On Wed, 2019-05-01 at 09:15 +0200, Pierre Labastie via blfs-dev wrote:
> On 01/05/2019 08:38, Bruce Dubbs via blfs-dev wrote:
> > On 4/30/19 10:07 PM, Douglas R. Reno via blfs-dev wrote:
> > > Hi folks,
> > >
> > > While updating Subversion (and SWIG) to their latest versions on my
> > > development machine before I built nmap for gnome-nettool, I ran into a
> > > problem building the python bindings:
> > >
> > > /usr/bin/swig -I/sources/subversion-1.12.0/subversion-1.12.0/subversion
> > > -I/sources/subversion-1.12.0/subversion-1.12.0/subversion/include
> > > -I/sources/subversion-1.12.0/subversion-1.12.0/subversion/bindings/swig
> > > -I/sources/subversion-1.12.0/subversion-1.12.0/subversion/bindings/swig/include
> > > -I/sources/subversion-1.12.0/subversion-1.12.0/subversion/bindings/swig/proxy
> > >
> > > -I/sources/subversion-1.12.0/subversion-1.12.0/subversion/bindings/swig/proxy
> > > -I/usr/include/apr-1 -I/usr/include/apr-1 -I/usr/include -python
> > > -classic
> > > -o subversion/bindings/swig/python/svn_client.c
> > > ./subversion/bindings/swig/svn_client.i
> > > Deprecated command line option: -classic. This option is no longer
> > > supported.
> > > make: *** [build-outputs.mk:292:
> > > subversion/bindings/swig/python/svn_client.c] Error 1
> > >
> > > This seems to be because SWIG removed the "-classic" flag from their CLI
> > > options. This was originally introduced to provide support for Python-2.2
> > > bindings, and since we're on Python-2.7, it no longer affects us. However,
> > > because of the removal of this switch, the Python bindings for SWIG will
> > > no
> > > longer build properly.
> >
> > Will Python bindings work if -classic is not removed?
> >
> > > Originally, I worked around this by using "sed -i /-classic/d"
> > > build-outputs.mk ,
> >
> > Shouldn't that be:
> >
> > sed -i "s/-classic//" build-outputs.mk
> >
> > however when validating it for the book before
> > > introducing it, I noticed it failed building the ruby bindings. I'm not
> > > sure
> > > if it's related or not, but I'll probably downgrade my system to SWIG-3.x
> > > tomorrow and press on, unless you guys think I should go ahead and
> > > investigate it further. Can someone confirm/reproduce this? It might be
> > > worth an upstream bug report, as I looked there and couldn't find any
> > > changes related to it (last change to Subversion's SVN repo in the related
> > > folder was 2 weeks ago, which was the APR patch).
> >
> > We do need to get this straightened out, but it can wait a bit. Better for
> > you to speed time on gnome and systemd right now.
> >
> >
> > > It's subversion-1.12.0 and SWIG-4.0.0 that seems to be the magical
> > > combination for problems here.
> >
> > I did build both of those, but I did swig4 after subversion-1.12.
> >
>
> Hmmm, I've never seen a use for subversion python binding... Perl bindings can
> be used by git for the "git svn" commands, but nothing uses python nor ruby
> bindings.
>
> So this should be low on our to-do list.
>
> Pierre
>
'./autogen.sh --release' fixes this issue.
BTW, has anyone considered building python bindings with python3?
i.e.
make swig-py \
swig_pydir=/usr/lib/python3.7/site-packages/libsvn \
swig_pydir_extra=/usr/lib/python3.7/site-packages/svn
make install-swig-py swig_pydir=/usr/lib/python3.7/site-packages/libsvn \
swig_pydir_extra=/usr/lib/python3.7/site-packages/svn
Regards,
Wayne.
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
