[00:04] <sinzui> wgrant: mumble?
[00:07] <sinzui> lifeless: I thought expiration was 60 days. That bug was commented 33 days ago
[00:08] <lifeless> sinzui: I thought bugs went 'eligible for expiry in X days' and then 'marked for epxiry Y days ago'
[00:08] <lifeless> sinzui: I may be misunderstanding
[00:09] <lifeless> sinzui: https://bugs.launchpad.net/launchpad/+bug/616704 is a better example
[00:09] <_mup_> Bug #616704: Security package copy should copy packages to -updates immediately <lp-soyuz> <Launchpad itself:Incomplete> < https://launchpad.net/bugs/616704 >
[00:09] <sinzui> lifeless: Yes that should be expired
[00:09] <lifeless> sinzui: (not that that should be in this state, because bigjools was just asking a question)
[00:09] <lifeless> but  it demonstrates the issue well
[00:10] <sinzui> lifeless: The days marked is from the moment it qualifies to be expired...incomplete, un assigned, only one task etc...
[00:10] <lifeless> sinzui: so its not 60 + count, its 'will expire when count == 60' ?
[00:11] <sinzui> oh, did our db merging of projects break expiration for us?
[00:11] <sinzui> yes equality
[00:11] <lifeless> the phrasing is confusing then
[00:17] <poolie> o/ lifeless, sinzui
[00:21] <lifeless> hi poolie
[00:21] <lifeless> sinzui: I thought it used to say 'will expire in N days'.
[00:21] <lifeless> sinzui: in fact, I'm pretty sure it does.
[00:23] <sinzui> lifeless: it does say that
[00:24] <lifeless> sinzui: at what point does it switch to 'marked for expiry' ?
[00:25] <sinzui> when all the conditions are met when the status incomplete is set
[00:26] <sinzui> https://dev.launchpad.net/MaintenanceRotationSchedule
[00:28] <lifeless> sinzui: I don't see anything on that page relevant to this
[00:28] <sinzui> lifeless: sorry, that was for teal
[00:28] <lifeless> ;)
[00:32] <lifeless> sinzui: I think the 'marked for X days ago' turns up on the day it can expire
[00:32] <lifeless> sinzui: so I think X days ago is last-changed + 60
[00:32] <lifeless> 702022	Once a project is modified you can no longer modify the status		2011-01-14	
[00:33] <lifeless> is 91 days ago
[00:33] <lifeless> and it shows 'marked 31 days ago'
[00:33] <lifeless> so, if its marked X days ago, it should have expired if bug expiry is working
[00:33] <sinzui> lifeless: Changed could be ambiguous. I think there is a specific date, like date_incomplete_without_followup
[00:33] <lifeless>  date_last_updated      | timestamp without time zone | not null default timezone('UTC'::text, now())
[00:34] <lifeless>  date_made_private      | timestamp without time zone |
[00:34] <lifeless>  who_made_private       | integer                     |
[00:34] <lifeless>  date_last_message      | timestamp without time zone |
[00:34] <lifeless> date_last_message is the one used
[00:34] <lifeless> but note that adding a message sets date_last_updated
[00:34] <lifeless> so date_last_message can only ever be older than date_last_updated
[00:41] <sinzui> lifeless: These three methods look like the issues affecting expiration, but I do not see what is a miss yet: http://pastebin.ubuntu.com/593809/
[00:42] <sinzui> Sorry. I was missing the forth method from bgutaskset:http://pastebin.ubuntu.com/593811/
[00:54] <lifeless> days-before-expiration is unset in prod, so default value of 60
[00:57] <lifeless>     enable_bug_expiration = BoolCol(dbName='enable_bug_expiration',
[00:57] <lifeless>         notNull=True, default=False)
[00:57] <lifeless>  is a simple lookup on product
[00:57] <lifeless> no fancy check-group stuff
[00:59] <lifeless> sinzui: it may be something simple, like not running the cronjob
[00:59] <sinzui> I was thinking the same
[01:00] <lifeless> or
[01:01] <lifeless> it may be that its running in a very slow loop over one project at a time
[01:01] <lifeless> with 20K+ projects that will perform very poorly.
[01:01] <sinzui> I see it in loganberry-launchpad to run "17 04 * * * "
[01:01] <sinzui> 10,000 projects have nothing in them
[01:01] <lifeless> yeah
[01:01] <lifeless> but its still 10K lookups
[01:01] <lifeless> when the DB knows
[01:01] <lifeless> and can do it in one hit.
[01:17] <lifeless> sinzui: it must be getting late for you
[01:17] <lifeless> sinzui: when did you want to talk ?
[01:18] <sinzui> lifeless. Nothing I have nothing conclusive to talk about it seams. I think teal needs to do a better job watching scope of the tasks it undertakes
[01:18] <sinzui> lifeless: Are you available tomorrow?
[01:20] <lifeless> sinzui: sure am
[01:21] <sinzui> I will ping you tommorow then
[01:21] <lifeless> cool
[01:21] <sinzui> or maybe tomorrow.
[01:21] <lifeless> Tomorrow, when the OOPS ended
[01:27] <cinerama> lifeless: lol
[01:33] <wgrant> lifeless: :( CPU graphs are odd this morning.
[01:33] <lifeless> wgrant: oh?
[01:33] <lifeless> cinerama: :)
[01:34] <wgrant> lifeless: There was a big drop right after the new appserver was activated, but not long later they returned to normal :(
[01:34] <lifeless> wgrant: well, thats to be expected
[01:34] <lifeless> wgrant: when the new appserver was turned off
[01:37] <wgrant> lifeless: Oh, it was turned off?
[01:37] <wgrant> Maybe there was another incident I haven't read about yet.
[01:37] <lifeless> ~2am your time
[01:37] <lifeless> yeah
[01:37] <wgrant> Aha.
[01:38] <lifeless> linked in -ops
[01:38] <wgrant> Yeah, found it.
[01:38] <wgrant> It's still down? :/
[01:39] <wgrant> Hmm.
[01:39] <wgrant> Odd.
[01:39] <lifeless> we should be able to run the smoketest and so on and bring it up
[01:42] <sinzui> lifeless: wgrant: mthaddon: was at the EOD when he decided to take the new server offline. He had made changes to give the server access to the internal librarian, but jelmer and I did not see them take affect
[01:43] <wgrant> sinzui: Do you know which rules?
[01:43] <lifeless> sinzui: yah, it was the right thing to do
[01:43] <wgrant> I suspect the restricted ports were forgotten.
[01:43] <wgrant> (I wish we had transparency here)
[01:44] <sinzui> wgrant: I think the same, but I do not know for certain
[01:45] <wgrant> Hopefully we will run out of fires in a few hours and be able to debug.
[02:09] <huwshimi> lifeless: I'm looking at a bunch bugs related to user profiles. Would it be appropriate to tag these all with "profile" or something?
[02:09] <lifeless> huwshimi: if you like
[02:09] <lifeless> huwshimi: basically, if its useful to you, do it.
[02:10] <huwshimi> lifeless: Just in regards to your email, I don't want to add more clutter (I know this is not an official tag).
[02:10] <lifeless> huwshimi: folk that are trying to avoid scope creep will often not want to fix clusters of bugs (because the cluster can include a ringer)
[02:10] <lifeless> huwshimi: I have no objection to large numbers of tags
[02:10] <lifeless> huwshimi: official tags are special because they always show up everywhere, even if not used much (or used so much they are not valuable, like the lp-* ones)
[02:11] <lifeless> huwshimi: I hope that helps
[02:12] <huwshimi> lifeless: Yes, thanks.
[02:21] <lifeless> mwhudson: hey
[02:21] <mwhudson> lifeless: hello
[02:21] <lifeless> mwhudson: does the branch rewriter script have a db statement timeout ?
[02:21] <lifeless> I suspect it doesn't
[02:22] <mwhudson> lifeless: i believe not, i think there may be a bug for this already
[02:22] <lifeless> there is
[02:22] <lifeless> its critical
[02:22] <lifeless> I'm proposing to make it 500ms
[02:22] <lifeless> how badly will the rewriter handle a timeout exception ?
[02:22] <lifeless> as in, will it die
[02:22] <lifeless> or shrug and move on?
[02:23] <mwhudson> lifeless: i don't know, i'd have to look at the code
[02:23] <lifeless> mwhudson: it looks like it will be ok to me.
[02:23] <mwhudson> actually, i'm pretty sure it will shrug
[02:44] <lifeless> mwhudson: do you think this needs tests? http://pastebin.com/3HHjiQkL
[02:44] <lifeless> mwhudson: [please say no]
[03:02] <mwhudson> lifeless: i dunno, could you monkey patch _getBranchIdAndTrailingPath and check that there is a request when it's called?
[03:03] <lifeless> mwhudson: I could, but it seems like a comment at the call site is as good insulation
[07:47] <huwshimi> lifeless: Is that bug about strikethroughs on closed bugs now a duplicate of bug #1924 now that we've changed the description?
[07:56] <lifeless> huwshimi: indeed
[08:12] <huwshimi> Should this bug be unassigned:https://bugs.launchpad.net/launchpad/+bug/58784 ?
[08:12] <_mup_> Bug #58784: Spec page refers to non-existent "Specifications in grey" <feature> <lp-blueprints> <ui> <Launchpad itself:Triaged by sabdfl> < https://launchpad.net/bugs/58784 >
[08:17] <lifeless> heh
[08:17] <lifeless> I suspect so
[08:40] <lifeless> wgrant: dunno about you but today was a write off
[08:43] <StevenK> I managed to get 1.5 branches done. I was aiming for 2.5
[08:46] <wgrant> lifeless: Yeah, it was pretty hopeless.
[08:46] <wgrant> I'll hopefully get some stuff done tonight, though :/
[08:49] <lifeless> StevenK: nice
[09:01] <LPCIBot> Project devel build #638: FAILURE in 5 hr 5 min: https://lpci.wedontsleep.org/job/devel/638/
[09:09] <adeuring> good morning
[09:12] <wgrant> Not really.
[09:13] <huwshimi> wgrant: Everyone's a bundle of laughs tonight :)
[09:13] <bigjools> morning
[09:34] <jml> huwshimi: hi
[09:35] <huwshimi> jml: Hey mate
[10:36] <bigjools> anyone know a way of making pdb break on a class?
[10:38] <lifeless> on definition?
[10:38] <lifeless> or construction ?
[10:46] <bigjools> lifeless: running any code inside a class
[10:46] <bigjools> it seems to want either filename:lineno or a function
[10:46] <lifeless> yeah
[10:46] <bigjools> the best I can do is put the class in its own file
[10:47] <lifeless> probably won't do base classes for you
[10:47] <bigjools> which is exactly what I want :/
[11:07] <bigjools> If I have an object that inherits from IHasOwner, I can't seem to use a custom launchpad.Edit security adapter because the EditByOwnersOrAdmins adapter is used - is there a way around that or do we have to just not use lp.Edit?
[11:28] <henninge> jtv: Hi!
[11:28] <jtv> hi
[11:28] <henninge> jtv:  I forgot: what is needed to trigger pottery processing on a branch, i.e. automatic template creation?
[11:28] <henninge> jtv: I know the branch needs an intltool setup. What else?
[11:29] <henninge> jtv: Is that usable for projects?
[11:29] <jtv> henninge: template imports enabled, POTFILES.in present, can't think of anything else right now.
[11:29] <jtv> Also, I'm not here.
[11:29] <henninge> jtv: thanks ;)
[11:45] <lifeless> bigjools: sorry, I don't know the answer. perhaps curtis/benji or gary would be good folk to ask
[11:46] <bigjools> lifeless: thanks, we can work around it my removing IHasOwner
[11:47] <lifeless> bigjools: perhaps justhave a more specific class first ?
[11:48] <bigjools> lifeless: yeah it's almost certainly down to ordering, but inherited classes come first it seems, not much you can do
[12:18] <jtv> bigjools: I marked the DSDs-with-identical-versions fix as qa-ok since it's safe to deploy, but haven't done proper Q/A on it.
[12:18] <jtv> Ah.
[12:18] <bigjools> jtv: I literally just tested them ok
[12:18] <jtv> There it goes.
[12:18] <bigjools> :)
[12:18] <jtv> Thanks.
[12:25] <jtv> Damn I hate it when this happens.  Spend hours chasing down a report of a tiny memory leak in libpqxx against one particular postgres version, and it's just valgrind crying foul murder over libnss caching two layers down.  :/
[12:25] <danilos> allenap, hi, are you OCR today? ;)
[12:25] <allenap> danilos: Yep. Whatchagot?
[12:26] <danilos> allenap, a nice little JS branch: https://code.launchpad.net/~danilo/launchpad/bug728370-descriptions/+merge/57656 :)
[12:27] <allenap> danilos: Tip top.
[12:27] <danilos> allenap, thanks :)
[12:44] <danilos> allenap, I'll be logging out and back in, should be back shortly (if xchat keeps working with gnome3 running :)
[12:44] <allenap> danilos: Hehe :) No worries.
[13:17] <wallyworld> henninge: hi. there's a few translation imports that need reviewing. i'm not sure how to do that. are you or jtv still the best ones to ask?
[13:18] <henninge> wallyworld: probably. and danilos ;)
[13:19] <wallyworld> henninge: the last time i checked i don't think the required doco on what to do for us noobs had been fully developed?
[13:20] <henninge> wallyworld: I thought jtv had written something up.
[13:22] <wallyworld> henninge: yes, right you are. i'll take another look through it
[13:26] <wallyworld> henninge: "For now, if you get a request to review and approve a translations upload, it's still best to forward it to one of the former Translations crowd: Danilo, Henning, and Jeroen". i'll give it a go but be prepared that i may need to as questions :-)
[13:41] <danilos> allenap, hi, having any questions about the branch?
[13:42] <allenap> danilos: None yet; I just had my lunch :) I'm going through it now.
[13:42] <danilos> allenap, ok, thanks
[13:45] <henninge> wallyworld: we can look at a couple of them now if you want but I'll have my stand-up call in 15 min
[13:46] <wallyworld> henninge: it's ok. i'm reading through all the doc i can find. i'll see how i go and perhaps ask tomorrow if i get stuck. i just don't want to do something dumb
[13:47] <henninge> wallyworld: there are 13 entries in the queue so it should not be too hard
[13:48] <wallyworld> henninge: yep. and some are not that old so may yet be done by the cron job
[13:49] <henninge> wallyworld: possibly. Let's have a look at it together tomorrow.
[13:50] <wallyworld> kk
[14:05] <jml> beuno: did you write up UI guidelines for Launchpad?
[14:06] <jml> mpt: did you write up UI guidelines for Launchpad? Do they still exist? (I know of UserInterfaceWording only)
[14:07] <beuno> jml, nothing besides: https://dev.launchpad.net/UI/
[14:07] <jml> beuno: thanks.
[14:08] <mpt> jml, I wrote <https://launchpad.canonical.com/MatthewPaulThomas/BetterDesign/> (sorry, Canonical-only link) in 2008, but didn't have the opportunity to propose them as LP guidelines
[14:09] <jml> mpt: ok, thanks.
[14:09] <mpt> (Actually, most of that's from 2006)
[14:29] <allenap> jam: Would you mind reviewing https://code.launchpad.net/~jstpierre/loggerhead/css-changes/+merge/56393? I don't really understand what he's doing :-/
[15:01] <adeuring> allenap, jcsackett: could one of you please review this mp: https://code.launchpad.net/~adeuring/launchpad/bug-746460-productseries/+merge/57685 ?
[15:02] <jcsackett> adeuring: sure
[15:02] <adeuring> jcsackett: thanks!
[15:04] <abentley> jml: can we chat about bug #758857?
[15:04] <_mup_> Bug #758857: Linked branch in translations sharing details page shows branch URL without lp:// prefix <exploratory-testing> <upstream-translations-sharing> <Launchpad itself:Triaged by abentley> < https://launchpad.net/bugs/758857 >
[15:07] <jml> abentley: sure
[15:07] <abentley> jml: mumble?
[15:07] <jml> abentley: yep.
[15:08] <jam> allenap: I'll give it a shot, though if his screenshots are accurate, something is wrong
[15:08] <allenap> jam: Cheers :)
[15:15] <rvba> jcsackett: Hi, could you have a look at https://code.launchpad.net/~rvb/launchpad/fix-package-diff-equal-versions/+merge/57686 ?
[15:16] <jcsackett> rvba: sure, it's next in the queue. :-)
[15:16] <rvba> jcsackett: great.
[15:17] <jml> abentley: https://launchpad.net/ubuntu/natty/+source/ssss/+edit-packaging
[15:18] <jml> abentley: https://bugs.launchpad.net/testtools/+subscriptions
[15:19] <sinzui> bigjools: do you have any insight into question https://answers.launchpad.net/launchpad/+question/151985
[15:20] <bigjools> looking
[15:20] <wgrant> sinzui: I could not reproduce that locally with what I believe to be a broken sudo.
[15:20] <wgrant> I may try harder to verify the bad umask.
[15:21] <bigjools> it is certainly very coincidental with the broken sudo
[15:25] <bigjools> sinzui: since it's happening as far back as lucid, I suspect it's not the broken sudo actually.
[15:26] <bigjools> he's probably found a new bug, but my knowledge of recipes is not good enough to help without spending more time on this
[15:26] <sinzui> bigjools: I was thinking this permission issue related to the umask problem last week, but lamont believes that is fixed
[15:26] <bigjools> sinzui: the umask was done to paper over the sudo problem AFAIK
[15:27] <bigjools> and yes it was fixed
[15:27] <bigjools> fairly quickly :)
[15:27] <bigjools> it was natty only
[15:27] <sinzui> bigjools: understood
[15:28] <jcsackett> adeuring: r=me.
[15:28] <bigjools> sorry I can't be of more help
[15:31] <wgrant> bigjools: It wasn't natty-only :(
[15:31] <wgrant> bigjools: It's the system sudo.
[15:31] <bigjools> ah ok, thanks, I misunderstood then
[15:32] <bigjools> so it was the sudo outside the chroot?
[15:32] <wgrant> Right.
[15:32] <bigjools> oh dear
[15:32] <wgrant> In the original incident a few weeks ago only natty was majorly affected, because it was the majority of the builds.
[15:40] <jcsackett> rvba: r=me.
[15:41] <rvba> jcsackett: thanks.
[15:48] <danilos> allenap, hi, responded to your review :) I'd like to get your comments before I proceed to land this, because I mostly haven't changed anything (did say why not :)
[15:49] <danilos> allenap, also, I just checked, the slide-out/in effects take 0.4s, which is why 0.5 worked for me consistently
[15:51] <allenap> danilos: Cool, I'll take a look.
[15:58] <danilos> allenap, sorry if I missed anything you said, gnome3 crashed with keyboard layout change :/
[15:58] <allenap> danilos: Hah, no.
[15:59] <allenap> danilos: setup_bug_subscriptions() calls fill_in_bug_subscriptions() which appends to a list that *is* in the page's DOM.
[15:59] <danilos> allenap, right, but it doesn't get called before show_* function is called, and that one is called in 'domready'
[15:59] <allenap> danilos: Oops, I get it.
[16:00] <danilos> allenap, I originally had it inside 'domready' but figured there's no need... perhaps a comment would do :)
[16:00] <danilos> with "would do" == "is a must" ;)
[16:03] <allenap> danilos: Thanks for the response though. I'm still not a fan of using IDs as much as we do, but it's interesting to hear other people's approaches :)
[16:04] <danilos> allenap, heh, yeah  I'd love them even more if web page somehow failed to render if you've got duplicate IDs
[16:07] <adeuring> gary_poster: do you have time for a pre-imp call?
[16:08] <gary_poster> adeuring: yes, in...15 min or so?  Is that ok?
[16:08] <allenap> danilos: You said that there's no way for a browser to use a selector, except when a browser has the Selectors API. Can they not fall back to XPath? I assumed that's what YUI, jQuery, et al did.
[16:08] <adeuring> gary_poster: sure, thanks!
