[00:09] <mwhudson> RAOF: did that change help mir?
[00:34] <RAOF> mwhudson: Still finishing it to test!
[00:34] <RAOF> Turns out there's quite a lot of places where we assumed the 3-parameter open(), and it's annoying to forward varargs calls.
[01:44] <RAOF> mwhudson: For more in the “how did that ever work?” department - we were only ever implementing the 3-parameter open() call, but virtually all of the calls we were trying to intercept were of the 2-parameter open().
[02:46] <RAOF> Seriously, gcc? ‘<anonymous>’ may be used unitialised?
[02:46] <RAOF> On ppc64el, but not x86-64?
[02:46] <RAOF> C was a mistake 🔥
[04:04] <mwhudson> RAOF: yay C
[04:04] <RAOF> Actually, this is on C++.
[04:04] <mwhudson> "it's annoying to forward varargs calls" i think it's impossible generically but i guess in this case it's not very generic
[04:05] <RAOF> Yeah, impossible in the general case and merely annoying in this case. :)
[04:05] <RAOF> Apparently gcc (just on ppc64el!) thinks the default constructor for a std::optional accesses an anonymous uninitialised variable.
[04:05] <mwhudson> that's... special
[04:05] <mwhudson> and smells like a gcc bug tbh
[04:06] <RAOF> I think it must be.
[04:06] <RAOF> Oh! It's going to be an interaction between the optimiser and the C++ frontend, isn't it. That's how it can be ppc64 specific.
[04:12]  * mwhudson types things into gcc bugzilla
[04:13] <RAOF> Helpfully, the diagnostic identifies it in entirely the wrong function, too.
[04:13] <mwhudson> RAOF: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80635
[04:14] <mwhudson> although that's old
[04:15] <mwhudson> looks similar though
[04:18] <RAOF> It does, yes.
[04:19] <RAOF> It also suggests that we might have hit the problem on a !ppc64 architecture, essentially at random?
[05:09] <RAOF> 🎉 That fixes the stack smashing!
[05:10] <RAOF> Leaving only a bunch of failing tests 😬
[06:24] <tobikoch> o/
[06:25] <tobikoch> We bootstrap a bunch of containers every night and this was working well until this morning, when I woke up to:
[06:25] <tobikoch> Setting up ubuntu-advantage-tools (27.2.1~18.04.1) ...
[06:25] <tobikoch> ERROR: File not found '/run/cloud-init/instance-data.json'. Provide a path to instance data json file using --instance-data
[06:25] <tobikoch> dpkg: error processing package ubuntu-advantage-tools (--configure):
[06:25] <tobikoch>  installed ubuntu-advantage-tools package post-installation script subprocess returned error exit status 1
[06:27] <tobikoch> Which apparently gets pulled in by ubuntu-minimal. A bit annoying.
[06:40]  * RAOF checks.
[06:40] <RAOF> Hm. Is that a recent SRU?
[06:40] <RAOF> Yup! Looks like that was published 14 hours ago.
[06:40] <RAOF> blackboxsw: Around? It looks like your recent ubuntu-advantage-tools upload introduced a regression ^^^
[06:42] <RAOF> tobikoch: Have you filed a bug about this yet?
[06:43] <tobikoch> No yet, but can do.
[06:44] <RAOF> It's easier to link to :)
[06:44] <tobikoch> k
[06:50] <tobikoch> RAOF: https://bugs.launchpad.net/ubuntu-advantage/+bug/1938128
[06:50] <tobikoch> rcj: this might bite you, as well^
[07:04] <RAOF> tobikoch: Thanks. Did you deliberately mark that bug as proprietary? It means that I can't re-assign it to the correct place :)
[07:04] <RAOF> If it is meant to be proprietary I can work around that.
[07:04] <RAOF> I need to head off now, so I hand this over to…
[07:04] <RAOF> !regression-release
[07:04] <RAOF> Oh, no ubotu in here? Have I forgotten that?
[07:04] <tobikoch> RAOF: I thought I had changed it, but maybe it is private because UA is proprietary?
[07:05] <RAOF> Maybe. I can just copy the information out into the non-proprietary place and subscribe you to the new bug if that's ok.
[07:06] <tobikoch> Sure, sorry for the hassle :|
[07:07] <RAOF> No problem!
[07:08] <RAOF> rbasak, you should be on soon. If you're back before I am back from collecting my daughter, please see above!
[09:48] <GunnarHj> vorlon: Thanks for fixing gtk4 on i386!
[10:15] <mwhudson> huh i fixed that bug in git
[10:15] <mwhudson> the ubuntu-advantage-tools one
[10:16] <mwhudson> https://bugs.launchpad.net/ubuntu/+source/ubuntu-advantage-tools/+bug/1936833
[10:20] <rbasak> Thanks
[10:20] <rbasak> Which series are affected?
[10:26] <rbasak> sil2100: could you halt phasing of ubuntu-advantage-tools please? Bug 1936833 and the discussion above.
[10:26] <rbasak> What version of apt started doing phasing again?
[10:26] <rbasak> But still, desktops may be affected too.
[10:42] <sil2100> rbasak: trying o/
[10:48] <mwhudson> rbasak: all of them, i think
[10:48] <mwhudson> i thought the SRU just landed but it looks like it was a week ago?
[10:49] <mwhudson> ah no was about 16 hours ago
[10:50] <sil2100> Ok, overrode the phasing, but that's only for update manager
[10:50] <mwhudson> apt started phasing in hirsute but image builds generally ignore that anyway
[10:52] <laney> any reason not to roll it back?
[10:58] <rbasak> I had thought it had been released for a while, but perhaps not.
[10:59] <rbasak> Ah that's what Michael said :)
[11:03] <mwhudson> the date on https://launchpad.net/ubuntu/+source/ubuntu-advantage-tools is confusing
[11:03] <mwhudson> oh is it showing the date it was published in proposed?
[11:03] <sil2100> I don't have much context re: this upload, but we can indeed pull it from -updates if needed
[11:04] <laney> I would
[11:06] <rbasak> sil2100: OK. Please pull it from -updates.
[11:10] <rbasak> sergiodj: ^
[11:23] <sil2100> ACK
[11:23] <sil2100> Will take a moment
[11:24] <sil2100> Since I need to remove the current package and reinstate the previous version in -updates
[11:24] <sil2100> rbasak: can you mention on the regression bug that we're pulling this update back to -proposed?
[11:25] <laney> you'll want to un-fixed-released the original SRU bugs too I guess
[11:26] <laney> fix-released*
[11:43] <rbasak> sil2100: done
[11:50] <rbasak> un-fixed-released the original SRU bugs too> also done
[11:55] <rbasak> otp now for the next hour
[13:08] <dbungert> If anyone is interested in reviewing some update-notifier, I'm working on an interesting issue where a crash happens, you get the dialog to report or not, click the report button, and nothing happens.  https://code.launchpad.net/~dbungert/update-notifier/+git/update-notifier/+merge/405628