[00:00] does anyone know the rdf ontology used for our project metadata? [00:03] found it, doap. [00:33] mwhudson, fwiw, I have an ec2test run going that runs the code doctests in the DBFunctionalLayer [00:33] jml: which layer is that? [00:34] mwhudson, the DatabaseFunctionalLayer [00:34] jml: is that the one that doesn't start the librarian? [00:34] mwhudson, the one with zcml & the db and not much else [00:34] mwhudson, yeah [00:34] jml: some code import tests may well fail [00:34] i guess you'll find out :) [00:34] mwhudson, I expect so. :) [00:34] some diff related ones too, perhaps [00:39] mwhudson, I'm also running tests that delete a lot of columns from ProductSeries [00:39] jml: \o/ for that [00:39] mwhudson, all of which are related to vcs-imports, afaict. [00:39] yeah [00:39] they're long dead [00:40] mwhudson, if the tests pass, can you please review the patch? [00:40] jml: sure [01:15] thumper: i just wrote https://bugs.edge.launchpad.net/launchpad-code/+bug/118625/comments/12 btw [01:15] Bug #118625: codebrowse sometimes hangs [01:19] mwhudson: fwiw, sounds sane to me; and would seem match observation. logs tho... hrm. [01:21] wget -O- https://edge.launchpad.net/launchpad/+rdf | wc -l [01:21] 1634 [01:37] thumper, btw, I have your 'active-branch-counts' still in my active reviews page [01:47] jml: yes, I've not done anything with it recently [01:56] * mwhudson lunches [02:05] thumper, write better cover letters. [02:44] jml: yeah, I suppose [02:44] normally do [02:44] just frustrated [02:45] thumper, page migration getting you down? [02:46] just a bit... [02:46] given that the breadcrumbs were broken for branches [02:46] again [02:47] s/were/are [02:59] mwhudson, ping [03:20] jml: hi [03:21] mwhudson, I'd like a second opinion on something. Got a moment for a short skype call? [03:21] jml: sure thing [03:49] mwhudson, lib/lp/code/tests/../doc/branch-merge-proposal-notifications.txt [03:49] lib/lp/code/tests/../doc/codeimport-result.txt [03:49] mwhudson, those are the only ones that failed. [03:49] jml: my predictive powers are clearly working reasonably well today [03:50] mwhudson, don't jinx it :) [03:50] (i just wish they were predicting more success with buildbot :/) [03:52] haha [03:53] mwhudson: you're having op trouble? [03:53] lifeless: op ? [03:53] operational [03:53] ah [03:53] no, not really [03:55] lifeless: mostly trying to bash the model buildbot uses into my brain [03:55] mwhudson: pb jelly + bidirectional callbacks around steps of activity. [03:55] mwhudson: If you need a clue at any point gimme a shout [03:56] mwhudson: I doubt its changed radically since I climbed all through it years ago [03:56] lifeless: well, it has changed somewhat because gary added 'latent builders' to it [03:56] they being ec2 slaves [03:56] which it can activate [03:57] ? [03:57] yes [03:57] Arbitrary number, or a slave-factory? [03:57] [does this make buildbot a slave owner?] [03:57] the 'force build' button doesn't appear for them unless a build is running [03:58] and thats the bug you're working on? [03:58] because the slave isn't 'connected' according to buildbot in this state [03:58] well [03:58] yes [03:58] i have this niggling suspicion that if i can make it appear it then won't work [03:58] thats to be expected. Uhm, I'd decouple that assumption [yes I realise thats what you're looking for the knobs of] [03:58] yes [03:59] are the latent slaves all manuall configured? [03:59] or does it act as a factory for them? [03:59] lifeless: i don't understand the question [03:59] lifeless: it starts an ec2 instance for each build [04:00] so does the config say 'slave FOO; is latent; ', 'slave BAR; is latent...', and then you can have 1 and only 1 build on FOO at a time, ditto BAR [04:00] -or- [04:00] slaves are ~always configured manually in buildbot aren't they? [04:00] lifeless: yes [04:01] this is part of the terminology thing [04:01] does the config say 'slave FOO is latent ', and then new slaves appear each time ec2 is prodded to do stuff, called FOO- [04:01] i have a hard time remembering the difference between builders and slaves [04:01] and build factories [04:01] mwhudson: IIRC builders are 'code that knows how to perform an operation' [04:01] build factories are 'code that