[02:14] good morning [03:34] Morning callmepk [03:34] morning duflu [04:42] gnome-control-center ubuntu/master Marco Trevisan * [open] merge request !19: ubuntu panel and display: Monitor and fractional scaling fixes * https://deb.li/3mlwF [04:53] Awesome Trevinho [04:54] * duflu wonders if it's late night or early morning [05:02] duflu: well both :D [05:34] good morning === cpaelzer__ is now known as cpaelzer [06:04] Hi didrocks and seb128 [06:07] hey duflu [06:08] good morning desktoppers [06:09] Hi oSoMoN [06:23] salut oSoMoN [06:23] hey duflu, lut oSoMoN, didrocks, how are you doing today? [06:23] seb128, OK. Slightly less overwhelmed than yesterday but still keen on a weekend, and maybe a week off some time soon. Not sure... How are you? [06:28] I'm alright but days are buuusy and I struggle to do what I want, my todolist goes in the wrong direction as a result :/ [06:28] (and I'm on +1 maintainance today so probably going to get even more behind) [06:29] what are you working on atm? [06:29] seb128, now going deep into why 4K performance is poor (and CPU usage is low at the same time) [06:29] which is a general performance problem for everyone. Only I can see it better now [06:31] Close to my heart, because without a fix I would have to upgrade my desktops [06:36] good good :) [06:38] duflu, good luck figuring it out! [06:40] It's interesting. I don't remember ever having exhausted all CPU problems and being left with only GPU bottlenecks. Not since university at least [06:40] Though technically we didn't have GPUs back then. They were only graphics cards [06:44] hey duflu [06:44] salut didrocks, seb128 [07:47] fixing cryptsetup FTBFS which is in proposed for a week :/ [07:47] seems the proper install was only done for risc [07:48] weird, the .install list the missing files: scripts/cryptsetup.conf /usr/lib/tmpfiles.d/ [07:48] and it’s not arch-specific [07:54] dh_missing: warning: usr/lib/tmpfiles.d/cryptsetup.conf exists in debian/tmp but is not installed to anywhere [07:54] didrocks@casanier:/tmp/cryptsetup-2.3.3$ find . -name cryptsetup.conf [07:55] ./debian/cryptsetup-bin/usr/lib/tmpfiles.d/cryptsetup.conf [07:55] ./debian/tmp/usr/lib/tmpfiles.d/cryptsetup.conf [07:55] O_o oh dh_missing, what are you doing to us [07:55] (and no diff in the file) [07:59] ok, dh_missing is basing now on helper by the various tools [07:59] not a diff itself [08:00] and in debian/.debhelper/generated/cryptsetup-bin/installed-by-dh_install, dh_install is using the original file path [08:00] not the one moved in the .install [08:02] \O\ |O| /O/ [08:06] hey Laney [08:07] greetings didrocks [08:07] having fun? [08:07] yeah, didn’t know dh_missing changed behavior that way [08:10] hey Laney, how are you? [08:10] changing to usr/lib/tmpfiles.d/cryptsetup.conf works and now builds on all arch (which are not risc :p) [08:14] moin seb128 [08:14] i'm good, enjoying watching the rain [08:14] you? [08:15] thanks for looking at the team report thing [08:15] I changed that url to be compressed yesterday [08:15] Laney, I'm alright thanks! [08:17] Laney, np for the team report, now I need to try to understand the report changes, is the format of that file documented somewhere btw? [08:18] no [08:20] presumably the problem with update-excuse bugs is that now the yaml contains the verdict (PASS) for those and the code doesn't expect that [08:21] https://git.launchpad.net/~ubuntu-release/britney/+git/britney2-ubuntu/tree/britney2/policies/lpexcusebugs.py?h=ubuntu/rebased#n73 [08:23] right, the problem is https://bazaar.launchpad.net/~ubuntu-archive/ubuntu-archive-scripts/trunk/view/head:/generate-team-p-m#L260 [08:23] would it be better if that actually put the bugs in a named entry? [08:23] so it was like [08:23] update-excuse: [08:23] it expect item['policy_info']['autopkgtest'] to include a package name [08:23] bugs: [...] [08:23] verdict: PASS [08:23] ah, you are speaking about the bug problem there, sorry, I was on the exception one [08:24] ah [08:24] I didn't look at the bug problem yet, just saw it on the page generated when I did skip over the exception