[Bioperl-l] Updated Makefile.PL

[Chris Fields]

> Hi,
> I've just committed an updated Makefile.PL to HEAD for bioperl-live.
> Could some people test it on multiple platforms and confirm it is ok
> (try out the different possible options as well)?
> 
> (NB. in the below, 'pre-reqs' are things the makefile considers optional
> dependencies)
> 
> Note that some pre-reqs have been removed:
> # DBD::mysql (nothing in Bioperl-live uses it; Bioperl modules may end
> up requiring it but only after the user makes an explicit choice by
> typing 'DBD::mysql' in their own code to supply as an option to Bioperl
> code)
> # File::Temp (standard in 5.6.1)

I'll try it out on WinXP and Mac OS X.  BTW, do any of Lincoln's Bio::DB*
use DBD::mySQL?  Bio::DB::GFF comes to mind.  I don't think it should be an
absolute requirement, though.

If we plan on removing those, then we should also remove them from
Bundle::Bioperl (if they are present).

> This pre-req was wrong:
> # Data::Stag::Writer
> and has been replaced with:
> Data::Stag::XMLWriter
> 
> 
> Also, I note that very many Bioperl modules need IO::String, including
> Bio::SeqIO, so I'm not sure to what extent we can pretend it is an
> optional module. I didn't make any change though.

Do they all require IO::String or is it an option?  There are a few
instances (WebDBSeqI-implementing, for instance) where this is presented as
an option for most OS's (along with the default, pipeline, and tempfile).
However, it is currently used by default with Windows due to lack of
pipe/fork support at the time.

BTW, the latter may now work with WinXP ActivePerl.  ActiveState has been
working on WinXP fork() emulation for a while, but I think it is still
somewhat experimental.  

> I don't know if these changes affect the Windows ppm Nathan, or anything
> else (Bundle?)?
> 
> The INSTALL docs need updating with these new and improved pre-reqs
> (note that some pre-reqs had wrong/not enough Bioperl modules listed as
> needing them); does someone want to correct the wiki (based on the new
> Makefile.PL) and then Chris can re-create the text version?

Easier to just modify the text version based on what is changed in the wiki,
at least for the time being.  The text dumping from elinks/lynx isn't
full-proof re: tables and such, which is one reason I think we should move
the prereqs to a separate file as it's easier to maintain long-term (this
seems to be where most changes occur anyway).  

Christopher Fields
Postdoctoral Researcher - Switzer Lab
Dept. of Biochemistry
University of Illinois Urbana-Champaign