[00:01] <EdwinGrubbs> rockstar: it looks like it is oopsing still. about 50% of the time.
[00:02] <rockstar> EdwinGrubbs, got it to OOPS, and I have nfi what's going on.  That's really odd.
[00:03] <rockstar> EdwinGrubbs, looks like a problem with the librarian.
[06:03] <kfogel> who wants to do an _absolutely trivial_ merge proposal review?
[06:03] <kfogel> https://code.edge.launchpad.net/~kfogel/launchpad/cc-script-2010-01-28-updates/+merge/18183
[06:03] <kfogel> barry: ayt?
[06:04] <kfogel> thumper: (see above -- trivial review invite :-) )
[06:18] <cody-somerville> kfogel, I reviewed it for you. ;)
[06:21] <adiroiban> jtv: Hi. Looks like the branch from bug 359180 was release, even thou the tag was qa-bad
[06:21] <mup> Bug #359180: Missing keyboard shortcut for navigation <qa-bad> <trivial> <ui> <Launchpad Translations:Fix Released by adiroiban> <https://launchpad.net/bugs/359180>
[06:21] <jtv> adiroiban: ah, I now realize I neglected to update you on this: it was too late to do anything about it in the original rollout, so we've arranged for the fix to go out with a re-roll.
[06:22] <adiroiban> jtv: np. I just wanted to make sure everything is ok
[06:23] <jtv> adiroiban: it's a shame we found out about the problems so late, but no disasters.  It does illustrate how far we've come in terms of quality—a few years back we'd have worse things than this all the time.
[06:26] <jtv> adiroiban: I think the main culprit in this case is the confusion at PQM close...  it was known well in advance that something needed fixing, but not that PQM would close in that state—after the bad branch landed but before the one that fixed it.
[06:26] <adiroiban> jtv: yes. there is no test for loading a +translate page for a language with no plural forms
[06:27] <jtv> Well, not for the single-message view—which is where the oopses happened.
[06:29]  * jtv looks for any pagetests for this case
[06:33] <jtv> adiroiban: looks like you dug up a hidden body there.  The _browser code_ is well-tested for this case, but we don't have the integration-testing we'd get from a story.
[06:35] <adiroiban> I have not looked for pagetests... but I assume since we had one it would fail as a simple HTTP get is enought to  triger the OOps
[06:39] <jtv> adiroiban: a valid assumption in theory, but the problem with improving practices all the time is you're always carrying around code from back when practices weren't as good.  I've filed bug 513625.
[06:39] <mup> Bug #513625: Pagetest for translations without plural-forms information <Launchpad Translations:New> <https://launchpad.net/bugs/513625>
[06:41] <adiroiban> jtv: ok. I was planning to write this test togheter with the fix for bug 359180.
[06:41] <mup> Bug #359180: Missing keyboard shortcut for navigation <qa-bad> <trivial> <ui> <Launchpad Translations:Fix Released by adiroiban> <https://launchpad.net/bugs/359180>
[06:41] <jtv> adiroiban: <hug>
[06:41] <adiroiban> new bug assigned to me :)
[07:39] <noodles775> Morning
[07:46] <jtv> hi noodles775
[07:47] <jtv> wgrant: can you point me to the branch that implements SourcePackageRecipeBuild.build?
[07:47] <jtv> wgrant: it's not the one I looked at, out of the myriad you have lying around.  :)
[07:56] <noodles775> I obviously missed the start of this conversation, but SourcePackageRecipeBuild *is a* build (inheriting BuildBase), why would it have a build attribute?
[07:56] <noodles775> (hi jtv :) ).
[07:58] <jtv> noodles775: ah sorry, I copy-pasted.  I meant SourcePackageRecipeBuild itself, which would be referred to as something the lines of SourcePackageRecipeBuildJob.build.
[07:59] <noodles775> jtv: and it is? (looking at db-devel, SPRBJob.build exists?
[07:59] <jtv> noodles775: gah, I was grep'ing devel.  Silly me.  Thanks.
[07:59] <noodles775> np!
[08:04] <jtv> henninge: we got an assertion failure, as far as I can make out because someone tried to post translations while we were in read-only.
[08:05] <jtv> henninge: I'm filing a bug about it now, unless you already had.
