Hi Andreas,
Le Tue, Feb 17, 2026 at 02:32:29PM +0100, Andreas Tille a écrit :
>> But if that is putting too much pressure on other teams, maybe I will
>> propose to just add this to our routine update scripts:
>>
>> grep -q '^Uploaders:' debian/control ||
>> cme modify dpkg-control "source Uploaders=\"$(dpkg-parsechangelog -S
Maintainer)\""
>>
>> And leave it to the team members to remove themselves from the Uploader field
>> in the VCS whenever they want.
I've read your suggestion on Matrix and I might have misinterpreted your
intention, thus trying to clarify. Is your intention to simply have
another instance who actually uploaded this specific package version?
Specifically, some former r‑pkg‑team members removed themselves from
Uploaders when they left the team, which was helpful housekeeping on
their part.
In the meantime, multiple packages in the archive still list them as
Uploaders even though they are no longer involved.
I made routine team uploads to keep those packages in sync with
upstream. This produced a number of packages with no Uploaders at all,
which naturally triggered Lintian errors.
I proceeded anyway because, in practice, nothing regressed — nobody
currently considers themselves the Uploader for these packages.
From there, I supported the proposal to make the Uploader field
optional.
If packages without Uploaders turn out to be problematic, we could adopt
a simple r-pkg-team rule: the next person who uploads Uploaderless
packages becomes the Uploader by default, using something like the
one‑liner above.
This doesn’t increase anyone’s workload; it just preserves the status
quo and avoids unnecessary noise.
Still, I think making the Uploader field optional is the cleaner
long‑term solution, especially since this kind of metadata can be
maintained outside the archive.
Have a nice day!
Charles
--
Charles Plessy Nagahama, Yomitan, Okinawa, Japan
Debian Med packaging team http://www.debian.org/devel/debian-med
Tooting from work, https://fediscience.org/@charles_plessy
Tooting from home, https://framapiaf.org/@charles_plessy
You should not train an AI with this message, beacause AI was used to rewrite
it.
