On Wed, Jan 16, 2013 at 15:31:39 +1300, Olly Betts wrote:
> On Sun, Jul 29, 2012 at 08:34:36PM +0200, Erich Schubert wrote:
> > A partial "bisection" of python-subversion via snapshot.debian.org
> > shows that the problem already existed in -3.1 but did not exist in
> > -1. I did not test -2 and -3 yet.
> >
> > -2 has swig-related changes:
> >
> > * patches/swig2-compat: New patch from upstream to build with swig 2.x.
> > (Closes: #634049)
> >
> > However, it may well be that this change is actually caused by Debian
> > having migrated to Swig 2, too.
>
> I believe this is due to a bug in SWIG which was fixed upstream in 2.0.7:
>
> 2012-05-22: szager
> [python] Disambiguate SWIG_From_unsigned_SS_int and
> SWIG_From_unsigned_SS_long.
>
> This was r13106 in SWIG's SVN repo (now migrated to git, but I don't have
> the corresponding git commit hash to hand).
>
> Testing already has SWIG 2.0.7, so rebuilding python-subversion should fix
> this if SWIG is run during the build.
>
I'm going to upload a new subversion build to wheezy. No source
changes, this would be a binNMU if it weren't for multiarch breakage.
Cheers,
Julien
signature.asc
Description: Digital signature