[16:08] <gary_poster> np, talk to you then
[16:09] <allenap> danilos: Perhaps I should code up a cheeky checkers that does a document.write("<h1>Boom!</h1>") when duplicate IDs are found ;)
[16:09] <danilos> allenap, yeah, selectors API is the querySelector* stuff; I assume that gets much easier for the platform when you narrow the scope because it's not just IDs it has to go through
[16:09] <danilos> allenap, heh, I wouldn't mind :)
[16:10] <danilos> allenap, anyway, I personally feel better asking in a more constrained set, perhaps it's totally irrelevant, but please don't make me do it in the entire document scope! pleaaaseee! ;)
[16:11] <danilos> allenap, also, doing it on a node is a sanity check at the same time: you haven't moved the element someplace else
[16:12] <allenap> danilos: I agree, always start from the narrower scope. Although I suspect that query by ID might be quicker in document scope (because the browser may optimize for that case), I think the difference is going to be minimal.
[16:13] <danilos> allenap, yeah, but there's also getElementById on the DOM node, so that should be even faster :)
[16:15] <danilos> allenap, btw, I've added a comment explaining the sequence of events, and now I am going to ec2 land the branch :) thanks again for the review!
[16:15] <allenap> danilos: Cool, and you're welcome :)
[16:16] <gary_poster> adeuring, actually I'm ready now.  I was overestimating for once, it seems. :-) I'm on Skype (garyposter) and mumble.  mumble is sometimes weird for me, but sometimes its fine, so either is worth a try. :-)
[16:16] <adeuring> gary_poster: cool. shall we  mumble?
[16:16] <gary_poster> sure
[16:17] <gary_poster> I'm in Yellow: 1-on-1.  Join me or drag me somewhere
[16:26] <jcsackett> benji: ping.
[16:38] <jcsackett> does anyone know how to restrict version when exporting a webservice collection? i know you can use as_of when exporting an entry, but that's not supported for collection.
[17:00] <jcsackett> sinzui: can you chat for a few minutes?
[17:00] <sinzui> yes
[17:08] <abentley> jcsackett: Do you mean a resource type?  Or an attribute of an object that is a collection?
[17:08] <jcsackett> abentley: the former, i think. e.g. exporting IQuestionSet using "export_as_webservice_collection"
[17:09] <abentley> jcsackett: It's a known issue that there is no way to restrict that to a particular version.
[17:09] <jcsackett> abentley: dig. thanks.
[17:12] <jcsackett> abentley: is there a bug already filed for it, do you know? i can't seem to find one for on lazr.restful
[17:13] <abentley> jcsackett: don't know.
[17:13] <jcsackett> abentley: okay, thanks. i'll dig around a bit.
[17:31] <jcsackett> sinzui: i have put up the MP and requested you as reviewer. https://code.launchpad.net/~jcsackett/launchpad/api-wants-questionset/+merge/57723
[17:31] <sinzui> jcsackett: thanks
[17:51] <rvba> jcsackett: another small fix ... could you have a look? https://code.launchpad.net/~rvb/launchpad/fix-packagediff-request-link/+merge/57727
[17:52] <jcsackett> rvba: sure
[17:59] <jcsackett> rbva: r=me. i mention in my comment there's an alternative way to do some of your testing, but it's just a suggestion, not block to landing.
[17:59] <rvba> jcsackett: thanks, I'll have a look right now.
[18:01] <rvba> jcsackett: it's a nice suggestion ... I think I'll keep it this way because I've factored the 2 asserts in a method that does a little bit more than just testing ... but I'll definitely do it next time.
[18:02] <jcsackett> rvba: dig.
[18:30] <LPCIBot> Project windmill build #174: STILL FAILING in 1 hr 5 min: https://lpci.wedontsleep.org/job/windmill/174/
[19:01] <adeuring> gary_poster: about the canAccess(), canWrite() methods we discussed: Any suggestions for an exception that should be raised if the number of participations is wrong?
[19:01] <adeuring> gary_poster: ValueError sounds quite generic to me... but I have no better idea
[19:06] <gary_poster> adeuring: if the number of participations in the current interaction is wrong, that is a NotImplementedError from the perspective of what we discussed before
[19:06] <gary_poster> Because if you make your own interaction, you can get the answer irrespective of the current one.
[19:06] <gary_poster> So there is a fix for this
[19:06] <adeuring> gary_poster: ah right, seems that it is too late here to ask such questions ;)
[19:07] <gary_poster> We just are not bothering with it for now.
[19:07] <gary_poster> :-) np at all
[19:52] <sinzui> jcsackett: do you have a moment to mumble?
