[00:56] <doko> vorlon: already building
[00:57] -queuebot:#ubuntu-release- New: accepted ghmm [amd64] (noble-proposed) [0.9~rc3-5.1ubuntu3]
[00:57] -queuebot:#ubuntu-release- New: accepted ghmm [armhf] (noble-proposed) [0.9~rc3-5.1ubuntu3]
[00:57] -queuebot:#ubuntu-release- New: accepted ghmm [riscv64] (noble-proposed) [0.9~rc3-5.1ubuntu3]
[00:57] -queuebot:#ubuntu-release- New: accepted ghmm [arm64] (noble-proposed) [0.9~rc3-5.1ubuntu3]
[00:57] -queuebot:#ubuntu-release- New: accepted ghmm [s390x] (noble-proposed) [0.9~rc3-5.1ubuntu3]
[00:57] -queuebot:#ubuntu-release- New: accepted ghmm [ppc64el] (noble-proposed) [0.9~rc3-5.1ubuntu3]
[00:58] <doko> also started a local build, dh_makeshlibs exits after the first error ...
[01:13] <vorlon> doko: comparing your debdiff to mine I expect a build failure, and also it's not clear to me the symbols you've marked as optional should be optional as opposed to arch-specific.  I'm going to go ahead and upload my version
[01:21] <doko> vorlon: these are all template instantiations ...
[01:31] <vorlon> doko: 'sandbox::localtime_override(long const*)@Base' <- this is not a template
[01:31] <vorlon> maybe it shouldn't be exported, but it's not a template
[01:31] <vorlon> anyway the missing symbol was definitely part of the ABI
[01:31] <vorlon> QtWebEngineCore::ProfileAdapter::determineDownloadPath(QString const&, QString const&, long const&)@Base
[01:32] <doko> which already is marked as optional
[02:21] -queuebot:#ubuntu-release- New binary: calamares-settings-ubuntu [amd64] (noble-proposed/universe) [1:24.04.19] (lubuntu, ubuntustudio)
=== chris14_ is now known as chris14
[12:09] <jbicha> mwhudson: We suspect that new vala will fix ftbfs for evolution-data-server, folks & telepathy-glib. Will sync from Unstable later today
[13:19] <LocutusOfBorg> vorlon, libfm/menu-cache are chickening and egging
[13:20] <LocutusOfBorg> but libfm can be built without menu-cache with stage1 profile
[13:21] <LocutusOfBorg> (appstream-glib/boinc are fixed)
[14:55] -queuebot:#ubuntu-release- New binary: kronosnet [amd64] (noble-proposed/main) [1.28-2ubuntu1] (i386-whitelist, ubuntu-server)
[14:55] -queuebot:#ubuntu-release- New binary: kronosnet [i386] (noble-proposed/main) [1.28-2ubuntu1] (i386-whitelist, ubuntu-server)
[14:56] -queuebot:#ubuntu-release- New binary: kronosnet [s390x] (noble-proposed/main) [1.28-2ubuntu1] (i386-whitelist, ubuntu-server)
[14:57] -queuebot:#ubuntu-release- New binary: kronosnet [ppc64el] (noble-proposed/main) [1.28-2ubuntu1] (i386-whitelist, ubuntu-server)
[14:58] -queuebot:#ubuntu-release- New binary: kronosnet [armhf] (noble-proposed/main) [1.28-2ubuntu1] (i386-whitelist, ubuntu-server)
[15:00] -queuebot:#ubuntu-release- New source: dfx-mgr (noble-proposed/primary) [2023.2-0ubuntu1]
[15:00] -queuebot:#ubuntu-release- New binary: kronosnet [arm64] (noble-proposed/main) [1.28-2ubuntu1] (i386-whitelist, ubuntu-server)
[15:18] -queuebot:#ubuntu-release- New binary: kronosnet [riscv64] (noble-proposed/main) [1.28-2ubuntu1] (i386-whitelist, ubuntu-server)
[15:54] -queuebot:#ubuntu-release- Unapproved: software-properties (xenial-proposed/main) [0.96.20.12 => 0.96.20.13] (desktop-core, ubuntu-server)
[15:57] <jbicha> dpkg was merged over the weekend to include Debian's -Werror=implicit-function-declaration . Do we want to revert that at this point for noble?