parameterises a builder for a job' [04:02] wuu, -1 actually seems to work these days. [04:02] (it shows only the first failure in a doctest) [04:02] mwhudson: so if its the first case, then all the slaves are static, the only interesting thing is that they connect on-demand. [04:02] mwhudson: I'd consider two things; [04:02] - a 'connect this slave' button. [04:03] this would work, unless the slave auto-disconnects instantly [04:03] (and a connecting slave would want 'force build' to be available as a callback on connect completing) [04:03] - making force build trigger a connect implicitly [04:04] hmm. the lp-dev-utils kerfuffle is stopping me from opportunistically coding. [04:04] something to deal with when I return. [04:04] jml: comefrom? [04:04] lifeless, return from running a bunch of errands, which is what I'm about to do. [04:42] mwhudson, thumper: Is lp:~wgrant/launchpad/structural-subscription-security cursed, or just unlucky? [04:42] Three have tried to land it so far. [04:42] I never heard back from the third attempt. [04:54] wgrant, hmm. [04:55] wow, the "Request another review" widget is broken as. [04:55] wgrant, commit message? [04:55] (I'll update the field on the MP page to have the commit message so you don't get asked again) [04:56] * jml writes out, "[r=gary][ui=none] (wgrant) Move structural subscription security to the model, rather than the view." [04:57] jml: That sounds about right. [04:57] Thanks. [04:57] Hopefully this time it will work. [05:01] You know, I really hate tests that fail due to obscure errors in a three-level-deep exception handler. [05:01] I also really hate three-level-deep exception handlers, but particularly when they crash. [05:02] * jml sympathises [05:02] sometimes it's a sign that there need to be unit tests for that third level layer. [05:03] particularly in the Launchpad code base. [05:14] mwhudson, thanks for replying to the RFC about package branchse. [05:14] * jml needs to reply to slangasek's email. [05:14] jml: np [05:14] * mwhudson doesn't think he's getting all the mails on this topic [05:15] mwhudson, slangasek's mail was sent to launchpad-dev on the 29th. [05:16] jml: subject? [05:16] i don't seem to have it [05:17] mwhudson, Re: [RFC] Taking source package branches to the next release [05:17] jml: nope, nothing from him [05:17] mwhudson, hmm. [05:17] It wasn't on the list. [05:17] Or is stuck in moderation. [05:18] stuck in moderation, I'll bet [05:18] * jml moderates in anger [05:22] * mwhudson receives lots of fairly old mail\ [05:23] Ah. That's filling in the gaps. [05:24] mwhudson: jml: which list? I did the int one this morning, nought but spam. [05:24] spm, the launchpad-dev list [05:25] spm, as found on http://launchpad.net/~launchpad-dev [05:25] hrm. nothing there either. [05:26] spm, that's because I just moderated them all :) [05:26] awesome. my problems just went away. ;-) [05:29] heh. just noticed (and fixed) the lp-eng team membership of same was set to expire later this year. that would have been amusing to watch. [05:34] spm: You seem to want https://launchpad.net/+announcements. [05:35] That page was almost removed a week or two ago. [05:35] wgrant: rofl. which just goes to show how well... advertised(?) the service that provides is :-) [05:35] spm: I've never seen it linked. [05:35] But it has a useful Atom feed too. [05:36] (in fact, I only discovered that page by guessing the URL when I saw people pondering its removal) [05:36] lol [05:36] now why the bleep isn't db_lp building?? [05:37] mwhudson: buildbot hates you. [05:37] wgrant: i hate it more [05:37] alright. I am very confused. [05:37] I have a test here that really should fail in python2.4, but only fails in python2.5. [05:37] AFAICT. [05:37] Basically, it's logging an exception. [05:38] Said exception contains '%s' in the traceback. [05:38] exceptions are new style in 2.5, could it be that? [05:38] So when the logger function attempts to use % on the string, it crashes. [05:38] Possibly. Let's see what 2.4 exceptions look like. [05:38] python2.4 exceptions might look a little bit more like sequences to some parts of the code [05:39] Hmmm. [05:39] or maybe not [05:39] wgrant: can you pastebin come code? [05:41] mwhudson: http://paste.ubuntu.com/262311/ [05:42] The problematic exception is http://paste.ubuntu.com/262312/. [05:42] wgrant: oh gosh, mocklogger is terrible [05:42] mwhudson: Heh, there is a comment in one of the callsites saying that it should be merged with FakeLogger. [05:43] * wgrant looks for any other callsites. [05:43] FakeLogger? is that like an alias to /dev/null? [05:43] Mm, there are a few. [05:43] there are so many dummy loggers lying around [05:43] it's really cruddy [05:43] *why*!?!?!? [05:43] wgrant: i don't see why that would behave differently 2.4/2.5 [05:44] mwhudson: Neither. [05:44] mwhudson: But the same exception is raised in both. [05:44] wgrant: but the mocklogger code is broken, please fix it :) [05:44] And somehow crashes on 2.5... [05:44] wgrant: maybe format_exception does different things? [05:44] seems unlikely though [05:44] mwhudson: Possibly. Let's see.. [05:44] pdb ftw [05:52] Oh. [05:52] No. [05:52] Surely not. [05:53] mwhudson: I now know why. [05:53] mwhudson: 2.4 prints the first line of the statement. [05:53] 2.5 prints the last. [05:53] wgrant: ha ha [05:53] How utterly stupid. [05:54] I almost didn't bother to check that, it was such a silly idea. [05:54] its bytecode data I think [05:56] So, I think that was the only mysterious Python 2.5 failure. === thumper is now known as thumper-afk [06:32] Anybody want to review https://code.edge.launchpad.net/~wgrant/launchpad/fix-mocklogger-format-string-crash/+merge/10902? [06:54] wgrant, has anyone answered yet? [06:56] * mwhudson eods [06:56] jml: No. [06:57] jml: (has my accursed branch killed ec2test yet?) [06:57] wgrant, ec2test went headless, still waiting for a reply [06:57] jml: Great, thanks. [06:58] wgrant, I don't fully grok your patch yet... [06:58] wgrant, [06:58] * jml kicks xchat [06:59] wgrant, specifically, I don't know what "2.4 seems to render the first line of each statement in tracebacks, while 2.5 renders the last" has to do with only doing string interpolation if arguments are provided. [06:59] jml: Well, I half expected the reviewer to be somebody who had read the earlier discussion. [07:00] jml: Basically, in the case of that test the last line of the statement contains '%s'. [07:00] So MockLogger gets a string with a '%s' in it, without any arguments. [07:00] Boom. [07:00] ahh ok. [07:01] (also, I'm extremely hungry) [07:01] It is approximately way past lunchtime. [07:06] jml: How would I best go about testing it? I can't exactly just use a mock logger and check the output for this one... [07:08] wgrant, construct a MockLogger, temporarily replace sys.stdout with a StringIO object, call log() without args, check the contents of the StringIO, replace sys.stdout [07:09] jml: OK. I was hoping there was some less bad way than replacing sys.stdout. [07:09] by 'without args', I really mean "log('%s')" or something similar. [07:09] Yep. [07:09] wgrant, well I was about to get to that :) [07:09] wgrant, the nicer way is to give MockLogger a constructor that takes an optional file-like object [07:09] wgrant, if provided, use that; if not, use sys.stdout [07:10] jml: Ah, of course. [07:10] that would mean changing the print call to self._output.write or whatever, naturally. [07:10] Yep. [07:10] Thanks. [07:10] np [07:15] wgrant: presumably the caller is buggy, and should be doing "log('%s', msg)" instead of "log(msg)" ? [07:16] hmm... [07:16] (Making the logger robust against this misuse is still a good idea, though) [07:16] spiv: stdlib logging says otherwise. [07:16] Huh? Ok. [07:16] spiv: Anyway, the problematic caller is actually the logger itself. [07:17] spiv: The standard library logging module allows that kind of call. [07:23] loggers are meant to lazy evaluate [07:23] out of a sense of cheapness [07:23] this is arguable [07:34] * jml jfdis [07:39] Howdy [07:50] hi noodles775! [07:50] * noodles775 waves at jtv [07:52] noodles775: and since it's Monday morning for you... the "this site is running demo code" bar seems to be horribly broken on both 3.0 and pre-3.0 pages, but differently on dev and edge. Saw something about it on the reviews list; is it being worked on? [07:52] jtv: https://bugs.launchpad.net/bugs/421454 [07:52] Bug #421454: Filebug link on edge/staging header needs structure [07:53] yep, it's landed in devel... should