[07:12] <pitti> hello all
[07:12] <pitti> any idea why systemd is blocked for promotion? I closed the block-proposed bug yesterday (bug 1346199) as it's now ready to land and past traincon-0
[07:13] <pitti> i. e. there's still a release team blocker for it
[07:13] <pitti> oh wait, that's just the general alpha-2 freeze I suppose
[08:42] <infinity> pitti: Self-diagnosing developers.  I like it.  We need more like you.
[08:46] <pitti> infinity: you mean people with the nasty habit of only seeing the solution *after* asking? :-)
[08:46] <pitti> 'cause that happens to me all the time, annoyingly
[08:47] <infinity> pitti: That happens to all of us.  I think it's the act of asking that kicks the brain into better diagnostic modes.
[08:48] <infinity> pitti: I'm almost positive cjwatson now has a habit if delaying any response to me by half an hour to make sure I actually want an answer, rather than just wanting someone to ask to trick myself into finding the answer. :P
[08:49] <infinity> pitti: For me, I think it's the split second when I hit <enter> on a question where I think "wait, is this going to make me look like an idiot?... IT SURE IS", and then the answer comes to me.
[08:51] <pitti> infinity: I'm glad I'm not the only one to embarass oneself :)
[08:57] <infinity> cjwatson: base-files and livecd-rootfs could use a quick once-over for precise.
[08:58] <infinity> stgraber: Or you, if you're around, since this is your point release.  Not sure what timezone you're in, or if you're around.
[08:58] <cjwatson> let me delay that for half an hour so you can decide whether you really need them
[08:58] <pitti> infinity: oh, and when would be a good time for my daily nag about postgresql SRUs?
[08:58] <infinity> cjwatson: Lollerfrigginskates.
[08:58] <infinity> pitti: Oh, perhaps nowish.  Which release(s)?
[08:59] <pitti> infinity: lucid, precise, trusty
[08:59] <pitti> IOW, "all"
[08:59] <infinity> *nod*
[08:59] <infinity> Looking.
[08:59] <pitti> quite nice to have a time with no supported non-LTS releases :)
[09:00] <infinity> pitti: Yeah.  Won't last.  But I guess it'll recur for the three month period after each LTS.
[09:00] <pitti> ROLLING RELE*cough* *cough*
[09:00] <infinity> pitti: Bad pitti.  No biscuit.
[09:01]  * pitti sobs
[09:02] <infinity> pitti: Hrm.  The 8.4/lucid warning probably belongs in NEWS, not changelog.  Also, it would be less sketchy if you suggested people upgrade to 12.04/14.04 before suggesting that, if they must run lucid, they upgrade to non-archive packages.
[09:03] <pitti> infinity: ah, I thought of apt-listchanges
[09:03] <pitti> infinity: but I can reupload with adding that to NEWS too
[09:03] <infinity> pitti: The default apt-listchanges config only shows NEWS.
[09:04] <infinity> pitti: People (like me, and probably you) have to actually reconfigure to show changelogs.
[09:04] <pitti> "Consider upgrading to Ubuntu 12.04 LTS or 14.04 LTS, or upgrading to PostgreSQL 9.x from http://apt.postgresql.org/"
[09:04] <pitti> better?
[09:05] <pitti> FWIW, that's really the place that we send all people to who need backports; these are officialy supported by upstream, and maintained by Christoph and me with a nice CI and QA
[09:05] <infinity> pitti: Yeah.  Or maybe even "Consider upgrading to PostgreSQL 9.x on Ubuntu 12.04 LTS or 14.04 LTS or, if you must stay on lucid, the PostgreSQL 9.x packages from http://apt.postgresql.org/"
[09:05] <infinity> pitti: Since precise has 8.4 as well...
[09:06] <infinity> pitti: I don't doubt that they're well-supported, but they still don't have our seal of approval, per se, plus any chance to get people off lucid is one worth taking. :P
[09:06] <pitti> infinity: yeah, as an upgrade shim; in trusty, the equivalent postgresql-9.1 only builds -plperl, but we didn't think of that yet in precise
[09:06] <pitti> heh, yes
[09:08] <infinity> pitti: Also, don't want to seem too sky-is-falling about it.  No more upstream support and point releases, but our security team will still patch psql 8.4 until lucid EOL.
[09:08] <infinity> pitti: Which is probably worth mentioning, so people don't panic.
[09:10] <pitti> infinity: http://paste.ubuntu.com/7903085/
[09:10] <pitti> proposed debian/postgresql-8.4.NEWS
[09:11] <infinity> pitti: That looks reasonable, but I'd move the last line above the paragraph instead of below.  Good(ish) news before bad.
