[04:48] bashfulrobot: seems like you're in charge of coordinating beta 2? we going to have images to test tomorrow? [04:56] I'm actually drafting the wiki page and sending the email to the mailing list. So we actually consider this beta 2? Or Jitsu a delayed beta one? [04:56] wxl: ^^ [04:57] sorry, i think i meant 1 [04:57] yes i did :) [04:57] beta 2 is supposed to be next month [04:58] wxl: did you ever figure out your seed issue? [04:58] i'm wearing the lubuntu release manager hat from off tsimonq2's head for the better part of the week, just in case you need me, bashfulrobot [04:59] nacc: no, but i think we have what we need to figure it out. tsimonq2 has a lot more experience with germinate than i do so i left it in his hands. [04:59] wxl: ack [05:00] wxl: I appreciate the offer for help. Simon always guided me through. Lost a hard drive with my notes, but. Prepping wiki page, will email list to get the flavors signed up. Then I know we need to track down someone to help with Nusakan. [05:01] I'll confirm in here as well once the wiki and email list has been completed [05:02] bashfulrobot: he said sil2100 was going to do it but i guess your fall back would be poor old infinity [05:02] * wxl pats adam's well-worn head [05:03] * bashfulrobot plans to hit up sil2200 when the time is right. [05:24] bashfulrobot: I've done the flavour bit of this in the past too - another name you can ping if needed :) [05:25] Great flocculant ! Appreciate the support [05:25] no problem :) [05:32] I plan on making detailed notes this time so I don't need to bug anyone next time. [05:34] best laid plans of mice and men ... [05:45] too true! [05:46] added Xubuntu to the wiki page [05:46] The one I just created? [05:48] Also on past pages there were links from the beta/alpha pages that went to something like: https://wiki.ubuntu.com/BionicBeaver/Beta1/CommonInfrastructure - but many of those pages were never created. I imagine I should just yank that (as it is uselesss) [05:50] yea the one you just created [05:51] and yea - I'd yank that - only time it really works is the one that Ubuntu joins in with, when they've created the main wiki page [05:51] In the past there were also links to release announcements - was that always something official or the release manager wrote it? [05:52] flocculant: Cool - I'll comment that out [05:52] same [05:52] perfect [05:52] Commenting up a storm! [05:53] pre-Final Beta the only announcement is what whoever does the flavour side writes for the mailing list [05:54] bashfulrobot: see https://wiki.ubuntu.com/ArtfulAardvark/Beta1/ReleaseAnnouncement [05:57] Cool. I had peeked at that one earlier. I'll leave the page as is for now. [05:58] yea [05:58] personally when I last did it - I didn't do any of that stuff :) [05:59] originally B1 was supposed to be Mar 1 - so I guess we can just push out to Mar 8? 1 week later. Or is that date too close if I am about to email the mailing list? [05:59] beta 1 will be out of date as soon as it updates anyway - I try not to advertise it's existence [06:00] ha ha [06:00] release schedule says 8th [06:00] Well, I'll just get the flavors to pipe in on that wiki page. And great for the 8th. [06:01] assuming you can get a Canonical bod involved yea [06:01] * bashfulrobot pilfering a previous email to send to the email list. [06:01] bashfulrobot: yep - mail release list pointing to it and wait for the deafening silence :D [06:11] * mwhudson stares at freeipa [06:12] does mwhudson want other's to join in with freezing stare? [06:13] oh i see there has been lots of talk about this already [06:14] flocculant: more like a "begone from -proposed" stare i think? [06:14] :) [06:16] pg-repack [06:16] oops [06:16] although has anyone looked at that? [06:18] ok - email sent to the mailing list [06:18] * bashfulrobot waiting to hear the crickets chirp [06:19] for those super keen people that want to sign up for beta one...https://wiki.ubuntu.com/BionicBeaver/Beta1 [06:21] sil2100 and infinity [06:22] knocking on the Nusakan door [06:22] * bashfulrobot slinks away into the shadows [06:25] * bashfulrobot turning in. Zzz [06:27] bashfulrobot: thanks for doing this :) [06:36] flocculant: my pleasure! Like getting involved. [06:36] \o/ [06:37] * flocculant wanders off for work [07:45] oh, and linux-meta-kvm [07:45] slangasek, hwe, hwe-edge, kvm still to do, it looks like. [08:03] -queuebot:#ubuntu-release- New: accepted linux-azure [amd64] (bionic-proposed) [4.15.0-1002.2] [08:10] -queuebot:#ubuntu-release- New source: curl3 (bionic-proposed/primary) [7.58.0-2ubuntu2] [08:18] -queuebot:#ubuntu-release- New: accepted pyxs [source] (bionic-proposed) [0.4.1+dfsg1-0ubuntu1] [08:20] -queuebot:#ubuntu-release- New binary: pyxs [amd64] (bionic-proposed/universe) [0.4.1+dfsg1-0ubuntu1] (no packageset) [08:21] slangasek, curl3 is in, i guess best debdiffed against curl from release pocket.... [08:22] -queuebot:#ubuntu-release- New: accepted rax-nova-agent [source] (bionic-proposed) [2.1.12-0ubuntu1] [08:23] xnox: indeed [08:27] morning. slangasek or anyone else: is anyone able to give attention to the gstreamer plugins-good issue before beta 1 isos get done? [08:31] acheronuk: I've flagged it to the MIR team; I'm not on the MIR or security team so approving this into main is not up to me. If we can't get the MIRs done in time for image mastering, we have two other options: pre-promote these packages into main while the MIR is in progress, with the possibility that they might be rejected and have to be dropped again; or revert the package in -proposed and [08:31] accept the ongoing transition without the new upstream gst-plugins-good (<-- Laney?) [08:32] xnox: not a NEW blocker, but maybe drop the XSBC-Original-Maintainer from curl3 [08:32] meh, ok. [08:33] xnox: or put your own name in a maintainer field for a change ;P [08:33] -queuebot:#ubuntu-release- Unapproved: google-cloud-sdk (artful-proposed/partner) [183.0.0-0ubuntu1~17.10.0 => 191.0.0-0ubuntu1~17.10.0] (no packageset) [08:33] slangasek, i'm so over it. [08:33] slangasek, i'm thinking to start uploading binNMUs with -- Ubuntu Developers [08:34] -queuebot:#ubuntu-release- Unapproved: google-cloud-sdk (trusty-proposed/partner) [183.0.0-0ubuntu1~14.04.0 => 191.0.0-0ubuntu1~14.04.0] (no packageset) [08:34] -queuebot:#ubuntu-release- Unapproved: google-cloud-sdk (xenial-proposed/partner) [183.0.0-0ubuntu1~16.04.0 => 191.0.0-0ubuntu1~16.04.0] (no packageset) [08:34] xnox: why is debian/libcurl3.links dropped? It may not help anyone in practice, but it also doesn't hurt [08:35] slangasek, reject, let me reupload. It was dropped when i thought i had to provide libcurl3 co-installable with libcurl4, and started to thinking into shipping it as libcurl.so.3 / libcurl.so.3.4.5.1 or some such [08:35] xnox: too late, already accepted, you can do a -2 [08:35] -3 [08:35] ok [08:35] -queuebot:#ubuntu-release- New: accepted curl3 [source] (bionic-proposed) [7.58.0-2ubuntu2] [08:37] slangasek: ok. fwiw I have not yet seen any confirmed reports of the alleged issue with upgrades to mesa 18 without the bugfix in -proposed Qt, and by its nature the bug should not in theory affect a fresh install from iso or the live session. so if Qt does not move, at the moment seems a 'this might be buggy on upgrading a current system' in beta release notes could just about be acceptable [08:38] acheronuk: well, I think it is important for us to resolve this transition in time for beta, let's continue aiming for that [08:38] slangasek: great. no complaints from with that. thank you [08:39] *from me [08:39] -queuebot:#ubuntu-release- New: accepted linux-aws [amd64] (bionic-proposed) [4.15.0-1001.1] [08:40] -queuebot:#ubuntu-release- New: accepted linux-gcp [amd64] (bionic-proposed) [4.15.0-1001.1] [08:41] -queuebot:#ubuntu-release- New: accepted pyxs [amd64] (bionic-proposed) [0.4.1+dfsg1-0ubuntu1] [08:47] -queuebot:#ubuntu-release- Unapproved: accepted google-cloud-sdk [source] (artful-proposed) [191.0.0-0ubuntu1~17.10.0] [08:47] -queuebot:#ubuntu-release- New