=== JanC_ is now known as JanC [11:17] rbasak: some time ago you said that you would probably enable the fix for deleted directories automatically; when would it be triggered? would it be post-commit? [12:25] I have asked in #ubuntu-bugs, but requesting here also - can anyone please mark LP: #1987073 as affecting Jammy please [12:25] -ubottu:#ubuntu-devel- Launchpad bug 1987073 in linux-show-player (Ubuntu) "Failure to start on 22.04" [Undecided, Confirmed] https://launchpad.net/bugs/1987073 [13:18] sudip: . [13:19] thanks ginggs [13:21] ginggs: also, in your glibc 2.39 snapshot rebuild report I can see many packages failed to build, but there are no bug reports for many of them. is it expected that if someone is working on those ftbfs then they will file the bug first? [13:21] * sudip apologises in advance for the newbie question [13:25] adrien: I don't recall that, sorry. Are you OK using the current workaround for now? [13:34] sudip: i do not know. schopin? ^^ [13:35] sudip: I'll send out a report tomorrow, but we've looked at those failures, and most of them are either unrelated to glibc or solved by a fresher glibc snapshot. [13:36] sudip: if you're impatient you can look at https://bugs.launchpad.net/ubuntu/+source/glibc/+bugs?field.tag=test-rebuild-glibc-noble for those we've identified as glibc-related. [13:37] And it's not a newbie question, it's perfectly legitimate to ask this :) [13:38] yeah, I have seen that bug list with the tag which had the link to ginggs's rebuild report and that led me to my question. so even if they are unrelated, they are still ftbfs, and quite a lot of them. [13:39] sure, but they were already FTBFS in the initial report for most of them, and for the others I believe ravikant_ filed bugs, just without the tag. [13:41] ok, thanks schopin [13:58] rbasak: yeah, definitely, I was just curious and also wondering if I would do such a change on my side too [14:31] can I ask for a review of https://code.launchpad.net/~adrien-n/ubuntu/+source/ruby3.1/+git/ruby3.1/+merge/459015 ? the issue is present in ruby3.0 < 3.0.7 (unreleased), ruby3.1 < 3.1.6 (unreleased), ruby3.2 < 3.2.3 (released 2024/01/18), ruby3.3 < 3.3.1 (unreleased) [14:32] all ruby tests fail due to that AFAIK; it might also make sense to put that in other ruby versions [14:33] just checked: jammy has 3.0 and subsequent versions have 3.1 [15:03] LocutusOfBorg: \o We have current Ubuntu 24.04 Cloud Azure image build failures due to a recent change you uploaded for libtss2-mu-4.0.1-0 which depends on libtss2-mu0 (< 4.0.1-4) but libtss2-mu0 4.0.1-3ubuntu1 is the greatest package version in Ubuntu currently . I don't see anything in the uplaod queue. Is this something you are working on? === pushkarnk1 is now known as pushkarnk [15:27] philroche, I wasn't [15:27] but I am now :D [15:32] Thank you [15:32] just question, what are you trying to install? [15:33] I see 4.0.1-7 in Debian queue, I might merge it [15:33] but I don't know if it will wolve or not [15:33] Attempted install of the following triggered the error: `libtss2-esys-3.0.2-0 libtss2-mu0 libtss2-rc0` [15:33] Error was: `The following packages have unmet dependencies: libtss2-mu-4.0.1-0 : Breaks: libtss2-mu0 (< 4.0.1-4) but 4.0.1-3ubuntu1 is to be installed` [15:34] libtss2-mu0 [15:34] who is installing that libtss2-mu0? [15:34] the package got renamed into libtss2-mu-4.0.1-0 [15:34] so no direct installations please [15:35] reverse-depends -r noble libtss2-mu0 [15:35] shows nothing [15:35] well that helps. Thank you. [15:36] Was there a similar recommendation for libtss2-rc0 ? [15:37] nope [15:37] I only see libtss2-rc0 for libtss2-rc* in noble [15:37] that one didn't get renamed [15:37] cool. [15:37] Affected: .depends ~ /\b(libtss2\-mu\-4\.0\.1\-0|libtss2\-mu0)\b/ [15:37] this is what the transition says [15:38] That's me unblocked. [15:38] https://release.debian.org/transitions/html/auto-tpm2-tss.html [15:38] reason is that upstream broke ABI by removing one symbol used in python package [15:38] so, transition to make sure everything