be fixed on edge on the next rollout (hopefully soon). [07:53] noodles775: does that also cover the bar not showing up at all on dev? [07:53] jtv: huh? it's never shown up on dev afaik? [07:54] noodles775: oh, that might well be it... It's just that dev seemed to reserve space for it. [07:54] (ie. you need to add the site_message config option to dev. [07:55] noodles775: sorry for the false alarms, but as I said, this _is_ Monday morning so bad things are supposed to happen. :) [07:55] jtv: np! fwiw, it was a case of a missing test (we'd never had a test for markup in the site_message, although we'd always used it on edge). When the site_message was moved to a template it was auto-escaped. [07:56] noodles775: tricky! [08:02] noodles775, hello [08:02] noodles775, did you manage to get that branch reviewed? [08:02] Hi jml [08:03] Yep, it's been through buildbot and waiting for the edge update. [08:03] cool. [08:03] jml: sorry, I was assuming that you were talking about the branch above - rather than your permissions branch? [08:04] noodles775, the branch above. [08:04] noodles775, bigjools got to my permissions branch. [08:04] great. [08:04] * jml is eagerly awaiting the British morning. [08:04] jml :( [08:04] it's a public holiday there. [08:04] they are always having public holidays [08:05] I shall write a sternly worded letter to the Daily Mail. [08:05] lol [08:26] good morning [08:41] jml: whats the name of the plugin/command you wrote to find merge, no-shelve, no-edit local branch|trees ? [08:41] lifeless, bzr-removable [08:42] is it packaged? [08:42] lifeless, it probably doesn't even work any more [08:42] lifeless, no, it's not. [08:42] lp:bzr-removable? [08:43] lifeless, yeah, I think so. [08:44] thats doing something [08:44] you should upgrade it to 2a [08:45] Does anyone know if there is something similar to https://staging.launchpad.net/successful-updates.txt but for edge? [08:46] spm: I don't suppose you could check why the daily update of edge hasn't happened yet... or am I too early? [08:47] I don't think bzr-removable has been updated for the new shelf in bzr. [08:47] lifeless, I should certainly do _something_ to it. [08:53] * wgrant returns. [08:53] jml: If you haven't EODed yet, can you have a look at my branch with added tests once I push it? [08:53] sure [08:54] spm: Either thanks or don't worry - edge just updated :) [08:55] * wgrant finds the new header a bit boring and colourless. [08:55] I stumbled upon some UI 1.0 screenshots yesterday, and they are much more inviting than I remembered or 2.0/3.0 are. [08:56] wgrant: I *think* the idea is that projects themselves can emphasize their own branding, like https://edge.launchpad.net/firefox [08:57] noodles775: Right, and that's a good thing. But the facet links are all the same. They're boring, and hard to look through. [08:57] Is it the lack of colour on the app-buttons specifically that you think is missing? [08:57] yep. [08:57] Right. [08:58] I thought maybe facet colours were being eliminated, but the involvement portlet and titles suggest otherwise. [08:58] jml: well, something yes. Anyhow I've just sent you a merge request. [08:59] wgrant: yeah, I'm not sure what the reasoning is for removing them in the app-buttons, but guessing it's to avoid clashing with the product branding... [08:59] noodles775: I thought that might be it. [08:59] I'll check with beuno later - (or maybe he's still around)? [08:59] noodles775: But it makes it all really really boring. [09:00] noodles775: The bottom has that nice blue fade. [09:00] s/fade/gradient/ [09:00] ... and a nice colourful lp branding (which isn't linked :/ ). [09:01] My only other complaint about the branding/watermark is the alignment of the heading wrt. a selected Overview tab. [09:01] Heh... so I know about that one - that's a specific request from Martin... that the first-letters line up (rather than the outline). [09:02] noodles775: It looks pretty strange. [09:02] Hm. although if the -0.5em margin is removed, it looks slightly strange when deselected. [09:02] How unfortunate. [09:02] wgrant: I agree - when the overview tab is selected - and for that reason the initial version had the background aligned. [09:03] noodles775: I guess it will become less of a