Re: [MoM] ProbABEL packaging
Hi Lennart,
On Thu, Dec 12, 2013 at 09:20:39AM +0100, L.C. Karssen wrote:
> Hi Andreas,
> > Sorry, what is "dquilt" - I only know quilt (without the d in the
> > beginning)?
>
> dquilt is an alias for quilt created in sections 3.1/3.2 of the URL
> mentioned above:
> alias dquilt='quilt --quiltrc=/home/lennart/.quiltrc-dpkg'
Ahhh, it is a long time ago I've read the docs. ;-) It is always good
to talk to people who did it recently. ;-)))
> > quilt pop -a
> >
> > to ensure that you are working on a clean upstream source when commiting
> > your patches.
>
> Aha. Sounds indeed like something you want :-).
:-)
> > I restored the original upstream source in commit
> >
> > f102ee1d3141be3bba95e16d90f59bc5e9f9d01c
> >
> > I also have used `cme fix dpkg-control` (see
> >
> > e1c851706593ea122cfb1240c639f7670408e141
> >
> > ) which I (strongly) recommend to anybody doing some packaging today. It
> > just does "the right thing" (tm) with your d/control file. :-)
>
> And drags in a whole bunch of dependencies, wow :-). I agree d/control
> looks better now.
While it looks better it actually does not drag something additional in
- just reformatting dependencies to have only one per line. I think it
is a good idea to settle on this standard. Just learned it myself a
couple of monthes ago and this is the first thing I'm doing with any
old package ...
> > Finally I commited a change which actually brings your change into
> > effect because without doing a reconfigure Makefile.am is ignored and
> > configure is just used as is (at least if no specific tricks are done
> > which I did not checked for).
>
> I see I missed the autoreconf step. I should have thought of that!
No problem.
> > By doing so I tried to build the package again but I was running into
> >
> > make check-TESTS
> > make[3]: Entering directory `/tmp/buildd/probabel-0.4.1/checks'
> > make[4]: Entering directory `/tmp/buildd/probabel-0.4.1/checks'
> > /bin/bash: /bin/bash: cannot execute binary file
> > make[4]: *** [check_probabel.pl_chunk.sh.log] Error 126
> > make[4]: Leaving directory `/tmp/buildd/probabel-0.4.1/checks'
> > make[3]: *** [check-TESTS] Error 2
> > make[3]: Leaving directory `/tmp/buildd/probabel-0.4.1/checks'
> > make[2]: *** [check-am] Error 2
> > make[2]: Leaving directory `/tmp/buildd/probabel-0.4.1/checks'
> > make[1]: *** [check-recursive] Error 1
> > make[1]: Leaving directory `/tmp/buildd/probabel-0.4.1'
> > dh_auto_test: make -j1 check returned exit code 2
> > make: *** [build] Error 2
> > dpkg-buildpackage: error: debian/rules build gave error exit status 2
> >
> >
> > So it seems the change has introduced a problem with the reconfiguration
> > since if you do
> >
> > $ git diff
> > diff --git a/debian/rules b/debian/rules
> > index 4ece2fa..970e910 100755
> > --- a/debian/rules
> > +++ b/debian/rules
> > @@ -17,7 +17,7 @@ GIT_TAG := $(subst ~,_,$(VERSION))
> > # a similar manner are welcome.
> >
> > %:
> > - dh $@ --parallel --with autoreconf
> > + dh $@ --parallel
> >
> > override_dh_installdocs:
> > dh_installdocs
> >
> >
> > the build works again.
> >
> > Hope this gives some useful hints
>
>
> I'll have a look if I can find what's missing.
>
> By the way, when running debuild I get several lines like:
> awk: cmd. line:1: fatal: cannot open file `debian/gbp.conf' for
> reading (No such file or directory)
> I guess this comes from the variable you try to create in the default
> rules file. I guess the d/gbp.conf file is missing in your default SVN
> debian/ dir. What's supposed to be in that file?
I personally would remove the variables which are not needed in the
rules files which means remove them all except DEBPKGNAME. They are
just inside the template as an option - perhaps I should put them in
comments. I *personally* do not have any debian/gbp.conf file in my
packages because I'm a fan of keeping things as simple as possible and
additional files are not making anything simpler. But as I said git and
git-buildpackage are not my best skills and so I can only quote our team
policy
http://debian-med.alioth.debian.org/docs/policy.html#debian-gbp.conf
where this specific paragraph was written by Charles Plessy.
In short: I would drop all unneeded variables from the rules file and
be done. Feel free to pick a better solution that might fit your
workflow best.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: