[03:26] vorlon: the generated ISO for Ubuntu Unity doesn't seem to boot (either kernel panics, or gets stuck on 'Mounting POSIX Message Queue File System...') [06:08] vorlon, yes, it happens every ghc migration [06:08] too much to be copied in one shot? [06:08] shouldn't britney block during publishing? [06:27] RAOF: hey my local friendly SRU team member :) - could you please sponsor intel-microcode from bionic/focal/jammy-proposed to -updates for LP: #1984166? thanks - once it is fully phased in -updates then I will copy it to -security [06:27] Launchpad bug 1984166 in intel-microcode (Ubuntu Jammy) "Update to latest upstream 20220809 to fix CVE-2022-21233" [Undecided, Fix Committed] https://launchpad.net/bugs/1984166 === guiverc2 is now known as guiverc [07:58] rs2009: The Ubuntu Unity daily ISO booted up just fine for me... into a GNOME shell. [09:01] * LocutusOfBorg is syncing some node-* packages mainly bugfixes and autopkgtests fixes [09:01] and compatibility with newer nodejs fixes [09:10] -queuebot:#ubuntu-release- Unapproved: accepted mypy [source] (jammy-proposed) [0.942-1ubuntu1] [09:11] hello SRU team please have a look at virtualbox in jammy/focal queue [09:11] the current version is breaking images!!! [09:11] its critical issue with newer kernel [09:30] o/ [09:31] ON IT! (in a moment) [09:32] -queuebot:#ubuntu-release- Unapproved: accepted loguru [source] (jammy-proposed) [0.6.0-1ubuntu1] [09:35] <3 [09:51] LocutusOfBorg: what tests do you usually perform on virtualbox new upstream releases validation? [09:52] start vm, check dkms modules build, check if guest addition installs, check ext pack installation [09:52] ask the people who reported bug to do the same [09:52] and using virtualbox daily :) [09:53] most important, monitoring new bugs, since the package is widely used [09:55] Alright! The GNOME MRE script is now ready for anyone who feels like reviewing it, or its output: https://discourse.ubuntu.com/t/scope-of-gnome-mru/18041/50?u=raof [09:56] Also, if anyone feels like a quick FFe review, https://bugs.launchpad.net/ubuntu/+source/mir/+bug/1988702 should be pleasantly easy ;) [09:56] Launchpad bug 1988702 in mir (Ubuntu) "FFe: Upstream 2.9.0 release" [Undecided, New] [09:56] LocutusOfBorg: can you do the same for the new version after I accept it? And add info about that in the SRU bug Test Case ;) [09:57] RAOF: looking [09:59] RAOF: approved o/ [10:02] :) thanks [10:02] sure [10:02] sil2100: 🎉 [10:04] -queuebot:#ubuntu-release- Unapproved: accepted virtualbox [source] (jammy-proposed) [6.1.38-dfsg-3~ubuntu1.22.04.1] [10:15] -queuebot:#ubuntu-release- Unapproved: accepted virtualbox-ext-pack [source] (jammy-proposed) [6.1.38-1~ubuntu1.22.04.1] [10:15] -queuebot:#ubuntu-release- Unapproved: lttng-modules (bionic-proposed/universe) [2.10.8-1ubuntu2~18.04.3 => 2.10.8-1ubuntu2~18.04.4] (kernel-dkms) [10:17] -queuebot:#ubuntu-release- Unapproved: accepted virtualbox-guest-additions-iso [source] (jammy-proposed) [6.1.38-1~ubuntu1.22.04.1] [10:19] -queuebot:#ubuntu-release- Unapproved: accepted virtualbox-hwe [source] (jammy-proposed) [6.1.38-dfsg-3~ubuntu1.22.04.1] [10:28] -queuebot:#ubuntu-release- Unapproved: accepted virtualbox [source] (focal-proposed) [6.1.38-dfsg-3~ubuntu1.20.04.1] [11:54] sil2100: was wondering what the blacklist file is for in a seed (is it the list of packages to be removed during the ISO build after installing the packages in the desktop file?) [12:11] arraybolt3: ah, that's because gnome-session too installed due to it being present in the depends/recommends of some of the packages present in the Ubuntu Unity seed [12:12] rs2009: I previously tried installing Unity on top of Ubuntu 22.04 and 22.10 and I was unable to log in to Unity on either version. Either a missing dependency or something essential was crashing [12:13] jbicha: that's because unity also requires dbus-x11, which isn't installed on Ubuntu by default [12:13] (I'm adding it to the new seed) [12:13] I'm doing some checks, and syncing stuff that is stuck in proposed due to build failures/autopkgtest failures [12:13] and syncing whenever it might be useful, at the worst scenario, we don't make migration status worse [12:14] rs2009: if it's required by Unity, then Unity