[15:59] <jbicha> https://wiki.debian.org/qa.debian.org/FTBFS#A2024-03-13_-Werror.3Dimplicit-function-declaration
[16:15] <juliank> jbicha: I think this is a very important feature and I'm surprised this was not on by default in Ubuntu, it seems important to ensuring correctness of the t64 transiton
[16:27] <jbicha> it's just going to trigger lots of ftbfs that aren't directly related too
[16:31] <Eickmeyer> ubuntu-release: VERTY LOUD PING w.r.t. bug 2058161: Ubuntu MATE is requesting help moving to ubuntu-desktop-provision for their new installer and has requested my help. Need an FFe ack ASAP so I can get started on the work! ******
[16:31] -ubottu:#ubuntu-release- Bug 2058161 in ubuntu-mate-meta (Ubuntu) "FFe: Switch installer to ubuntu-desktop-provision" [Critical, New] https://launchpad.net/bugs/2058161
[16:47] <seb128> juliank, as an example (which I guess triggered the question from Jeremy), gnome-control-center started to fail to build in noble
[16:47] <seb128> https://launchpad.net/ubuntu/+source/gnome-control-center/1:46.0-0ubuntu1/+build/27938036
[16:48] <juliank> Yes that's an unrelated build failure because it happens to not return time_t, but if it did return time_t you'd have silently broken gnome-control-center on armhf
[16:49] <seb128> juliank, unrelated to?
[16:49] <juliank> unrelated to t64
[16:49] <juliank> sorry
[16:49] <juliank> the flag was added in Debian to ensure that t64 is reasonably safe
[16:50] <juliank> imagine if it was cc_common_get_time()  or something returning a time_t
[16:50] <seb128> right, Jeremy wasn't asking about t64 though but about that flag being default now on other archs
[16:50] <vorlon> LocutusOfBorg: the stage1 changes to debian/control in libfm are a bit eyebrow-raising, but let's see
[16:50] <seb128> juliank, whatever the rational is, that's an
[16:51] <seb128> that's a
[16:51] <juliank> seb128: But arguably it makes no sense to only enable that on armhf
[16:51] <seb128> - change which has an impact on buildability of other packages
[16:51] <seb128> - landed post feature freeze and doesn't seem to have related Ffe attached
[16:51] <vorlon> juliank, jbicha: enabling it on armhf means there will be build failures on armhf, possibly of packages we can afford to jettison the armhf binaries of.  enabling it on all archs means build failures on all archs
[16:51] <seb128> - hasn't been announced
[16:53] <seb128> it doesn't seem unreasonable to question whether that change is right, if it is wanted it feel like it should have been at least announced and that we might need another archive test rebuild to see the impact? (or did we just get unluck that one of the first desktop package to get uploaded fails to build due to it?)
[16:53] <vorlon> I am going to be posting to ubuntu-devel-announce today about chucking armhf binaries out
[16:54] <vorlon> we simply do not have time/capacity to fix all of these
[16:54] <vorlon> and as a corollary, I think it does make sense to not error on implicit declarations on the 64-bit archs
[16:55] <vorlon> seb128: the reason for the change to dpkg in Debian is that it's a surprise to maintainers when their packages which don't use time_t now ftbfs on 32-bit archs ony
[16:55] <vorlon> only
[16:56] <vorlon> but I think the calculus in Ubuntu now is different given the release deadline
[16:56] <juliank> +1
[16:56] <vorlon> LocutusOfBorg, doko: ^^ you probably want to be aware of this discussion
[16:56] <seb128> vorlon, I understand the rational, it's still the kind of change I would expect to require a ffe to land at this stage of this cycle
[16:56] <seb128> or at minimal to be announce on the devel list
[16:57] <vorlon> LocutusOfBorg: ok well the trivial stage1ification of libfm ftbfs? https://launchpad.net/~canonical-foundations/+archive/ubuntu/archive-bootstrap/+build/27938214
[16:57] <seb128> but yeah, if it's armhf only and we decide to start removing binaries when needed on that arch that lowers the impact...
[16:57] <seb128> ups
[16:58] <vorlon> seb128: we should consider -Werror=no-implicit-function-declarations required for armhf; without it you get misbuilds when using time_t-sensitive glibc functions
[16:58] <vorlon> and on armhf it didn't land post-FF
[16:58] <vorlon> (it landed as part of the time_t change)
[16:59] <seb128> ack, I still wonder how many packages it's going to impact... we didn't get a test rebuild with that flag before the transition did we?
