Re: Doubt on debian/changelog version history
On Sun, Apr 25, 2021 at 09:37:49AM +0200, José Luis Blanco-Claraco wrote:
> Hi dear mentors!
>
> I'm the upstream author and Debian maintainer of a package, and after
> uploading a couple of versions to different releases (unstable and
> experimental), now I'm unsure about what should be the *order* of the
> different changelog entries in a forthcoming release to experimental,
> I'm sure you'll be able to quickly solve my doubt.
>
> Here is the order of events (from [1]):
>
> [2021-01-03] Accepted mrpt 1:2.1.7-1 (source) into unstable
> (New release, unstacle is frozen, so upload to experimental)
> [2021-03-31] Accepted mrpt 1:2.2.0-1 (source amd64 all) into experimental
> (Then we had to upload a patch to solve a serious bug in 2.1.7-1):
> [2021-04-13] Accepted mrpt 1:2.1.7-2 (source) into unstable
> Next, I want to release 2.3.0, into (I guess) experimental.
>
> If I reflect the order of events above in debian/changelog, it would read:
> mrpt (1:2.3.0-1) experimental; urgency=medium
> ...
> mrpt (1:2.1.7-2) unstable; urgency=medium
> ...
> mrpt (1:2.2.0-1) experimental; urgency=medium
> ...
> mrpt (1:2.1.8-1) unstable; urgency=medium
(I dont see that version on tacker… typo?)
> ...
> mrpt (1:2.1.7-1) unstable; urgency=medium
> ...
> mrpt (1:2.1.6-1) unstable; urgency=medium
> ...
>
> Is it ok to have non-consecutive versions in the changelog? If not:
> what's the correct way to handle it?
>
> Extra question: How is a port from experimental to unstable supposed
> to happen? Making a new upload with a new changelog entry?
>
> Thanks!
>
> [1] https://tracker.debian.org/pkg/mrpt
My bikeshed color on this would be not to order by date, but how the relations
of the versions are; likw in anchestor/successor.
I mean, for example 2.2.0 never saw sid. So 2.1.7-2 is not based on 2.2.0, and following
the path, 2.3.0-1 is not based on 2.1.7-2, but on 2.2.0-1
IOW, I would order them this way:*
1:2.3.0-1
1:2.2.0-1
1:2.1.7-2
1:2.1.7-1
1:2.1.6-1
* There'd be a corner case when 2.3.0 is not based on 2.2.0, but thats unlikely (I didnt check)
Reply to: