=== dan is now known as Guest95690 [03:37] Just heard the news about infinity. :-( Very sorry to hear it, and my condolences to his family and all the team at Canonical. A great loss for the community. [03:38] <3 it hurts :( === antoine5 is now known as antoine === ItzSwirlz_ is now known as ItzSwirlz === sem2peie- is now known as sem2peie === nicolasbock_ is now known as nicolasbock [03:55] he'll be sorely missed :( === guiverc2 is now known as guiverc === tjaalton_ is now known as tjaalton [08:39] juliank, hey, did you see bug #1912718 which claims to be a regression from the recent oem changes [08:39] bug 1912718 in update-manager (Ubuntu) "update-manager: clicks do not toggle checkboxes for "Remove old kernels"" [Undecided,New] https://launchpad.net/bugs/1912718 === sem2peie- is now known as sem2peie [10:42] seb128: people claim a lot, half the time it's wrong. I don't have any insight into this feature and what it does [10:43] you can see it's been tagged incoming, so it'll be discussed next thursday [10:44] like why blame the oem changes and not the hwe kernel grouping changes [10:47] SRU team ^ we might want to back that update out of focal-updates or .2 until the regression is sorted out [10:47] sil2100, ^ fyi [10:48] juliank, one week to wait is really suboptimal with the LTS .2 coming [10:49] the update is already phased at 0% [10:49] it's not like anyone gets it [10:50] i just autoremoved my kernels yesterday, so I'm not going to be helpful [10:51] juliank, did we land apt phasing in focal? if not any apt user would still get the updated version... [10:52] no [10:52] also the .2 ISO doesn't care about the phasing % [10:52] yes [10:52] it's going to include the current versions [10:53] I don't have insight into this, but holding back the update from .2 seems like the wrong approach [10:53] it's fine if you don't want to handle that regression, still I think it should be on the release team list to check before .2 [10:53] right, shipping a regression in a LTS point release is also the wrong approach [10:57] seb128: I can confirm the bug in hirsute [10:59] seb128: yeah this is a regression from the linux hwe change, as it now matches versioned packages rather than meta packages [11:01] So I guess we gotta add a tag to bug 1902025 [11:01] bug 1902025 in update-manager (Ubuntu Groovy) "hwe kernel not scoped under "Ubuntu Base"" [Medium,Fix released] https://launchpad.net/bugs/1902025 [11:06] seb128: So I can revert those changes, in SRUs at least, the devel change history is harder to revert... [11:07] but maybe bdmurray and JawnSmith want to actually fixi t [11:07] We do have another regression report that causes phasing to go to 0% which I don't understand [11:07] https://errors.ubuntu.com/problem/eac836e467681c6b96247949de5b356d6c3da968 [11:08] people apparently using the synaptic backend (despite aptdaemon being installed) [11:08] did they remove their aptd binary? [11:08] weird [11:09] (they certainly don't set the environment variable forcing aptdaemon as we can see in ProcEnvironment file) [11:09] Should be fixed too [11:10] but arguably we have no way to collaborate on SRUs of update-manager [11:10] there is no repo, so I can't commit a fix and have the others commit their fixes :/ [11:11] juliank, those reports are from LXQt desktops [11:11] so that flavor is probably not installing aptdaemon? [11:11] seb128: Dependencies says "aptdaemon 1.1.1+bzr982-0ubuntu32.3" [11:12] that's why I said, they both have aptdaemon installed and not set the variable, so how do they enter the codepath :D [11:13] juliank, you don't know if they have not set the variable, I think the ProcEnviron is only listed a selected set of variables [11:13] otherwise you would have things like XDG_CURRENT_DESKTOP in there [11:13] or XDG_DATA_DIRS [11:13] ah [11:13] juliank, that's the environment variable that sets the backend? [11:14] UPDATE_MANAGER_FORCE_BACKEND_SYNAPTIC=1 [11:14] I don't have context, what regression are we talking about? [11:14] I think making it "return" or "pass" instead of "raise NotImplementedError" [11:14] sil2100: We have two [11:15] sil2100, https://bugs.launchpad.net/ubuntu/+source/update-manager/+bug/1912718 is the one I'm concerned about for desktop [11:15] Launchpad bug 1912718 in update-manager (Ubuntu) "update-manager: clicks do not toggle checkboxes for "Remove old kernels"" [Undecided,New] [11:15] sil2100: A regression from bug 1902025 means kernels are not autoremoved anymore (bug 1902025), and another regression from the oem stuff causes crashes for users forcing synaptic backend [11:15] bug 1902025 in update-manager (Ubuntu Groovy) "hwe kernel not scoped under "Ubuntu Base"" [Medium,Fix released] https://launchpad.net/bugs/1902025 [11:15] sil2100, I think it should be resolved for .2 [11:16] I don't know what the right answer for bug 1902025 is [11:16] but certainly matching all kernel packages when we're trying to find meta packages isn't [11:16] Reverting that change is easy which means hwe kernels won't be scoped under Ubuntu base again, but meh [11:17] I don't know if we can just revert the upload [11:18] that'd be something to discuss with the OEM team, as I don't know what people agreed upon [11:19] Well, I'd revert the change if anything if we know this particular change caused the regression [11:19] Since backing off the package from focal-updates is really not that helpful [11:19] Laney can reproduce the other issue - the NotImplementedError, but I don't have luck so far [11:20] ah I need to be on an enabled system [11:20] Reverting all the changes, well, as juliank mentioned, this upload has more changes than just this so it's troublesome [11:20] We can wait a couple hours and give bdmurray and jawnsmith a chance to fix it [11:21] the hwe kernel regression [11:21] but I'm not sure what the right solution is there [11:23] Let's see [11:25] * UpdateManager/Dialogs.py: do not mention livepatch settings when the [11:25] interface is not installed, thanks to Sebastien Bacher for the patch. [11:25] (LP: #1807900) [11:25] Launchpad bug 1807900 in update-manager (Ubuntu) "update-manager suggests to use Livepatch, which is not available" [Low,Fix released] https://launchpad.net/bugs/1807900 [11:25] ugh [11:25] extracting those individual changes and _not_ reverting them is hard [11:26] There is a ton of test changes that I can't trace to anything [11:27] make a branch this time [11:27] seriously [11:27] Laney: I'm mostly talking about hirsute, which has a branch :D [11:27] a bzr branch [11:28] and everything is in "release" merge commits [11:29] How do I revert a commit in bzr even? [11:30] As I said, fixing the SRUs is easy, it's not split over 20 commits there [11:33] I want to revert changes 2900.2.* [11:34] which come after 2901 [11:34] I think you bzr merge the inverse of them, or something like that [11:35] juliank: https://git.launchpad.net/~laney/update-notifier/commit/?id=b6f277875d95130ab0b5bc9a5e4d26511f25a3d4 any comment? I didn't realise apt.Cache() could fail, but it can and this error has stopped phasing :> [11:36] No matter what I do, it also reverts the change in 2900 [11:37] ugh it was a silent reformatting in a release commit [11:38] need to get better at vcs discipline ¬_¬ [11:39] I think I got it now [11:40] Laney: Well that's the way bzr is supposed to be used and that's why git won :D [11:41] rebasing and producing clean commits rather than just have huge numbers of fixups in merges [11:43] running autopkgtest now for hirstue [11:43] juliank: give me a sign once you have something for update-manager! I'm leaving it in your hands o/ [11:45] sil2100: Please remove update-manager 20.10.2 from groovy, it only contained the broken change [11:46] sil2100: Or should I upload higher-versioned revert so people get update? [11:51] juliank: I think it makes more sense to do a versioned revert, because this one has been around for quite a while now [11:51] sil2100: ack [11:54] focal is running autopkgtests now [11:55] look at my ping now >:| [11:55] though to be fair, I compared against the 20.04.10.3 upload I created (which was replaced) [11:56] huh [11:56] git-ubuntu is missing the 20.10.3 upload [12:01] juliank: which package? [12:01] update-notifier? [12:01] rbasak: update-manager [12:02] It's in the queue. I can bump priority if you'd like? [12:03] Queue size is 189 - mostly language-pack-* [12:03] rbasak: Nah, I worked around locally [12:03] :D [12:03] OK! [12:03] sil2100: update-managers for focal, groovy ready [12:04] juliank: thanks! Will look in a moment! === nikow4 is now known as nikow [12:39] * Laney times out on feedback from juliank and uploads [12:39] Laney: I did not have any comment, all was good, fwiw [12:39] thanks, same effect then :> [12:40] silly bugs are silly [12:47] heh we have two different tagging schemes going on in update-notifier git [12:57] doko: just saw https://fedoraproject.org/wiki/Changes/PythonNoSemanticInterpositionSpeedup on LWN, have you/we considered doing the same? [12:59] Laney: how would that help? the interpreter is not linked with the shared libpython [13:00] k thanks for the feedback [13:02] 13:57 < sil2100> juliank: hey, do we have a bug for the commit_oem issues with the synaptic backend? [13:03] 13:58 < sil2100> Since I'd like that linked in with a separate tracking bug if anything [13:03] 14:00 < sil2100> The error tracker report doesn't seem to have a bug yet [13:04] Hello, long-time Ubuntu contributor here. I've always had an interest in LTS and I'd like to become an LTS dev. Alas, my wiki page says immutable. Is this a common problem? I thought a wiki page was still part of the application. [13:04] https://wiki.ubuntu.com/RolfLeggewie [13:04] Laibsch: try logging out and