[00:03] tsimonq2: done and rolled out [00:03] slangasek: Thanks. [01:42] slangasek: So the build logs for Lubuntu daily-live timestamped 20180104 show that the *implementation* is correct but the repositories aren't. [01:43] slangasek: I certainly wouldn't have expected this... [01:45] mm? [01:45] slangasek: It errors out with this: fatal: could not read Username for 'https://git.launchpad.net': No such device or address [01:45] ok then [01:45] So it just needs git+ssh? [01:46] (Do the builders have the capability to do that? Stupid question but I think it's worth asking...) [01:46] well, git+ssh *definitely* requires a username... [01:47] Right, but apparently so does https [01:47] https://people.canonical.com/~ubuntu-archive/cd-build-logs/lubuntu/bionic/daily-20180104.log - there it is [01:47] tsimonq2: yes, if I browse to https://git.launchpad.net/~lubuntu-dev/ubuntu-seeds/+git/platform.bionic I get an Ubuntu SSO prompt [01:48] That's not what I'm referring to. [01:48] and then I log in, and then I get Repository '~lubuntu-dev/ubuntu-seeds/+git/platform.bionic' not found. [01:48] Besides, that doesn't exist (thus the SSO prompt) [01:48] Ohhh [01:48] I'm misreading this [01:49] (or am I? dunno) [01:49] what I see is that https://git.launchpad.net/~lubuntu-dev/ubuntu-seeds/+git/lubuntu/ clones fine, then it fails to get platform.bionic [01:49] In fact, line 59 of that log I just linked is probably the culprit [01:49] RIght [01:49] s/59/60/ [01:50] * Checking out https://bazaar.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/platform.bionic/ [01:50] bzr: ERROR: Not a branch: "https://bazaar.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/platform.bionic/". [01:50] Right [01:50] I can now repro locally [01:51] fwiw your bzrpattern edit changed http to https [01:51] and I can branch from http [01:51] - elif self.prefer_bzr: [01:51] - pattern = "http://bazaar.launchpad.net/~%s/ubuntu-seeds/" [01:51] you assumed https was correct? :) [01:51] Right, and I thought that was harmless, because I assumed when everything in Canonical's sites switched over to https, so did bazaar.launchpad.net [01:52] (there was an announcement and everything) [01:52] ok. I'm switching it back [01:53] Also change it back in the tests (when you fixed them I saw you changed to https as well) [01:55] tsimonq2: done. by all rights I should've demanded you change this already when reviewing because you were making an unrelated change, but I was being lenient ;) [01:56] slangasek: Right, and to be completely fair, I thought HTTPS Bazaar was a thing. [01:59] slangasek: Also, you apparently believed me, you did it in your commits fixing the tests ;P [02:02] slangasek: ubuntu-archive-tools> To get this part fixed, I could do this in multiple ways... I'm leaning towards trying to programatically detect if it's a git or bzr repo and then adding a --vcs argument if the user wants to override, would that be alright? [02:03] tsimonq2: sorry, I don't actually know the context of what needs to be fixed on that branch, and my attention is elswhere right now [02:05] slangasek: Alright. When you do have a minute though, branch-seeds in lp:ubuntu-archive-tools is what I'm referring to. [02:05] (Just needs "Gitifying") [02:18] tsimonq2: ok, I don't know that I've ever used this command, but I see that it's referenced in https://wiki.ubuntu.com/NewReleaseCycleProcess. I don't see a use case for a --vcs option to override - it seems to me that if a flavor is migrating their seeds, they should do so for historical releases all at the same time [02:18] tsimonq2: I'm fine with programmatically detecting bzr vs git (which should be as straightforward as a string match on the URI) [02:22] slangasek: Adam does this at the