[07:31] Hi jtv - do you know if anyone's working on the current buildbot failure? lp/translations/doc/gettext-check-messages.txt [07:31] noodles775: no idea; let me check my irc backlog [07:32] noodles775: nothing in my backlog... was late logging in. [07:32] jtv: thanks - I can't see a failure email either... do you? If not, I'll check why the email didn't get sent. [07:33] I'm checking my mailbox, but I may already have deleted it if I got it. [07:33] failure on jscheck... [07:33] yeah, saw that one. [07:35] checking trash now [07:35] jtv: do you mind doing the testfix? It looks like it was danilos r8643... we need to get it fixed asap in case the other later changes have issues too :/ [07:36] noodles775: oh, it's a real test breakage? Been a while since we saw one of those... [07:37] yep [07:37] on db-devel? [07:37] https://lpbuildbot.canonical.com/waterfall [07:41] noodles775: a sucky one, but definitely one in my neck of the woods [07:41] Thanks jtv! [07:49] noodles775: bug 476218 [07:49] Bug #476218: Buildbot failure on db-devel: unique violation on POTranslation [07:49] Thanks jtatum [07:49] jtv (sorry) [07:50] jtatum: if I were you I'd figure out what IRC client(s) these folks are using, and file a bug for over-aggressive nick completion! [07:53] noodles775: bad news. The unique constraint that's being violated is on the primary key. [07:53] :/ [07:53] In other words, somehow it became possible to create objects with ids that were already taken. [07:54] Where's stub when you need him ;) [07:55] * noodles775 pulls db-devel [08:02] jtv: running `bin/test -vv -t fix_translation_credits.txt -t gettext-check-messages.txt` gives a different error again. [08:03] as if these tests are dependent on being run in some order. [08:03] as if they forget to mark the database as dirty maybe? [08:04] al-maisan: I just noticed this is a branch that you reviewed (not that that implies anything other than you might be familiar with it... do you mind taking a look too? It's very urgent that we get this fixed). [08:05] noodles775: where's the branch you are alluding to? [08:06] al-maisan: r8643 db-devel [08:06] for bug 474874 [08:06] Bug #474874: Missing DB permissions in fix_translations_credits script

[08:06] ah, OK. [08:08] noodles775: I get exactly the same error when I run that combination [08:08] jtv: I just realised why though... I assume we should be running make schema (as there is a db schema security change). [08:09] ah, that's what I did, yes [08:10] yep, my bad. [08:10] I texted stub [08:10] Thanks. [08:11] jtv: this is very funny .. danilo's branch merely changes schema security settings.. [08:11] I haven't spotted any recent changes so far that might've caused this [08:11] al-maisan: I don't think there's any particular reason to suspect that branch in particular, apart from the fact that this happens to be rosetta [08:12] hmm.. [08:12] jtv: but it's 100% reproducible isn't it? which means it *has* to be that branch... or did I miss something? [08:12] stub has an idea [08:13] yay [08:13] there was a zcml branch from a community member that he saw a review pass by for few days ago [08:13] sorry, a branch that affects zcml [08:14] ah, he saw a patch fixing a problem like this (which may still not have gone in!) [08:14] the problem seemed to be something along the lines of objects being created twice [08:14] I did a ZCML-related duplicate primary key thingy. [08:14] But it only affects "make harness". [08:14] wgrant: that sounds like the one! [08:14] wgrant! tell us more... [08:15] wgrant: now it's affecting our test-run... [08:15] https://code.edge.launchpad.net/~wgrant/launchpad/fix-duplicate-harness-zcml/+merge/14464 is the one. [08:15] * noodles775 gets a patch [08:15] paste [08:15] may be same cause [08:15] It's unlikely to be the same. [08:15] Hmm. [08:15] wgrant: http://pastebin.ubuntu.com/311315/ [08:16] You can see it with db-devel r8643 by doing: [08:16] Ew. [08:16] bin/test -vv -t fix_translation_credits.txt -t gettext-check-messages.txt [08:16] Hm, it's not in a script, though. [08:16] What was the diff that killed it? [08:18] * wgrant narrowly avoids swap-death. [08:18] wgrant: we don't know... looks like the last build that _succeeded_ was triggered by a branch of yours, on Wednesday around 18:xx UTC [08:18] OK, pull nearly done. [08:18] wgrant: it's turned up on a buildbot run of r8643 - diff http://pastebin.ubuntu.com/311319/ [08:19] so looks like everything >8640 is suspect [08:19] hmm .. running gettext-check-messages.txt alone passes [08:19] I went through all of those though and saw nothing that might cause this. :-( [08:19] al-maisan: yep. [08:19] wgrant: notice that the diff has a new doctest which *does* run a script. [08:20] Ah. [08:20] I wondered what gettext-check-messages.txt had to do with anything. [08:20] so, something in fix_translation_credits.txt is breaking things, no? [08:20] Does fix_translation_credits.txt itself fail? [08:20] Or does it just break tests run after it? [08:20] * al-maisan checks [08:20] wgrant: no, it does not fail. [08:21] * noodles775 applies wgrant's diff. [08:21] Probably not related, then. [08:21] oh, that's for harness. [08:23] * jtv tries forcing a dirty database [08:24] yeah .. it would seem fix_translation_credits.txt only updates existing records [08:24] but why does running it together with gettext-check-messages.txt cause the exact failure? [08:25] noodles775: probably because something updates the database in a way that our test framework doesn't notice, but without calling force_dirty_database either [08:25] Known, nasty problem. Thing is, that shouldn't get us into this particular situation. [08:25] As long as the test runs in DatabaseLayer, it should be reset every time, right? [08:25] Because this particular situation smacks of an id sequence being reset, [08:26] but the objects that have been created not being cleaned up [08:26] I don't see code in DatabaseLayer that checks if the DB is dirty. [08:26] Hm, but it does run in DatabaseLayer. [08:26] Damn. [08:26] good morning [08:26] hi adeuring [08:26] oh, wait, I see `TranslationMessage` instances being created in potmsgset.updateTranslation() [08:26] hi adeuring [08:26] hi jtv! [08:27] hi noodles775! [08:27] al-maisan: yes, that's what it does [08:27] and that in turn will create a POTranslation (unless one for the exact same string already exists) [08:27] i.e. this is not an update-only action [08:28] but we should be looking at fix_translation_credits.txt [08:28] jtv: does run_script() flush the db after completion? [08:28] It runs a script. [08:28] yes [08:28] right. [08:28] It flushes the db, but that's not the issue. The issue is, there's no way for anyone to know that that script just modified the database. [08:29] So what I'm trying right now is to mark that test as dirtying the db [08:29] (I was a bit slow about this earlier because I was also still on the phone with stub :-) [08:30] yup, that fixes it [08:30] jtv: hmm .. sounds promising. [08:30] I don't see other tests that *just* run_script. [08:30] Ah. [08:30] wgrant: we may have them here or there... but they should tell the test setup that the db is dirty [08:30] al-maisan: yes, more than just "promising." :) [08:30] jtv: :) [08:31] Thankyou jtv, wgrant and al-maisan! [08:31] and stub [08:31] de nada [08:31] * noodles775 's blood pressure drops drastically [08:31] (yes, and stub of course!) [08:34] When's the reroll? [08:34] wgrant: tentatively scheduled for 16:00UTC today [08:36] Ah. [08:39] who wants to review the fix? [08:39] (still working on the mp) [08:40] I'll r-c it :) [08:40] noodles775: that was my next question—"are you authorised to RC?" [08:40] adeuring is todays OCR, but it might be quicker for al-maisan given that he's already familiar with the situation? [08:40] jtv: yep. [08:41] Do I request a "release-critical" review for that? [08:41] yes please. [08:42] noodles775: requested [08:42] turns out searching for "noodles" gives 3 little pages of results [08:42] none of which seems to be you [08:43] still, if I did that on google it'd probably be a whole lot worse :) [08:43] yep, I've only started using that as a nick since starting here (although it was a nickname in high school). [08:44] Why the 775? [08:45] noodles is already in use, needed a number [08:45] Ah. [08:45] (and i was born in 75 which is probably why a 'random' number ended up being 775 ;) ). [08:48] noodles775: maybe run_script (and equivalents, I think we have 2 of that name already) should just force a dirty db? [08:49] jtv: yeah, might be worth chatting with