[19:52] <jcsackett> i do.
[19:54] <lifeless> morning
[20:02] <adeuring> abentley: lp:~adeuring/launchpad/api-query-permissions-on-object (not yet submitted for review; will do that tomorrow -- it's alread 9pm for me)
[20:03] <abentley> adeuring: cool.
[20:06] <jcsackett> sinzui: http://paste.ubuntu.com/594169/
[20:32] <LPCIBot> Project windmill build #175: STILL FAILING in 1 hr 4 min: https://lpci.wedontsleep.org/job/windmill/175/
[20:37] <lifeless> flacoste: hi
[20:38] <flacoste> hi lifeless
[20:39] <lifeless> so, major outage
[20:39] <lifeless> I think we should talk about this on the blog
[20:44] <flacoste> lifeless: the switch thing? sure, care to write something? otherwise, we could ask mrevell, but he will need guidance
[20:44] <lifeless> I'll write something in mail and send it to a couple of internal lists for commentary
[21:21] <jcsackett> sinzui: changes pushed to the review branch.
[21:21] <jcsackett> i've also started in on the launchpadlib thing.
[21:24] <sinzui> jcsackett: thanks
[22:02] <thumper> morning
[22:24] <sinzui> jcsackett: r=me
[22:28] <poolie> jcsackett, the questionset api might be good to mention on the blog as a new feature
[22:32] <jcsackett> poolie: indeed. but first, finishing it. :-P
[22:33] <jcsackett> sinzui: thanks for the review. :-)
[22:35] <sinzui> poolie: We cannot do anything with the questions, nor can you search for them. It is only useful for hiding a comment on a question you know exists :(. I hope to hack on the seaarch and workflow methods of the next two weekends to make something I can personally use
[22:38] <poolie> heh, ok
[22:38] <poolie> you can't even look at them if you know the number?
[22:39] <sinzui> poolie: let me check. I think we can do that in prod right now
[22:40] <poolie> it looks like you can, if you guess the url
[22:40] <poolie> (which is not so hard)
[22:41] <sinzui> yep
[22:41] <sinzui> I want to get the questions for a project or a group of distribution source packages
[22:43] <poolie> that would be useful
[22:43] <sinzui> Indeed. I cannot see all the open questions for all the things I am an answer contact for
[22:45] <flacoste> sinzui: do you know how the ProductWithLicense SQL expands to?
[22:45] <flacoste> iow, how to i get the related rows as an array column in the main query
[22:46] <sinzui> I do not know
[22:46] <sinzui> flacoste: I think I want to make that query obsolete.
[22:46] <flacoste> ok
[22:46] <lifeless> flacoste: more graphs?
[22:46] <flacoste> lifeless: yep :-)
[22:46] <lifeless> flacoste: oh, and did you have a chance to look at the uncommitted changes stub removed ?
[22:46] <lifeless> from the PPR tree
[22:47] <flacoste> lifeless: not yet, but i'll get to it
[22:47] <lifeless> flacoste: no rush :)
[22:47] <sinzui> I am sitting on data that provides the missing licenses for about 800 projects. I then want to set the remaining projects to have I_DONT_KNOW
[22:47] <flacoste> i think it's fine
[22:47] <lifeless> monday hopefully we can change the timeout again
[22:47] <flacoste> since i recall commiting it at some point
[22:48] <flacoste> i want to show # of unreviewed projects
[22:48] <flacoste> and track the 'special' licenses one separately from the regular ones
[22:48] <lifeless> flacoste: I remember what I wanted to ask you
[22:48] <lifeless> flacoste: bug count aggregates, what were your thoughts
[22:49] <flacoste> lifeless: ideally, i think exact below 50 is desireable
[22:50] <flacoste> above that it can be fuzzy
[22:50] <flacoste> i liked the render fuzzy, get exact number async
[22:50] <flacoste> approach
[22:50] <flacoste> but i'm also fine with fuzzy only
[22:51] <lifeless> I think there are a few ways we can be really precise
[22:51] <lifeless> I think its not worth it in the short term
[22:51] <lifeless> - the polls are pretty seriously weighted towards speed
[22:52] <lifeless> - precision will definitely still have a cost
[22:54] <james_w> lifeless, will the system know if the number is fuzzy?
