2012/02/01/#launchpad-dev.txt

[00:32] <wgrant> wallyworld___: ppa:launchpad should work on precise now
[01:00] <lifeless> man once you start looking for dupes
[01:00] <lifeless> there are tonnnnes
[01:00] <lifeless> (e.g. bug 132300)
[01:00] <_mup_> Bug #132300: mail for new bugs inconsistent with other bug mail (ignores mail-on-my-actions disabled, missing footers, duplicate mail vs structural subscriptions) <email> <lp-bugs> <notifications> <Launchpad itself:Triaged> < https://launchpad.net/bugs/132300 >
[01:01] <wgrant> lifeless: Well
[01:01] <wgrant> lifeless: They're not necessarily dupes
[01:02] <wgrant> They would all be solved by making them consistent, but they're really separate bugs.
[01:03] <lifeless> wgrant: I think there is one that isn't necessarily a dupe
[01:04] <lifeless> wgrant: anyhow, we have limited facilities for saying 'X symptom', 'Y symptom' of a given root cause, so meh:)
[01:17] <wgrant> Bah
[01:17] <wgrant> lxc has been apparmored
[01:20] <wgrant> Ahhhh
[01:20] <wgrant> And that explains why my syslog has been all broken.
[01:21] <wgrant> rsyslog from a container sometimes takes hold of the same stream
[01:21] <wgrant> So /var/log/kern.log only gets the occasional byte
[01:25] <lifeless> ok so I need help finding a damn bug :)
[01:25] <lifeless> there was one about bugs like bug 1
[01:25] <lifeless> which get flashcrowded
[01:26] <_mup_> Bug #1: Microsoft has a majority market share <ubuntu> <Clubdistro:Confirmed> <Computer Science Ubuntu:Confirmed for compscibuntu-bugs> <LibreOffice Productivity Suite:New> <dylan.NET.Reflection:Invalid> <dylan.NET:Invalid> <EasyPeasy Overview:Invalid by ramvi> <Ichthux:Invalid by raphink> <JAK LINUX:Invalid> <LibreOffice:In Progress by bjoern-michaelsen> <Linux:New> <Linux Mint:In Progress> <The Linux OS Project:In Progress> <metacity:In Prog
[01:26] <lifeless> I cannot find
[01:26] <wgrant> Bug #73122?
[01:26] <_mup_> Bug #73122: individual bug reports can become very noisy and confused when unrelated comments, tasks and other links are created <lp-bugs> <ubuntu-platform> <Launchpad itself:Triaged> < https://launchpad.net/bugs/73122 >
[01:27] <lifeless> aiie thanks
[01:27] <wgrant> Hmm
[01:27] <wgrant> So I guess by apparmoring lxc-start you actually end up apparmoring an entire userspace, which is probably unprecedented.
[01:27] <wgrant> Not surprising it doesn't work.
[01:28] <lifeless> wgrant: so it looks like it works but fails hard?
[01:28] <wgrant> Most stuff works
[01:28] <wgrant> But eg. dpkg-divert sometimes doesn't.
[01:29] <lifeless> 'nice
[01:29] <lifeless> '
[01:35] <wgrant> Yeah, that's it.
[01:35] <wgrant> Breaks ephemeral entirely too.
[01:35] <wgrant> Probably because overlayfs doesn't like LSM much.
[01:58] <lifeless> wgrant: filed a bug ?
[01:59] <wgrant> Nah, mailing gary about it.
[01:59] <wgrant> Needs some investigation to see what's actually going on.
[02:00] <wgrant> And I've got two further non-feature things to do today
[02:00] <lifeless> work? zomg
[02:08] <wgrant> I really hate Erlang.
[02:13] <wgrant> wallyworld_: Do you still have the rabbitmq-management-agent issue?
[02:14] <wallyworld_> wgrant: yes, not sure exactly what the root cause is, but the 4 or 5 rabbit packages don't install
[02:15] <wgrant> wallyworld_: sudo rm -r /usr/lib/erlang/lib/rabbit_management_agent
[02:15] <wgrant> There's something wrong with part of the rabbitmq-management-agent package
[02:15] <wallyworld_> wgrant: and try reinstalling now?
[02:15] <wgrant> Yes.
[02:15] <wgrant> And you can remove the libapt-pkg symlink as well
[02:15]  * wallyworld_ fires up apt-get
[02:15] <wgrant> I've fixed the PPA up.
[02:16] <wallyworld_> and the debversion package too?
[02:16] <wgrant> The fixed version is built.
[02:16] <wallyworld_> too slow at typing, yay, thanks. trying it now \o/
[02:16] <wgrant> So just upgrade
[02:17] <wgrant> A fresh installation of launchpad-developer-dependencies works in precise now, apart from the rabbitmq issue
[02:24] <wallyworld_> wgrant: yes, it indeed works. thanks, that's awesome
[02:25] <wgrant> Huh
[02:25] <wgrant> rabbitmq-management installs lots of files that should be in rabbitmq-management-agent
[02:29] <wgrant> Ahhh
[02:29] <wgrant> lynxman typoed it
[02:29] <wgrant> --- rabbitmq-management-2.6.1+hg20110927.orig/debian/install
[02:29] <wgrant> +++ rabbitmq-management-2.6.1+hg20110927/debian/install
[02:29] <wgrant> @@ -0,0 +1,2 @@
[02:29] <wgrant> +build/app/rabbitmq_management-2.6.1/* usr/lib/erlang/lib/rabbit_management_agent
[02:36] <wgrant>        85 /   69  MaloneApplication:+bugs
[02:36] <wgrant> Why don't I believe that.
[02:36] <wgrant> lifeless: Any idea about those numbers?
[02:36] <wgrant> They seem implausibly low.
[02:36] <lifeless> 144 /    0  MaloneApplication:CollectionResource:#bugs
[02:37] <lifeless> mm
[02:37] <lifeless> 2K total timeouts
[02:37] <lifeless> wgrant: '54 /  205    '
[02:37] <lifeless> wgrant: might be hiding something
[02:38] <lifeless> wgrant: the numbers seem consistent with the total timeout count
[03:31] <StevenK> wgrant: Do you mind QAing r14737?
[03:32] <wgrant> StevenK: Done
[03:34] <wgrant> lifeless: "may never" lolol
[03:36]  * StevenK reaches for mawsonm
[03:36] <StevenK> s/m$//
[03:37] <StevenK> wgrant: Plan is to get r14738 deployed to NDT as soon as mawson behaves, and hopefully FDT it onto ppa/ftpmaster tonight
[03:37] <wgrant> Correct.
[03:38] <StevenK> wgrant: I just came up with that plan, how can it be correct?
[03:38] <wgrant> It is the right plan.
[03:38] <wgrant> The only plan, even.
[03:40] <lifeless> hey
[03:40]  * StevenK waits for mawson to finish WADLing
[03:40] <lifeless> does poppy need DB access any more ?
[03:42] <wgrant> No.
[03:42] <wgrant> Well
[03:42] <wgrant> In 5.5 hours it won't.
[03:42] <wgrant> But it may still connect.
[03:43] <lifeless> cool, lets rip it out
[03:44] <StevenK> lifeless: Your bullet points should be seperate points
[03:44] <StevenK> lifeless: "or of bugs filed against launchpad-project "
[03:44] <StevenK> Which doesn't make sense either
[03:45] <lifeless>  * Increasing the number or difficulty of bugs filed against launchpad-project
[03:45] <lifeless>  ?
[03:46] <StevenK> Yes, which requires it and the previous point to be read together, and it scans awkwardly
[03:48] <wgrant> lifeless: Bug #246022: team admins/owners (and possibly members?) can contact all members (or the mailing list if one is configured). Everyone else can only contact the owner.
