Re: Bug#383349: ITP: nifticlib -- IO libraries for the NIfTI-1 data format
Hi,
Thanks for looking at the package.
On Thu, Aug 17, 2006 at 07:53:07AM +0200, Andreas Tille wrote:
> On Wed, 16 Aug 2006, Michael Hanke wrote:
>
> >Grmpf, I knew something was missing...
>
> No problem.
>
> >It's at mentors.
> >
> >http://mentors.debian.net/cgi-bin/maintainer-packages?action=details;package=nifticlib
>
> OK.
>
> Some remarks:
>
> 1. You "Added a lintian override concerning a single CVS directory in the
> upstream sources." This is an appropriate method to get rid of the
> lintian warning. I personally would prefer to repack the upstream
> tarball (and advise upstream how to release source tarballs).
You're right. I will do so.
>
> 2. In debian/compat you can use 5 as debhelper version. It does not
> change anything (except that you will have to increase the debhelper
> version in debain/control as well) but it would be reasonable to start
> with the latest available compatibility version once you start a
> new package.
The reason for this is backporting. I do not really not debhelper 5 and
as it is not part of sarge I sticked to debhelper 4.
>
> 3. I have no ral experience in packaging dynamic libraries. That's why
> I was happy about d-shlibs which takes care for implementing the
> policy correctly. If I where you I would give d-shlibmove a try.
Thanks for the pointer.
>
> 4. I wonder what might be the sense of your debian/pycompat file.
I have a python interface for the libraries (almost) ready, that I
originally wanted to include in the package. But now I think it might be
better to keep it separate from the upstream parts. This file is
obviously a left over. Removed.
>
> 5. Lintian keeps on shouting
> W: libnifti0: package-name-doesnt-match-sonames libfslio0 libniftiio0
> libznz0
> because you did not moved the override file to
> /usr/share/lintian/overrides.
I'm still not sure about this. While znz and niftiiio look pretty stable
it might happen that fslio is undergoing major changes. Perhaps I should
move this lib into a separate binary package.
> All these remarks are more or less cosmetical - no reason to hold back the
> package. I admit that I'm unable to do some reasonable tests whether the
> package works or not.
I will incorporate all necessary modifications and try to provide an
update before the weekend.
Thanks,
Michael
--
GPG key: 1024D/3144BE0F Michael Hanke
http://apsy.gse.uni-magdeburg.de/hanke
ICQ: 48230050
Reply to: