[Bioperl-l] Updated Makefile.PL
Nathan S. Haigh
n.haigh at sheffield.ac.uk
Sun Oct 22 19:37:07 UTC 2006
Chris Fields wrote:
> On Oct 22, 2006, at 5:34 AM, Nathan S. Haigh wrote:
>> Is it possible to make a distinction in Makefile.PL between those
>> modules that are an absolute must for Bioperl-core and those which are
>> optional and should go into Bundle::BioPerl?
>> Once I'm sure what should be "option" I'll do the Bundle::BioPerl
>> package and PPD's.
> We probably should steer this way eventually. Do you aim on placing
> prereqs required for bioperl core in the bioperl PPD and the
> 'optional' ones with the bundle?
That's correct. However, PPM will always try to update packages to the
latest available. Therefore, if at some point in the future, a
dependency is removed, and thus removed from Bundle::BioPerl, a
situation may arise where an older version of BioPerl is running with
the a recent version of Bundle::BioPerl and could have missing
dependencies - not ideal but it is how things currently stand. The
process of making the Bundle::BioPerl PPD would be simplified if these
"optional" dependencies are separated from the "core" dependencies. If
one of the following solutions is possible (i'm not sure if they are),
it would be very useful:
1) Maintain 2 hashes in Makefile.PL that contain the "core" and
"optional" dependencies. In unsure of the way dependencies are ordered
during a "make ppd", but it may be possible to pass hash references of
both to PREREQS_PM in MakeMakefile and have the "optional" depenencies
grouped separately from "core" depenedcies in the ppd file - thus making
it easy to stip them out into a Bundle::BioPerl ppd.
2) Again, maintain 2 hashes in Makefile.PL that contain the "core" and
"optional" dependencies. Have some Makefile setup that allows the
generation of a Bundle::BioPerl ppd separately from the main Bioperl ppd.
Like I said, these are just some thoughts and I'm not sure if they are
even viable options.
More information about the Bioperl-l