concern when the Overview tab dies. [09:03] wgrant: well, the same issue will then hit the Code tab. [09:03] noodles775: right, but that's not the default view. [09:03] Right. [09:04] And first impressions do count, a lot. [09:07] I actually had the -0.5em margin removed only when the first li was selected - so it's possible to have both, but the decision was for the first letters to always align (and I just do what I'm told there ;) ). [09:10] micro freshmeat [09:10] http://simal.oss-watch.ac.uk/index.html [09:18] lifeless, thanks for the patch. [09:18] lifeless, the shape of it looks about right :) [09:19] * jml needs to figure out what all the words mean though ... [09:28] jml: it works. Use it. Love it. Commit it. [09:28] oh,and I didn't reduce test coverage. [09:30] lifeless, heh heh [09:30] lifeless, in terms of percentage, no. [09:30] \o/ [09:36] noodles775: Any ideas on how PPA download stats could be shown? [09:38] wgrant: sure - let's collect them at https://dev.launchpad.net/SoyuzPPAIndexPage [09:39] wgrant: they'll be useful immediately to both the owner and potential users, so we'll probably want some general stats on the index page, and then developer/detailed stats on the /+packages page? [09:40] noodles775: I suspect so. [09:40] wgrant: what are your thoughts? [09:41] noodles775: Not sure. It gets rather complicated, because there are different semantics associated with different sets of binaries. Some sources will produce binaries that will be installed disjointly, while others will always be installed together. [09:42] So numbers are probably not meaningful except at the binary level. [09:44] wgrant: maybe it would be helpful to identify (on the wiki page - or a separate one even, or even it's own blueprint spec) what we think the target users will want to see, and then adjust that with what's actually possible/meaningful. [09:44] Once we've got that info, a lot more people would be able to comment too. [09:45] wgrant, started the run on the mystery meat branch [09:46] I'm not sure I caught how you'd get around the download count simply reflecting how often people update a package (without storing ip's etc.) [09:46] s/you'd get/you got/ [09:46] jml: Thanks... I can't see anything even mildly suspicious [09:47] wgrant, me neither. [09:48] wgrant, not being able to reliably land code bothers me quite a lot, so I hope we can get to the bottom of this. [09:48] noodles775: Except for daily builds, it should be fairly good, as every user should get every version if it's there for more than a few days. [09:48] noodles775: And for daily builds, it is at least a reasonable way to judge between other PPAs and packages. [09:48] noodles775: The raw numbers are probably not useful on their own. === thumper-afk is now known as thumper [09:50] wgrant: yes, but what I meant is, I'm assuming a user of a PPA is most interested in how many downloads the ff package in PPA X has - irrespective of the version. [09:51] tee hee [09:51] I don't even have bzr-removable on my laptop :) [09:51] oh wait, it's in my dropbox [09:52] noodles775: Right. It should be possible to do that quite reliably for packages updated less frequently than weekly. [09:52] Great. [09:53] I have most of the backend code done, it just need to be split into threeish branches and tested properly. [09:53] whoohoo! great stuff :) [09:53] And then needs to be poked into the UI. [09:54] lifeless, also, https://code.edge.launchpad.net/~lifeless/bzr-removable/merge hasn't got branch data :( [09:54] wgrant: have you tried bzr pipeline? It's great for multiple-branch-spanning-functionality. [09:54] noodles775: I started using it about a week ago for my latest four-branch effort. [09:54] It is excellent. [09:54] :) [09:59] jml: code review merge-by-mail bug I guess. [09:59] jml: I replied with the url i did push to [09:59] which is ...treeless [10:01] gah pycs and directory removal and bzr [10:01] jml: +1 [10:09] maxb: So, the mysterious soyuz-set-of-uploads.txt failure is pretty much resolved. [10:09] maxb: That just leaves the rest of the obvious (largely benign) failures. [10:24] lifeless: ping [10:25] hi [10:25] lifeless: re twitter [10:25] hai yes, [10:25] the thing that does not work was bin/jstest [10:26] as in "the test script