[17:02] <vorlon> we did not
[17:02] -queuebot:#ubuntu-release- New binary: halide [amd64] (noble-proposed/universe) [17.0.1-1] (no packageset)
[17:03] -queuebot:#ubuntu-release- New: accepted halide [amd64] (noble-proposed) [17.0.1-1]
[17:04] <juliank> just remove the entire desktop on armhf, problem solved :)
[17:07] <vorlon> that's not necessarily the wrong answer. But the line between "desktop" and "not desktop" is blurry, and we have to start at the edges of the graph and be pragmatic
[17:08] <vorlon> the largest complaints will probably be from raspi users who want to be able to upgrade even if it's unsupported, but yo, you don't even have a kernel, what are you doing
[17:09] <vorlon> ok how did this build locally, I definitely don't have distutils installed in my test build env :| https://launchpad.net/ubuntu/+source/mongo-cxx-driver-legacy/1.1.3-3.3ubuntu1/+build/27938257
[17:18] <LocutusOfBorg> thanks vorlon for highlighting
[17:18] <LocutusOfBorg> of course we can revert dpkg change, or reduce to armhf only
[17:18] <LocutusOfBorg> indeed we discussed this, but not *that* deep when we agreeded to merge dpkg
[17:19] <jbicha> bdmurray: GNOME is preparing their 47 schedule. Any chance we can get a draft 24.10 schedule to compare?
[17:20] <LocutusOfBorg> seb128, the discussion was somewhere above, like 5 days ago, but turned out the merge was a little bit more trickier due to MoM failing to do some proper job in keeping the delta from proposed pocket (and my simplification of delta in newer release)
[17:27] <LocutusOfBorg> vorlon, maybe the gtk-doc-tools shouldn't be !stage1?
[17:32] <LocutusOfBorg> yeah adding it works
[17:32] <LocutusOfBorg> sadly I can't upload on that ppa :/
[17:33] -queuebot:#ubuntu-release- Unapproved: initramfs-tools (jammy-proposed/main) [0.140ubuntu13.4 => 0.140ubuntu13.5] (core, i386-whitelist)
[17:37] <Eickmeyer> ubuntu-release: VERTY LOUD PING w.r.t. bug 2058161: Ubuntu MATE is requesting help moving to ubuntu-desktop-provision for their new installer and has requested my help. Need an FFe ack ASAP so I can get started on the work! ******
[17:37] -ubottu:#ubuntu-release- Bug 2058161 in ubuntu-mate-meta (Ubuntu) "FFe: Switch installer to ubuntu-desktop-provision" [Critical, New] https://launchpad.net/bugs/2058161
[17:44] <juliank> this change makes britney use all-proposed for noble armhf until beta: https://code.launchpad.net/~juliank/britney/+git/britney2-ubuntu/+merge/462607
[17:44] <juliank> the similar patch has just landed in autopkgtest-cloud's request.cgi for requesting retries - they will get all-proposed=1 set automatically even if not specified
[17:44] <juliank> (on armhf noble)
[17:44] <juliank> I expect this to land shortly and then things will improve a bit in terms of not getting wrong stuff scheduled :)
[17:51] -queuebot:#ubuntu-release- New source: involflt (noble-proposed/primary) [0.1.0-0ubuntu6]
[17:52] <LocutusOfBorg> release team please NBS-cleanup libnode108t64 (all archs)
[18:28] <juliank> the britney change has been merged now, all future uploads and retries will use all-proposed on noble armhf.
[18:28] <juliank> all-proposed will stop being enabled on beta day
[18:30] <Eickmeyer> ubuntu-release: VERTY LOUD PING w.r.t. bug 2058161: Ubuntu MATE is requesting help moving to ubuntu-desktop-provision for their new installer and has requested my help. Need an FFe ack ASAP so I can get started on the work! ******
[18:30] -ubottu:#ubuntu-release- Bug 2058161 in ubuntu-mate-meta (Ubuntu) "FFe: Switch installer to ubuntu-desktop-provision" [Critical, New] https://launchpad.net/bugs/2058161
[18:37] <vorlon> LocutusOfBorg: gtk-doc-tools> thanks, I've added that back in
[18:41] <vorlon> LocutusOfBorg: ftr NBS cleanup requires archive admins rather than the release team. Anyway, doing a round of -proposed NBS cleanup now
[18:43] <vorlon> LocutusOfBorg: libfm built, I'll copy menu-cache over to that ppa in a bit to finish the bootstrap
[18:52] <Eickmeyer> vorlon: Thanks for the FFe. This is my priority over the next couple of days, expect an urgent MP for livecd-rootfs.
[19:21] <seb128> LocutusOfBorg, backlog of this channel isn't a proper way to communicate late changes to the members of the projects...
[19:22] <seb128> LocutusOfBorg, my point is that if you change a default compiler flag in dpkg at the end of the cycle there should at minimal go with a FYI email to -devel and/or equivalent post to discourse
[19:56] <Eickmeyer> vorlon: This was easy to layer; they already had a minimal seed. https://code.launchpad.net/~eeickmeyer/livecd-rootfs/+git/livecd-rootfs/+merge/462626
[19:57] <vorlon> Eickmeyer: I have an itch that I want to refactor the livecd-rootfs code for all of the ubuntu-desktop-installer flavors to reduce the code duplication; would you like me to land that before or after your change?