logging in again to the wiki please [13:04] Sure [13:04] Good to see some of the old names still around ;-) [13:05] Laney: there's also #948802, now replied to [13:05] rbasak: It's still immutable. [13:05] Last touched in 2014. [13:06] There were some spam issues in the wiki AFAIK, which meant that people had to be approved, which was basically done for every real person. [13:06] Let me see if I can find that. [13:07] juliank: I created https://bugs.launchpad.net/ubuntu/+source/update-manager/+bug/1913732 via the error tracker [13:07] Launchpad bug 1913732 in update-manager (Ubuntu) "/usr/bin/update-manager:NotImplementedError::on_button_install_clicked:start_install:_start_pane:start:commit_oem" [Undecided,New] [13:08] rbasak: FWIW, https://wiki.ubuntu.com/DeveloperMembershipBoard/Agenda is immutable for me as well. [13:08] I'm logged in via Ubuntu One. [13:09] Laibsch: sorry I can't find a reference. I think either your login is screwed up somehow (bug in wiki.ubuntu.com?) or you need to belong to some "I'm not a spammer" team. Could you ask in #canonical-sysadmin please? [13:09] ocf [13:09] ofc [13:09] thx [13:09] Laibsch: You need to be a member of ~ubuntu-wiki-editors. Please see https://help.ubuntu.com/community/WikiGuide#Contributing [13:09] Ah, thanks. [13:09] I was looking at my own memberships. I'm not a member! [13:09] I guess not many people applying for elevated privileges these days anymore [13:10] rbasak: Haha, that's classic. [13:10] Spam makes everything terrible [13:10] nice! ;-) [13:10] Yes, spam sucks [13:10] It's just anti-spam. I believe the bar for membership of that team is low :) [13:10] I suspect that the actual permissions are the union of ~ubuntu-wiki-editors and a few other teams [13:10] rbasak: As a dev you are a member of ubuntu-members, which also gives edit access to the wikis. [13:10] There we go [13:10] That explains it. Thanks :) [13:11] juliank: ok, added it to your changelog entry and re-uploaded [13:11] popey: You are listed as the admin of that team. Maybe you can have a look and approve r0lf for wiki editing? (notice the zero instead of o) [13:12] hello there [13:12] Hi, that was fast! [13:12] Done! [13:12] I need wiki edit rights for https://wiki.ubuntu.com/RolfLeggewie [13:12] Thanks! [13:13] Might need to logout / back in, and refresh the page [13:13] Sacrificing chickens in chalk pentagrams is optional and may also help. [13:13] Yup, that worked. Thanks a bunch. [13:14] Silly, chickens are for SCSI [13:14] Wiki needs goats [13:24] bdmurray: while we are at it, would you be so kind to renew my membership in https://launchpad.net/~bugsquad (It's been pending for a while, I've been a member for about a decade or so, but dropped out last year) [13:24] https://launchpad.net/~r0lf is my LP handle [13:41] hggdh: you were the last one to renew my membership in November 2019. Maybe you can pull the switch again this time for me for bugsquad? [13:43] juliank, seb128: can one of you confirm that the reverted update-manager in focal-proposed fixes the regression? [14:06] sil2100: Sorry I somehow did not get notifications [14:07] Laibsch: hey! We see that you have added your name to both the SRU developer and MOTU applications to the DMB Agenda [14:07] ah it seems i turned them off [14:08] Laibsch: sadly this is not enough information for us to consider your applications - please read https://wiki.ubuntu.com/DeveloperMembershipBoard/ApplicationProcess and prepare all necessary steps [14:08] Laibsch: thank you! [14:09] I'm in the process of doing that, just securing my slot [14:10] Laibsch: there's not too many applicants usually, so just feel free to add your name once you have the base application and a date at which you'd want to apply [14:11] For now I have removed your name from the agenda - but just re-add it with the application link and date [14:12] Laibsch: as a DMB member I would rather you get your application in order first before you put it on the agenda, taking into account sil's statements :) [14:12] sometimes getting the apps in line takes longer than you expect ;) [14:13] *points at Eickmeyer as an example* took him 3 months to get his application in order :P [14:13] ('course i use that as an extreme example - i know why there was a delay there) [14:13] FWIW, I'd prefer to give some latitude here. If an application is ready soonish, I'm not bothered what order things happen in. [14:15] *shrugs* guess i'm a stickler for getting ducks in order :P [14:15] Getting an application in ends up annoyingly circular sometimes. [14:15] I remember being asked by potential endorsers "What's the meeting date" "I don't know I'm supposed to get endorsements first" etc. [14:16] So to avoid sending applicants around in circles, even if the formal process doesn't require that, I prefer just to let them sort out their ducks in whatever order they want :) [14:30] Well, yes, but in this case there was no date and no application started [14:47] sil2100, sorry I can't easily do that today, I don't have a focal setup to trigger the bug handy and that's not a quiet day :-/ [15:39] Laibsch: done [15:40] hggdh: Thank you so much!! [15:49] juliank: so will you be able to somehow easily verify LP: #1912718 today? [15:49] Launchpad bug 1912718 in update-manager (Ubuntu Hirsute) "update-manager: clicks do not toggle checkboxes for "Remove old kernels"" [High,Fix committed] https://launchpad.net/bugs/1912718 [16:00] sil2100: gotta boot a focal vm and install some old kernels [16:00] mark them auto, then run update-manager [16:00] Of course, if somebody else already has a VM that would save some effort :D [16:00] Ugh guess I dcan just boot live disc, too [16:18] sil2100: I verified that bug, maybe Laney can check the other one? I fail getting OEM packages to appear in the first place atm [16:18] it's weird [16:19] you can just write oem-qemu-meta to $XDG_RUNTIME_DIR/ubuntu-drivers-oem.package-list probably [16:20] Laney: nvermind, my update-manager was too old :D [16:20] /o\ [16:23] Laney: Oh well, this crashed too [16:23] ah [16:23] ugh [16:23] no synaptic installed [16:23] I should really drink some tea [16:23] hahahaha [16:23] having a good afternoon over there [16:24] I have no idea how much the synaptic backend actually works [16:24] I'm going to kill that backend in hirsute soon [16:24] Not sure if anything is happening, but oh well, it's not crashing [16:26] It doesn't seem to do anything [16:28] The one in release pocket works but throws an AttributeError [16:31] damn it [16:32] uh [16:34] I alt+f4d gedit in the vm, but I actually alt+f4d the vm [16:34] Today is really not the day for this [16:35] This is going to remain broken one way or the other, we can't really fix it [16:36] as in, you see the oem packages but they do nothing [16:40] I guess you get the stuff next time [16:40] or is it really nothing, like they're in the list but not installed? [16:41] Laney: So current commit fixes the crash but never launches synaptic, gets stuck idling in main loop [16:41] Laney: Have to call action_done [16:41] from commit_oem [16:41] Then it basically ignores installing oem pacakges [16:43] oh I see, yeah, you have to move from PRE_INSTALL to INSTALL [16:43] annoying [16:46] Laney: https://paste.ubuntu.com/p/8mrT3tfWjr/ [16:46] Laney: copied from live focal system, so already pre-verified :D [16:47] uploaded [16:51] New update-manager [16:51] yay [16:52] sil2100: At least it's already verified, so we don't need more time than building it :D [16:52] Weeeeeeeeell, the process requires testing the built binaries! Though uh, this is python so whatever [16:52] :D [16:52] It's close enough [16:53] Next day we do something like this, please let it be a day where I'm not half dead [16:54] Sleep tracking looks great, but it was a terrible night [16:58] Laney: don't we also need your new update-notifier for .2? [16:59] ah I guess it's more meh than the crashes in update-manager [16:59] :D [16:59] no particular opinion there, do you think you can get into that situation from an image? [16:59] i seem to remember that update-manager is somehow disabled there actually but that could be wrong [16:59] probably the notifications are turned off [17:00] Anyhow I think I'm done for now, and will hop on Zwift to get some 20km virtual cycling in - 3 days left in this month and 54 kilometers left to reach my 400 km goal [17:01] Also need to find free space in the kitchen to actually make dinner :/ [17:01] Ok, the action_done stuff looks legit, accepting [17:01] :) [17:01] plz mark verified once it's all built and shiny [17:02] I'm out for like 1-2 hours with a break in the middle I can check in [17:02] but it's a busy evening :/ [17:13] Laney: Are you good with https://code.launchpad.net/~jawn-smith/ubiquity/+git/ubiquity/+merge/396616 now? [17:16] sure, I can approve [17:16] I can do the uploads just wanted to make sure you are happy with it. [17:17] I dunno what prompted the latest 'no system bus is present' stuff though, the desktop will not work at all if that's the case [17:17] but surrrrrrrrrrre [17:17] It failed to build [17:18] commented with a suggestion to squash it all and approved, cheers [17:19] thanks! [18:41] xnox: yeah, that's what I was going to resort to depending on what rick_h said (not sure how this is expected to work) [20:14] bloody alt+f4 hell in vms [20:20] sil2100: verified the bugs, actually reverified them properly :D [20:21] \o/ [20:22] it took a bit longer than planned, seems I was not fast enough making dinner or post-ride finish up :/ [20:22] juliank: thank you! Will release soonish, even though it's Friday - but uh, better this than the regression [20:22] No worries, I also went AFK for some exercise earlier ;) [20:22] woohoo - exercise :)