On 24/01/26 12:30, Rene Engelhard wrote:
Am 23.01.26 um 14:58 schrieb Gioele Barabucci:
Example of a simple `d/s/support` file that I would use for my own
packages:
```
debian/stable
debian/unstable
```
This doesn't make sense.
Assume I put debian/stable into it. Because I only care about backports
to the last stable.
Now that one becomes stable itself. Does it need to be changed to
debian/oldstable?
I don't understand your question.
`debian/stable` is a time-relative label. It refers to whatever is
stable at the time of reading that file. If I download today a source
package and read that it wants to keep compatibility with
`debian/stable`, then it is referring to Debian 13 trixie. In two years
that will be Debian 14 forky.
If the maintainer is interested in keeping compatibility with a specific
release, then they should not use `debian/stable` and then change it to
`debian/oldstable`, but rather use `debian/13-trixie` from the get go.
oldoldstable is by definition another developer.
Why is oldoldstable by definition another developer?
I have collaborated on packages that targeted old^3stable. Sure, the
backported packages where not uploaded to the archive, but the
maintainer team still had an interest in keeping (build) compatibility
with the previous three releases.
If that one is given to LTS one needs *another* upload to update this
field to the LTS people?
Yes, in a certain sense. But if one uses time-relative labels, then that
has to be one once and from that moment it will be valid in that form
without changes.
That said, I'm starting to think, also after discussing this with
others, that recording contact information will probably create more
problems than it would solve, so if this proposal ever gets turned into
a DEP, I will probably skip the part about the contact information.
I think the most open question is why you want to add another field
which will quickly get out of date, only fixable by (o-)o-p-u and put
more burden on maintainers for (almost) no gain.,
I think that there is a misunderstanding here. I hope my answers to your
remarks made things clearer.
Regards,
--
Gioele Barabucci