[03:48] <_mup_> Bug #246022: No way for team admins to contact all members of the team <contact-via-web> <feature> <lp-registry> <mailing-lists> <teams> <Launchpad itself:Triaged> < https://launchpad.net/bugs/246022 >
[03:48] <lifeless> StevenK: I'ma sking if that rephrasing is better
[03:49] <lifeless> wgrant: bah, I dropped the list condition from the summary
[03:51] <StevenK> lifeless: Oh, right. Sounds fine to me. There is a similar issue in the second group of bullets
[04:03] <nigelb> Mornin'
[04:04] <nigelb> StevenK: For a minute there, I wondered why lifeless was skiing.
[04:04] <StevenK> Hah
[04:04] <StevenK> nigelb: Looking forward to the first T20I? :_D
[04:05] <nigelb> Not really.
[04:05] <nigelb> I think they'll get crushed. :)
[04:53] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/ttb-no-chroot/+merge/91015
[05:26] <StevenK> wgrant: Can haz review?
[05:29] <wgrant> StevenK: Is there a test for _getDistroArchSeries?
[05:30] <StevenK> It just returns ubuntu.currentseries.nominatedarchindep, I suspect that is already well-tested
[05:49] <StevenK> lifeless: Didn't we kill oops prefixes?
[05:50] <StevenK> wgrant: Distracted much? :-P
[05:50] <wgrant> StevenK: Sure, but the method isn't tested any more.
[05:51] <wgrant> I would reinstate _getChroot and the test.
[05:51] <wgrant> They're doing not much harm.
[05:51] <wgrant> Or is dispatchBuildToSlave fairly well tested?
[05:52] <wgrant> Given what it does, it seems unlikely that its tests are stellar.
[06:00] <StevenK> wgrant: Why did you unassign yourself from bug 780835?
[06:00] <_mup_> Bug #780835: representation cache a pessimization <easy> <performance> <qa-ok> <tech-debt> <lazr.restful:Triaged> < https://launchpad.net/bugs/780835 >
[06:01] <StevenK> Oh, since it probably needs to be ripped out of lazr.restful now that LP doesn't make use of it?
[06:07] <wgrant> StevenK: Right.
[06:07] <wgrant> It was a lazr.restful task, which I didn't notice until after I'd landed the branch.
[06:10]  * StevenK waits for the diff to update. :-(
[06:14] <StevenK> wgrant: The diff isn't updating, but http://pastebin.ubuntu.com/824788/ is a diff against submit:
[06:16] <wgrant> StevenK: Diff looks updated to me.
[06:16] <wgrant> Note that AJAX updating is presently broken.
[06:16] <wgrant> Approvalised
[06:16] <StevenK> I've been refreshing anyway
[06:17] <nigelb> Wait. diff stuff is now queued?
[06:17] <nigelb> Or is it for launchpad team only for now.
[06:17] <StevenK> nigelb: Currently, you need to be a member of ~launchpad
[06:17] <StevenK> It needs a bit of polish
[06:17] <wgrant> And it's been broken for a couple of days, but will hopefully be back next week.
[06:18] <nigelb> Hrm, I suspect as much. But still, closer.
[06:18] <StevenK> nigelb: Jealous? :-P
[06:18] <nigelb> StevenK: You have *no* idea.
[06:18] <nigelb> 44
[06:18] <nigelb> bah irssi
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
[08:07] <lifeless> StevenK: so oops_prefix -> reporter in the new oops stack
[08:07] <lifeless> StevenK: we have glue in LP to dynamically allocate them for scripts etc
[08:07] <lifeless> StevenK: but they are still used in appservers (but with little reason, we can discriminate via hostname...)
=== almaisan-away is now known as al-maisan
[08:33] <wgrant> lifeless: Can we migrate to one-config-per-user yet?
[08:45] <lifeless> I don't see why not
[08:46] <lifeless> I only ever wanted us to just limit the rate-of-change, for auditability
[09:03] <adeuring> good morning
[09:16] <danhg> Morning
[10:16] <stub> wgrant: I see you rebuild postgresql-8.4-debversion for precise. Are you using the lucid packages atm. for the rest of things?
[10:17] <wgrant> stub: I grabbed slony (1) from oneiric in the primary archive, but had to rebuild debversion due to a libapt-pkg soname bump in precise.