binary: curl3 [s390x] (bionic-proposed/none) [7.58.0-2ubuntu2] (no packageset) [08:49] -queuebot:#ubuntu-release- Unapproved: accepted google-cloud-sdk [source] (xenial-proposed) [191.0.0-0ubuntu1~16.04.0] [08:50] -queuebot:#ubuntu-release- New binary: curl3 [amd64] (bionic-proposed/none) [7.58.0-2ubuntu2] (no packageset) [08:50] -queuebot:#ubuntu-release- New binary: curl3 [ppc64el] (bionic-proposed/none) [7.58.0-2ubuntu2] (no packageset) [08:50] -queuebot:#ubuntu-release- New binary: curl3 [i386] (bionic-proposed/none) [7.58.0-2ubuntu2] (no packageset) [08:51] -queuebot:#ubuntu-release- Unapproved: accepted google-cloud-sdk [source] (trusty-proposed) [191.0.0-0ubuntu1~14.04.0] [08:51] -queuebot:#ubuntu-release- New binary: curl3 [arm64] (bionic-proposed/universe) [7.58.0-2ubuntu2] (no packageset) [08:52] -queuebot:#ubuntu-release- New binary: curl3 [armhf] (bionic-proposed/universe) [7.58.0-2ubuntu2] (no packageset) [08:53] -queuebot:#ubuntu-release- Unapproved: apt (xenial-proposed/main) [1.2.25 => 1.2.26] (core) [09:02] -queuebot:#ubuntu-release- New: accepted curl3 [amd64] (bionic-proposed) [7.58.0-2ubuntu2] [09:02] -queuebot:#ubuntu-release- New: accepted curl3 [armhf] (bionic-proposed) [7.58.0-2ubuntu2] [09:02] -queuebot:#ubuntu-release- New: accepted curl3 [ppc64el] (bionic-proposed) [7.58.0-2ubuntu2] [09:03] -queuebot:#ubuntu-release- New: accepted curl3 [arm64] (bionic-proposed) [7.58.0-2ubuntu2] [09:03] -queuebot:#ubuntu-release- New: accepted curl3 [s390x] (bionic-proposed) [7.58.0-2ubuntu2] [09:03] -queuebot:#ubuntu-release- New: accepted curl3 [i386] (bionic-proposed) [7.58.0-2ubuntu2] [09:25] jamespage: you have some significant lintian warnings in python-diskimage-builder around update-alternatives; please have a look [09:26] -queuebot:#ubuntu-release- New: accepted python-diskimage-builder [amd64] (bionic-proposed) [2.11.0-0ubuntu1] [10:49] -queuebot:#ubuntu-release- Unapproved: accepted dpkg [source] (trusty-proposed) [1.17.5ubuntu5.8] [11:00] slangasek: would you be interested in approving https://bugs.launchpad.net/ubuntu/+source/xubuntu-default-settings/+bug/1753015 ? It's a very minor change and is the last feature missing from xubuntu bionic :) [11:00] Ubuntu bug 1753015 in xubuntu-default-settings (Ubuntu) "[FFe] Add thunar custom action to directly print certain file types" [Undecided,Confirmed] [12:29] -queuebot:#ubuntu-release- New binary: rax-nova-agent [amd64] (bionic-proposed/universe) [2.1.12-0ubuntu2] (no packageset) [14:05] hello, who is the one trying to remove ruby2.3 from the archive [14:05] ? [14:07] xnox, a lot of stuff is failing because autopkgtests drags in the ruby2.3 package http://autopkgtest.ubuntu.com/packages/r/ruby-allocations/bionic/amd64 [14:07] LocutusOfBorg, but... ruby-defaults still has not migrated =( [14:08] I tried with all-proposed and it is failing and dragging 2.3 too [14:08] maybe adding a conflict on ruby2.5 against 2.3? [14:09] same, ruby-defaults won't migrate, because autopkgtests are failing for the same reason [14:09] see: http://autopkgtest.ubuntu.com/packages/r/ruby-turbolinks/bionic/amd64 [14:09] autopkgtest [19:09:17]: test command1: ruby2.3 -e gem\ \"gitlab-turbolinks-classic\" [14:09] autopkgtest [19:09:17]: test command1: [----------------------- [14:09] bash: ruby2.3: command not found [14:14] where is that version picked up? [14:18] -queuebot:#ubuntu-release- New sync: python-bayespy (bionic-proposed/primary) [0.5.12-1] [14:24] sil2100: Were you contacted by any flavours with regard to spinning Beta 1 images? [14:25] flexiondotorg: haven't tried yet, as there are some transitions to fix 1st [14:26] if they can be [14:36] flexiondotorg: yeah, but infinity is leading [14:36] Thanks. [14:36] He said he'll take care of it so I'm leaving it in his hands [15:01] Looks like we have 4 flavours good to go for beta 1. Just getting up and going, so I'm just starting to digest the info. [15:01] infinity: I hear you are the man to