or unity-session should depend on it. People sometimes install desktops without installing the big metapackage [12:14] jbicha: ah, I'll add it to Unity7's depends too. thanks for the suggestion! [12:15] rs2009: you can try adding things to a blacklist file. It looks like the Ubuntu Budgie seed has a few extra things at the bottom of its blacklist file [12:16] jbicha: yep, that's why I was asking what the blacklist file is for (I added the GNOME packages to it just now) [12:16] I think it works by helping germinate pick the right alternative dependencies. It is less powerful than people would think but it can still help [12:17] we should hopefully have a completely functional Ubuntu Unity ISO in today's daily build (apart from the unity package, which Trevinho had some feedback for) [12:18] jbicha: the GNOME packages only get installed because of recommends, not depends, so that shouldn't be a problem either [12:18] it may require several tries to get things correct :) [12:19] ack :) [12:20] I'll probably ask v0rlon or sil_2100 (don't want to ping them) to manually trigger the UU ISO build manually multiple times today to work on these issues [12:21] they might be able to get you access to that rebuild ISO button [12:22] jbicha: yep, that would be great until I get a proper build env set up locally [12:24] I thought that after adjusting the seed, you'd have to wait for germinate to update, but I don't see Unity at https://people.canonical.com/~ubuntu-archive/germinate-output/ [12:25] that germinate page is very useful. It tries to explain what is being installed and why (and then there are lots of other relationships on that page like build-depends that don't matter for what you need) [12:28] there's a little documentation at https://wiki.ubuntu.com/Germinate [12:28] See also "man germinate" [12:28] The blacklist file is almost never what you want for anything at all [12:29] At best it might be able to cause seed expansion to fail rather than picking something undesired [12:29] But you should mostly regard it as a historical relic and ignore it [12:30] As "man germinate" says: "The purpose of a blacklist is to make it obvious when a package that is not supposed to be installed ends up in germinate's output, so that package relationships can be fixed to stop that happening. It is not intended for the purpose of working around buggy package relationships, and attempts to do so will not work because apt has no way to know about blacklist entries [12:30] in seeds." [12:30] cjwatson: ic, is there any other alternative for removing packages installed by recommends of packages in a seed? [12:30] The far superior way to pick the right alternative dependencies is to seed then [12:30] *themn [12:30] argh whatever [12:31] the GNOME packages actually aren't depends, just recommends of packages like nautilus-data [12:31] You can have a seed that has the no-follow-recommends feature set, but beware that that needs to be carefully synchronized with all tools that build images, notably livecd-rootfs [12:31] Seeds are just simply not normally the right place to do this [12:32] no-follow-recommends is probably a bad idea too [12:33] It's absolutely vital to understand that germinate is essentially an apt simulator. If you set up unusual situations in seeds then it becomes impossible for germinate to simulate apt correctly [12:33] So this sort of thing normally has to be resolved in collaboration with the people who maintain the relevant packages [12:33] I believe Lubuntu used the no-follow-recommends option years ago, but it was a lot of work [12:34] Or maaaaaybe sometimes by supplying alternative packages that Provides: the things you don't want to be installed, and arranging for those to be seeded instead (so that they end up in metapackages) [12:34] But that may not always be appropriate [12:36] cjwatson and jbicha: I think I could try using no-follow-recommends temporarily [12:42] rs2009: I do not believe setting no-follow-recommends temporarily will help [12:43] rs2009: I think one of your problems is that ubiquity-frontend-gtk depends on gnome-shell. Maybe if you depend on metacity for now, you can get metacity instead [12:44] jbicha: what about blacklist for now? [12:44] you might