BjornT about a longer term solution. [08:49] noodles775: I can think of some... I suspect it's mainly a matter of minimizing performance cost [08:50] also, nowadays, postgres doesn't allocate an xid until you've made changes. So that could be a way to keep track as well. [08:50] Or maybe the existing mechanisms should have been enough to prevent this particular failure, and there's a bug in that somewhere. [08:51] AIUI the id sequences should be reset to whatever is the highest that's still seen in the table, + 1 [08:51] in which case this problem _should_ never have happened. [08:52] yep. [09:02] noodles775: ec2 instance is setting up [09:04] jtv: um, for a test-fix, I assumed you'd land it directly on db-devel? [09:04] jtv: in fact, I think time-wise, we have no option but to do so? [09:04] noodles775: ah, of course. [09:05] noodles775: blame the 'flu or my total inexperience with [testfix], your choice [09:05] yes, please submit to pqm to land on db-devel. [09:07] the [testfix] comes before the [release-critical=], right? [09:09] Looking at the log, it seems so: [testfix][release-critical][r=kiko][ui=rs] Reverting the exposure of [09:09] Request and rc from kiko. Using testfix as there was a db init [09:09] [testfix][release-critical][r=edwin][ui=none] [bug 320889] Fixed a [09:09] Bug #320889: Can't retarget a blueprint [09:09] woops, sorry for pasting there. [09:09] hmm... got a regex error nonetheless. :( [09:09] Now to find why "[testfix][release-critical=noodles][r=adeuring][bug=476218] Fix\n\tbuildbot failure: test failed to force dirty db." [09:09] does not match "(?is)^\\s*\\[testfix\\]\\s*\\[(?:release-critical|rs?=[^\\]]+)\\]" [09:10] Morning [09:10] hi mrevell, had a chance to look at my scribblings? [09:10] (that's his morning ruined in a hurry :-) [09:11] will do this morning jtv :) [09:11] :) [09:11] jtv: it looks like that regex (ie. for testfix when in rc) doesn't want the =noodles bit. [09:11] that would also fit with the examples I pasted above too. [09:11] noodles775: it says =[^\\]+ before the closing brackets though... [09:11] oh, that's for rs only [09:11] I have to reboot, back in a sec [09:11] yep. [09:12] * jtv resubmits [09:12] ...and PQM accepts! [09:16] 3-points - thanks jtv! [10:43] bigjools: Is your bug #476316 not a dupe of bug #475172 which is also probably a dupe of bug #464161 which already has bug #475566 as a dupe? [10:43] Bug #476316: Slony-I: Table emailaddress is replicated and cannot be modified on a subscriber node [10:43] Bug #475172: EnsurePerson fails for new maintainers during upload processing [10:43] Bug #464161: authdb Store for scripts connecting to incorrect database [10:43] Bug #475566: Slony-I: Table account is replicated and cannot be modified on a subscriber node [10:45] wgrant: sort of, but I still need a local fix [10:45] Ah. [10:46] file first, dupe later I say [10:59] Morning, all. === matsubara-afk is now known as matsubara === ursula_ is now known as Ursinha [12:37] salgado: ping [12:37] hi sidnei [12:37] salgado: found the problem. thankfully i've been bitten by this before. [12:38] salgado: take a look at _bindUIPicker in picker.js [12:38] sidnei, yay! will do in a minute [12:38] salgado: there's a handler for the focus event that's doing e.halt() if e.target !== this._search_input [12:39] salgado: i think the order that the handlers get dispatched changed, so that handler, which is for the bounding box gets triggered before the more specific handler for the extra-input [12:40] and that handler is to make the focus go to the search input whenever I click on the bounding box? [12:41] salgado: apparently yeah. i think it should also check if the target isn't an input or some other control. that should fix your issue. [12:42] sidnei, cool, I'll give that a try. at least now I know where the problem is. :) [12:42] sidnei, thanks a lot! [12:42] iirc, all this mess is because firefox gets focus events wrong. there are lots of threads out there about it. === mrevell is now known as mrevell-lunch [12:53] bigjools, oi :) [12:54] Ursinha: oi [12:54] bigjools, there's one oops that only happens to one team, easily reproducible: https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1405K1314 [12:55] bigjools, but only you may know the cause :) [12:56] Ursinha: I am looking [12:57] Ursinha: it's a data inconsistency issue, I've no idea why it would get like that [12:58] basically the build state says it's building but there's no buildqueue record [12:58] we need to reset the build state [12:59] bigjools, yes, then it shows in the list and when you click, boom [12:59] bigjools, resetting it is easy to do? [12:59] needs LOSA + SQL [13:00] right [13:00] bigjools, I'll file a bug just to keep track [13:03] BjornT: can you approve the DB query I just added to LPS please === mrevell-lunch is now known as mrevell [13:40] lol jtv :) [13:44] oh, I lost a joke from jtv :/ === ursula is now known as Ursinha [13:47] for Ursinha : jtatum: if I were you I'd figure out what IRC client(s) these folks are using, and file a bug for over-aggressive nick completion! [13:47] lol [13:47] thanks noodles775 :) [13:48] :) [13:48] don't feel bad, Ursinha, it was ~6 hours ago ;) [13:48] jtatum, oh! I thought that it happened in the moment I lost connection here :) [13:50] Ursinha: we didn't say anything bad about you, honest [13:50] bigjools, oh, that I don't mind, kinda used to it [13:50] :P [13:51] surely not :) [13:53] jtatum: so... you still think I'm joking? Make a note to read it again a few months later. :-) [13:57] lol [13:57] jtv, noted ;) [14:23] barry, when you encountered that different rich-root problem, I assume you fixed it by creating a new repo? if so, what format did you use for it? I used pack-0.92 but that doesn't support stacking AFAICT [14:24] salgado: i used pack-0.92. i didn't notice if it stacked or not (i /think/ it did but i'm not positive) [14:24] salgado: yes, i created a new repo, and used diff/patch to get the change across. kinda sucked ;) [14:25] barry, indeed, I've just did that. I think the first format that supports stacking is 1.6, though [14:29] salgado: yeah. stacking's a lp optimization though. when the format bug is fixed i'll happily let lp consume less disk space ;) [14:30] barry, well, not just that, if you don't use stacking you'll have to push the whole history every time you push a new branch [14:30] salgado: true. lp-dev-utils is small though and i have a fat pipe === salgado is now known as salgado-lunch [15:15] danilos, what's the difference between product and project? [15:17] barry: do you know the procedure for adding a dependency to launchpad-dependencies (as in, python2.5-devel :-) ) [15:18] gary_poster: i don't, but salgado-lunch does [15:18] barry: I thought he did, and pinged him elsewhere, but was excited enough to ask you just in case while he is at lunch :-) [15:19] thx [15:19] gary_poster: ;) [15:24] I wish I could ask wgrant, he's a walking grep :) [15:38] Ursinha, product is a project and project is a project group. those are the old names [15:40] danilos, sure :) [15:40] matsubara, thanks! === danilo_ is now known as danilos [15:41] danilos, so, what's the difference between primary_translatable and translation_focus? [15:42] Ursinha, translation_focus is explicitely set (or not) and primary_translatable can be defined even if translation_focus is not set (it's calculated otherwise) [15:42] Ursinha, basically, translation_focus mirrors the DB value, and primary_translatable provides with a good default if translation_focus is not defined [15:42] danilos, right, that leads to the second question [15:42] danilos, if translation_focus is none, what should it be, development_focus? [15:43] Ursinha, well, what do you think it should be? (a serious question) [15:43] errr sorry [15:43] hm, no, that's right [15:43] danilos, I think it should be the dev focus [15:43] Ursinha, the situation is this: project maintainer has not defined any release as the focus for translations [15:44] Ursinha, and I agree, it makes most sense :) [15:44] danilos, and after that, keep the same as it was [15:44] Ursinha, and if that's not defined either, I'd pick the latest series [15:44] Ursinha, well, it should default to development_focus already [15:45] Ursinha, so, the change you are making should be "if there's translation_focus, then that, otherwise keep the same as it was" :) [15:46] danilos, that's what I thought was the right thing to do :) [15:46] danilos, other question [15:47] danilos, what's the secondary_translatable_serieses? [15:47] Ursinha, I don't