talk to regarding Nusakan. :-) [15:29] slangasek, does curl need hinting with curl3? [15:44] slangasek, RM scilab on i386?! fails test, better not ship broken software on i386, but let everything else to migrate? [15:48] xnox: revdeps of scilib? [15:48] lab [15:56] slangasek, twats [15:59] xnox: assign java failure to tdaitx_ for investigation? [15:59] xnox: skip documentation on i386? [16:00] slangasek, it's failing in debian too [16:00] slangasek, possibly, need to check if the docs are arch;all [16:01] xnox: could be omitted on a per-arch basis even if not arch:all [16:23] slangasek, huh, it already has the code to do that, lolz [16:23] doing so [16:24] bashfulrobot: wanting to go, but maybe not good-to-go yet [16:25] xnox: any guesses why abi-monitor alone is reported uninstallable, and only on amd64? [16:29] acheronuk: weekday kind of time line so you think you are looking at? [16:30] slangasek, hmmmmm....... because build-essential is not installable?! it does depend on curl... [16:30] Depends: build-essential, cmake, wget, curl, git, perl:any [16:45] bashfulrobot: depends on release team actions, so I can't say [16:49] Ok, sounds good [16:50] xnox: build-essential did not appear in the list of uninstallables, only abi-monitor [16:50] ok [17:18] -queuebot:#ubuntu-release- Unapproved: python-pyperclip (artful-proposed/main) [1.5.27-3 => 1.5.27-3ubuntu1] (no packageset) [18:03] bashfulrobot: Yo dawg, I heard you like milestones. [18:19] infinity: woof [18:26] infinity: milestones can be a good thing [18:28] bashfulrobot: Setting it up now. [18:36] infinity: thank you good sir! [18:40] infinity: Is there anything additional that you need from me? Info, etc? [18:41] bashfulrobot: you did see the topic, right? [18:43] good beer in Budapest [18:44] wxl: Missed it [18:44] (in the office multi tasking) [18:44] bashfulrobot: "We accept payment in cash, check or beer" [18:44] acheronuk: I'll be on a plane shortly? [18:45] Well the next time I am in a city with any one of you. Beers on me. [18:45] But that seems liek a cop out [18:45] unless there is a beer donation button. [18:45] ha ha [18:45] yeah i don't think release team have implemented that yet [18:45] wxl: my IRC client had truncated the topic. :-p But I just expanded it. [18:46] Should put the super important payment information at the front of the topic [18:47] you know. Beer 1st, beta 2nd??? ha [18:49] must be some app you can send international beer token with [18:49] if there isn't, there should be! [18:50] acheronuk: I'll setup the github repo shortly [18:50] beerchain. [18:50] there we go. [18:50] running ledger of beer owed [18:51] Free beer for miners. [18:51] (as opposed to minors, who should perhaps not be given free beer) [18:52] :) [18:52] rbasak: But kids love free stuff! [18:53] bashfulrobot: I don't need anything from you right now. Just waiting for the archive and britney to settle after putting blocks in place, then will spin up the first set of images. From there, it should be vaguely self-service until you're ready to release on Thursday. [18:53] perfect. Thank you infinity . [18:54] infinity: so gstreamer-plugin-good and Qt 5.9.5 is not getting unblocked? [18:54] Qt 5.9.4 [18:54] [08:38] acheronuk: well, I think it is important for us to resolve this transition in time for beta, let's continue aiming for that [18:57] acheronuk: Oh, are we still waiting on a Qt? :/ [18:58] acheronuk: gst-plugins-good1.0 looks broken? [18:58] [08:31] acheronuk: I've flagged it to the MIR team; I'm not on the MIR or security team so approving this into main is not up to me. If we can't get the MIRs done in time for image mastering, we have two other options: pre-promote these packages into main while the MIR is in progress, with the possibility that they might be rejected and have to be dropped again; or revert the package in -proposed and [18:58] Bunch of unsatisfiable deps. [18:59] infinity: apparently needs that MIR? [18:59] Evidently. [19:00] [08:31] accept the ongoing