problem [08:26] hmm I don't see the problem with autopkgtest, it looks the same to me [08:27] well it also has the verdict too [08:31] * Laney asks upstream about that, feels more sensible to fix the structure of the yaml to me [08:31] autopkgtest: [08:31] policy_results: [08:31] foo/1: ... [08:31] verdict: PASS [08:31] or something [08:43] oh, logged in and forgot to say hi [08:43] morning desktoppers [08:44] hey marcustomlinson [08:46] Laney, not always, see l12248 from your yaml [08:46] autopkgtest: [08:46] verdict: REJECTED_TEMPORARILY [08:46] block: [08:46] the old file didn't have instances with no package listed [08:47] hey marcustomlinson, how are you? [08:48] hey didrocks and seb128, doing alright thanks, you? [08:49] marcustomlinson, I'm alright, a bit tired but coffee solved part of it :) [08:49] I feel you [08:50] marcustomlinson: good good, thx :) [08:52] seb128: what about it? [08:53] moin marcustomlinson [08:53] Laney, that autopkgtest section has 0 package entry, is that legit? [08:53] as said there is no case of this happening before your rebase [08:55] I'm just trying to check what you're saying [08:55] Laney, sorry, I might be not clear, the by-team report fails on [08:55] for package, results in sorted(item['policy_info']['autopkgtest'].items()): [08:56] basically it expect the policy_info/autopkgtest sections to have a least a package/result combo [08:56] there are cases of that, it's just that they were actually empty rather than having a verdict only [08:56] ah [08:56] k, so I just need to teach the code to deal with having a verdict only [08:56] morgen Laney [08:56] thx, I didn't notice that there were empty section, those files are big and a bit tricky to navigate [08:57] np [08:57] if upstream agrees I'll change that to be better [08:57] so this might not be the final final structure [08:57] Laney, I just found another issue [08:57] Laney, https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_excuses.html#dovecot [08:58] missing build on all: mail-stack-delivery (from 1:2.3.7.2-1ubuntu4) [08:58] arch:all not built yet, autopkgtest delayed [08:58] where the current report has [08:58] missing build on i386: mail-stack-delivery (from 1:2.3.7.2-1ubuntu4) [08:58] Not considered [08:58] [08:58] we don't have "arch all' builds in Ubuntu [08:58] so the new summary is more confusing imho [08:59] fair [08:59] also let me know if IRC comments are not the best way to give the feedback, maybe it's easier to track by replies on the list [09:00] or bug reports? [09:00] this is ok for now [09:01] k [09:01] bug reports once it's live [09:01] ah, now it does trigger tests on archs that built even if one is failing [09:01] that's nice :) [09:01] (though probably waste some resources if the build is failing and the item is never going to be candidate) [09:21] seriously https://launchpadlibrarian.net/484788335/buildlog_ubuntu-groovy-riscv64.cryptsetup_2%3A2.3.3-1ubuntu2_BUILDING.txt.gz [09:21] so the file is installed everywhere, but risk [09:21] risc* [09:31] O_O [09:33] that explained why it built on risc previously and why the hack was done just to build there [09:40] didrocks: at a quick glance: missing build-dep on systemd / passing of --with-tmpfilesdir=/usr/lib/tmpfiles.d/? [09:42] Laney: it seems that Steeve chose the option to directly copy from source, so to minimize his diff, I will do this conditionally before dh_install on riscv [09:43] I think it's weird to be implementing a solution just because systemd randomly isn't installed on that arch [09:44] Passing the configure flag would bypass that requirement [09:47] Laney: yeah, I agree, I wanted to minimize the diff, but if we have risc machine I can access by pushing on a ppa, I’m fine trying this ^ [09:49] I find it sad this package to be blocked in proposed for a week, only building on one arch, blocking our encryption work because it’s not a priority to fix it apparently [09:50] didrocks: BI LE TOOOOOOOOOOOOO [09:51] yeah that is a bit sad [09:51] I need to find how it works without a branch [09:51] yeah, giving the work to fix to other people [09:51] bileto? [09:51] 'build', get a ppa, then you can dput to it [09:52] let’s give it a try [09:55] it's like god coming back to the civilisation he created many generations later [09:55] 'you did WHAT?' [09:55] * Laney strokes the spreadsheet [09:55] heh :) [09:55] uploaded, let’s see [10:49] Laney, btw are the autopkgtest discussions happening on an IRC channel (asking so I could join and keep en eye on those :) [10:50] seb128: in ubuntu-release if there are any, or on the trello board https://trello.com/b/tjbSeNWo/autopkgtest, or sometimes the Launchpad project https://launchpad.net/auto-package-testing/ [11:05] Belated good morning to Laney and anyone else I missed [11:05] And good nigh [11:06] t [11:06] Oops. It's not even morning there. Yeah I am late for dinner [11:08] heh [11:23] Laney, I meant the upstream discussions you mentioned [11:24] oh [11:24] #debian-release [11:24] or sometimes #debci [11:24] thx [11:42] gnome-control-center tags 1a911d5 Sebastien Bacher upstream/3.36.3 * Upstream version 3.36.3 * https://deb.li/8SzD [11:43] gnome-control-center pristine-tar 239a935 Sebastien Bacher gnome-control-center_3.36.3.orig.tar.xz.delta gnome-control-center_3.36.3.orig.tar.xz.id * pristine-tar data for gnome-control-center_3.36.3.orig.tar.xz * https://deb.li/3osUz [11:43] gnome-control-center upstream/latest 6373b71 Sebastien Bacher * pushed 18 commits * https://deb.li/miay [11:53] gnome-control-center signed tags 02eba19 Sebastien Bacher ubuntu/1%3.36.3-1ubuntu1 * gnome-control-center Debian release 1:3.36.3-1ubuntu1 * https://deb.li/3fsjg [11:53] gnome-control-center ubuntu/master 41950aa Sebastien Bacher * pushed 30 commits (first 5 follow) * https://deb.li/MYuN [11:53] gnome-control-center ubuntu/master 27e6b16 Cheng-Chia Tseng po/zh_TW.po * Update Chinese (Taiwan) translation * https://deb.li/ikiRx [11:53] gnome-control-center ubuntu/master 01b6985 Yi-Jyun Pan po/zh_TW.po * Update Chinese (Taiwan) translation * https://deb.li/7eUv [11:54] gnome-control-center ubuntu/master 4f53ea3 Yi-Jyun Pan po/zh_TW.po * Update Chinese (Taiwan) translation * https://deb.li/3ri1n [11:54] gnome-control-center ubuntu/master fdbbc2f Marco Trevisan (Treviño) debian/patches/0031-fingerprint-dialog-Don-t-limit-the-number-of-maximum.patch * account-fingerprint: Don't make flowbox children selectable * https://deb.li/asFs [11:54] gnome-control-center ubuntu/master 88bd021 sicklylife po/ja.po * Update Japanese translation * https://deb.li/3M7TU [11:54] gnome-control-center ubuntu/master Sebastien Bacher * [merge] merge request !19: ubuntu panel and display: Monitor and fractional scaling fixes * https://deb.li/3mlwF [11:56] seb128: mergin! [11:56] Trevinho, hey, good morning :) thanks for the fixes! [11:56] unfortunately they were not one-liners.... but well. [12:03] ricotz, FYI, I successfully built and used nodejs 12.18.0 to build firefox 78.0~b8 (groovy) in https://launchpad.net/~osomon/+archive/ubuntu/nodejs-12.18-groovy/+packages, will try to get it sponsored [12:07] glib ubuntu/focal c220732 Iain Lane debian/ control control.in gbp.conf * control, gbp.conf: Use the ubuntu/focal branch * https://deb.li/3kSd0 [12:07] glib ubuntu/focal 4ae4943 Iain Lane debian/changelog * Update changelog * https://deb.li/iWs32 [12:07] glib ubuntu/focal 8dc62bf Iain Lane debian/changelog * Finalise changelog * https://deb.li/peOC [12:07] glib ubuntu/focal e554968 Iain Lane debian/gbp.conf * gbp.conf: Add the ubuntu tag pattern too * https://deb.li/3jimG [12:07] glib signed tags ff12c4f Iain Lane ubuntu/2.64.3-1_ubuntu20.04.1 * glib2.0 Debian release 2.64.3-1~ubuntu20.04.1 * https://deb.li/1sPc [13:56] Laney: did you have time to look into ubiquity subtree/squash/submodules? i feel like merging subtree version, and just rolling with it. [13:57] Laney: if it becomes tricky, we can rewrite history and do something else later. [13:57] xnox: will do in a minute, subtree not the others [13:57] unless subtree is total shit but I'm sure it's not [13:58] Laney: ack. [13:58] Laney: clone it, build it, check git log -p in a d-i/sources/$project, check git log -p from