beginning of every release cycle. ;) [02:22] slangasek: http://bazaar.launchpad.net/~lubuntu-dev/ubuntu-seeds/lubuntu.bionic/revision/390 for example [02:23] (well, not *absolutely* sure, but it does look like this is the script he uses) [02:24] (I'd ask Adam himself but it seems like he's still on vac) [02:27] vcs> makes sense [04:47] slangasek: There, I think this should do it: https://code.launchpad.net/~tsimonq2/ubuntu-archive-tools/add-git-support-to-branch-seeds/+merge/335687 [07:01] slangasek: bzr merging always i bizzare to me (so it lives to its abbreviation) - but yeah now that the Delta is generated it obviously is the wrong target [07:01] slangasek: thanks for the hint [07:03] what I had set was the default LP gave me [07:03] I think I fixed the target, but will double check once the diff is up this time [07:11] hmm now it shows no diff, did you already merge and your pingwas only FYI? [07:12] oh you did, thank you slangasek [08:59] hello apw, please merge? https://code.launchpad.net/~costamagnagianfranco/britney/hints-ubuntu/+merge/335626 === Laney changed the topic of #ubuntu-release to: LP build farm and autopkgtest request.cgi disabled for maintenance; no ETA yet | Released: Xenial 16.04.3, Artful 17.10 | Archive: open | Bionic Release Coordination | Please don't upload things during freezes where you shouldn't, or be prepared to apologise to the release team | We accept payment in cash, check or beer | melius malum quod cognoscis [10:16] :'( [10:24] Oh shit [10:25] holiday, celebrate! === Ukikie is now known as OldManWinter [14:10] Right, unexpected holiday ... [15:17] -queuebot:#ubuntu-release- Unapproved: rejected sosreport [source] (artful-proposed) [3.5-1~ubuntu17.10.1] [15:17] -queuebot:#ubuntu-release- Unapproved: rejected sosreport [source] (zesty-proposed) [3.5-1~ubuntu17.04.1] [15:17] -queuebot:#ubuntu-release- Unapproved: rejected sosreport [source] (xenial-proposed) [3.5-1~ubuntu16.04.1] [15:18] -queuebot:#ubuntu-release- Unapproved: rejected sosreport [source] (trusty-proposed) [3.5-1~ubuntu14.04.1] [15:57] sil2100: hey [15:57] are you working on #1737640? [16:48] -queuebot:#ubuntu-release- Unapproved: accepted sosreport [source] (artful-proposed) [3.5-1~ubuntu17.10.1] [16:50] -queuebot:#ubuntu-release- Unapproved: accepted sosreport [source] (zesty-proposed) [3.5-1~ubuntu17.04.1] [16:54] -queuebot:#ubuntu-release- Unapproved: accepted sosreport [source] (xenial-proposed) [3.5-1~ubuntu16.04.1] [16:56] -queuebot:#ubuntu-release- Unapproved: accepted sosreport [source] (trusty-proposed) [3.5-1~ubuntu14.04.1] [17:19] -queuebot:#ubuntu-release- Unapproved: accepted openvswitch [source] (artful-proposed) [2.8.1-0ubuntu0.17.10.1] [17:22] -queuebot:#ubuntu-release- Unapproved: accepted openscad [source] (artful-proposed) [2015.03-2+dfsg-2ubuntu1.17.10.1] [17:25] -queuebot:#ubuntu-release- Unapproved: accepted cloud-initramfs-tools [source] (artful-proposed) [0.39ubuntu1.1] [17:42] -queuebot:#ubuntu-release- Unapproved: accepted cloud-initramfs-tools [source] (zesty-proposed) [0.35ubuntu2.1] [17:45] -queuebot:#ubuntu-release- Unapproved: rejected evolution-data-server [source] (artful-proposed) [3.26.3-1ubuntu0.1] [17:45] -queuebot:#ubuntu-release- Unapproved: rejected evolution [source] (artful-proposed) [3.26.3-0ubuntu0.1] [17:50] -queuebot:#ubuntu-release- Unapproved: accepted cloud-initramfs-tools [source] (xenial-proposed) [0.27ubuntu1.5] [17:57] ...are the Launchpad builders still down, even for Ubuntu builds? [17:58] Oh, hm, seems things might be back? [18:00] Oh, reading #lp, nvm [18:07] -queuebot:#ubuntu-release- Unapproved: rejected horizon [source] (xenial-proposed) [2:9.1.2-0ubuntu4] [18:13] tsimonq2: also the topic here :) [18:14] rbasak: I know, but