[09:12] <infinity> pitti: I almost missed it, cause I wanted to stop reading halfway through the paragraph that implied I was no longer supported. ;)
[09:13]  * infinity runs to the store for a beverage while you reupload.
[09:14] <pitti> http://paste.ubuntu.com/7903102/ ?
[09:14] <pitti> infinity: ^ better?
[09:19] <pitti> infinity: ^
[09:22] <infinity> pitti: Much better, thanks.
[09:22] <infinity> pitti: And precise doesn't need the same warning?  Is the psql 8.4 there neutered to only be useful for upgrading?
[09:23] <pitti> infinity: yes, when you upgrade to precise you get a big fat warning that 8.4 is unsupported and only for upgrades, and it's in universe
[09:23] <pitti> infinity: pretty much the only requirement there is that the version number must be bigger than lucid's
[09:23]  * infinity nods.
[09:23] <pitti> so that you don't lose plperl after upgrades (darn libperl5.x not being installable in parallel)
[09:24] <infinity> Eww, psql imports a copy of tzdata?
[09:24] <infinity> Tell me it doesn't use that on Debian/Ubuntu.
[09:25] <pitti> infinity: no, it doesn't
[09:25] <pitti> infinity: we've configured with --system-tzdata (or so) for the last 10 years (or more?)
[09:26] <infinity> Oh, good.  Cause gross.
[09:26] <pitti> infinity: well, not all target platforms of postgresql have that, so it's a configure option
[09:26] <pitti> (or provide updates to it regularly)
[09:26] <infinity> Keeping on top of tzdata updates is the bane of my existence, I can't imagine trying to track embedded copies.
[09:33] <infinity> pitti: In trusty, you drop 5 patches, but the changelog only explains 2 (or 3, I guess).
[09:34] <pitti> -60-pg_regress_socketdir.patch
[09:34] <pitti> -61-extra_regress_opts
[09:34] <pitti> -62-pg_upgrade-test-in-tmp
[09:34] <pitti> -63-pg_upgrade-test-bindir
[09:35] <pitti> those are the four patches which explicitly used /tmp/ for pg_regress, which is now solved in a better way upstream
[09:35] <pitti> the fifth is dropping 04-configure-tcl8.6 which has its own changelog entry
[09:35] <infinity> Ahh, kay.  Reading the fourth hinted at that.  The naming just didn't imply it related to the changelog entry.
[09:35] <infinity> Good enough for me.
[09:46] <pitti> infinity: cheers!
[09:46] <pitti> infinity: you're still looking at trusty/9.1, I suppose?
[10:40] <infinity> pitti: Oh, no, didn't notice there were two trusty versions.
[10:41] <infinity> pitti: Could be because you messed up the tasks on the bug. :P
[10:41] <pitti> infinity: oui, je suis désolé
[10:41] <pitti> infinity: we always need the previous major version as an upgrade shim for -plperl (see above)
[10:42] <pitti> so that you can use the pacakges from the old release for upgrading (but not install the old version any more)
[10:43] <infinity> pitti: Done.
[10:43] <pitti> cheers
[10:44] <infinity> wgrant: Does anyone know Malone well enough to fix that bug/misfeature where deleting a series task from a bug means you can never re-add it?
[10:47] <pitti> infinity: so for trusty the SRU verification is now a single "build" click in jenkins on the trusty-adt-postgresql-9.3 job; want this for older releases, too!
[10:47] <pitti> well, lucid will EOL in 7 months, one less problem
[10:48] <infinity> pitti: However you want to do your validation in a way that you feel would satisfy your MRE obligations and your users. :P
[10:48] <wgrant> infinity: Hm, are you sure that's not fixed/
[10:48] <pitti> infinity: well, running the postgresql-common test suite is what I do manually as well
[10:49] <infinity> wgrant: Positive.
[10:49] <pitti> infinity: there's nothing which requires particular brain powers about it
[10:49] <infinity> wgrant: https://bugs.launchpad.net/ubuntu/+source/postgresql-9.1/+bug/1348176/+nominate
[10:49] <infinity> wgrant: Note the empty list.
[10:49] <infinity> wgrant: But some tasks don't have all three series'.  In fact, none of them do.
[10:50] <wgrant> infinity: So, I think there's a workaround, but it's slightly awful. Let me test on staging.
[10:50] <pitti> oh, would it perhaps help to delete the stable tasks for all others as well, would it then reappear in +nomiate?
[10:50] <infinity> pitti: No.
[10:50] <infinity> pitti: Pretty sure I've tried that in the past.
[10:50] <wgrant> It's a side effect of bug #110195.
[10:51]  * wgrant is trying the workaround.
[10:52] <infinity> wgrant: If the workaround is an API workaround, it's academically interesting, but mostly useless.
[10:52] <wgrant> infinity: It's not.
[10:52] <infinity> If you can coax the web ui into letting one re-add a deleted task, that would be interesting.
[10:52] <wgrant> Come on staging, you can do it...
[10:53] <pitti> an API workaround would work for fixing a particular bug, too
[10:53] <infinity> pitti: Sure, I just mean in general, it's not something most people will do.
[10:53] <pitti> *nod*
[10:53] <infinity> For all its warts and weird design decisions, this is probably the only bug in Malone that really annoys me on a regular basis.
[10:54] <infinity> The rest, I've just learned to live with, given an "at least it's not bugzilla" attitude.
[10:58] <wgrant> OK, so there *is* a way through the UI, but it's not pretty.
[10:59] <wgrant> The ability to renominate is restored if you delete a task for the target that you want to nominate.
[10:59] <wgrant> Due to #110195, the target is always a DistroSeries, not a SourcePackage.
[11:00] <wgrant> Since you have tasks for each relevant series on that bug, you could remove the package name from each task, delete its series task which is now on the DistroSeries rather than the SourcePackage, set the package name again, and then move onto the next package.
[11:00] <wgrant> Eventually you'll have deleted a task for each DistroSeries, and all will be open for nomination.
[11:06] <pitti> wgrant: happy to do that, if I don't end up with having no tasks at all
[11:07] <wgrant> Heh, if we get into that situation then I'll just hack the DB :)
[11:07]  * pitti takes a deep breath and tries
[11:07] <wgrant> Just make sure that any task you delete doesn't mention a source package.
[11:07] <pitti> ack
[11:09] <pitti> wgrant: so, step 1 on bug 1348176: I should now be able to delete teh ubuntu/precise task, and get back precise nomination?
[11:09] <wgrant> pitti: Yep.
[11:10] <pitti> wgrant: thanks muchly, that worked!
[11:13] <pitti> infinity: how does that look now? :-)
[11:16] <wgrant> Comment #9 makes good use of the inline status changes...
[11:59] <infinity> pitti: Much better!
[13:05] <stgraber> infinity: I'm back to my usual timezone (if you can call it that), so back to Montreal. I assume you got all the reviews you wanted by now?
[13:25] <infinity> stgraber: Yeahp.
[15:46] <infinity> bdmurray: I think we're ready to flip the precise->trusty upgrde toggle.
[15:55] <jibel> infinity, there is still bug 1348067
[15:55] <jibel> for desktop users
[15:57] <jibel> u-m crashes immediately when the user press 'upgrade'
[15:57] <infinity> slangasek: ^
[15:57] <infinity> bdmurray: Belay that...
[15:59] <slangasek> hmm
[15:59] <bdmurray> I just saw that on errors this morning
[15:59] <bdmurray> seems only to effect 12.04 fwiw
[16:00] <infinity> bdmurray: Any idea what's going on there?  mvo's abandoned us for a week. ;)
[16:00] <bdmurray> I'll have a look
[16:02] <bdmurray> infinity: there's an upload in the precise queue already for it
[16:03] <bdmurray> https://launchpadlibrarian.net/180904636/update-manager_1%3A0.156.14.16_1%3A0.156.14.17.diff.gz
[16:03] <infinity> Oh, shiny.  The bug didn't imply there was.
[16:04] <bdmurray> infinity: really? 6. install update-manager from precise-proposed
[16:04] <infinity> jibel: If I review and accept that, how soon can you turn around a verification?
[16:04] <infinity> bdmurray: I assumed that was jibel being optimistic about a fix happening. :)
[16:04] <bdmurray> heh
[16:05] <bdmurray> I've put a link to the errors bucket in the bug too
[16:17] <jibel> I can do that later today
[16:17] <jibel> infinity, ^
[16:18] <infinity> jibel: Awesome.  If you could verify the other two bugs fixed in the upload in passing (they looks simpler than your bug), that would be helpful to expedite it.  If not, we'll hunt someone else down.
[16:19] <jibel> meh, I've a test VM ready with lts-trusty but not lts-saucy :(
[16:19] <jibel> infinity, I'll do the verification too
[16:20] <infinity> jibel: You're my hero.
[19:47] <jibel> infinity, I verified the 3 SRU of update-manager in Precise. It's successful, for bug 1349400 I verified that *-lts-trusty:i386 replaced *-lts-saucy:i386 and there is a display after upgrade and a reboot. Steps 4 and 7 "verify that it work/doesn't work" do not mean anything.
[20:21] <doko> mlankhorst, online?