[19:58] <Eickmeyer> vorlon: Can we do after? That way it's just one more flavor to 86 from that duplication.
[19:59] <vorlon> ok
[20:00] <Eickmeyer> vorlon: As an aside, I looked at Budgie's and it's still way more than necessary since they're not doing a legacy install anymore.
[20:10] <vorlon> LocutusOfBorg, doko: a straight revert, but eyes welcome: https://launchpad.net/ubuntu/+source/dpkg/1.22.6ubuntu3 (also, I missed reverting the test case oops)
[20:22] <vorlon> why does perl make you count your tests, that's such a dumb convention
[20:23] <seb128> Eickmeyer, vorlon, so with mate switching can we drop the ubiquity gui installer? (keeping the oem-setup part still)
[20:27] <Eickmeyer> seb128: Yeah, once I can land the change in the seed/meta. I'm blocked by not being part of ~ubuntu-mate-dev. Working on that.
[20:30] <Eickmeyer> seb128: I think we'll also need to push out another spin of `ubiquity-slideshow-ubuntu` since it has some generic wording changes on Edubuntu's part. Also, I think `oem-setup` needs a no-change-rebuild because the slideshows have a major graphical glitch.
[20:49] <vorlon> seb128: I was going to double-check whether all flavors had moved
[20:50] <vorlon> Eickmeyer: core-dev can land seed changes for ubuntu-mate-dev
[20:51] <vorlon> but the previous livecd-rootfs didn't migrate because of failing autopkgtests, let's see
[20:52] <vorlon> seb128, Eickmeyer: I still see ubiquity in ubuntu-unity
[20:53] <arraybolt3> vorlon: ack, I did a ton of work to fix that yesterday and am trying to fix it up ASAP
[20:53] <arraybolt3> Got swallowed by an insane workload on multiple fronts and it kept me very delayed
[20:54] <vorlon> 379s The $BUILDSTAMP variable is empty
[20:54] <arraybolt3> (r.e. ubiquity on Ubuntu Unity)
[21:01] <vorlon> ok well I'm going to go refactor this code then
[21:01] <vorlon> meanwhile I have pinged utkarsh and philroche about the autopkgtest regressio
[21:14] <ItzSwirlz> vorlon: what will I need to get nemo (and then cinnamon) out of the proposed pocket and into -release
[21:14] <ItzSwirlz> I know cjs has a build failure
[21:36] <vorlon> Eickmeyer: why have we ended up saying not to put casper in the live layer for all of these flavors, which don't have casper in their live seed?
[21:36] <vorlon> Eickmeyer: only ubuntucinnamon, ubuntu, ubuntustudio, and xubuntu appear to have it seeded
[21:37] <vorlon> ItzSwirlz: nemo's own autopkgtest is failing?
[21:38] <vorlon> ItzSwirlz: and cinnamon-desktop shows failing autopkgtests for cinnamon-control-center: https://ubuntu-archive-team.ubuntu.com/proposed-migration/update_excuses.html#cinnamon-control-center https://ubuntu-archive-team.ubuntu.com/proposed-migration/update_excuses.html#cinnamon-desktop
[21:38] <vorlon> ItzSwirlz: it will *also* block on our ability to get new glib2.0 through for the armhf time_t transition, but the above are all nemo/cinnamon-specific issues to be resolved
[21:40] <ItzSwirlz> cinnamon-control-center, nemo are failing
[21:40] <ItzSwirlz> and cjs
[21:41] <vorlon> Eickmeyer: ah because it's being added in the livecd-rootfs code later <grump>
[21:43] <Eickmeyer> vorlon: Correct.
[21:44] <vorlon> Eickmeyer: ok. it doesn't look like it's being added there for all flavors either; I'm going to do an audit and raise MPs against seeds as needed
[21:45] <Eickmeyer> vorlon: FWIW, I have it *pending* for mate. Also, FWIW, I'm not a core-dev (yet), hence can't land these changes on my own to the mate seed.
[21:45] <vorlon> yes, but you can point me at it and I can merge
[21:45] <Eickmeyer> vorlon: Ah, fair.
[21:51] <Eickmeyer> vorlon: https://code.launchpad.net/~eeickmeyer/ubuntu-seeds/+git/ubuntu-mate/+merge/462637
[21:52] <Eickmeyer> After that, I can do a meta update with no issues.
[23:29] <LocutusOfBorg> thanks vorlon !
[23:29] <LocutusOfBorg> btw vorlon doko abseil from Debian experimental works on 32bit with time64_t
[23:30] <LocutusOfBorg> maaaaaaaaaaaybe syncing it is better than disabling tests
[23:30] <LocutusOfBorg> (discussed in debian-release too)
[23:30]  * LocutusOfBorg will not sync it, sleep time