know [15:47] danilos, :) [15:48] Ursinha, """Return a list of IDistroSeries that aren't the translation_focus. [15:48] It only includes the ones that are still supported. [15:48] danilos, just want to be sure that's what I imagine it to be, a list of translatable series that not the translation focus [15:49] danilos, I see the comment, boss :P [15:49] Ursinha, well, it seems pretty clear to me :) [15:54] deryck: ping === EdwinGrubbs2 is now known as EdwinGrubbs [15:55] EdwinGrubbs, pong === matsubara is now known as matsubara-lunch [15:56] deryck: I just sent you an email that I hope you can respond to now, so that I can land the branch today before I go to the lazr-js sprint. [15:57] EdwinGrubbs, will look now. [15:58] EdwinGrubbs, the design is fine that you sent. I have no qualms at all. :) I actually like that better than what I proposed. [15:59] thanks === salgado-lunch is now known as salgado [16:24] EdwinGrubbs: I sent you a reply to the design [16:25] EdwinGrubbs: If you don't was to fix the CSS, I'll do it right now to fix bug 465524. I know the problem affects more than the two portlets we are discussin [16:25] Bug #465524: ~name Latest Memberships portlet joined date is rendered in the incorrect style === deryck is now known as deryck[lunch] === deryck[lunch] is now known as deryck [16:50] Ursinha: attagirl for your reply to danilos :) [16:51] bigjools, is that good? :) [16:51] yes very :) [17:03] sinzui: I'm fixing the registered CSS class not to use italics. Can you look at https://edge.launchpad.net/ubuntu/+mirror/ftp.iinet.net.au-archive ? It seems like the "registered by" info should go in the top right corner even though it is really long on this page. [17:04] EdwinGrubbs: does that registration info *look* like it is correct? Where should that information be on the page? [17:05] sinzui: I have no idea if it is correct. Normally, we put "registered by" just above the sidebar. [17:05] EdwinGrubbs: Yes it goes in the registration slot. Official belongs in the Archive information portlet [17:05] EdwinGrubbs: you should have an idea. Any time you want to know when an object was created and by whom, look to the right under the logout button [17:06] EdwinGrubbs: You do not need to fix this issue. I'll report and bug and get it fixed. There are several cases of these. [17:10] Ursinha, you'll know who to ask any questions from now on, now that you hurt my feelings ;) [17:10] danilos, lol, sorry boss [17:10] danilos, you're emacs, gnome guy, but I love you anyway [17:10] * Ursinha pets danilos [17:10] Ursinha, :) [17:10] :P [17:19] danilos, how do I find out the currentseries of a project? [17:19] err product [17:24] Ursinha, 'currentseries'? [17:24] Ursinha, do you mean development_focus series? :) [17:24] oh, wait [17:25] danilos, apparently product.currentseries is related to distros [17:27] Ursinha, I don't see product.currentseries [17:27] Ursinha, I only see distribution.currentseries [17:28] danilos, sorry, I see targetseries = ubuntu.currentseries [17:29] Ursinha, right... is there any reason why do you worry about the rest of that method? :) === EdwinGrubbs is now known as Edwin-lunch [17:30] danilos, no, I'm just trying to understand how it works :) [17:32] Ursinha, well... it's quite simple :) it returns either the development_focus productseries or the latest productseries which has translatable templates; if none of those exist, it returns any linked ubuntu sourcepackage instead, or if there's no ubuntu sourcepackage, any other sourcepackage [17:33] Ursinha, btw, did you get the DB patch reviewed? [17:34] danilos, nope :( [17:34] guess people are busy this weel [17:34] week [18:00] good weekend all === matsubara-lunch is now known as matsubara [18:47] abentley: around? I don't understand your assertion that "undecided" repositories would encourage pack-0.92? === Ursinha is now known as Ursinha-afk [19:11] flacoste, do you have a forecast as to when we will be dropping r-c? [19:15] mars, how's the yui3ification? [19:21] beuno, moving along. Everything works well, we've only had a few small changes to deal with. [19:21] beuno, I just fixed a bug that was preventing the production config from loading. With that fix we are one big step closer to being able to land the branch. [19:22] mars, that's