoy MetaCommunications Engineering