| -queuebot:#ubuntu-release- New binary: fenics-ffcx [amd64] (jammy-proposed/universe) [1:0.3.0-3] (no packageset) | 00:40 | |
| bdmurray | mwhudson: can you get the smooth_updates conversation restarted? | 08:40 |
|---|---|---|
| -queuebot:#ubuntu-release- Unapproved: gst-plugins-bad1.0 (focal-proposed/universe) [1.16.2-2.1ubuntu1 => 1.16.3-0ubuntu1] (i386-whitelist, kubuntu) | 09:00 | |
| -queuebot:#ubuntu-release- Unapproved: simplestreams (bionic-backports/main) [0.1.0~bzr460-0ubuntu1.1 => 0.1.0-48-gb936edd4-0ubuntu1~bpo18.04.1] (ubuntu-cloud, ubuntu-server) | 09:03 | |
| -queuebot:#ubuntu-release- New: accepted fenics-ffcx [amd64] (jammy-proposed) [1:0.3.0-3] | 09:59 | |
| rbasak | sil2100: have you seen this failing oem-meta-packageset-sync cronjob on snakefruit spamming devel-permissions@? | 11:00 |
| rbasak | sil2100: could you disable it for now please? | 11:00 |
| bdmurray | rbasak: I'll let him know when I see him | 11:21 |
| holmanb | not sure if there is time in SRU vanguard today or not. cloud-init has a queued for -proposed (22.1-14-g2e17a0d6) that we'd like to request review for upload into (bionic|focal|impish)-proposed per https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1961446 | 11:26 |
| ubottu | Launchpad bug 1961446 in cloud-init (Ubuntu Impish) "sru cloud-init (22.1 update) Bionic, Focal, Impish" [Undecided, In Progress] | 11:26 |
| holmanb | Please also reject the prior unaccepted cloud-init 22.1-1 uploads that are queued for bionic, focal and impish. | 11:26 |
| bdmurray | holmanb: Is it a priority to get it done this week? A couple of SRU team members are at a sprint. | 11:29 |
| bdmurray | holmanb: If it is a priority you might try pinging RA_OF | 11:31 |
| holmanb | bdmurray: it will push back the start of our aging to Monday if it doesn't go in today. It wouldn't be the worst thing, but sooner is better. | 11:38 |
| -queuebot:#ubuntu-release- New binary: fenics-dolfinx [ppc64el] (jammy-proposed/universe) [1:0.3.0-13] (no packageset) | 11:50 | |
| laney | rbasak: can you fix the owner of https://people.canonical.com/~ubuntu-archive/packagesets/jammy/canonical-oem-metapackages to be ~ubuntu-archive? I wonder why it got created with a different owner - does the initialisation process need to be fixed? | 12:11 |
| laney | ah, I see you already noticed the spam | 12:12 |
| -queuebot:#ubuntu-release- New binary: fenics-dolfinx [s390x] (jammy-proposed/universe) [1:0.3.0-13] (no packageset) | 12:39 | |
| sil2100 | rbasak: hey! You were pinging about the oem-meta-packageset-sync, right? | 12:51 |
| rbasak | o/ | 12:51 |
| rbasak | Yes, but laney suggests what's wrong above. I'm just looking into fixing that now. | 12:52 |
| sil2100 | I'll take a look, I think the creds might have expired. Will check up soon | 12:52 |
| -queuebot:#ubuntu-release- New binary: fenics-dolfinx [armhf] (jammy-proposed/universe) [1:0.3.0-13] (no packageset) | 12:52 | |
| sil2100 | Oh, ok o/ | 12:52 |
| sil2100 | No backlog here! | 12:52 |
| sil2100 | If you need any help or anything, just give me a poke. I'll get back to sprinting madness then | 12:55 |
| rbasak | Thanks | 12:55 |
| rbasak | OK owner changed to ~ubuntu-archive | 12:55 |
| rbasak | Thanks laney! | 12:56 |
| rbasak | Let's see if that fixes it. | 12:56 |
