[Bioperl-l] Installing Bioperl on Ubuntu and Mac OS X Snow Leopard (perl 5.10.0)
steffen_moeller at gmx.de
Fri Jul 2 12:39:00 EDT 2010
On 07/02/2010 05:46 PM, Chris Fields wrote:
> On Jul 1, 2010, at 11:21 PM, George Iatropoulos wrote:
>> Hi guys!
>> I'd like to thank you for this important project. I currently develop
>> a script that uses the SeqIO module to read fasta or files with other
>> bio-formats. However your installation instructions are too difficult
>> to be followed, or even outdated, especially for the mac platform.
>> After trying a week I really can't install bioperl on a Snow Leopard
>> machine (perl 5.10.0) and I tried the CPAN and the fink method. None
>> of them works! Especially cpan seems to mess up with itself and now
>> doesn't install anything!
> That's odd. Can you send me what you are seeing?
>> As for the ubuntu platform a simple command like: "sudo apt-get
>> install bioperl" simply does the trick!
>> However I couldn't find this simple command in your installation
>> guidelines, and I don't really understand why I should set up a
>> server, before installing bioperl and then follow this complicated
>> series of instructions, to achieve something that is so simple!
> That's b/c the bioperl that is installed with apt is not up to date. CPAN has v.1.6.1; apt for karmic had 1.6.0 last I looked, v1.5.2 for bioperl-run (where CPAN is at v.1.6.1).
this depends on the age of your distribution rather than on what is
apt-get-able from various places. Debian Squeeze has 1.6.1, and so does
Ubuntu since Lucid.
> I would like to get the releases to a point where they are easily packaged for apt/yum/ppm/fink/macports/whatever, so any help along those lines is greatly appreciated. Now, re: Ubuntu I haven't had any problems with 9.04 or 9.10 (haven't jumped into 10.04 yet), but I'm a core bioperl dev. There are a few recommended modules that require outside libraries (GD first among them). These modules are not required for most core functionality.
> I am CCing the Debian-Med list just in case they have some better idea here, but what I suggest if you don't want to
>> I understand that you try to give complete instructions for people who
>> may be pleased with more advanced stuff, but please don't forget
>> biology students who don't have a deep computational background. These
>> guys like me would appreciate some simple stuff to get their job and
>> their projects done, without needing to waste days to figure out
>> complicated-guru stuff.
> We're actually working on segmenting BioPerl into separate packages, something I intend on bringing up next week at BOSC.
The intention of us Debian- and Ubuntu-Helpers is very much also to help
you X-core guys to concentrate on whatever you want to concentrate on
and by taking as many of the "cannot build this or that" questions away
from you as we possibly can. From our point of view, karmic or lenny are
already rather ancient. As a BioPerl developer you probably don't want
to run any of the packages in your routine anyway, so there is not much
of a point to update for you. But I share your concern for production
sites that want to sync with your vibrant development. If one wants to
stay with binaries I see apt-pinning and backports.org. If one would
just add the Debian sources of unstable to the /etc/apt/sources.list,
then one could also very easily create distribution-adjusted packages
oneself by apt-get source -b bioperl, which would then execute "the
make" for you.
The original request for a pointer to the Debian/Ubuntu packages would
be nice, indeed. We'll see us at BOSC next week. If there is anything
you think we should do to help your experience with the distributions
then please drop us a line or grep for me in Boston. The segmentation is
no problem for the packaging, but possibly more so for the regular user
with a biological background.
More information about the Bioperl-l