Joining Python Modules Packaging Team

2013-06-06 Thread Brian May
Hello,

It has been suggested that I join the Python Modules Packaging Team.

I currently am the maintainer of the following packages:

django-tables
django-filters
django-ajax-selects

Plus a number of other packages that are not yet available in Debian.

My Alioth login is bam.

(Yes, I have been told that those package names should be prefixed with
python- to be policy compliant - just one of the things that needs fixing)

Thanks
-- 
Brian May 


Re: python3.3 status

2013-06-06 Thread Jakub Wilk

* Scott Kitterman , 2013-06-06, 12:46:

Did anyone file bugs for jwilk's FTBFS TODO packages?


I've file them all, with one exception:
I couldn't reproduce nuitka's FTBFS in a different build environment.

--
Jakub Wilk


--
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/20130606204443.ga8...@jwilk.net



Re: python3.3 status

2013-06-06 Thread Julian Taylor
On 06.06.2013 19:04, Jakub Wilk wrote:
> python-scipy FTBFS with new Cython: #707315

and needs fixed numpy.
a new upstream which works with new cython is ready in svn.


-- 
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/51b0cb5d.4050...@googlemail.com



Re: python3.3 status

2013-06-06 Thread Jakub Wilk

* Scott Kitterman , 2013-06-06, 12:46:
mpi4py, pyepr, pystemmer,  python-llfuse, python-scipy, and 
python-astropy all needed a fixed cython, which is now available.


pyepr builds successfully, but puts dbg extensions into wrong package: 
#708011

python-scipy FTBFS with new Cython: #707315
python-astropy needs fixed Numpy
mpi4py, pystemmer, python-llfuse are OK

--
Jakub Wilk


--
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/20130606170405.gb6...@jwilk.net



Re: python3.3 status

2013-06-06 Thread Scott Kitterman
On Thursday, May 23, 2013 06:19:06 PM Scott Kitterman wrote:
> I've been working on this a bit ...
> 
> On Tuesday, May 07, 2013 02:01:26 PM Jakub Wilk wrote:
> > python3-defaults maintainer(s?) decided to make python3.3 a supported
> > version without prior notice. Yay. Now we have dozens of packages
> > failing to build:
> > 
> > boost1.49 FTBFS TODO
> > boost1.50 FTBFS TODO
> > cairosvg FTBFS, needs fixed py3cairo
cssutils  - Fixed
> > distribute FTBFS TODO
> > flufl.bounce FTBFS #707086, needs new zope.interface
> > germinate FTBFS, needs fixed python-apt
> hivex FTBFS fixed, but still doesn't generate dependencies right #709516
> > libguestfs FTBFS TODO
> > mdp FTBFS, needs fixed python-numpy
mod-wsgi, Now fixed
> > mpi4py needs fixed cython
> > nuitka FTBFS TODO
> > objgraph FTBFS #686331
> > pyepr needs fixed cython
> > pyfits FTBFS, needs fixed python-numpy
pymarkups - Fixed
> > pyopencl FTBFS, needs fixed boost1.49
> > pyside FTBFS  TODO
> > pystemmer needs fixed cython
pytest - Fixed
> > python-apt FTBFS #681744
> > python-astropy needs fixed cython
> > python-cffi FTBFS TODO
> > python-csb FTBFS TODO
> > python-distutils-extra FTBFS needs rebuilt pygobject
python-docutils - Fixed
python-flexmock - Fixed
> > python-llfuse needs fixed cython
python-notify2  - Fixed
> python-numpy Fixed in experimental + SVN
> > python-scipy needs fixed cython, then likely FTBFS #691254
> > python-stem  FTBFS TODO
> > python-wadllib FTBFS #686332
> > shiboken FTBFS TODO
> > yapsy FTBFS TODO
> > yp-svipc FTBFS, needs fixed python-numpy
zope.interface - Needs update to 4.0.2 or later

> > "FTBFS TODO" means that a package fails to build, but there's no bug
> > filed yet. I plan to file bugs today or tomorrow, unless there are
> > other volunteers to do it.

I've updated this list by removing ones that were reported fixed in the last 
mail I sent and then updating again based on responses to it.  

mpi4py, pyepr, pystemmer,  python-llfuse, python-scipy, and python-astropy all 
needed a fixed cython, which is now available.  It would be nice if someone 
would check those.