transition without the new upstream gst-plugins-good (<-- Laney?) [19:00] acheronuk: OTOH, Steve saying he'd like it to happen isn't the same as someone making it happen. :P [19:01] acheronuk: Is the gst-plugin thing literally the only thing holding Qt up? [19:01] which is why I had said: [19:01] [08:37] slangasek: ok. fwiw I have not yet seen any confirmed reports of the alleged issue with upgrades to mesa 18 without the bugfix in -proposed Qt, and by its nature the bug should not in theory affect a fresh install from iso or the live session. so if Qt does not move, at the moment seems a 'this might be buggy on upgrading a current system' in beta release notes could just about be acceptable [19:02] acheronuk: If so, looks like the MIR is maybe sorted. Lemme dig deeper. [19:02] infinity: I can't 100% decipher update_output.txt, but if there is anything else, it's not obvious to me [19:02] ok [19:03] * infinity drops the freeze block again while sorting this. [19:04] thanks. I don't want to hold things up too much, but if that did drop into place now, that would be great [19:05] * acheronuk crosses fingers [19:11] Oh, hrm. That gst-plugins-good also accidentally pulls gst-qt5 back into main. I'll mangle that. [19:23] Alright, archive and seeds mangled. Will have to wait a bit to see how it all settles. [20:34] infinity: looks like gst-plugins-good rdep failing test here: https://autopkgtest.ubuntu.com/packages/ruby-gnome2/bionic/amd64 [20:35] :/ [20:36] but don't think the fault of gst-plugins-good [20:37] acheronuk: Indeed. I'll re-run a baseline on that. [20:37] Though, the baseline run 5 days ago looked good. [20:37] Oh, how quickly we break things. :/ [20:37] thanks [20:40] These are some curious tests. [20:41] It only installs rub2.3 sometimes. [20:41] And only sometimes does rub2.3 fail. [20:51] infinity: does this mean Qt would migrate? (from update_output_notest.txt) https://paste.ubuntu.com/p/dQ8xCZxJ5j/ [20:52] acheronuk: It does seem to imply that. [20:53] right. I have never quite grasped what that "Trying easy from autohinter:" is actually doing [20:55] Okay, hinted ruby-gnome2. Seems equally broken in the release pocket. :/ [20:57] ick. thanks though [20:58] acheronuk: The autohinter tries to group packages together based on interdeps and generates temporary "easy" hints. Easy hints are hints that say "try all these together, and if the result is that the archive is no worse than when you started, you win". [20:59] gotcha [21:00] There are three classes of grouping hints. hint-easy doesn't allow anything to regress, hint allows uninstallable trading (which is how KDE became briefly broken, cause vorlon did a 'hint' that ended up trading something else broken for your breakage instead), and force-hint, which allows you to regress the archive entirely and jam in broken bits. [21:12] infinity: are you or some other AA going to delete freeipa from proposed? i thought there was agreement to do that and whenever i chase anything complicated looking on excuses it comes down to freeipa blocking curl and libunistring [21:13] mwhudson: I'm not sure I was party to that discussion. [21:13] infinity: fair enough [21:16] I thought the freeipa issues were meant to be solved with the nss revert. I guess not? [21:17] oh maybe [21:17] mwhudson: Also, if deleting it were enough to make other things migrate, deleting it wouldn't be necessary. [21:18] mwhudson: So I assume you mean delete it and then upload a no-change rebuild of the release pocket version? [21:18] i find keeping track of this sort of issue basically impossible [21:18] infinity: yes, i think that is what i mean [21:19] when did the nss revert happen? [21:19] I could certainly do that and then undelete it. [21:19] The nss revert was in early February, so it clearly didn't fix the freeipa tests. :P [21:19] yeah [21:21] mwhudson: Deleted and rebuild thrown at the archive. [21:21] * infinity makes a mental note to undelete later. [21:24] infinity: thanks [21:26] infinity: well crap [21:26] mwhudson: Oh, but a no-change rebuild doesn't build. [21:26] infinity: throw it all into the sea? [21:26] what's DAL in this context [21:27] mwhudson: I'll put the other version back. And let Timo sort it out after the sprint and beta. [21:27] infinity: +1 [21:28] And done. [21:30] acheronuk: Qt looks migratey. [21:31] migratory? [21:31] bashfulrobot: You'll have Beta1 images after the archive settles from this Qt migration. [21:32] infinity: \o/ [21:32] thank you thank you thank you [21:34] Thanks! [21:40] nice! [21:40] infinity: You rule! [21:43] what's the process for getting a package to migrate when it drops a binary package? get an AA to delete the binary package from release? [21:44] * acheronuk slides beer tokens in infinity's direction [21:44] infinity: So we are looking good for the Thursday date? [21:49] mwhudson: iirc yeah. [21:50] infinity: The topic isn't clear, do y'all accept bitcoin? :P [21:51] bashfulrobot: Now that the transition seems to be done and over with, looks like it, yeah. [21:51] tsimonq2: The only fiat currencies I accept are USD and CAD. [21:51] infinity: USD and fake money? :P [21:51] * tsimonq2 runs [21:52] tsimonq2: And possibly a 124 Spider. [21:52] Hah :D [21:52] infinity: I thought you drove stick. :P [21:54] tsimonq2: Are you implying that in some weird part of the world you can't get a Fiat Spider with a manual transmission? [21:54] (You sure can here) [21:54] LOL [21:54] It's Canada, ofc you can. [21:54] stick is the way to go. Not that you can almost even buy a car with it anymore [21:54] Canada? [21:54] whaaaaa [21:54] ha ha [21:56] I can't even imagine driving an underpowered lawnmower like that if I couldn't downshift and hold it on the red line. [21:56] That's pretty much the *only* way to drive those go-karts. [21:57] :D [21:59] Wow, that's satisfying... https://i.imgur.com/nTd3A40.png [22:03] \o/ [22:04] :D [22:07] [M [22:07] F[M [22:07] (what) [22:15] tsimonq2: You just lost me there. ha [22:41] mwhudson, infinity: it's not about removing freeipa from -proposed, but from the release pocket. Which I would do, just as soon as I can figure out why britney claims abi-monitor and ros-simulators are uninstallable on amd64 [22:47] where is the vcs for our casper> [22:47] ? [22:47] slangasek: If it's on amd64-only, it's arch:all packages it's whining about, if that helps. [22:50] infinity: yeah, but ros-simulators only depends on ros-robot and gazebo9 [22:56] slangasek: And installing both of those pulls in 744 packages, so fair shot that there's a broken combo in there. [22:56] Oh, "only" 604 without recommends. [23:01] slangasek: Another hint is that updating curl3 alone *doesn't* break ros-simulators-dev or abi-monitor, but updating "curl3/7.58.0-2ubuntu2 xmltooling/1.6.4-1ubuntu2" together does. [23:02] ahahaha [23:04] slangasek: Did you desperately want this in for B1, or can we revisit Thursday? [23:04] (I don't really want to hold these folks up any longer unless you see the fix jumping out at you) [23:05] infinity: I don't think you should delay the beta; but also, per previous rants, I have adjusted the freeze hint to not freeze base system components for an opt-in milestone [23:07] slangasek: I think our rants on that topic will continue to be at odds until we just stop having opt-in milestones entirely. [23:08] (Honestly, I'd rather just tell flavours they can't have them than have them testing against moving targets because ermagerd can't wait two days to migrate stuff) [23:10] (I'd rather have them and wait two days for the things to migrate than have them taken away) [23:10] tsimonq2: Yes, Steve's telling you that combination isn't an option in his mind. [23:10] tsimonq2: Which leaves "no milestones" or "base changing during milestones". [23:11] tsimonq2: But we really should revisit the honest question of "what value do you get from milestones that you wouldn't get from dedicating one day a month to testing dailies?" [23:11] infinity: It may sound weird but I prefer the