rebuilds [15:38] but please don't use direct packages if possible [15:38] +1 [15:39] not sure if the -dev package can go in seeds, or if you can use some shlibs to detect such changes [15:39] sorry I don't really know that image building process :D [15:39] please ping in case I break something else [15:39] Using livecd-rootfs but the issue was in the closed source Azure binary hook [15:40] ack. Thanks for your help [15:41] so can it be adjousted, or does it need upstream to rebuild manually? [15:41] it can be adjusted - no action in archive required [15:42] nice to know! === rkratky__ is now known as rkratky === pushkarnk1 is now known as pushkarnk === arif-ali_ is now known as arif-ali [20:06] Long shot, but does anyone here have the ability to moderate the ubuntu-users list? A message of mine got stuck in moderation. [20:12] arraybolt3: oh my password still works [20:30] ogayot: kitty is back ;) https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#kitty [20:36] wait... that didn't redir to archive-team did it hmm [20:41] Hi. On the Debian side, we are mitigating the effects of the /usr-move (DEP17) via https://salsa.debian.org/helmutg/dumat. I have run this tool against Ubuntu today and that results in https://subdivi.de/~helmut/ubuntu-20240125.yaml . would there be anyone interested in turning these diagnostics into actions on the Ubuntu side? [20:43] Quite a bit of what is detected there is also detected (and mitigated) in Debian, so that'll naturally trickle into Ubuntu, but there also is a bigger part of findings that are specific to Ubuntu and not present in Debian at all. In particular the amount of empty directory loss and the undeclared file conflicts are not to be found in Debian in that quantity [20:43] Back in may, I had paride perform an early test run of it, but he didn't seem to pick it up ever since. [20:45] helmut: I think this is a great question. Could you please send an email to the ubuntu-devel list? [20:45] In fact, I'm really curious how you set that up for Ubuntu specifically, if there is any magic. [20:45] tsimonq2: I am very happy to walk you through that and fixing bugs as we encounter them. [20:46] tsimonq2: I can also take it to email, yes. [20:47] tsimonq2: hw-wise running dumat requires 1 core, 1gb ram, 4gb disk space, and an unmetered connection to a full mirror. [20:47] I've done the full ubuntu import on my laptop today. [20:48] Perfect, Lubuntu's sandbox server can definitely handle that. [20:48] if this is considered too niche for this channel, I am happy to accomodate ubuntu /usr-move discussions on oftc#debian-usrmerge [20:49] Depends, is the setup fairly easy? (Am I just pointing mmdebstrap towards Ubuntu URIs, or is it more involved?) [20:49] what definitely needs tweaking for ubuntu is a little function in the analysis part called "is_supported" and I simply made it "return True" for the purpose of this report, so the size definitely is exagerated. tweaking this function is key to making the report useful [20:50] tsimonq2: I hope it is easy, but then I wrote it. it does not require root. running it as a dedicated useraccount is recommended. you need user namespaces + package uidmap [20:53] regarding adapting it to ubuntu, the archive is configured via the sqlite database. populating the "release" and "component" tables (which are otherwise treated as read-only) is key. [20:53] you may source my attempt at this configuration from my database dump at https://subdivi.de/~helmut/ubuntu-20240125.sql.zst [20:54] Sounds good, I'll report back with something ASAP. :) [20:54] (Same page re: is_supported, by the way.) [20:55] notably, the analysis is split into two phases: a feature extraction phase where .debs are downloaded analyzed and the results are stored in this database, and the actual analysis phase that only consumes the database and generates the report. in particular, this allows you to tweak the analysis without downloading lots of crap [20:55] the analysis runs in less than a minute on a fast cpu. so you can really iterate that and that