[Debichem-devel] libodsstream packaging (Was: Bug#879556: ITP: beads -- 2-DE electrophoresis gel image spot detection)

[Olivier Langella]

Hi Andreas,

> 
> Yes, please.  The '-1' revision is simply wrong for an upstream tarball.
> It does not conflict with the Debian version - so if upstream would
> choose a versioning system containing '-' signs it is fine but upstream
> should not try to mimik the Debian revision.  The watch file keeps on
> reporting 0.4.10-1 - it seems not really be fixed at the download
> location.
ok, it is now online on :
https://sourcesup.renater.fr/frs/download.php/latestfile/963/libodsstream-0.4.11.tar.gz

I've modified the watch file, how does it works ?
>   
>>> dh_auto_build: cd obj-x86_64-linux-gnu && make -j1 returned exit code 2
>>> debian/rules:9: recipe for target 'build' failed
>>> make: *** [build] Error 2
>>> dpkg-buildpackage: error: debian/rules build subprocess returned exit status 2
>>>
>>> Hope this helps for a first shot.
>>
>> I see that there is already a lot of work.
>> Thanks a lot, I'm still struggling to get a correct pbuilder environment to
>> be able to compile it with sid (I'm working in stretch).
> 
> I'd recommend cowbuilder (which is a bit faster than pbuilder since it
> is using copy-on-write).  You can easily do it on stretch.  May be the
> Debian wiki helps:
>     https://wiki.debian.org/cowbuilder
Ok, thanks for the tip on cowbuilder.

> 
>> If I can solve it, I will push some fix.
> 
> Just let us know here if you keep on having trouble.
it seems that the dh_auto_build problem is fixed. from what I understood 
of the debian helper automatic mecanisms, overriding is not easy because 
one must keep the same logic (creating and naming temporary 
directories...), but it is possible to only inject cmake arguments to 
the configuration step or inject make targets in the build step.
I've also seen that the man page I've left in the package was the QuaZip 
one. I left it here in order to make my own but I forgot it. It's fixed 
now, but I guess I am yet far from the debian quality requirements.


> 
>>> BTW, you might like to drop the Qt4 package since Qt4 will be droped
>>> from Debian in this release cycle (if I'm not missleaded).
>>
>> Yes, this is planned. I've just finished the Qt5 port of my softwares, and I
>> have also a working version of beads in Qt5, so it is possible to drop it...
>> I had the hope that perhaps some people could be interested in libodsstream
>> : it can read and write ODS files using a kind of "SAX" approach. As many
>> people still use Qt4, I thought it might be usefull.
> 
> Feel free to keep Qt4 in your upstream source - but I think we should
> leave it out in the Debian packaging.

Ok, we can leave it in debian packaging, no problem.

kind regards,
Olivier

-- 
Olivier Langella
http://pappso.inra.fr/bioinfo
Génétique Quantitative et Évolution - Le Moulon
INRA - Université Paris-Sud - CNRS - AgroParisTech
Ferme du Moulon
91190 Gif-sur-Yvette, France
Tel : 01.69.33.23.68  Fax : 01.69.33.23.40
JID : olivier.langella at u-psud.fr