[libvirt] Release of libvirt-1.0.5
Daniel Veillard
veillard at redhat.com
Thu May 2 14:24:57 UTC 2013
On Thu, May 02, 2013 at 03:14:32PM +0100, Daniel P. Berrange wrote:
> On Thu, May 02, 2013 at 10:08:11PM +0800, Daniel Veillard wrote:
> > On Thu, May 02, 2013 at 03:35:50PM +0200, Guido Günther wrote:
> > > On Thu, May 02, 2013 at 06:39:07PM +0800, Daniel Veillard wrote:
> > > > I forgot to make the rc2 release on tuesday, but looking at the
> > > > git state it was looking like the problem reported on rc1 had been
> > > > fixed, so I decided to go ahead and roll the release as scheduled
> > > > today. It is tagged in git and the tarball is available as usual at:
> > > >
> > > > ftp://libvirt.org/libvirt/
> > > >
> > >
> > > The release tarball lacks python/sanitycheck.py (rc1) and I couldn't
> > > figure out why. Make dist in git adds it to the tarball. So this looks
> > > like a problem with tarball generation on your machine? This breaks
> > > running "make check".
> >
> > python/sanitycheck.py is indeed not in the tarball, no idea why
> >
> > thinkpad:~/libvirt -> grep EXTRA_DIST python/Makefile.am
> > EXTRA_DIST = \
> > EXTRA_DIST += sanitytest.py
> > thinkpad:~/libvirt ->
> >
> > make check which is part of 'make rpm' internal %check passed
> > and the python rpms were generated. So i'm not sure the issue is
> > really that big.
>
> Could your initial 'make dist' have been done in a libvirt checkout
> which did not have python enabled. That's the only way I see it could
> have not got into the dist. We should fix the python Makefile.am to
> set EXTRA_DIST outside the if HAVE_PTYHON conditional anyway.
Well I came to the same conclusion but could not understand why,
I do have python and all the devel stuff on that box, it built the
python rpm, etc... fine, i have really no warning.
I may have tweaked something at some point as i was porting all the
python scripts to python3 for libxml2 last month, but the setup on my
Fedora17 machine is really based on python2, i'm a bit puzzled...
Agreed the EXTRA_DIST += sanitytest.py ought to go outside the
HAVE_PYTHON conditional, I would even argue this should go to
the initial set of files.
Daniel
--
Daniel Veillard | Open Source and Standards, Red Hat
veillard at redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/
http://veillard.com/ | virtualization library http://libvirt.org/
More information about the libvir-list
mailing list