top level, check gitk => it should give you a good feel of what it feels like. [13:59] (especially what the history looks like) === alan_g is now known as alan_g_ === alan_g_ is now known as alan_g [15:55] xnox: sorry, meetinghs, looking now! [16:04] xnox: are you imagining people just commit to the d-i stuff in there or what? [16:05] Laney: they can, but they don't have to. [16:06] Laney: i.e. when merging from debian it might be easier to do it from lp:$project into lp:~ubuntu-installer/$project, and then subtree pull from lp:~ubuntu-installer/$project [16:06] right [16:06] Laney: but when patching in new ubuntu delta => it can be easier to do "commit d-i add foo api", "commit ubiquity use foo api" [16:06] Laney: in that case a merge proposal will modify things inline. [16:07] only easier up until the next merge from debian [16:07] that [16:07] on the other hand if you could do the merge right in there ... [16:07] but i also worry that salsa will move to tango, lp git <-> git salsa mirrors stop working, and we loose history. [16:07] Laney: so merging right there should work. [16:07] Laney: from top level of ubiqutiy $ git pull salsa:partman-base => should do the right thing [16:08] but i'm not sure if "rebase" will work. [16:08] might be good to write some quick recipes in debian/readme.how-to-work-this-thing [16:08] (git will realise that all of the partman-base is now under d-i/source/partman-base and merge things into there. [16:08] ) [16:08] Laney: well ideally, we would have had the new installer by now right =) [16:09] more readmes yes => but i was hoping to right them, as I go along and figure out the ropes of things. [16:09] you might say that [16:09] ok [16:09] make an empty one that says 'if you do a new thing, write it in here, so that the next person can copy what you did' [16:09] ack [16:09] the history of this is super weird [16:10] like git-ubuntu stuff and upstream stuff [16:10] yeah [16:10] cause we have some "ubuntu-only" projects, and i had to take something for them. [16:10] git-ubuntu is like the 'delta' I guess [16:10] unless you want to kill wubi support => because i think the partman-loop is the weird one. [16:11] yeah, i mostly took d-i upstream git repo, upto the current merge-base with ubuntu, and slapped a single commit of ubuntu delta on top of that, and then subtree merge that [16:12] d/rules update can go no? [16:12] it still generates debian/control B-D from d-i/sources/*/debian/control [16:12] well the get-sources stuff [16:12] (ie. see https://git.launchpad.net/~ubuntu-installer/partman-base/log/ ) [16:12] Laney: it's a no-op [16:13] i think i did patch the script that tries to update => but it does regenerate debian/control each time. [16:13] in some ways git submodules is easier to understand, cause it says that path over there is git repo lp:~ubuntu-installer/$project => go edit that. [16:14] but then it's submodules, and 20 repos. [16:17] right, neither way is perfect [16:17] pyflakes test fails btw :> [16:22] Yeah fixed in master [16:23] merge ittttt [16:29] kenvandine, hey, maybe you know, how can you set wayland as the default for a user in the greeter? so it is set on first boot? [16:31] not sure about as the default [16:31] abeato: you just need to select the wayland session for the user in gdm [16:31] abeato: and it'll remember that until you change it again [16:32] at first boot is probably a session file [16:41] kenvandine, right, but this is for a project where they want auto-login, so I actually need to know to do this by changing a config file [16:42] Laney: oh btw, your suggestion for cryptsetup works very well, thanks! [16:42] kenvandine, but I do not know where [16:42] 😘 [16:42] abeato: Laney probably knows [16:51] afaik gdm calculates the default session by itself, i don't know how to overrride it (if ou can) [16:59] Laney: "merge ittttt" just the pyflakes fixes into submodules/subtree/squash, or do you mean switching to subtrees? [17:05] xnox: I just approved the MP (but with a comment about documentation) [17:05] now I'm off until Monday, have fun burning it down [17:05] o/