[10:18] <stub> no signer listed for slony... don't know if that is a problem
[10:19] <stub> Are your PostgreSQL packages from precise?
[10:19] <wgrant> Nah, the lack of signer is just because it was synced from Debian.
[10:19] <stub> Just wondering if I should leave things as they are or try to push the pg8.4 packages into the lp ppa.
[10:20] <wgrant> postgres itself we're still using from the primary archive.
[10:20] <wgrant> Which, as you say, will probably go away in less than a month.
[10:20] <stub> I'd like devs on PG 9.1 in the next week or two in any case
[10:20] <stub> ok. I'll leave it as things are then and assume gary's missing PG packages are his problem.
[10:20] <wgrant> But we can grab 8.4 from precise once it vanishes, if we haven't moved to 9.1 by then.
[10:20] <wgrant> His problem was debversion and slony.
[10:21] <wgrant> Which I fixed a few hours back.
[10:21] <stub> oic. thanks.
[10:21] <wgrant> I also diagnosed the postgresql-common issue, as you may have seen.
[10:21] <wgrant> Unrelated to postgres, as it turns out.
[10:21] <stub> and now I see your email :)
[10:21] <stub> yes, saw that thread earlier ta.
=== matsubara-afk is now known as matsubara
[11:22] <stub> How is package syncing/package copyiHow is package syncing/package copying being done now? The sync_packages database user exists but I can't find anything in the code base that actually uses it.
[11:22] <stub> How is package syncing/package copying being done now? The sync_packages database user exists but I can't find anything in the code base that actually uses it.
[11:24] <wgrant> stub: It's used by IPlainPackageCopyJobs
[11:24] <wgrant> Which are run when IPlainPackageCopyJobSource is given to process-job-source.py
[11:24] <wgrant> Which gets the dbuser from schema-lazr.conf
[11:25] <stub> Which doesn't mention sync_packages, so must be using a different database user
[11:26] <wgrant> dbuser: sync_packages
[11:26] <wgrant> Line 1958
[11:27] <stub> pebkac. Was running bzr grep in the wrong directory.
[11:31] <wallyworld__> nigelb: watching the T20? :-P
[11:34] <nigelb> wallyworld__: Not you too :P
[11:35] <wallyworld__> oh yeah!!!
[11:35]  * wallyworld__ loves cricket
[11:35] <wallyworld__> especially when India lose :-P
[11:36] <wallyworld__> and another wicket falls!!
[11:36] <nigelb> fuuuuuuu
[11:36] <bigjools> wallyworld__: haha :)
[11:37] <wallyworld__> yep :-)
[11:39] <bigjools> nigelb, wallyworld__: new Indian interweb domain: .ball
[11:39] <wallyworld__> lol!!!
[11:39] <wallyworld__> that is hilarious
[11:39] <wallyworld__> i almost spit out my wine
[11:39] <wallyworld__> all over the keyboard
[11:40] <bigjools> my work here is done
[11:40] <wallyworld__> indeed it is
[11:40] <nigelb> bigjools: At least India didn't get bowled out for 72....
[11:40] <wallyworld__> ouch
[11:40] <wallyworld__> nigelb pwned bigjools
[11:45] <bigjools> nigelb: wanna talk about your tour to England last year? all those innings defeats? :)
[11:45] <nigelb> Still. 72. It happened even in T20 in a while.
=== al-maisan is now known as almaisan-away
[12:19] <StevenK> I was enjoying the '35 needed from 4 balls' counter
=== salgado is now known as salgado-afk
=== matsubara is now known as matsubara-lunch
=== salgado-afk is now known as salgado
=== abentley changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: gmb, abentley | Firefighting: - | Critical bugtasks: 4*10^2
[14:54] <abentley> deryck: qastaging's database appears at least a month out of date from production.
[14:55] <deryck> flacoste, is this known ^^ ?
[14:58] <flacoste> deryck: yes, it's a known problem
[14:58] <flacoste> we don't have a regular update of the qastaging db
[14:58] <flacoste> unlike staging which is reset every week-end
[14:58] <deryck> ah ok
[14:58] <abentley> flacoste: I see.  Thanks.
[14:58] <deryck> flacoste, do we schedule it on demand then?
[14:59] <flacoste> deryck: tbh, i have no idea :-/
[14:59] <flacoste> but something like that yes
[14:59] <deryck> ok, cool :)  thanks, flacoste
[14:59] <flacoste> mthaddon might know
[14:59] <mthaddon> yep, it's on demand
=== matsubara-lunch is now known as matsubara
[16:15] <rick_h> abentley: can you peek at https://code.launchpad.net/~rharding/launchpad/one_yui_instance/+merge/91111 when you get a sec?
[16:15] <abentley> rick_h: sure thing.
[16:16] <rick_h> abentley: ty
[16:21] <abentley> rick_h: Why "LPJS"?  Wasn't it something else before?
[16:21] <rick_h> abentley: it was LPS before but it was brought up that LPS stands for something and isn't as clear
[16:21] <abentley> rick_h: r=me.
[16:22] <rick_h> abentley: ty
[16:23] <deryck> actually, it was always meant to be LPJS.  But I can't type a J for some reason.  And my sed command was wrong. ;)
[16:23] <deryck> and no one caught it or complained until recently. :)
[16:34] <sinzui> jcsackett, do you have a few minutes to mumble?
[17:05] <deryck> abentley, I'm triaging bugs.  Bug 921213 looks like Bug 760735 to me.  Do you agree?
[17:05] <_mup_> Bug #921213: recipe does not build: make: Cannot open: Permission denied <Launchpad itself:Confirmed> < https://launchpad.net/bugs/921213 >
[17:05] <_mup_> Bug #760735: Daily recipe: failure because of Permission denied accessing build/patch directories <recipe> <soyuz-build> <Launchpad itself:Fix Released by jelmer> < https://launchpad.net/bugs/760735 >
[17:07] <jelmer> deryck: they are
[17:07] <abentley> deryck: yes
[17:08] <deryck> abentley, jelmer thanks!
[17:28] <deryck> abentley, adeuring, rick_h -- I did my hour.  yay me! :)  I cleared project review, open bugs, and answered a couple questions.
[17:28] <adeuring> deryck: congrats!
[17:28] <rick_h> deryck: woot!
[17:29] <abentley> deryck: cool
[17:42] <jcsackett> sinzui: i can mumble now, sorry i was on the phone when you pinged earlier.
[17:42] <jcsackett> sinzui: are you still free?
[17:42] <sinzui> I am
[17:43]  * jcsackett fires up mumble
[18:13] <lifeless> morning
[19:36] <mhall119> can I check if a person is a canonical  employee or not through the launchpadlib api, using the anonymous login?
[19:36] <james_w> no
[19:55] <mhall119> ok, then I need to get an app authorized to access that
[19:59] <dobey> mhall119: authorized in what sense? you will only be able to see members of that team, if you are also a member of it, as it's a private team.
[20:00] <mhall119> dobey: I'm adding to the script that builds bug stats for unity
[20:01] <mhall119> I need to show the number MPs coming from community
[20:01] <jelmer> mhall119: have you seen the contributors script in Launchpad?
[20:02] <mhall119> jelmer: no, does it involve zope?
[20:02] <lifeless> no, why would it?
[20:02] <jelmer> mhall119: IIRC it doesn't; it just looks at bzr commits in the mainline rather than merge proposals though
[20:02] <mhall119> lifeless: because it's launchpad?
[20:03] <mhall119> where can I see this?
[20:03] <jelmer> utilities/community-contributions.py in lp:launchpad
[20:03] <dobey> mhall119: you can probably check that the owner of the source branch with is_valid_reviewer() on the target branch, in a proposal
[20:04] <jelmer> mhall119: the result of that script is at http://dev.launchpad.net/Contributions
[20:04] <dobey> jelmer: if it returns true, the owner is part of the unity team, if it's false, the owner is community :)
[20:04] <mhall119> jelmer: where can I see the script itself?
