Hi, On 22-12-2024 19:08, Richard Lewis wrote:
Currently the page has several that relate to bookworm (and some before): should anything relating to bookworm or earlier be closed? is it still desirable/possible to amend the bookworm release-notes?
bookworm release-notes can be updated until bookworm is no longer supported by the Release-Team and security support. So, only bugs only applicable for bullseye and before should be closed at this stage.
these bugs look like valid things which could have been in bookworm release-notes: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1037144 - Please add bug #1037142 in the "5.4 Known severe bugs" section for bookworm the Debian - that bug is https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1037142 which is now closed - is some issue specific to a particular graphics card - if bookworm wasnt released it would be worth doing
I generated the list of "Known severe bugs" around the day of the release. Was that bug already open when we released? We can add it, but my intention of that list was mostly to make it clear that every release has bugs, and even known ones.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1038151 - document rename of the "ssh" group to "_ssh" in bookworm as people might have relied on 'ssh' eg with 'AllowGroups to "root ssh"' - if bookworm wasnt released it would be worth doing
I missed this one, because it was reassigned without CC and it was after the release. Indeed, might be worth still doing.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1037116 - says that debian-edu bookworm was not ready -- this is, i assume, valid, but anyone affected has presumably worked close? - if bookworm wasnt released it would be worth doing
I'm sorry that I missed this one too. It's not relevant anymore.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1037203 - aide is broken by the upgrade. The proposed text says what the workaround is - if bookworm wasnt released it would be worth doing
I didn't check, but this is fixed by now, right? Then it's no longer worth changing the release notes I'd say.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1037430 - exim upgrade now fails if tainted data is used - if bookworm wasnt released it would be worth doing
Might still be worth mentioning.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1055027 - links to https://salsa.debian.org/ddp-team/release-notes/-/merge_requests/201 - if bookworm wasnt released it would be worth doing
My questions never got answered, hence it fell through the cracks. In my opinion, this should just be fixed if it can instead of being documented.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1030119 is tagged trixie, but is about the annoying 'pam_env(sshd:session): deprecated reading of user environment enabled' messages which you get in bookworm. However the ssh maintainer has fixed this for trixie: no need to document that in trixie, so suggest closing this?
I think it can be closed (I didn't read the full bug report), but Colin mentioned a MR at the end and that one was merged.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1012174 -- is about mentioning https in apt lines: I think this is fixed, so close?
In both places? Otherwise a reassign is the best option.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1035065 - states that "the sssd cache becomes invalid on the upgrade to Bookworm due to a new format", - no new info provided on what this means since request in May 2023 - suggest closing as it doenst seem to have been an issue in practice and there is nothing actionable in it
We could do that, but in the past I kept these bugs open until the release they were ment for was no longer supported.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1036876 - states that "debian was previously able to boot from lvm volume groups that were not complete while running in the initramfs and that bookworm will no longer support this" - porbably valid, but not a usual setup -- close?
If there was an agreed proposal for text, I would add it even now.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1037121 - The salt-stack was removed from bookworm. there is proposed text, and i assume valid, but im not sure it's high-profile enough to warrent inclusion now?
If no objections to the text, we could just add it.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1050833 - changes to network naming: it is not clear to me if this was actually a change in bookworm or an old change. i think this is the user not reading previous release-notes correctly - close?
I don't have the energy now to read carefully enough to see if I agree with you.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1070258 - user did not follow the instrcutions to remove non-debian software. close?
Not sure. We already document that people should clean up before upgrading. Maybe we can add a section with a comment on things like pip, cargo, etc.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=987068 - says cgroup2 support is incomplete for containers in bullseye. - no response from submitter since 2021. close?
bullseye is no longer supported, so indeed the bug can be closed.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=987069 - another cgroup2-related thing from 2021. last message says it's too long for release-notes. close?
Same.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=991809 - says new instalations of nis should follow a nis-howto document, but upgrades work - no response to "what do users upgrading need to know", seems out of scope for release-notes. close?
Agree.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1036907 - something to do with dash. unclear to me... doesnt seem to have been an issue. close?
Actually this might be still valid to document.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=927679 - is about docbook syntax. release-notes now use markdown. close?
Absolutely. Except if the markdown setup also has something like "entities".
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=930318 - unclear - i think they are suggesting to rename i386 to "PC". i dont think that is actually any clearer, "PC" would have made sense in the 1990s, but i dont think people really use it any more. close?
We should have the discussion in the bug instead of here and come to a conclusion in the bug.
PaulPS: will you forward this discussion to the bugs? It's a bit weird to discuss bugs but not in the bug report.
Attachment:
OpenPGP_signature.asc
Description: OpenPGP digital signature