latter. [23:11] tsimonq2: Because, other than publishing them, which encourages people to install from stale ISOs (ick), that's all you're doing. [23:11] given how many dozens of hours have been spent trying to get this transition ready, and the risk of setbacks to the transition from stray uncoordinated uploads, I don't really have any sympathy for the view that not allowing a milestone freeze to trump the transition is unreasonable impatience [23:12] slangasek: I'm not disagreeing with you. [23:12] tsimonq2: I know /you're/ not, but infinity is :) [23:13] slangasek: Yeah, I absolutely am. Because you're using *this* transition to argue for a global policy. If intersecting packages aren't frozen, we're saying "yeah, don't worry, Qt, glibc, glib, and GTK all changed between your first spin and your third, but that probably won't set back your testing any". [23:13] Y'know, cause those are all possiblities in the intersecting base set. [23:14] infinity: I continue to believe that testing ISOs one day a month doesn't gather as much momentum as a "Beta" because that signals we're in a certain stage of the cycle. [23:14] slangasek: Then again, you and I agree (I think) that these early milestones are about as valuable as "pick a daily and test it", so what we have is a strangely unresolvable venn diagram of agreement between you, me, and the flavour leads. :) [23:14] heh, quite [23:14] Right, heh [23:14] tsimonq2: So, call your fourth month of daily testing "beta". [23:15] and those are all packages with CI via autopkgtests, and while there is always risk of regression, I don't accept the notion that freezing these packages to ensure the beta candidates are stable in between, only to have massive changes drop the day after beta, is actually valuable to anyone [23:15] (To be honest, I'd be happier if we did exactly what we do now, including with the ISO tracker, but then didn't release the result... It's the publishing that I greatly disagree with) [23:16] Milestones have a very negative effect socially, as people are more inclined to install the "tested" Alpha2 than the daily that comes 3 weeks later and fixed a bunch of bugs. [23:17] Or, as is often the case, the daily that comes two days later and fixed all the bugs found but not deemed RC enough during the ISO testing. :P [23:17] infinity: "ZOMGWTFROFLCOPTER, this beta wasn't tested at all, why did they let this thing release?!?" or "These flavors didn't release at the same time, what's the point?!?" are two possibilities. I don't personally care about having a set date, or an ISO that stays there for longer than the time it takes for the next ISO to release, for that matter. Just that we have those stages because I'm [23:17] told it's good for marketing. :P [23:19] Hell, I wouldn't even care if we did like a biweekly testing coordination thing and just eliminate what we have now. As long as we have stages and not just "prerelease" and "release" [23:20] infinity, slangasek: I mean, can we find common ground *there*? [23:21] tsimonq2: Coordinated testing is something I wouldn't even need to get involved in. [23:21] infinity: It depends on how we do it, but yeah. [23:21] tsimonq2: (Though I could put it on the release schedule for you, like I do with milestones today, so you don't bikeshed about the date) [23:22] tsimonq2: Then, on the same date that would have been Alpha2, instead you say "this is Month2 ISO testing week", and flavours flood the ISO tracker with test results, and action said results after the facgt. [23:22] tsimonq2: All that removes from the current state of affairs is the "blessed image", but my contention has always been that the blessed image does more harm than good. [23:23] And that's not just guessing, we field a lot of bugs from people who install from milestone media, just to close them with "yes, that bug was fixed three days later". [23:23] Super not helpful. [23:23] I'm open to trying it for 18.10 and then evaluating how things went from there. [23:24] So, consider