[08:05] <henninge> jtv: hm, how could they have done it?
[08:05] <henninge> jtv: no, I have not
[08:05] <jtv> henninge: load the page before RO mode, submit after.
[08:05] <jtv> Actually I think that's probably a generic problem.
[08:05] <henninge> Ah, I see.
[08:05] <henninge> yes
[08:06] <jtv> Maybe LaunchpadFormView should check for that as part of standard validation.
[08:06] <jtv> goedemorgen jelmer!
[08:06] <jelmer> mogguh :-)
[08:07] <jtv> jelmer: I looked at the Tuesday LCA schedule the other day and found that _all_ the presentations I had really wanted to see in other rooms overlapped with mine!
[08:11] <jtv> stub, question about read-only mode: do we have any kind of generic check in form validations to deal with submissions during RO spells?
[08:13] <stub> any post generates a 'you can't do that at the moment' error page. Any page that tries to write to the db generates the same page. Any page requiring anything but read permission generates the same page.
[08:14] <stub> So you shouldn't be able to get to the form as it should be protected with Edit or similar permission, and if you do, you can't submit it since we use POST for our forms, and if you could, it couldn't make any database changes.
[08:15] <stub> As far as I can recall anyway :) Rules are in lib/canonical/launchpad/webapp/publisher.py I think.
[08:17] <adeuring> good morning
[08:26] <jtv> stub: thanks...  it does make this oops I'm looking at a bit of a mystery though  :)
[08:26] <jtv> hi adeuring
[08:27] <adeuring> hi jtv!
[08:30] <jtv> stub: oh, I see read-only state seems to be signaled outside of the db... in which case it may appear after validation and during processing of the request.
[08:31] <stub> Which raises an exception, that it rendered as a pretty Try Again Later error page.
[08:32] <jtv> Which is fine, except it shows up as an oops.
[08:38] <jtv> I think we shouldn't be failing this assertion just because we went into read-only mode.
[08:40] <jml> good morning launchpadders
[08:45] <jtv> good morning jml
[09:02] <jml> noodles775, I've just looked over the BuildBranchToArchiveUI page -- good stuff! when would you like to talk about it?
[09:03] <noodles775> jml: now?
[09:03] <noodles775> jml: or in 5mins?
[09:03] <jml> noodles775, 10 minutes :)
[09:03] <noodles775> great.
[09:04] <stub> Yay segfault trying to activate the Firebug console
[09:06] <mrevell> Hi
[09:15] <jml> mrevell, hello.
[09:15] <mrevell> Hi there jml
[09:16] <jml> noodles775, what's your skype id?
[09:32] <jml> noodles775, http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/lucid/firefox-3.5/lucid/files
[09:32] <jml> noodles775, http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/lucid/python-launchpadlib/lucid/files
[09:38] <jml> noodles775, http://pastebin.ubuntu.com/364454/
[09:51] <noodles775> jml: https://edge.launchpad.net/~michael.nelson/+archive/pocketsphinx/+builds?build_text=&build_state=all
[10:03] <jml> mrevell, hi
[10:03] <mrevell> hey there jml, same number as yesterday?
[10:03] <jml> mrevell, yes please
[10:06] <jtv> wgrant: something I don't get about Build._handleStatus_OK: if any of that long, complex method fails, it never gets to buildqueue_record.builder.cleanSlave() or buildqueue_record.destroySelf().  Doesn't that leave garbage?
[10:08] <jtv> ...did we just get spammed?
[10:27] <gmb> jtv: Heh. That seems to be the case, yes.
[10:28] <jtv> I wonder if we have a way of getting it out of our IRC logs?
[10:33] <gmb> jtv: No idea. The logs live on canonical servers, though, so I guess IS would be able to help.
[10:35]  * jtv trundles in that general direction
[10:46] <jtv> elmo suggests: "consider making launchpad-dev +r or +R"... anyone who understands that as something else than a chmod option?
[10:47] <jml> yes
[10:47] <jml> but I need chanop to do anything about it
[10:48] <jml> oh look
[10:51] <jml> That might well have worked.
[10:55] <wgrant> jtv: If it fails, the status will not be set, so it will just be reprocessed.
[11:01] <deryck> Morning, all.
[11:03] <jml> deryck, good morning.
[11:07] <adiroiban> danilos: Hi, can I continue with bug 340664 ? or you think we should wait to fix the +template timeouts ?
[11:19] <stub> So I'm generating reports from the zope logs for performance analysis. I'm now confused where things should live.
[11:20] <stub> Should the script live in scripts or utilities (I think scripts, but I've confused myself).
[11:20] <stub> And that script imports the bulk from a module in lib/lp/scripts/foo.py or lib/lp/scripts/???/foo.py or what?
[11:32] <jml> stub, I had thought that scripts was for actual scripts executed as part of Launchpad's normal operations
[11:32] <jml> stub, and that utilities is for stuff that developers & admins use.
[11:32] <jml> but, tbh, until someone writes an actual guide... meh.
[11:42] <stub> jml: I think I'm confused because I don't know if this is mainly adhoc or going to be automated ;)
[11:49] <jml> stub, ahh ok.
[11:49] <jml> stub, put it in utilities and move it when you automate it?
[11:50]  * stub shrugs. Its more admin related anyway so it can probably stay in utils
[11:56] <jml> sounds good to me.
[14:12] <kfogel> jml: want to do an utterly trivial code review?  https://code.edge.launchpad.net/~kfogel/launchpad/cc-script-2010-01-28-updates/+merge/18183
[14:13] <kfogel> intellectronica, beuno: https://code.edge.launchpad.net/~kfogel/launchpad/506018-patch-report/+merge/18181  when you have time (I'm not blocked on it landing).
[14:13] <beuno> kfogel, I have the tab open, I will get to it within the next few hours
[14:13] <kfogel> beuno: awesome, thanks
[14:17] <intellectronica> kfogel: cool. i'll review this latest after standup
[14:19] <kfogel> intellectronica: that'd be great, thank you.  Note it doesn't have your windmill tests yet, because I want to look at them / grok them / maybe extend them first.  So I'll include them in another branch.
[14:20] <kfogel> jml: please ignore above, cody somerville reviewed it
[14:21] <intellectronica> kfogel: sure. and 'my windmill tests' are nothing but a stub, as i left them on the last day of the sprint. they still need heaps of work
[14:21] <jml> kfogel, ok. I missed it anyway
[14:22] <kfogel> intellectronica: *nod*
[15:03] <intellectronica> anyone remembers the trick to make storm spit out the sql it's running?
[15:09] <intellectronica> ah, the answer is in https://dev.launchpad.net/Debugging#Tracing SQL statements through STORM :)
[15:37] <intellectronica> and now i can't run lp. i get the following error: http://pastebin.ubuntu.com/364635/
[15:37] <intellectronica> any idea what's going on and how to fix>
[15:37] <intellectronica> barry: maybe you know? ^^^
[15:37]  * barry looks
[15:38] <barry> intellectronica: do a 'make clean' then 'make'
[15:38] <barry> intellectronica: it looks like your master-qrunner.pid file got corrupted perhaps
[15:42] <barry> intellectronica: do a 'make clean' then 'make'
[15:42] <intellectronica> barry: yes, you're right, the pid file is empty!
[15:42] <intellectronica> the gnomes ate my pid
[15:43] <barry> tasty!
[15:44] <sinzui> gary_poster: I can now speculate with some certainty why closing bugs while making a release passed on staging, but not in production
[15:44] <gary_poster> sinzui: really, I'm definitely curious
[15:45] <sinzui> gary_poster: I was able to use the feature to create releases for all the launchpad projects yesterday because some developers closed the bugs...reducing the number of bugs that needed updating...
[15:45] <sinzui> gary_poster: staging's data is more than two weeks old, so it has fewer bugs in the fix committed state.
[15:47] <gary_poster> sinzui: ah-ha.  So that's another data update issue to some degree, though it's also indication of the fact that QA at a given point in time is never entirely indicative of actual performance at another point in time
[15:47] <sinzui> agreed
[15:49] <intellectronica> barry: strangely, after running make clean and make i get: http://pastebin.ubuntu.com/364646/
[15:49] <intellectronica> what's supposed to start it?
[15:50] <barry> intellectronica: make run_all
[15:50] <barry> intellectronica: but that may be misleading output.  look higher up (might need SHHH= ) for some other error
[15:51] <intellectronica> it looks like it's unhappy about something called qrunner not running
[15:52] <barry> intellectronica: what it's really saying is that something else went wrong and when killservice tried to kill mailman, it wasn't started.  that is normal output in that case (i.e. trying to shutdown mailman when it hasn't been started)
[15:52] <intellectronica> oic
[15:57] <matsubara> Chex, rockstar bigjools danilo sinzui allenap: LP production meeting in 3 min @ #launchpad-meeting
[15:57] <al-maisan> matsubara: bigjools is sick
[15:58] <al-maisan> I'll stand in for him
[15:58] <matsubara> thanks al-maisan
[15:58] <al-maisan> you are welcome :)
[16:06] <barry> intellectronica: ping me by nick if you have more information.  i'm minimizing my irc window
[18:31] <jml> flacoste, hi
[19:20] <EdwinGrubbs> beuno: ping
[19:22] <beuno> EdwinGrubbs, hi
[19:24] <EdwinGrubbs> beuno: nm, I think I answered my own question. I couldn't find individual images matching sprites, but that is probably just due to images being re-used for multiple css classes.
[19:24] <EdwinGrubbs> beuno: Is there any existing template that generates the css, or do you just add them by hand based on the existing sprite file?
[19:26] <bac> barry: ping
[19:26] <kangarooo> https://bugs.launchpad.net/bugs/297239 maybe this can be solved?
[19:27] <kangarooo> ups sory- its now reported as done
[19:27] <beuno> EdwinGrubbs, by hand, yes
[20:02] <beuno> anyone from the bugs team around?
[20:02] <beuno> gmb, intellectronica, allenap?
[20:02] <beuno> I have questions  :)
[20:19] <EdwinGrubbs> beuno: do you know anything about the bzr-favicon or person-tabs css classes. Those are the only two that don't have matching files, and they don't appear to be used anywhere.
[20:19] <beuno> EdwinGrubbs, they sounds like left over cruft
[20:20] <EdwinGrubbs> I'll add them to the garage sale
[20:23] <beuno> no bug heat in sample data?
[20:51] <beuno> sinzui, I admire how well you handle the blueprints situation
[20:51] <beuno> laying out the plan like that is really helpful
[20:52] <beuno> shows the genuine willingness to help out
[20:53] <wgrant> I agree -- comments like that one are very useful and encouraging.
[20:55] <sinzui> beuno: I think I gave the same comment to statik
[20:55] <sinzui> The work is daunting
[20:55] <beuno> good to hear Canonical has so many great people  ;)
[20:56] <sinzui> beuno: Does the u1 web team need private blueprints?
[20:56] <sinzui> or public API?
[20:57] <beuno> sinzui, I can't say for sure, but the requests came from U1 hackers
[20:58] <sinzui> beuno: Those two issues (and structural subscriptions) do not have a lot of bug heat, but thoses are the 3 most requested blueprint features. Users ask me on sprints, irc, and private email for them
[20:59] <beuno> sinzui, get them to me-too them!
[20:59] <beuno> soon, it will be up there with wikis
[21:00] <sinzui> No it wont. We need to invent something sexier than wikis. We seem to have missed the google wave
[21:01] <beuno> wiki wave?
[21:02] <sinzui> hmm
[21:03] <jml> blueprints situation?
[21:03]  * jml is interested
[21:05] <sinzui> beuno: Whether you are editing in some annoying markup syntax or a gui, wikis are just bug chunks of pages in a tree. I do not like them because they take a lot of maintenance. I have pondered if these pages would be more useful if they were composed of smaller piese that could be reused (edit it section, not this page) and I could tag them to build alternate views hierarchies. Or even construct a page from a section with 
[21:05] <beuno> sinzui, yes, I've heard your theory around chunks of data, and I think it has a lot of potential
[21:05]  * ajmitch sees the word blueprints & hides
[21:06] <beuno> smart man, ajmitch
[21:07] <ajmitch> well I saw him mention API stuff & blueprints in the same sentence & am interested to know what people really want there
[21:07] <sinzui> jml: we were discussing blueprint bug 208539. This lead to other pondering
[21:08] <sinzui> ajmitch: Launchpad's reporting and search of blueprints is very poor, and the notifications are very poor too. Users want API access to build better reports and notifications
[21:09] <ajmitch> ok, since I have a branch for bug 146389 which is pretty basic & I promised jml that I'd submit it
[21:10] <jml> \o/
[21:10] <sinzui> ajmitch: tags would help search, but reports are more complex. Many users want to see the bugs and the branch work in the same report
[21:10] <jml> sinzui, I'm thinking of proposing something to get us using blueprints more.
[21:10] <ajmitch> I looked at tags the other day & was worried at how tied to bugs they are
[21:10] <wgrant> I still do not see a compelling reason for Bugs and Blueprint to be separate.
[21:10] <ajmitch> making tags available across launchpad will be a little more work, I think
[21:11] <sinzui> ajmitch: right, I would make bugtags generic first, instead of copying the implementation
[21:12] <sinzui> wgrant: I still agree with you. I think a single good implementation would provide 90% of what bug or blueprints needs. Their differentiation is small I think.
[21:13] <ajmitch> the few unique things about blueprints are to do with sprints at the moment
[21:13] <wgrant> And the reviewer vs drafter vs assignee, and dependencies.
[21:14] <ajmitch> dependencies are something I'd love to have in bugs
[21:14] <wgrant> Particularly the latter is needed on bugs as well.
[21:14] <thumper> bug dependancies would be good
[21:14] <wgrant> They were promised at UDS Jaunty...
[21:14] <thumper> haha
[21:14] <sinzui> ajmitch: yes. Very few sprints are held to work on bugs, but sprints/meetings/conferences are certainly used by teams for other purposes that working on blueprints
[21:14] <jml> wgrant: by whom?!
[21:15] <sinzui> blueprints roles and statues are too complex. They prevent them from being widely adopted.
[21:15] <wgrant> jml: I don't remember. But it was presented as being one of the big things that would most probably be coming in 3.0.
[21:16] <sinzui> wgrant: that is an extraordinary promise given that blueprints has not had dedicated developers since 2007
[21:17] <wgrant> sinzui: Bug dependencies -- little to do with Blueprints.
[21:17] <jml> wgrant: it's always been a controversial feature, I have never heard anyone even close to promising it being delivered.
[21:17] <sinzui> wgrant: sorry, lost my concentration
[21:19] <ajmitch> jml: almost as controversial as having version information for bugs on packages as metadata, rather than text?
[21:19] <sinzui> yet we hack dependencies every 6 months using the tags, blueprint-to-bug-links, and BjornT_'s progress grid
[21:19] <jml> ajmitch, I've never heard of that one.
[21:19] <jml> ajmitch, wgrant: there's a bug in malone about dependencies
[21:20] <intellectronica> beuno: still have a question?
[21:20] <ajmitch> jml: that's surprising, the bug version information is something that debian has
[21:20] <wgrant> version information is very useful.
[21:20] <ajmitch> I'll look up a bug # for it, I'm sure it's in there :)
[21:20] <wgrant> It's one thing that debbugs has over Launchpad that makes Launchpad look pretty pathetic.
[21:21] <jml> ajmitch, it's surprising I haven't already heard of every single major feature request for Launchpad :)
[21:21] <ajmitch> jml: given that complaints that roll in, sure ;)
[21:21] <ajmitch> bug 424
[21:22] <ajmitch> so really early :)
[21:22] <beuno> intellectronica, hey!  yes, how do I see bug heat on launchpad.dev?
[21:23] <beuno> I bring up trunk and get no flames!
[21:23] <wgrant> db-devel
[21:23] <beuno> I wanted to take a crack at making a lighter icon and such
[21:23] <intellectronica> beuno: you wait until db-devel gets merged back in?
[21:23] <beuno> aha
[21:23] <intellectronica> (or branch from db-devel)
[21:24] <intellectronica> beuno: let me guess, you're creating better icons for us? :)
[21:24] <beuno> intellectronica, that was my intention, lets see how much past that it goes
[21:24] <beuno> I also sent an email about the feature as a whole
[21:25] <intellectronica> beuno: fountains of beer await you
[21:25]  * beuno stores away his drivers license, won't use it for sure this trip
[22:13] <wgrant> What happened to the post-3.0 notifications focus?
[22:14] <wgrant> Is that still happening at some point, but just pushed back by the upstream linkage work?