great news [19:22] and most widgets work with 3.0.0 and the updated lazr-js [19:22] beuno, we will at minimum have an integration branch to build off of for the sprint [19:22] beuno, most do [19:23] mars, that's awesome, thank you [19:36] mars: it's already dropped, since ~4h ago [19:37] flacoste, ok, thanks === mars changed the topic of #launchpad-dev to: This is Launchpad Development Channel | Week 4 of 3.1.10 | PQM is open - release manager is noodles775 (flacoste, BjornT) | I am Zero OOPS and So Can You! http://is.gd/4fkLl | https://dev.launchpad.net/ | Get the code: https://dev.launchpad.net/Getting | On-call review in #launchpad-reviews | Use http://paste.ubuntu.com/ for pastes | This channel is logged: http://irclogs.ubuntu.com/ [19:38] sidnei, ^ it should be ok to start landing lazr-js branches, if you dare tempt the wrath of Friday :) [19:40] mars: i would prefer waiting until monday :) [19:40] hehe [19:40] sidnei, ok, we will get everything straightened out monday morning during the sprint opener === Edwin-lunch is now known as EdwinGrubbs [19:41] sidnei, could you add the branch URL to the sprint page? [19:42] beuno, we'll have to land on trunk. salgado and I have made a number of fixes on top of sidnei's and Bjorn's work [19:43] mars, can we have the latest there? it will take a while to get it into trunk, and it would be ideal not to stall on it [19:43] beuno, we can. I can push my integration branch with everyone's changes rolled into it. [19:44] mars, that would rock. Add it to the agenda wiki and we're set! === matsubara is now known as matsubara-afk === salgado is now known as salgado-afk [21:29] * wgrant dislikes the new c-i-p UI. [21:40] wgrant, c-i-p? [21:42] beuno: canonical-identity-provider [21:42] wgrant, the one with the launchpad UI, or the ubuntu UI? [21:43] beuno: The Ubuntu one doesn't seem to be serving anything yet. Does that UI exist? [21:43] Maybe that one is less utterly sucky... [21:43] But I presume not, since 3.1.10 brought some changes to the Launchpad UI. [21:43] wgrant, it's completely different [21:43] and should stop sucking [21:44] beuno: Will it supplant the Launchpad UI entirely? [21:45] wgrant, it will be a separate instance with the same db [21:45] so it will be used for all ubuntu-related stuff [21:45] at some point, we will either re-work the launchpad one, or replace it with the ubuntu one [21:46] Ahh. [21:50] Bug #476963, anyway. [21:50] Bug #476963: UI is pretty awful [21:53] :) === Ursinha-afk is now known as Ursinha [22:26] * thumper is at kiwipycon [22:26] even has web access :) [22:28] Not once everyone else gets there. [22:28] * wgrant has bad memories of UDS Jaunty. [22:29] wgrant: it's not that bad, honest [22:30] (yet) [22:30] Until 128 people get into the hotel, and the DHCP range fills up. [22:30] we're not all at a hotel, thankfully [22:31] yes, hotels insisting on doing the networking are the bane of conferences [22:41] sinzui: so, do you just want me to move the milestone badge to be first, since it is the most often present? [22:42] EdwinGrubbs: I expect it to have the same order as this: https://edge.launchpad.net/launchpad-registry/+milestone/3.1.11 [22:42] sinzui: I'm also wondering if I should implement beuno's 0.5em padding on the or on the .registered stylesheet. [22:43] EdwinGrubbs: I think right-aligned is all that is needed. [22:44] EdwinGrubbs: I need to find my children. I'll be back in a bit. [22:44] sinzui: huh, beuno asked for 0.5em padding below each item. [23:44] I want to use bazaar to track source packages with modifications in the launchpad ppa. Would it make sense to register a launchpad-ppa project? [23:46] maxb: People use distro branches for that. [23:46] for packages in a ppa? [23:46] Yes. [23:46] Not exactly semantically correct. [23:46] hmm [23:46] But probably better than a project. [23:47] After all, they are (in most cases) branches of an Ubuntu package. [23:47] lp:~maxb/ubuntu/karmic/python-defaults/launchpad-ppa ? [23:47] Right. [23:55] mm, neat [23:56] Hm? [23:57] oh, I'm just sitting here grinning that having twigged that package branches are the right way to do it, I can just push and all is happy :-) [23:57] Yup. [23:58] I'm not sure if you can manually stack on LP without playing with bzrlib. [23:58] * wgrant tries.