| rbasak | FTR: | 12:58 |
| rbasak | pkgset = lp.packagesets.getByName(name='canonical-oem-metapackages', distroseries=lp.distributions['ubuntu'].getSeries(name_or_version='jammy')) | 12:58 |
| rbasak | pkgset.owner = lp.people['ubuntu-archive'] | 12:58 |
| rbasak | pkgset.lp_save() | 12:58 |
| laney | merci rbasak | 12:59 |
| laney | you should find out in 45 seconds I guess :-) | 12:59 |
| rbasak | Ah just in time :) | 12:59 |
| laney | Worked 👍 | 13:00 |
| rbasak | \o/ | 13:01 |
| laney | how come that all only just happened? | 13:01 |
| laney | I'd have expected the packageset to have existed all along | 13:01 |
| rbasak | Thanks laney. I couldn't determine the problem from the backtrace | 13:01 |
| laney | no worries, the 401 unauthorized gave it away to me | 13:02 |
| laney | probably an 'if you know, you know' thing | 13:02 |
| -queuebot:#ubuntu-release- New binary: fenics-dolfinx [arm64] (jammy-proposed/universe) [1:0.3.0-13] (no packageset) | 13:04 | |
| rbasak | Looks like the jammy packageset was created on 2021-10-15 | 13:08 |
| rbasak | But since then the emails suggest that the script never tried to add any entries to the jammy packageset | 13:09 |
| rbasak | So this is the first time it got triggered to actually use the packageset perhaps? | 13:09 |
| laney | it must be that this function wasn't returning jammy up until now https://git.launchpad.net/~developer-membership-board/+git/oem-meta-packageset-sync/tree/oem-meta-packageset-sync#n77 | 13:11 |
| laney | perhaps snakefruit only just got an update to distro-info-data | 13:11 |
| RikMills | kubuntu daily ISO build is not getting copied to cdimage | 13:13 |
| rbasak | Just updating the DMB docs on this. | 13:13 |
| rbasak | The archive admin team runs/manages the script, right? Not the release team? | 13:14 |
| laney | yes, but the DMB owns it in that they're supposed to gatekeep all code changes - that's why it is under that team on LP | 13:15 |
| rbasak | Thanks. I updated https://wiki.ubuntu.com/DeveloperMembershipBoard/KnowledgeBase#Canonical_OEM_metapackage_packageset - I noticed that while the packageset was documented the automation was not. | 13:17 |
| rbasak | I've glossed over packageset/per-series packagesets | 13:17 |
| laney | looks great | 13:17 |
| -queuebot:#ubuntu-release- Packageset: 97 entries have been added or removed | 13:21 | |
| doko | Eickmeyer: you uploaded onetbb, do you continue with the transition? | 13:48 |
| -queuebot:#ubuntu-release- New: accepted fenics-dolfinx [arm64] (jammy-proposed) [1:0.3.0-13] | 13:50 | |
| -queuebot:#ubuntu-release- New: accepted fenics-dolfinx [ppc64el] (jammy-proposed) [1:0.3.0-13] | 13:50 | |
| -queuebot:#ubuntu-release- New: accepted fenics-dolfinx [armhf] (jammy-proposed) [1:0.3.0-13] | 13:50 | |
| -queuebot:#ubuntu-release- New: accepted fenics-dolfinx [s390x] (jammy-proposed) [1:0.3.0-13] | 13:50 | |
| sil2100 | rbasak, laney: thanks! | 14:01 |
| Eickmeyer | doko: I'm not sure. It seems to me that the old tbb needs to be removed first (I did file a bug report about that), but the new onetbb seems to introduce some regressions according to britney. I was going to leave it up to ubuntu-archive. | 15:49 |
| ginggs | Eickmeyer: these packages need rebuilding before the old tbb can be removed | 16:48 |
