[Debian-med-packaging] Bug#1104044: Bug#1104044: orthanc-wsi: Plugin OrthancWSI fails to start on arm64, ppc64el, riscv64 and s390x

[Étienne Mollier]

Hi all,

Chris Hofstaedtler, on 2025-05-05:
> On Mon, May 05, 2025 at 09:56:41AM +0300, Sébastien Jodogne wrote:
> > It looks like there is a deadlock between orthanc and orthanc-wsi that
> > still prevents both packages to migrate to testing.
> 
> Looking at https://qa.debian.org/excuses.php?package=orthanc and
> https://qa.debian.org/excuses.php?package=orthanc-wsi it seems to me 
> orthanc should have added Breaks: orthanc-wsi (<< 
> 3.2+dfsg-something).
> 
> [quoting for context]
> > I think that this might be a consequence of the dependency of orthanc-wsi
> > on "liborthancframework-dev (>= 1.12.7+dfsg-2)". Couldn't this explicit
> > requirement on the version be removed, given that the dependency on
> > "libdcmtk-dev (>= 3.6.9-5)" implies a proper version of
> > "liborthancframework-dev"?
> 
> In case the test failure found by autopkgtest is a _test_ failure 
> only and cannot be observed by users, the release team could hint 
> both packages together into testing.  CC'ing them for awareness.
> 
> Sébastien, Étienne, maybe you can reply here if you think the 
> autopkgtest is indeed a test failure only, or if users can see it if 
> they do a partial upgrade.

If we're speaking this regression[1], then I _believe_ this
could be a user visible issue, assuming they would restart their
orthanc service on a half upgraded Debian testing machine.  That
being said, affected package versions do not involve stable
releases.

[1]: https://ci.debian.net/packages/o/orthanc-wsi/testing/amd64/60380965/

I see that the regression does not appear anymore on the excuse
tracker, so I guess there could have been an intervention from
the release team already.  If still deemed necessary, I believe
that the appropriate point in time to introduce the Breaks would
be against orthanc-wsi (>= 3.2+dfsg-2), although at this point
-7 might probably do as well.  I'm holding my horses though,
since I also understood too tight specifications might give
difficulties to the dependencies resolver.

I'm sorry I had difficulties to keep track of the situation this
time.  Thank you for your help with finishing this migration!

Have a nice day,  :)
-- 
  .''`.  Étienne Mollier <emollier at debian.org>
 : :' :  pgp: 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/20250505/4a27d6c5/attachment-0001.sig>