Getting zope.interface updated would be helpful as well, but seems to need 
sponsoring.  If someone has some sponsoring cycles available, please have a 
look.

Most of the red in http://release.debian.org/transitions/html/python3.3.html 
can be binNMUed after python-numpy is uploaded (promised this weekend, AIUI).  
It would be worth doing some investigation of the unknown packages to see if 
they are generating dependencies correctly.

Any other progress?  Did anyone file bugs for jwilk's FTBFS TODO packages?

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/3082334.MxQK4dTUX1@scott-latitude-e6320



Re: RFS: bunch, kitchen, grapefruit, fabulous, stomper, txws, txzmq, moksha.common, moksha.hub

2013-06-06 Thread Jakub Wilk

* Simon Chopin , 2013-06-06, 14:19:

fabulous/fonts/cmr10.ttf license doesn't look free to me.
I've removed all the fonts from the tarball since their license wasn't 
specified anyway — I had to go look in other packages to find them. I 
guess the FTP masters disagree with you on the license though, as one 
can find this font at least in fonts-lyx.


I wouldn't assume that just because something is in main, it's 
necessarily blessed by ftp-masters. Two reasons:

1) mistakes happen;
2) not every upload goes through NEW.

I don't see a point of making fabulous-xtermspeedup a separate 
package. It's tiny, and doesn't bring any extra dependencies.

Yes, but it's arch:any whereas the rest is arch:all.


That would be a valid reason to keep them separated _if_ the arch:all 
package were big. But it's tiny, too.



The package FTBFS when I build only arch-dependent packages:
|dh_sphinxdoc -a
| dh_sphinxdoc: Sphinx documentation not found
| make: *** [binary-arch] Error 2

Shouldn't dh_sphinxdoc -a be a NOP?


In general case, no: it's possible that the documentation is included in 
an arch:any package.


In this case: maybe. Here's an excerpt from #debian-mentors from a few 
days ago on this subject:


< helmut> jwilk: while working on doxygen, a question on --with 
sphinxdoc appeared. maybe you have an answer: as far as I can see the 
sequence causes dh_sphinxdoc to be executed on all builds, but it fails 
when it does not find any sphinx documentation. does this work when the 
docs reside in arch:all packages and building arch-only?

[...]
< jwilk> helmut: It doesn't.
[...]
< jwilk> The alternative is to do nothing if you don't find docs. (I'm 
not saying I like this idea.)
< jwilk> Oh, and at some point I was considering doing this: 
http://paste.debian.net/6076
< jwilk> But since nobody bothered to file a bug about the current 
behavior so far, I'm assuming that people are mostly content with it. :P


(The paste has expired since then, so I attached the patch I had in 
mind.)


If I were the maintainer, I would remove (with a Debian-specific 
patch) this junk from setup.py:

| from ez_setup import use_setuptools
| use_setuptools(version='0.6c11')
I always thought this was only a convoluted way to check for the 
version and never bothered to actually look at ez_setup. Now I 
understand why you'd remove it, and will do the same from now on.


You could drop the "The information above should ..." part from the 
patch header. :)



lintian4python emits (among others):
x: python-fabulous: except-without-exception-type 
usr/share/pyshared/fabulous/logs.py:86
x: python-fabulous: except-without-exception-type 
usr/share/pyshared/fabulous/utils.py:95
x: python-fabulous: except-without-exception-type 
usr/share/pyshared/fabulous/xterm256.py:104

Patched and reported upstream.


Now it emits:
e: python-fabulous: except-shadows-builtin 
usr/share/pyshared/fabulous/xterm256.py:104: AssertionError

--
Jakub Wilk
Index: debian/control
===
--- debian/control	(revision 18535)
+++ debian/control	(working copy)
@@ -20,6 +20,7 @@
   ${sphinxdoc:Depends}, libjs-sphinxdoc (= ${source:Version})
 Recommends: python (>= 2.6) | python-simplejson, python-imaging
 Suggests: jsmath
+Breaks: debhelper (<< 7.3.16~)
 Description: tool for producing documentation for Python projects
  Sphinx is a tool for producing documentation for Python projects, using
  reStructuredText as markup language.
Index: debian/changelog
===
--- debian/changelog	(revision 18552)
+++ debian/changelog	(working copy)
@@ -2,8 +2,11 @@
 
   * Export PYTHONWARNINGS=d in debian/rules to enable all warnings in Python
 code.
+  * dh_sphinxdoc: when run via the dh addon and acting on a strict subset of
+all binary packages, don't fail if no documentation was found.
++ This feature requires debhelper 7.3.16, add Breaks for older versions.
 