I see builds going, so I wondered... [18:47] rbasak: this brings up an obvious question, but does this require alteration to the release schedule to adjust freeze dates, etc. because of the downtime? [18:47] since we can't build things anymore :P [19:04] tsimonq2: Only test rebuilds, which have built on Launchpad already and so are negligible risk [19:04] (well, mostly only test rebuilds) [19:09] teward: I'm not on the release team. But I think it probably doesn't make sense to adjust schedules twice, so we might as well defer any discussion until we know when the downtime will be over. [19:10] -queuebot:#ubuntu-release- Unapproved: horizon (xenial-proposed/main) [2:9.1.2-0ubuntu3 => 2:9.1.2-0ubuntu4] (openstack, ubuntu-server) [19:10] cjwatson: Alright, thanks [19:25] jbicha: hrm, do you know the history of the gvfs delta that runs the autopkgtests through 'sudo', without taking any steps to ensure that the user running the tests has passwordless sudo access? [19:27] jbicha: (yours is the oldest Ubuntu upload before the merge changelog gets truncated, so you get to field the question ;) [19:33] apparently this dates back to gvfs 1.20.1-1ubuntu3 by pitti [20:12] -queuebot:#ubuntu-release- New source: opengcs (bionic-proposed/primary) [0.3.4+dfsg2-0ubuntu1] [20:19] I suggest skipping a January alpha (there's still a 17.10 respin that needs to be done too) and most flavors weren't participating in the January alpha [20:25] unless miraculous updates appear in the VERY near future, I would agree with that ^^^ [20:30] a year ago, we didn't have an "Alpha 1" and things were fine, so I don't think it's very useful [20:35] 'we've' not had an alpha for some cycles and things are fine :) [21:16] -queuebot:#ubuntu-release- Unapproved: accepted google-cloud-sdk [source] (artful-proposed) [183.0.0-0ubuntu1~17.10.0] [21:20] -queuebot:#ubuntu-release- Unapproved: accepted google-cloud-sdk [source] (xenial-proposed) [183.0.0-0ubuntu1~16.04.0] [21:21] -queuebot:#ubuntu-release- Unapproved: accepted google-cloud-sdk [source] (trusty-proposed) [183.0.0-0ubuntu1~14.04.0] [21:47] jbicha: Right, that was the plan. [21:47] jbicha: (I've had my eye on this, everything *could* magically work out but it's unlikely) [21:47] flocculant: Let's not get started on that debate again ;) [21:48] I totally see the use in it [21:48] (So, I'll continue pushing to do it.) [21:49] If a flavor doesn't want to do it, it's simple, they opt out. [21:49] ¯\_(ツ)_/¯ [21:51] I think you could get the same benefit without the milestone overhead [21:53] I disagree, it's a nice occasion for testing, and we've found (and squashed) bugs because of A1 testing before. [21:54] you can schedule testing without imposing an archive freeze and publishing milestone images [21:57] Sure you can, but milestones seem to get the most traction, because there's a sense of urgency attached to releasing images as a milestone. [21:58] if you wanted to do an "alpha" that was just coordinating testing, instead of imposing a milestone freeze on the archive that causes drag for other teams due to the buggy way per-flavor freezes are implemented, then I would have no objections [22:00] What would it take to fix the buggy implementation of the per-flavor freezes so that other teams wouldn't be dragged down by it? [22:15] (I ask because rather than avoiding milestones only because the per-flavor freeze implementation is buggy could be fixed by correctly implementing the per-flavor freeze... I'd rather do that than go without milestones.) [22:16] ((That sentence didn't make gramatical sense but you probably got the point anyways...)) [22:23] -queuebot:#ubuntu-release- Unapproved: rejected libzstd [source] (zesty-proposed) [1.3.1+dfsg-1~ubuntu0.17.04.1]