[Bioperl-l] Bioperl NT update

[Jason Stajich]

> Peter Schattner wrote:
> > >
> > > What is the problem with precomputed results?
> >
> > Perhaps I don?t understand what is meant by precomputed results.  I
> > assumed you meant making a file with the answers that the test program
> > expects and using those answers if the OS is non-unix.  That approach
> > seems less maintainable than simply skipping the tests if the OS is
> > non-unix.
> >
>
> I'm not sure either what you mean. Trying to be clear, I do not mean
> setting up a file with the output you expect from the test script. I
> meant, instead of calling out to an executable clustalw, precompute the
> output said executable is supposed to produce, capture it in a file and
> feed this file into your module. This way your module can be tested on
> any platform, and it can also be used on any platform whether or not
> clustalw and friends are there, provided the user has means to obtain
> the results from another machine than the one he runs bioperl on.
>
> This of course requires that your modules can accept either a stream
> containing input or a pipe fed by one of the external executables.
>
> Am I missing something?

The StandAloneBlast, ClustalW, and TCoffee modules are intended to be
wrappers around actually running these programs (one of my problems with
Bio::Tools is that we do not have a directory structure that differentiates
processing results and running programs!) rather than processing results.
The results processing for TCoffee and ClustalW (MSA tools) is done in
Bio::AlignIO, while the StandAloneBlast result is sent to Bio::Tools::BPlite
or Bio::Tools::Blast.  So there really is not way to run these on Win32 or
Mac so these tests should just be skipped (except for blast which can be run
on Win32).

-jason

>
> Hilmar
>
> --
> -----------------------------------------------------------------
> Hilmar Lapp                                email: hlapp@gmx.net
> GNF, San Diego, Ca. 92122                  phone: +1 858 812 1757
> -----------------------------------------------------------------
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l@bioperl.org
> http://bioperl.org/mailman/listinfo/bioperl-l
>