not need blacklist since it's not nearly as useful as it sounds [12:45] I'm aware of ubiquity-frontend-gtk, but we're okay with gnome-shell being installed for now, just not gnome-session [12:45] Blacklist for now will probably do precisely nothing. Seriously. [12:45] It certainly won't have the effect of causing apt not to install the package when building your images. [12:46] -queuebot:#ubuntu-release- Unapproved: systemd (focal-proposed/main) [245.4-4ubuntu3.18 => 245.4-4ubuntu3.19] (core, i386-whitelist) [12:46] At absolute most it will confuse you by causing the package in question not to show up in germinate output even though it shows up in images anyway, which would be counterproductive [12:46] rs2009: having gnome-shell installed will cause you problems! it recommends gdm3 and you don't want gdm3 running and as the gnome-shell maintainer, I'm not going to add an alternate on unity-greeter [12:47] sil2100, sorry forgot some focal virtualbox packages? [12:49] jbicha: I mean, we are aware of this, and we got around it in the 22.04 by installing budgie-core as I recall there being some problems with metacity being installed. I'll take a look at adding metacity to the seed again [12:54] -queuebot:#ubuntu-release- New binary: mir [amd64] (kinetic-proposed/main) [2.9.0-0ubuntu1] (ubuntu-desktop) [12:57] -queuebot:#ubuntu-release- New binary: mir [s390x] (kinetic-proposed/main) [2.9.0-0ubuntu1] (ubuntu-desktop) [12:58] LocutusOfBorg: yeah, still on it, got distracted [12:58] It's on my list tho! [13:14] -queuebot:#ubuntu-release- New sync: llvm-toolchain-15 (kinetic-proposed/primary) [1:15.0.0-1~exp1] [13:14] -queuebot:#ubuntu-release- New: rejected llvm-toolchain-15 [sync] (kinetic-proposed) [1:15.0.0-1~exp1] [13:14] -queuebot:#ubuntu-release- New: accepted mir [amd64] (kinetic-proposed) [2.9.0-0ubuntu1] [13:14] -queuebot:#ubuntu-release- New: accepted roger-router [riscv64] (kinetic-proposed) [2.4.2-3] [13:14] -queuebot:#ubuntu-release- New: accepted llvm-toolchain-15 [sync] (kinetic-proposed) [1:15.0.0-1~exp1] [13:14] -queuebot:#ubuntu-release- New: accepted mir [s390x] (kinetic-proposed) [2.9.0-0ubuntu1] [13:16] -queuebot:#ubuntu-release- New binary: golang-github-iovisor-gobpf [amd64] (kinetic-proposed/universe) [0.2.0-4] (no packageset) [13:29] -queuebot:#ubuntu-release- New binary: mir [armhf] (kinetic-proposed/main) [2.9.0-0ubuntu1] (ubuntu-desktop) [13:37] -queuebot:#ubuntu-release- Unapproved: libreoffice (jammy-proposed/main) [1:7.3.5-0ubuntu0.22.04.1 => 1:7.3.6-0ubuntu0.22.04.1] (ubuntu-desktop) [13:44] sil2100, hello :), please take a look at libreoffice in jammy/queue when time permits - https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1988744 [13:44] Launchpad bug 1988744 in libreoffice (Ubuntu Jammy) "[SRU] libreoffice 7.3.6 for jammy" [High, In Progress] [13:46] -queuebot:#ubuntu-release- Unapproved: accepted virtualbox-guest-additions-iso [source] (focal-proposed) [6.1.38-1~ubuntu1.20.04.1] [13:47] -queuebot:#ubuntu-release- Unapproved: accepted virtualbox-ext-pack [source] (focal-proposed) [6.1.38-1~ubuntu1.20.04.1] [13:48] -queuebot:#ubuntu-release- Unapproved: accepted virtualbox-hwe [source] (focal-proposed) [6.1.38-dfsg-3~ubuntu1.20.04.1] [13:55] ricotz: I'm reviewing it right now, a quick question in the SRU bug: https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1988744/comments/1 [13:55] Launchpad bug 1988744 in libreoffice (Ubuntu Jammy) "[SRU] libreoffice 7.3.6 for jammy" [High, In Progress] [14:06] sil2100, commented [14:09] sil2100, this could be an intermediate problem in Debian following every 7.3.x-x update, basically libreoffice-common "lost" the ownership of the directory leading to its removal in certain conditions [14:11] sil2100, fyi https://git.launchpad.net/~libreoffice/ubuntu/+source/libreoffice/log/?h=wip/jammy-7.3 [14:15] (updated the count of bug fixes) [14:18] -queuebot:#ubuntu-release- Unapproved: s390-tools (jammy-proposed/main) [2.20.0-0ubuntu3.1 => 2.20.0-0ubuntu3.2] (core) [14:20] -queuebot:#ubuntu-release- Unapproved: accepted libreoffice [source] (jammy-proposed) [1:7.3.6-0ubuntu0.22.04.1] === dbungert1 is now known as dbungert [14:23] -queuebot:#ubuntu-release- Unapproved: s390-tools-signed (jammy-proposed/main) [2.20.0-0ubuntu3.1 => 2.20.0-0ubuntu3.2] (core) [14:25] sil2100, thank you :) [14:27] vorlon, https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/costamagnagianfranco-ppa/+build/24346068 [14:27] Finished 14 minutes ago (took 4 hours, 30 minutes, 56.5 seconds) [14:27] what do you think about? :) === schopin_ is now known as schopin [15:43] LocutusOfBorg: what does #debian-haskell think of these changes? [15:43] (I have no idea what "unregistered" is) [15:44] sil2100: There's a gnome-tweaks SRU thats been opene for a little longer than a week. Could you please take a look when you get a chance? https://bugs.launchpad.net/ubuntu/+source/gnome-tweaks/+bug/1988276 [15:44] unregistered is the "not optimized version" of the compiler [15:44] Launchpad bug 1988276 in gnome-tweaks (Ubuntu Jammy) "GTK3 is not considered legacy in Ubuntu yet" [Undecided, New] [15:44] sil2100: bugging you because you're listed as the vanguard today [15:44] something "best effort" that works when an architecture is not officially supported [15:44] at least from what I understand [15:45] hellsworth: o/ [15:45] Sure [15:45] thanks! [15:50] schopin: librsvg build failure on i386 is a regression in rustc, the failing tests also exist in the release version of the package and it fails to build per https://people.canonical.com/~ginggs/ftbfs-report/test-rebuild-20220617-kinetic-kinetic.html . How do we handle rustc floating point handling having gotten slopper on i386? [15:51] it's arguably ok for us to ignore the fp errors on i386 but I also don't know how to do that for i386-only in this test framework [15:57] I'm investigating the amd64 queue back log [16:00] jbicha: I made a few changes to the seed to eliminate all the broken GNOME packages (other than nautilus, as Trevinho needs to approve all the unity 7.6 changes before we can go ahead with uploading a new version of unity without nautilus as a Recommends package) [16:00] would it be possible to update the ubuntu-unity-meta package in universe with the updated seeds? [16:03] vorlon: first, the easy answer: to ignore a specific test that uses the standard Rust testing framework (which seems the case in librsvg), replace the #[test] directive by #[cfg_attr(not(target_arch = "x86"), test)] on the test [16:03] schopin: will target_arch = "x86" match only i386, not amd64? [16:03] Yes, I just checked :) [16:03] perfect [16:04] But more globally, we're basically hitting a regression in rustc, right? [16:05] We should at least investigate to see if it's been reported upstream, because we might have other software silently breaking because of it :/ [16:11] Assuming we find a fix *and* it's in rustc (could be in LLVM if only found in Jammy & Kinetic), we can rebuild packages that have i386 binaries with Built-Using: the bad version of rustc [16:11] schopin: sure, probably worth reporting upstream, but I doubt we care much about fp accuracy on i386 in packages built with rustc: https://paste.ubuntu.com/p/fTKYC78MwV/ [16:12] basically it's just librsvg and mozjs [16:12] fair point :P [16:12] schopin: also, well, it looks like seb128 already uploaded a patch to ignore the test failures on i386 [16:12] "would it be possible to update the ubuntu-unity-meta package in universe with the updated Ubuntu Unity seeds?" [16:12] ^ vorlon [16:14] (we should have an almost fully-functioning ISO today if this gets uploaded) [16:17] once the unity package is approved by Trevinho and uploaded to universe (the last package to be uploaded), we would be ready for the beta [16:17] rs2009: updating now [16:18] vorlon: thanks! [18:04] vorlon, [18:04] The term "unregisterised" really means "compile via vanilla C", [18:04] disabling some of the platform-specific tricks that GHC normally uses to [18:04] make programs go faster. When compiling unregisterised, GHC simply [18:04] generates a C file which is compiled via gcc. [18:04] from ./docs/users_guide/codegens.rst [18:05] When GHC is build in unregisterised mode only the LLVM and C code [18:05] generators will be available. The native code generator won't be. LLVM [18:05] usually offers a substantial performance benefit over the C backend in [18:05] unregisterised mode. [18:21] LocutusOfBorg: so it gets faster if we *disable* ghc's attempts at optimization? :P [18:22] wow the list of non-existent packages that librust-rust-embed-dev depends on is just ridiculous [18:36] rust packaging is chaotic. The -dev packages provide long lists of virtual packages. And then those virtual packages shown up in Build-Depends and I'm unaware of good tools to understand it all [18:37] building random things and uploading them to the archive without first verifying that the built packages are installable sounds like a maintainer problem, not a tooling problem [18:37] vorlon: oh, I'll sync rust-rust-embed 6.4.0-2 which might help [18:37] jbicha: I've removed it from -proposed, no need to manually sync now [18:38] random broken rust package can wait until the next release [18:38] got it [19:37] hello, kindly pinging regarding the FFe bug for sssd: https://bugs.launchpad.net/ubuntu/+source/sssd/+bug/1988615 [19:37] Launchpad bug 1988615 in sssd (Ubuntu) "[FFe] Merge sssd 2.7.3-2 from Debian unstable" [Undecided, New] === arraybolt3_ is now known as arraybolt3 [20:21] -queuebot:#ubuntu-release- New binary: ormar [amd64] (kinetic-proposed/universe) [0.11.3-1] (no packageset) [20:22] -queuebot:#ubuntu-release- New sync: rust-crdts (kinetic-proposed/primary) [7.2.0+dfsg-1] [20:27] vorlon, I put ghc on kinetic with the armhf unregistered mode [20:28] lets see how bad is the rc debian bug [20:33] so you can remove the ghc sync block now [20:34] -queuebot:#ubuntu-release- New binary: onetbb [i386] (kinetic-proposed/universe) [2021.5.0-15] (i386-whitelist) [20:41] -queuebot:#ubuntu-release- New binary: onetbb [amd64] (kinetic-proposed/universe) [2021.5.0-15] (i386-whitelist) [20:42] -queuebot:#ubuntu-release- New binary: onetbb [armhf] (kinetic-proposed/universe) [2021.5.0-15] (i386-whitelist) [20:46] vorlon: regarding nrepl-clojure the switch from arch all to amd64 and arm64 only and so the armhf and ppc64el regressions are not really regressions. Should those be hinted given that arm64 is a regression? [20:48] bdmurray: I think they should be hinted as the tests will never pass without binaries on those architectures; the arm64 will block it still [20:48] -queuebot:#ubuntu-release- New binary: onetbb [arm64] (kinetic-proposed/universe) [2021.5.0-15] (i386-whitelist) [20:49] vorlon: ack, thanks [20:51] -queuebot:#ubuntu-release- New: accepted onetbb [arm64] (kinetic-proposed) [2021.5.0-15] [20:51] -queuebot:#ubuntu-release- New: accepted onetbb [armhf] (kinetic-proposed) [2021.5.0-15] [20:51] -queuebot:#ubuntu-release- New: accepted golang-github-iovisor-gobpf [amd64] (kinetic-proposed) [0.2.0-4] [20:51] -queuebot:#ubuntu-release- New: accepted onetbb [i386] (kinetic-proposed) [2021.5.0-15] [20:51] -queuebot:#ubuntu-release- New: accepted onetbb [amd64] (kinetic-proposed) [2021.5.0-15] [20:51] -queuebot:#ubuntu-release- New: accepted ormar [amd64] (kinetic-proposed) [0.11.3-1] [20:55] -queuebot:#ubuntu-release- New binary: onetbb [s390x] (kinetic-proposed/universe) [2021.5.0-15] (i386-whitelist) [20:56] # dpkg -S /usr/lib/powerpc64le-linux-gnu/libhs.so.5.4.7 [20:56] libvectorscan5:ppc64el: /usr/lib/powerpc64le-linux-gnu/libhs.so.5.4.7 [20:56] # [20:56] right, perfectly normal library naming [20:57] as opposed to amd64 where it's provided by libhyperscan5. :P [20:58] vorlon: to be clear what did you mean by "hint"? force-badtest or something else? [20:58] bdmurray: yeah, badtest [21:06] vorlon: I'm looking at ~ubuntu-core-dev +invitations and see invites from ubuntucinnamon-dev and ubuntu-unity-devs. Should those be for ~ubuntu-core-dev or ~motu? === arraybolt3_ is now known as arraybolt3 [21:29] bdmurray: ~ubuntu-motu [21:42] vorlon: ack, FYI it's really ~motu [21:42] bdmurray: right, sorry [21:44] seb128: what's this invitation for ~ubuntu-core-dev to join ~bluetooth about? [21:46] bdmurray: bluez is maintained there https://code.launchpad.net/~bluetooth/bluez/+git/bluez [21:47] jbicha: is the purpose of that instead of one of the ubuntu-dev teams to grant morphis commit access despite not being an Ubuntu Developer? [21:50] I don't know why it was done that way. duflu has been maintaining bluez and isn't a core dev [23:47] -queuebot:#ubuntu-release- New: accepted onetbb [s390x] (kinetic-proposed) [2021.5.0-15]