[Debian-med-packaging] Bug#988643: [RFS] vtk-dicom/0.8.12-3

[Étienne Mollier]

Hi Andreas,

Andreas Beckmann, on 2021-05-22:
> On 22/05/2021 12.03, Étienne Mollier wrote:
> > consecutive to changes in multi-arch hints I had to bring per
> > lintian regression, I attempted to install concurrently the
> > package libvtk-dicom-dev both in native and foreign
> > architecture, and it would now fail with the following:
> 
> I'd leave out the M-A: same, that is usually not considered appropriate
> during the freeze. Especially if it does not lead to a M-A co-installable
> package.

Thanks for your advice!  This is not something I found in the
freeze policy, nor the FAQ, but given the effects, I must agree
with you.  I reverted my changes and will go for the removal of
the libvtkDICOMPython*.so symlink from any vtk-dicom packages
for the moment, since there haven't been much breakage with the
dangling symlink anyway.  It may not be entirely ideal, but
there is no risk of regression this way.

> This should be fixed for bookworm.
> 
> I find it a bit curious that a -dev package depends on a python module, but
> then again we seem to be dealing with something closely related to python
> here.

It's a bit curious for me too; double checking the source code
and shared objects relations, I believe libvtkDICOMPython*.so
only provides primitives for the Python module, which are not
relevant for the C++ VTK DICOM libraries, so probably not for
headers as well.

Have a nice day,  :)
-- 
Étienne Mollier <etienne.mollier at mailoo.org>
Fingerprint:  8f91 b227 c7d6 f2b1 948c  8236 793c f67e 8f0d 11da
Sent from /dev/pts/2, please excuse my verbosity.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/debian-med-packaging/attachments/20210522/0d377bd0/attachment.sig>