"Ian" == Ian Jackson <ijackson@chiark.greenend.org.uk> writes:
Hi,
For reference, the resolution in #1007717 was:
"""
Therefore, using its powers
under constitution 6.1.5, the Technical Committee issues the following advice:
1. It is not a bug of any severity for a package with a non-native
version number to use a native source package format.
2. Thus, we think that dpkg shouldn't issue warnings, or otherwise
complain, when a non-native version number is used w/ 3.0 (native).
3. We suggest that the wontfix tag on #737634 be reconsidered.
4. We believe that there are indeed circumstances in which
1.0-with-diff is the best choice for a particular source package,
including, but not limited to, git-first packaging workflows.
However, we recommend discontinuing use of 1.0-with-diff in other
circumstances, to simplify the contents of the archive.
This is because there is currently no other source format which is
such that avoid both (i) uploading the whole source, including
upstream, for every upload; and (ii) having to maintain
debian/patches/ inside the package tree.
5. We decline to comment on the recent source package format MBF.
"""
As Ian points out, dpkg-source was changed to reject 3.0 quilt with native version and 3.0 native with non-native version in dpkg (in response to #700177) on the basis that it was considered to be a mistake to have either of those combinations. In discussion on the resulting #737634 Guillem said that he thought it was a failing in dpkg that 1.0 packages were sloppy in enforcing native vs non-native version numbers; he later added that he thought that such use of version numbers was contrary to policy[0].
The TC was not asked to consider over-ruling the dpkg maintainer in its consideration of #1007717; the question of whether 3.0 native should allow non-native version numbers arose during discussion.
I think Ian is right that we could declare as he wishes under 6.1.1 that:
dpkg-source should be able to build "3.0 (native)" source packages
with a non-native version number.
...although if the dpkg maintainers continue to decline to allow a fix of #737634 to implement such a policy I'm not quite sure where it would leave
us constitutionally.
I incline to the view that we should probably make such a declaration, and suggest to the release team that such a fix should be allowed into trixie (though obviously it's up to them).
Disclaimer: I've known Ian for years. I think there is plenty of evidence that this doesn't stop me disagreeing with them ;-)
Regards,
Matthew
[0] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=737634#77
NACK, it's way too late into the freeze. Toolchain freeze (which dpkg
belongs to) started on 2025-03-15.
I think Ian is right that we could declare as he wishes under 6.1.1 that:
dpkg-source should be able to build "3.0 (native)" source packages
with a non-native version number.
...although if the dpkg maintainers continue to decline to allow a fix
of #737634 to implement such a policy I'm not quite sure where it would
leave us constitutionally.
I incline to the view that we should probably make such a declaration,
and suggest to the release team that such a fix should be allowed into
trixie (though obviously it's up to them).
Sebastian Ramacher writes ("Re: Bug#1106402: dpkg-source, native source package format with non-native version"):
NACK, it's way too late into the freeze. Toolchain freeze (which dpkg belongs to) started on 2025-03-15.
Of course the final decision is with the Release Team, but looking at
the freeze policy I see that "small, targeted fixes" are acceptable
and also that changes to toolchain packages require pre-approval.
Dimitri argues that the native package format is useful even for...
non-native packages
If I have mispresented or omitted important facts, please do not
hesitate to correct me.
Guillem argues in the original bug that the versioning scheme is an important part of the distinction between native and non-native packages
"Timo" == Timo Röhling <roehling@debian.org> writes:
My reading of the TC's decision is that the TC supported the claim that
a native package with a non-native version number was not a bug--thus
not a policy violation.
Like Ian, I do not wish to see the underlying technical issue
relitigated.
However, I doubt the TC will want to try to clear a path through the nontechnical obstacles between this improvement and trixie. So, to be
clear, I'm withdrawing my request for the TC to authorise an NMU,
or to express any opinion to the Release Team.
I'd like to take over this request then. We have also the policy
proposal of #1049406 open, which will make secure boot package building impossible in Debian.
I still get the feeling that any exercise of governance or oversightSpeaking with my TC hat on: Yes, I believe they were absolutely
is Debian is widely regarded as cataclysmic. I have been criticised
here and in #1107137, as if my requests were unreasonable or obviously
beyond the pale. This is very uncomfortable for me, and this kind of
thing is hardly going to encourage others to try to resolve
longstanding problems using our official processes.
So I would appreciate at the very least some explicit recognition that
my requests to to the TC, in this bug, were appropriate.
It appears that dpkg is now going to be fixed, albeit with a bunch of >unpleasantness in the documentation, and only for forky.
[...]
I still think it would have been good to have this fixed in trixie.
The original one-line revert would have been very low risk from a
purely technical point of view.
Guillem, would you be willing to backport the change to trixie or
consent to an NMU?
* Ian Jackson <ijackson@chiark.greenend.org.uk> [2025-06-03 16:36]:
It appears that dpkg is now going to be fixed, albeit with a bunch of unpleasantness in the documentation, and only for forky.
[...]
I still think it would have been good to have this fixed in trixie.
The original one-line revert would have been very low risk from a
purely technical point of view.
Release Team, would you be willing to reconsider your position on a trixie fix, given that Guillem agreed to implement the change? I concur with Ian's risk assessment, and I believe your main objection was the nonconsensual nature of the upload.
Guillem, would you be willing to backport the change to trixie or consent to an NMU?
Cheers
Timo
--
⢀⣴⠾⠻⢶⣦⠀ ╭────────────────────────────────────────────────────╮
⣾⠁⢠⠒⠀⣿⡁ │ Timo Röhling │
⢿⡄⠘⠷⠚⠋⠀ │ 9B03 EBB9 8300 DF97 C2B1 23BF CC8C 6BDD 1403 F4CA │
⠈⠳⣄⠀⠀⠀⠀ ╰────────────────────────────────────────────────────╯
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 491 |
Nodes: | 16 (2 / 14) |
Uptime: | 120:38:27 |
Calls: | 9,687 |
Calls today: | 3 |
Files: | 13,728 |
Messages: | 6,176,524 |