- -- Jakub Wilk   Thu, 15 Sep 2011 18:48:37 +0200
+ -- Jakub Wilk   Sat, 17 Sep 2011 14:07:34 +0200
 
 sphinx (1.0.7+dfsg-2) unstable; urgency=low
 
Index: debian/dh-sphinxdoc/dh_sphinxdoc
===
--- debian/dh-sphinxdoc/dh_sphinxdoc	(revision 18535)
+++ debian/dh-sphinxdoc/dh_sphinxdoc	(working copy)
@@ -306,8 +306,11 @@
 return 1;
 }
 
-init();
+init(options => { "dh" => \$dh{FLAG_DH} });
 
+my $all_packages = @{$dh{DOPACKAGES}} == getpackages();
+my $mercy_mode = $dh{FLAG_DH} && $all_packages;
+
 load_packaged_js();
 
 my @paths = @ARGV;
@@ -340,7 +343,7 @@
 }, $pkgpath);
 }
 }
-if ($done == 0)
+if ($done == 0 and not $mercy_mode)
 {
 my $message = 'Sphinx documentation not found';
 $message .= " at $path" if defined $path;
Index: debian/dh-sphinxdoc/sphinxdoc.pm
===
--- debian/dh-sphinxdoc/sphinxdoc.pm	(revision 1

Re: RFS: bunch, kitchen, grapefruit, fabulous, stomper, txws, txzmq, moksha.common, moksha.hub

2013-06-06 Thread Simon Chopin
Quoting Jakub Wilk (2013-05-08 12:29:58)
> * Simon Chopin , 2013-05-07, 20:04:
> >(or RFR, for Jakub :P)
> 
> Oh hi!
> 
> >fabulous: Makes your terminal output totally fabulous
> 
> fabulous/fonts/cmr10.ttf license doesn't look free to me.

I've removed all the fonts from the tarball since their license wasn't
specified anyway — I had to go look in other packages to find them.
I guess the FTP masters disagree with you on the license though, as one
can find this font at least in fonts-lyx.

> 
> I see this in debian/patches/build_xterm256_ext:
> -library = expanduser('~/.xterm256.so')
> +library = '/usr/share/lib/xtermspeedup.so'
> Errr, /usr/share/lib/?
Ooops. Fixed
> 
> I don't see a point of making fabulous-xtermspeedup a separate package. 
> It's tiny, and doesn't bring any extra dependencies.
Yes, but it's arch:any whereas the rest is arch:all.

> The package FTBFS when I build only arch-dependent packages:
> |dh_sphinxdoc -a
> | dh_sphinxdoc: Sphinx documentation not found
> | make: *** [binary-arch] Error 2
Shouldn't dh_sphinxdoc -a be a NOP?

> 
> If I were the maintainer, I would remove (with a Debian-specific patch) 
> this junk from setup.py:
> | from ez_setup import use_setuptools
> | use_setuptools(version='0.6c11')

I always thought this was only a convoluted way to check for the version
and never bothered to actually look at ez_setup. Now I understand why
you'd remove it, and will do the same from now on.

> 
> lintian4python emits (among others):
> x: python-fabulous: except-without-exception-type 
> usr/share/pyshared/fabulous/logs.py:86
> x: python-fabulous: except-without-exception-type 
> usr/share/pyshared/fabulous/utils.py:95
> x: python-fabulous: except-without-exception-type 
> usr/share/pyshared/fabulous/xterm256.py:104
Patched and reported upstream.

Cheers,
Simon


signature.asc
Description: signature


RFS: python3-sympy

2013-06-06 Thread Thomas Kluyver
Please can someone upload the new package python3-sympy

Package name : python3-sympy
Version : 0.7.2-2
URL : http://sympy.org/
Binary packages: python3-sympy

It's already in the team svn:
svn://svn.debian.org/python-modules/packages/python3-sympy/trunk/

>From a previous discussion, it's OK to have separate source for this,
because upstream doesn't support building Python 2 and 3 packages from the
same release tarball:
http://lists.debian.org/debian-python/2012/10/msg00041.html

Thanks,
Thomas