Bits from the (Debian) Release Team: frozen hard to get hot

[Moderator Brian Masinick]

Debian mailing list announcement:

Hi all,

According to schedule, we froze bullseye a bit more last week
(12-03-2021). This means that we are one step closer to the release of
bullseye and we expect everyone to follow the freeze policy [1]. This
means that from now on key packages and packages without significant
autopkgtest coverage need to be unblocked by the release team to be able
to migrate from unstable to testing. If you need to request an unblock,
check that your request is in line with the freeze policy and use
$(reportbug release.debian.org) in order to get the meta data correct
and get the template that helps us get the right information.

Note that packages that are not on the key-packages list, with
non-superficial autopkgtests that succeed on all architectures and don’t
cause regressions will migrate without an unblock after 20 days. If
you’re unsure about the current migration status of your package, you
can check it at [3].

Now is a good time to help testing installations and upgrades to find as
many issues as possible and fix rc bugs [4]. Please be aware of bug
984533 (libc6), 974552 (libcrypt1) and 953562 (libcrypt1) when you
do. Add information to those bug if you can help them forward.

Please also remember to file bugs against the release-notes pseudo
package if you know of issues that deserve being mentioned in the
release notes. Ideally with proposed text, or even via a merge request
on salsa [2], but even if you can only point us at issues that’s already
appreciated.

On behalf of the Release Team,
Paul

[1] https://release.debian.org/bullseye/freeze_policy.html#hard
[2] https://salsa.debian.org/ddp-team/release-notes/
[3] https://qa.debian.org/excuses.php
[4] https://deb.li/rcbugs

--
Brian Masinick

[Moderator Brian Masinick]

Helpful

Up

0

::

This means that (in theory) the next Debian release is approaching. Except for defect resolution for release blockers or other significant defects, the next release is effectively frozen.

For most Linux distributions, this would mean a release within a few months. Debian, over the past couple of releases, has been improving their schedule. It’s POSSIBLE that Debian Bullseye could be ready by the Summer. If we want to release, I’d say that releasing on Bullseye at this point, formal release or not, is a reasonable choice because no additional functionality will be added to Bullseye; any further changes are for stability and defect resolution.

For my simple use cases, Bullseye has been solid for several months, actually more “stable” than quite a few proprietary commercial systems and even a few freely available systems. This is merely my opinion, nothing more.

--
Brian Masinick

[Member ex_Koo]

Helpful

Up

0

::

Brian Masinick wrote:

For my simple use cases, Bullseye has been solid for several months, actually more “stable” than quite a few proprietary commercial systems and even a few freely available systems. This is merely my opinion, nothing more.

I totally agree Brian it is a pleasure to use. Had firmware updates tonight which is a good sign too, along with 5.10.22 kernel the future is looking bright indeed.

[Moderator Brian Masinick]

Helpful

Up

0

::

Our antiX 21 implementation of Bullseye has been solid too; I am looking forward to our antiX 21 runit release in particular!

--
Brian Masinick

[Anonymous]

Helpful

Up

0

::

Debian Bullseye could be ready by the Summer.
[..]
releasing on Bullseye at this point, formal release or not, is a reasonable choice because

asking for clarification:
“this point” ~= “the Summer”
or
“this point” ~= “20 March, 2020”
???

Any date prior to the official Debian release, and switchover of the named (“testing”, “stable”) repositories — with time to allow propagation to the various mirrors — seems ill-advised. Even if the sources.d lists explicitly state “buster”, vs “stable”, across versions we have learned to expect many users will be “caught off-guard” and will experience (confusion and) breakage.

Our antiX 21 [alpha] implementation of Bullseye has been solid

Betatesting so far has been too limited (compared to prior antiX version releases)
and, months later, some of the reporatage is likely now “stale” (irrelevant//inaccurate).
Also, to-date we’ve had ZERO testing against a 32bit version.

[Member seaken64]

Helpful

Up

0

::

I am testing v21 a1 on my 64-bit laptop using VirtualBox. But I am awaiting the 32-bit version so I can test in on my older machines.

seaken64

[Moderator Brian Masinick]

Helpful

Up

0

::

We had some recent exchanges indicating problems with things like the c library (considered significant). Still, since I have not recently built any code, I didn’t notice any problems in this regard. Clearly this will have to be fixed before either Debian Bullseye or antiX 21 can be formally released. It certainly doesn’t put ALL testing to a halt, even if it does affect some final development. I suspect that the final library changes will be fixed soon, including the encryption libraries and whatever other libraries need improvement. Just how soon they propagate to a location where we can effectively use them is the question that will ultimately impact our final efforts, as well as our platform testing. Those of you who can test the existing builds that have been produced will help the cause for our release too,

--
Brian Masinick

[Author]

Posts