[20:04] <jelmer> mhall119: it's linked from the wiki page
[20:05] <mhall119> dobey: what object has is_valid_reviewer?
[20:06] <mhall119> dobey: ok, I see <branch>.iPersonTrustedReviewer
[20:06] <abentley> deryck: I'm OCR.  Could you possibly review https://code.launchpad.net/~abentley/launchpad/fix-index-download/+merge/91160 ?
[20:06] <dobey> mhall119: yeah, sorry, isPersonTrustedReviewer
[20:07] <mhall119> but that only tells me if someone is on the unity team, not if they're non-employees
[20:07] <dobey> mhall119: why does it matter if they're an employee or not?
[20:07] <mhall119> dobey: because that's the metric my manager wants
[20:08] <deryck> abentley, sure
[20:08] <mhall119> besides, I'm not even sure if everyone in the dx team would be isPersonTrustedReviewer on unity's trunk
[20:08] <abentley> deryck: thanks.
[20:08] <deryck> np!
[20:09] <dobey> mhall119: possibly not, but not everyone on dx works on unity, either
=== matsubara is now known as matsubara-afk
[20:10] <mhall119> dobey: right, so I don't think jono wants DX members who don't work on unity counted in this metric
[20:11] <dobey> *shrug*
[20:31] <deryck> abentley, r=me
[20:32] <abentley> deryck: thanks.
[20:32] <deryck> abentley, np!
[20:39] <lifeless> -> l'hopital
[20:39] <lifeless> bbiab
[20:41] <mhall119> dobey: ok, I've got it working, now to optimize it
[20:41] <mhall119> dobey: is there a way to pre-fetch all the people who can review lp:unity, and check against that for each MP against lp:unity
[20:45] <dobey> mhall119: you can get the default reviewer that's set for the branch, and branch owner, and compare against those (they might be different)
[20:46] <mhall119> dobey: right, but it'll always be the same team as reviewer
[20:46] <mhall119> so can I hit LP once to get all of the members of that team, and then check locally?
[20:47] <dobey> yes; your script should check both in case they are different. if you've already got the branch object, it shouldn't be expensive to see if they are the same.
[20:48] <dobey> (to save having to fix the code later, if you decide to run the script against a project which does have different values set) :)
[20:51] <mhall119> how far back does <branch>.landing_candidates go?
[20:53] <mhall119> and is that the best way to get a historical list of merge proposals
[20:59] <abentley> deryck, rick_h: Did my interrupt duties.  Project review, some Questions, spam deletion, and I updated https://wiki.canonical.com/Launchpad/FeedbackMonitoring
[21:00] <deryck> abentley, awesome, thanks.
[21:31] <rick_h> StevenK: ping when you're around to chat make/js build dir stuff
[22:06] <StevenK> rick_h: Hmmmm?
=== abentley changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: - | Firefighting: - | Critical bugtasks: 4*10^2
[22:21] <StevenK> abentley: If you tag a bug as 'qa-bad', could you also add 'bad-commit-<revno>' ?
[22:22] <StevenK> abentley: Not trying to pick on you, it just keeps happening. I think a mail to -dev is in order.
[22:22] <abentley> StevenK: It seems absurd that I should need to.
[22:25] <StevenK> abentley: Why? Its needed for the qa-tagger if you rollback the branch
[22:26] <abentley> StevenK: Because there's only one revno I could possibly be talking about.
[22:26] <StevenK> abentley: No, there could be more -- land, qa-bad, rollback, land, qa-bad ....
[22:26] <StevenK> It has happened like that.
[22:27] <abentley> StevenK: I'm saying there *aren't* more.
[22:28] <abentley> StevenK: And at the time something is marked qa-bad, I don't think there's ambiguity either.
[22:41] <deryck_> Later on, everyone.
[23:02] <wallyworld_> rick_h: have you submitted your LPJS branch to ec2?
[23:29] <lifeless> StevenK: I agree, a mail to -dev is in order.
[23:30] <lifeless> StevenK: if the process is wrong / needs more automation, we can/should do that; but driving the tools inconsistently confuses folk and that consumes time and energy.