is what contains "is_supported" [20:57] you'll notice that dumat will fail to import zstd-compressed .debs. this is fixed in a feature branch of the submodule, but cannot be merged yet, because I cannot require >= bookworm yet. just checkout switch the submodule to the feature branch [20:58] I expect that the feature extraction part (import_mirror.py + import_diversion.py) will just work for ubuntu, but the analysis.py (which also associates reported bugs) will need changes [21:00] before setting up a dumat instance, you might want to check that example yaml given first whether you find the report useful [21:01] maybe some of the problems that I see on the Debian side are considered non-problems for Ubuntu or you consider certain upgrade-paths as unsupported and so on. [21:01] The only improvement I would do right off the bat (I'll send patches ... somewhere ... hopefully :) ) is adding a packageset. [21:01] Plus, Lunar is EOL as of today, so I'll exclude that from my end. [21:02] sure, it can be deleted. I wasn't sure, so I included it [21:03] also it may not be clear to the uninitiated what these short "what: ..." descriptions mean in detail. I am happy to help with that [21:03] "what: undeclared file conflict" should be fairly clear though ;) [21:05] you may wonder why these are reported. undeclared file conflicts typically result in the addition of Replaces and Replaces may be rendered dysfunctional by the /usr-move. thus, missing Replaces may shadow a /usr-move problem [21:06] tsimonq2: any remaining questions for now? [21:07] helmut: Where does the hash come from in the component table? [21:07] insert NULL. it's the hash of the current list from the dists/ folder [21:07] I lied about components being read-only [21:08] okay :) then no, nothing else from me at the moment [21:08] good luck. do you still need a ubuntu-devel mail? [21:08] I'll check back before tomorrow beginning of day with a link for you, I'd like someone from Canonical QA to help me put that somewhere more permanent. I'll likely be writing something easier to populate these two tables. [21:09] yeah, the plumbing isn't the best. dumat hasn't been set up that often (-: [21:09] I'm really curious where you get your code style from ;) [21:10] mostly black + pylint [21:11] ah okay :)( [21:11] *:) [21:11] I am happy to rework it as needed, but thus far this code mainly had only one user. [21:13] tsimonq2: please highlight me for question or ask them over at the debian channel. I won't be following this channel actively. [21:14] helmut: Will do :) thanks [22:19] Is there a way to trigger an autopkgtest to run with a specific kernel? i.e. install the kernel, reboot, and then start the test? Or does that require modifying the test to use the reboot protocol? [22:31] I know there are no `pilots` in atm, but if anyone's got cycles can I get an upload to noble please? The bug's here: https://bugs.launchpad.net/ubuntu/+source/google-guest-agent/+bug/2050956 [22:31] -ubottu:#ubuntu-devel- Launchpad bug 2050956 in google-guest-agent (Ubuntu Noble) "Update d/control file with dependencies" [Undecided, New] [22:40] kajiya: I gotcha, sec [22:41] Cheers tsimonq2 :D [22:43] kajiya: I'll sponsor, could you just throw this ( https://wiki.ubuntu.com/google-guest-agent-Updates ) in the desc like last time? https://bugs.launchpad.net/ubuntu/+source/google-guest-agent/+bug/2040945 [22:43] -ubottu:#ubuntu-devel- Launchpad bug 2040945 in google-guest-agent (Ubuntu Mantic) "[SRU] Update and backport google-guest-agent to 20231004.02" [Undecided, Fix Released] [22:43] :) [22:44] Ah my bad, I'll do that now :D [22:48] All done! [22:48] kajiya: And same :D https://launchpad.net/ubuntu/+source/google-guest-agent/20231004.02-0ubuntu2 [22:48] Sorry, I'm still less-fast with Git-based packaging ;) [22:49] Thank you so much :D And don't worry, you're still faster than me xD [22:50] Happy to help, I'll be floating around if you need help getting this to the release pocket :) [22:51] I'll keep that in mind, thanks again for the help :D [22:51] Of course!