me (somewhat) convinced. [23:25] infinity: Is this a decision the Release Team is "just going to make" or should I (or you! :P) send something to the ML? [23:26] tsimonq2: I mean, it's a decition that Canonical *could* "Just Make", as it's about our human resource involvement in the community process. But no, I'm not threatening to do that. [23:26] tsimonq2: But it's no secret that we've wanted to back away from this process for years, and we've been trying to gently (and sometimes not) argue in that direction, and lead by example. [23:27] decition? Did apw infect my fingers? [23:27] decision. [23:27] infinity: Now to be completely clear here, is this a Canonical decision or an Ubuntu Release Team (or cdimage team, whatever Ubuntu hat you'd like to wear) decision? [23:27] tsimonq2: The release team, as a community entity could absolutely make this decision (and that's where I'd prefer it be made). [23:27] If no resource allocation would change (less resources needed for a Blessed Image), then why would it need to be a Canonical decision? [23:27] RIght [23:27] *Right. [23:27] I agree. [23:28] tsimonq2: Cause then it's not "Canonical is taking away toys", but "the community has decided this is more sane". [23:28] That's exactly where I was going with that. [23:28] tsimonq2: The community release team can't force Canonical to spend time/money, but they sure can ask them to spend less. :) [23:28] Heh, true. [23:32] -queuebot:#ubuntu-release- Builds: Lubuntu Alternate amd64 [Bionic Beta 1] (20180306.1) has been added [23:32] -queuebot:#ubuntu-release- Builds: Lubuntu Alternate i386 [Bionic Beta 1] (20180306.1) has been added [23:33] tsimonq2: I think with something as complex as the Debian/Ubuntu archive (and moreso us, because we generate multiple flavours), no matter how good our CI gets (and it can get better), it's a pipe dream to be "always releasable", but we very much should be "always testable". [23:33] tsimonq2: So, having flag days where we say "okay, for the next two days, spend your time testing and fixing your accumulated cruft" helps keep things in shape, and that's really what milestones have been doing for us for years. [23:34] infinity: Right, and I can agree with that. [23:34] infinity: (I'm also curious to see what ideas you have for how Ubuntu's ProposedMigration CI can be improved. :P) [23:35] I don't have a bunch of stellar ideas, but it's obviously imperfect, thus can be improved. [23:39] That's for sure. [23:41] slangasek: aah ok [23:41] acheronuk: no vcs for ubuntu casper as far as i'm aware [23:42] acheronuk: and as it was me that made the last upload, if there is a vcs, it's out of date :-p [23:43] "no vcs" aww come on :/ [23:43] tsimonq2: ok, the canonical vcs is the ubuntu archive!! [23:44] err, something got busted because there definitely was one [23:44] but regardless, not that big a deal, the archive is your vcs. [23:45] mwhudson: Lowercase "c" I hope. :P [23:46] cyphermox: oh hmm [23:52] cyphermox: debian/control doesn't know about it, so there isn't one. [23:52] (Pretty sure I never committed to anything but the source package) [23:52] 2018/03/06 23:43:39.397222 cmd.go:156: cannot read /proc/self/exe: readlink [23:53] wut [23:53] * debian/control: Drop Vcs-Bzr:, which does not exist any more. [23:53] -- Martin Pitt Tue, 19 Apr 2016 15:53:13 +0200 [23:53] infinity: lp:casper seems to think otherwise, and I thought I used that before [23:53] cyphermox: Blame pitti, then. :) [23:56] yes [23:56] cyphermox: lp:casper trunk last changes 2010-09-28 ... I can see why, 6 years later, pitti decided it might be dead. [23:57] no, lp:ubuntu/casper was dea [23:57] but also grossly outdated, for sure [23:57] https://code.launchpad.net/casper [23:57] I saw. [23:57] That's lp:casper, not lp:ubuntu/casper [23:57] ah, it's the channel thing [23:57] yes [23:58] infinity: there was also a "comment" on that page. but whatever, this isn't really important [23:59] I think I need a big bowl o' vietnamese. [23:59] lucky.