| ginggs | https://people.canonical.com/~ubuntu-archive/transitions/html/onetbb.html | 16:48 |
| Eickmeyer | ginggs: I see. I also see that some of them are incompatible with a newer libtbb? | 16:49 |
| ginggs | but at least half of those in level 2 will FTBFS | 16:50 |
| ginggs | so I think we should revert onetbb for now | 16:50 |
| ginggs | doko: ^ | 16:50 |
| Eickmeyer | ginggs: If that's the case, we lose numba, which will cause all sorts of issues. | 16:50 |
| ginggs | Eickmeyer: we already removed numba, what kind of issues will removing it again cause? | 16:51 |
| Eickmeyer | ginggs: I uploaded a newer version, and part of the reason python3-defaults wouldn't migrate was the lack of numba, and a bunch of packages depended on it. | 16:51 |
| doko | Eickmeyer: then repackage tbb to have non-conflicting packages? | 16:51 |
| ginggs | Eickmeyer: numba was not a reason why python3-defaults would not migrate | 16:52 |
| doko | no, migration of python3-defaults was independent of that | 16:52 |
| Eickmeyer | OK, let's remove numba and revert tbb. | 16:52 |
| Eickmeyer | This is what I get for trying to help. | 16:52 |
| doko | help is appreciated if it's complete =) | 16:53 |
| Eickmeyer | Well, I had no idea this was going to happen. | 16:53 |
| mwhudson | bdmurray: ack | 19:49 |
| LocutusOfBorg | ubuntu-archive: can we please kick myspell out from jammy? removed in sid, bug: #1006134 | 20:26 |
| ubottu | Bug 1006134 in Stellarium "Ocular plugin: minor issue when adding a telescope" [Low, Fix Released] https://launchpad.net/bugs/1006134 | 20:26 |
| LocutusOfBorg | ifrench-gut mighe | 20:27 |
| LocutusOfBorg | *might need to migrate first, others are false positives | 20:27 |
| jochensp | Hi, could someone help with getting the latest open3d into jammy? Maybe by simply dropping the old ppc64el build? | 20:54 |
| RikMills | jochensp: looks like a LTO related build error, and when trying the proposed amd64 build in a VM it segfaults on launch | 21:11 |
| jochensp | RikMills: interestingly it build in Debian, will try to run it | 21:12 |
| RikMills | jochensp: debian does not build with LTO enabled by default yet | 21:12 |
| jochensp | but open3d is build with -flto in Debian: https://buildd.debian.org/status/fetch.php?pkg=open3d&arch=amd64&ver=0.14.1%2Bdfsg-7&stamp=1645372419&raw=0 | 21:14 |
| jochensp | (though I guess the incompatible compiler options are from Ubuntu: ‘-mrelocatable’ and ‘-mcall-linux’ are incompatible) | 21:16 |
| jochensp | small correction: only the Python plugin is build with -flto and Ubuntu building everything with it is probably the reason for the segfault | 21:27 |
| jochensp | would you still accept a patch to disable LTO for Ubuntu? | 21:27 |
| mwhudson | jochensp: we've certainly disabled lto on other packages for this sort of reason | 21:35 |
| jochensp | I guess DEB_CFLAGS_MAINT_APPEND=-fno-lto should do it, should we push that through Debian or do you want to do a +ubuntu1 version? (I would really like to have the new version in jammy) | 21:36 |
| mwhudson | jochensp: there is the lto-disabled-list package too | 21:37 |
| mwhudson | i assume open3d is in universe | 21:37 |
| sarnold | hopefully helpful https://wiki.ubuntu.com/ToolChain/LTO | 21:38 |
| jochensp | universe, yes | 21:38 |