[22:54] <lifeless> james_w: in principle yes
[22:55] <james_w> so the rendering could include that information? I think that would help alleviate some concerns
[22:55] <james_w> "3 bugs" vs "about 500 bugs"
[22:55] <lifeless> yes, thats one way to handle it
[22:55] <lifeless> I agree its not ideal to present a wrong count as fact; however...
[22:56] <lifeless> note that /all/ our counts are usually wrong by the time someone clicks through.
[22:56] <james_w> still doesn't help if you want to know the exact number, but avoids confusion if you click what looks like a precise number to find a different number of results
[22:56] <james_w> well
[22:56] <james_w> on some projects
[22:56] <lifeless> james_w: we have broken memcached rules
[22:56] <lifeless> james_w: before you guess at the cause.
[22:56] <lifeless> james_w: we cache private results as publically visible, and anonymous likewise.
[22:57] <lifeless> james_w: for these portlets specifically.
[22:57] <lifeless> We noticed last week.
[22:57] <lifeless> Its been like that for ... some time.
[22:57] <james_w> ok
[22:57] <lifeless> james_w: I'm trying to separate out:
[22:57] <lifeless>  - things we can do to make it better than it is now
[22:57] <lifeless>  - things we must to do to increase performance
[22:58] <lifeless> So while I accept and agree that putting tasteful caveats around the place, (?) popups and so on are all good things
[22:58] <lifeless> I am not convinced that they are necessary to make switching to precalculated aggregates a net win.
[22:59] <lifeless> And if something is a net win, I think we should take an iterative approach rather than biting off more than needed.
[22:59] <lifeless> drive velocity up and get out of the 200+ critical bug zone.
[23:00] <lifeless> bac:  around? bug 753965 needs qa
[23:00] <_mup_> Bug #753965: An IStructuralSubscriptionTarget that does not use LP for bug tracking should not present a subscription link <qa-needstesting> <Launchpad itself:Fix Committed by bac> < https://launchpad.net/bugs/753965 >
[23:00] <lifeless> abentley: likewise - Fixes: Bug:702477
[23:03] <flacoste> http://pastebin.ubuntu.com/594235/ is what i came with
[23:03] <flacoste> any other suggestions
[23:03] <flacoste> ?
[23:13] <lifeless> flacoste: looking
[23:13] <LPCIBot> Project devel build #639: STILL FAILING in 5 hr 35 min: https://lpci.wedontsleep.org/job/devel/639/
[23:14] <lifeless> flacoste: seems reasonable to me
[23:15] <lifeless> mtaylor: poolie mentioned a jenkins issue you're having?
[23:15] <mtaylor> lifeless: yeah- I can't remember what it was right now
[23:15] <lifeless> he showed me a java log
[23:16] <lifeless> looks like slave reuse race condition to me, not a bzr plugin issue
[23:16] <lifeless> IMBW
[23:26] <LPCIBot> Project windmill build #176: STILL FAILING in 1 hr 10 min: https://lpci.wedontsleep.org/job/windmill/176/
[23:36] <wallyworld_> lifeless: i was going to test "recipe request daily build script triggers OOPS on disabled archive" but getting the make_daily_builds script to be run on qas and check the log. you don't think this is needed?
[23:36] <lifeless> wallyworld_: I think its fine not to - I looked at the change
[23:36] <wallyworld_> lifeless: np. saves me some work :-)
[23:48] <lifeless> sinzui: hi
[23:48] <lifeless> sinzui: shall we chat?
[23:49] <sinzui> yes
[23:50] <lifeless> skype?
[23:50] <sinzui> yes
[23:51] <abentley> lifeless: I'm missing context.
[23:51] <lifeless> abentley: ah there is a patch with you as assignee marked needing qa
[23:51] <lifeless> abentley: there were two, but I wasn't sure about this one
[23:54] <abentley> lifeless: yeah, that one's tricky and the steps to reproduce never went into the bug report.
[23:58] <lifeless> abentley: I would like to do a nodowntime deploy during asiapacs day; if you can qa that we'll know whether we can deploy everything or just up to the commit before yours
[23:59] <thumper> ah fuck
[23:59] <thumper> :(