Questions regarding a package that creates a python module (written in C)

2010-05-26 Thread Rogério Brito
Dear Python developers,

I am a newbie with Python (though I like the language) and I help
maintain a package called fontforge, which is a font ("letters") editor
that creates two modules [*] written in C so that some tasks with fonts
can be performed en masse.

Unfortunately, even when depending on python-dev-all (or whatever I
should call it nowadays), I only get the modules created for python2.5
and I would like (in principle, barring technical limitations) to
provide such modules for any version of python that Debian supports,
agnostically.

It is my understanding from the build system of fontforge that it only
picks up python2.5 since it is the first version that is encountered.

With that said, what is the recommended course of action? Should I
manually build the package twice in a row in debian/rules telling it
first to configure with python 2.5 and then 2.6 or is there any solid,
recommended approach to this problem (which I am sure should appear in
many other packages that provide python modules written in C)?


Any guidance is welcome, Rogério Brito.

P.S.: Please, keep me on CC'ies as I am not currently subscribe to -python.

[*]
WARNING: WARNING: /usr/lib/pyshared/python2.5/fontforge.so is linked but does 
not belong to any package.
WARNING: WARNING: /usr/lib/pyshared/python2.5/psMat.so is linked but does not 
belong to any package.

but

,[ dpkg -S /usr/lib/pyshared/python2.5/fontforge.so ]
| python-fontforge: /usr/lib/pyshared/python2.5/fontforge.so
`
-- 
Rogério Brito : rbr...@{ime.usp.br,gmail.com} : GPG key 1024D/7C2CAEB8
http://rb.doesntexist.org : Packages for LaTeX : algorithms.berlios.de
DebianQA: http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br


-- 
To UNSUBSCRIBE, email to debian-python-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20100526172853.ga19...@ime.usp.br



Re: Questions regarding a package that creates a python module (written in C)

2010-05-26 Thread Jakub Wilk

* Rogério Brito , 2010-05-26, 14:28:

I am a newbie with Python (though I like the language) and I help
maintain a package called fontforge, which is a font ("letters") editor
that creates two modules [*] written in C so that some tasks with fonts
can be performed en masse.

Unfortunately, even when depending on python-dev-all (or whatever I
should call it nowadays), I only get the modules created for python2.5
and I would like (in principle, barring technical limitations) to
provide such modules for any version of python that Debian supports,
agnostically.


That's very nice of you. :)


It is my understanding from the build system of fontforge that it only
picks up python2.5 since it is the first version that is encountered.


Yes, most likely this is the case. I haven't seen any upstream build 
system that supports building/installing module for multiple Python 
version on its own initiative.



With that said, what is the recommended course of action? Should I
manually build the package twice in a row in debian/rules telling it
first to configure with python 2.5 and then 2.6


Yes. Please don't hardcorde version numbers in debian/rules, however; 
use `pyversions -r` to query which Python versions are supported.


or is there any solid, recommended approach to this problem (which I am 
sure should appear in many other packages that provide python modules 
written in C)?


For distutils-based build systems (i.e., not for fontforge), dh(1) 
builds modules for all supported Python versions (almost) out of the 
box. I'm sure that cdbs has a way to automatize this task, too (though I 
don't recommend cdbs to anyone).


--
Jakub Wilk


signature.asc
Description: Digital signature


Python and python3 as separate runtime systems

2010-05-26 Thread Scott Kitterman
This has been discussed a bit, but I'd like to see where Debian Python 
consensus is on this.

I think users who don't care about Python 3 yet, should be able to have 
systems that don't pull any Python 3 elements on by accident.  Python 3 is 
primarily of interest to developers right now and most users won't care to 
have it.

It is very unusual for code to be able to run unmodified in both Python 2 and 
Python 3.

I think it makes sense for them to be kept separate at runtime.

This would mean separate python-foo and python3-foo binaries where both are 
supported from the same source.

It would mean that binaries should not depend on both Python 2 and Python 3 
interpreters.

I means that the support tools for Python 3 should run in Python 3 As an 
example, I've done a first cut at py3versions:

http://www.kitterman.com/debian/

This will also, in the very long run, make it easier to transition to a Python 
3 version as the default Python.

Thoughts,

Scott K


-- 
To UNSUBSCRIBE, email to debian-python-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/201005261743.43894.deb...@kitterman.com



RFS: pyevolve

2010-05-26 Thread Christian Kastner
Hello,

I am looking for a sponsor for my package "pyevolve".

* Package name: pyevolve
  Version : 0.6~rc1~svn397+dfsg-1
  Upstream Author : Christian S. Perone 
* URL : http://pyevolve.sourceforge.net
* License : Python License (with upstream subsituted for PSF)
  Section : python

It builds these binary packages:
python-pyevolve - Complete genetic algorithm framework written in pure
python
python-pyevolve-doc - Documentation for the Pyevolve genetic algorithm
framework

The package appears to be lintian clean.

The upload would fix these bugs: 580924


Pyevolve is one of the most complete genetic algorithm frameworks
currently available. It is trivially easy to use out-of-the-box and yet
easily extendable. It also provides extensive documentation.

For a good overview of Pyevolve's feautres and capabilities, see ACM
SIGEVOlution Volume 4 Issue 1, available at http://www.sigevolution.org.

Upstream is very responsive.


The package can be found on mentors.debian.net:
- URL: http://mentors.debian.net/debian/pool/main/p/pyevolve
- Source repository: deb-src http://mentors.debian.net/debian unstable
main contrib non-free
- dget
http://mentors.debian.net/debian/pool/main/p/pyevolve/pyevolve_0.6~rc1~svn397+dfsg-1.dsc

I would be glad if someone uploaded this package for me.


Thanks,
Christian



signature.asc
Description: OpenPGP digital signature


Re: Python and python3 as separate runtime systems

2010-05-26 Thread Jakub Wilk

* Scott Kitterman , 2010-05-26, 17:43:

I think users who don't care about Python 3 yet, should be able to have
systems that don't pull any Python 3 elements on by accident.  Python 3 
is primarily of interest to developers right now and most users won't 
care to have it.


It is very unusual for code to be able to run unmodified in both Python 
2 and Python 3.


I think it makes sense for them to be kept separate at runtime.

This would mean separate python-foo and python3-foo binaries where both 
are supported from the same source.


It would mean that binaries should not depend on both Python 2 and 
Python 3 interpreters.


I means that the support tools for Python 3 should run in Python 3


Seconded.

--
Jakub Wilk


signature.asc
Description: Digital signature


Re: Python and python3 as separate runtime systems

2010-05-26 Thread Barry Warsaw
On May 26, 2010, at 05:43 PM, Scott Kitterman wrote:

>I think it makes sense for them to be kept separate at runtime.
>
>This would mean separate python-foo and python3-foo binaries where both are 
>supported from the same source.
>
>It would mean that binaries should not depend on both Python 2 and Python 3 
>interpreters.

+1
-Barry


signature.asc
Description: PGP signature


Re: Python and python3 as separate runtime systems

2010-05-26 Thread Luca Falavigna
Il 26/05/2010 23.43, Scott Kitterman ha scritto:
> This would mean separate python-foo and python3-foo binaries where both are 
> supported from the same source.
> 
> It would mean that binaries should not depend on both Python 2 and Python 3 
> interpreters.

ACK.
Having a safety new provided by NEW processing would introduce lesser
bugs and a smoother transition.

-- 

  .''`.
 : :' :   Luca Falavigna 
 `. `'
   `-



signature.asc
Description: OpenPGP digital signature