| jochensp | can I do something to have it fixed in jammy? | 21:43 |
| mwhudson | jochensp: file a bug in launchpad? | 21:46 |
| jochensp | against which package? | 21:46 |
| mwhudson | jochensp: open3d pls | 21:47 |
| mwhudson | i can do the archive bits, they're trivial enough | 21:47 |
| jochensp | mwhudson: https://bugs.launchpad.net/ubuntu/+source/open3d/+bug/1963556 (thanks :) ) | 21:50 |
| ubottu | Launchpad bug 1963556 in open3d (Ubuntu) "Open3D segfaults when compiled with LTO" [Undecided, New] | 21:50 |
| mwhudson | til https://en.cppreference.com/w/cpp/language/siof | 21:51 |
| mwhudson | jochensp: uhhh open3d is already in here | 21:55 |
| mwhudson | and has been since march 2021 | 21:56 |
| jochensp | but I see a lot of -flto here: https://launchpadlibrarian.net/587715572/buildlog_ubuntu-jammy-amd64.open3d_0.14.1+dfsg-7build1_BUILDING.txt.gz | 21:56 |
| mwhudson | where is this coming from then | 21:57 |
| jochensp | I don't see them in the Debian build | 21:57 |
| mwhudson | time to do some debugging i guess | 21:58 |
| mwhudson | why does open3d have its own copy of dpkg-buildflags?? | 22:00 |
| mwhudson | ah wait no | 22:00 |
| mwhudson | jochensp: the lto build flags are not the ones ubuntu uses though hm,m | 22:05 |
| mwhudson | this is going to be cmake somehow isn't it | 22:07 |
| mwhudson | ah, pybind | 22:09 |
| jochensp | pybind uses LTO in Debian as well, but not the rest | 22:09 |
| mwhudson | jochensp: all the hits for flto in https://launchpadlibrarian.net/587715572/buildlog_ubuntu-jammy-amd64.open3d_0.14.1+dfsg-7build1_BUILDING.txt.gz are for python bits afaics | 22:10 |
| jochensp | seems like, yeah | 22:12 |
| mwhudson | in particular the link of libOpen3D.so.0.14.1 doesn't use anything exotic, so maybe this isn't https://github.com/isl-org/Open3D/issues/4747 after all | 22:13 |
| ubottu | Issue 4747 in isl-org/Open3D "Link Time Optimization (-flto) triggers static initialization bug in Open3D" [Open] | 22:13 |
| jochensp | maybe.. do you have an idea where ‘-mrelocatable’ and ‘-mcall-linux’ are incompatible is coming from in the ppc64el build? (maybe that gives us a hint) | 22:16 |
| jochensp | here: https://launchpadlibrarian.net/588854642/buildlog_ubuntu-jammy-ppc64el.open3d_0.14.1+dfsg-7build1_BUILDING.txt.gz | 22:17 |
| mwhudson | jochensp: no | 22:26 |
| mwhudson | but i can try a ppc64el build and see what i can find i guess | 22:26 |
| mwhudson | -Bsymbolic-functions is always a good bet for "things go wrong in strange ways" | 22:27 |
| mwhudson | what is *this* person doing https://stackoverflow.com/questions/50746045/how-to-disable-mcall-linux-in-gcc | 22:29 |
| mwhudson | "It is not a full OS. We do not use RAM. All the executables are loaded in the processor cache." | 22:30 |
| sarnold | mwhudson: *wild* | 22:30 |
| jochensp | the unstable open3d packages work in jammy so seems to be something in the Ubuntu build | 22:33 |
| mwhudson | jochensp: what is the default gcc / g++ in debian currently? | 23:19 |
| mwhudson | but that's interesting, means it's not glibc 2.35 i guess | 23:19 |
| jochensp | 11.2.0-18 | 23:19 |
| jochensp | and libc6 is 2.33-7 | 23:20 |
| mwhudson | Ubuntu 11.2.0-16ubuntu1 | 23:20 |
| mwhudson | close enough, you'd have thought | 23:20 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!