[Bioperl-l] problem with installation of Bioperl1.4 on Windows XPPC using ActivePerl PPM

[Nathan Haigh]

Quoting Chris Fields <cjfields at uiuc.edu>:

> > Did you get round to trying this? Did you find anything?
> 
> Yes I have.  PPM4 is interesting, but I find the lack of a full command-line
> version a bit frustrating (though I can see why they are moving in this
> direction).  It doesn't require the use of a package.lst in the ppm
> directory, but it looks for it first.  If it isn't present it looks for the
> ppd files.  This would be nice for Bioperl except we want to continue
> support for Perl 5.6.1, so we must include package.lst in the repository.
> 
> I can definitely replicate the error reported before when adding the
> repository using ppm4:
> 
> ERROR: Installing File-Spec-0.82 would downgrade File::Spec from version
> 3.12 to 0.82 and File::Spec::Functions from version 1.3 to 1.1 and
> File::Spec::Mac from version 1.4 to 1.2 and File::Spec::OS2 from version 1.2
> to 1.1 and File::Spec::Unix from version 1.5 to 1.2 and File::Spec::VMS from
> version 1.4 to 1.1 and File::Spec::Win32 from version 1.6 to 1.2
> 
> I think this comes from the various versioning requirements for those
> modules, which probably arise from the v 1.4 Makefile.PL.  The new version
> doesn't have those requirements but a few others are present.  When a new
> PPM is made we probably need to take that into consideration.

Not sure I understand :o(

> 
> I noticed that the GMOD repository (http://www.gmod.org/ggb/ppm/) carries a
> newer bioperl version (1.512) which is needed for GBrowse.  This is a
> minimal installation w/o any requirements added and is accessible from PPM4;
> it installs w/o a hitch.
>  
> > I'm still working on getting Bioperl 1.5.2 installed on ActivePerl 5.6.1
> > and as many dependencies as I can in order to run the test suit -
> > however, I'm have a lot of headaches! I'm about to work through it a bit
> > more systematically, and note the problems I'm having. One that comes to
> > mind straight away is:
> > 
> > To run the tests, I need nmake. However, nmake fails if commands break
> > the command line length limit - which is shorter on Windows than other
> > OS's. This was solved by the authors of ExtUtils::MakeMaker when I
> > informed them, some time ago. So, to run the tests, I need
> > ExtUtils::MakeMaker >=6.06 installed. However, I'm having problems with
> > installing this! I've informed the authors, but they resorted to
> > manually copying over the files and suggested using ActivePerl 5.8.8!
> > This is not good as far as making testing easy - although I hope that
> > once i've installed the latest version of ExtUtils::MakeMaker and the
> > Bioperl dependencies, I'll be able to run the tests.
> 
> If you download the free MS Visual C++ Express Edition, it comes with a much
> newer version of nmake:
> 
> G:\Program Files\Microsoft Visual Studio 8\VC>nmake -help
> 
> Microsoft (R) Program Maintenance Utility Version 8.00.50727.42
> ...
> 
> It's available here:
> 
> http://msdn.microsoft.com/vstudio/express/
> 
> That may resolve the issue.

Thanks - I'll give it a try! I did download the Borland compilier v5.5 (i think) and I didn't seem to have much luck with that either - I may have
become delierious by then though!

> 
> > Another thing is the lack of the correct implementation of some ppd
> > modules, but that hopefully, should be reasonable straight forward.
> > Nathan
> 
> Which ones were you thinking of?  GD::SVG and Text::Shellwords are both
> available on the bioperl site; we could add others as needed if they aren't
> available through ActiveState or the other repositories.
> 
> The critical issue is that of binary compatibility between 5.6.1 and the
> various 5.8.x versions of perl.  However, there are no binary components for
> Bioperl.  One could probably modify the ppd file to deal with both versions
> (similar to the older ppd files in bioperl.org/DIST.

There were several modules but can't remember which off hand. Some were not direct dependents of Bioperl, but dependencies of denpendencies! Most (i
think) were because only 1 implementation tag was in the ppd file - for Perl 5.8 but non for 5.6. Most of these should be easy to rectify, but others
may not be.

I'll have another bash after the weekend!

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