freeze to start with, unless I fix an import bug, then it actually runs to completion but reports 3 failed test suites" [10:26] ah, nifty. In a sad sad way. [10:26] I thought you were referring to something in the web app :) [10:26] fix for the import bug: http://python.pastebin.com/m34a9f134 [10:27] what python version? [10:27] does not matter [10:27] indeed, I see [10:27] uhm, care to file a bug, or would you like meto? [10:28] if I were you, I'd just trivial it [10:28] Lp [10:28] LP's landing process is a bit byzantine [10:28] its easier for me to hand it off [10:28] :( [10:28] wow [10:29] sounds to me like the process went a bit wrong if you can no longer trivial one-liners like that [10:29] one can [10:30] anyway, I do not care about the launchpad bug in itself [10:30] I need to do a bunch of setup to be able to [10:30] my laptop overflowed, and i've not cut code for lp in a couple years [10:30] I was only checking this because I wanted to learn from how launchpad integrated with windmill [10:30] ah :) [10:30] and have you learnt? [10:30] "not very well" [10:31] "does not run headless, so is not part of the test suite run by pqm" [10:31] anyhow, bug files. [10:31] "still requires painful hard-coded waits in some places" [10:31] [once cron takes care of mail] [10:32] ddaa: Somebody here had it running headless. [10:32] ddaa: It's close to buildbot integration. [10:32] And hard-coded waits are meant to be removed. [10:32] Lots were two weeks ago. [10:32] wgrant: THAT sounds very interesting [10:33] wgrant: anything else of the same kind that you think I might be interested in knowing about? [10:34] ddaa: Mmm. Not sure. [10:34] I don't know much more about that stuff. [10:34] Fine then. I was just probing. [10:35] wgrant: Yay! And congratulations on your amazing detective work :-) [10:35] maxb: It was more than a bit of a puzzle. [10:35] maxb: Particularly as it was buried inside unlogged exception handlers... [10:36] It does strike me that the LP tests seem to have a tendency to throw away useful logging [10:37] It's normally easy to get it back. [10:37] Just not in this case. [10:45] bug 421840 [10:45] Bug #421840: trivial bug in test support [10:45] jtv: ^ if you're around, care to peek? [10:45] lifeless: peeking... [10:46] lifeless: greek to me [10:47] * thumper taps fingers while 'make build' runs [10:47] well, not greek, but... [10:47] oh look thumper [10:47] that bug is in your tests ;0 [10:47] where? [10:47] lifeless: afaik we're still not running these tests as part of the regular test suite; mars was working on that [10:48] https://launchpad.net/bugs/421840 [10:48] Bug #421840: trivial bug in test support [10:48] lifeless: I'll add it to my branch [10:51] ddaa: ^ [10:51] thanks thumper [10:51] lifeless: I even added (ddaa) to the revision message :) [10:52] thumper: thanks pal [10:52] ddaa: np [10:54] hi carlos! [10:54] hi ddaa! [10:54] jtv: morning [10:54] jtv: hey buddy, how's life? [10:54] ddaa: overrated :-P [10:55] jtv: watch you for your attitude pal ;-) [10:55] s/watch you/watch out/ [10:55] oh, did it bite someone while I wasn't looking? [10:55] thumper: https://edge.launchpad.net/pyjunitxml/+download [10:55] search on that page for zz [10:57] lifeless: heh [10:58] ok, now I can't download keys from the GPG server :( [11:00] thumper: I figured you'd like it [11:00] in a head slapping kinda way === henninge_ is now known as henninge [11:05] Morning, all. [11:07] deryck: hi [11:07] deryck: I'd like to organise a short weekly call with you sometime [11:07] deryck: do you ever work *normal* hours for your TZ? [11:07] deryck: 'cause if you did, we could talk during the day time :) [11:08] thumper, a call would be cool. And sorry, but I'm always 10-7 GMT. But I could hang around later one day for a call. [11:09] may be a plan [11:09] deryck: lets organise this later though [11:09] I'm done [11:09] thumper, ok, cool. have a good night. === thumper is now known as thumper-zzz === henninge is now known as henninge-afk [11:43] Ooh. lp-dev-utils stuff. [11:44] * jml looks all innocent [11:45] Thanks jml. [11:46] np. [11:48] jml: Are you running that accursed branch without going headless, or should I try? [11:49] wgrant, I am. === henninge-afk is now known as henninge [11:49] jml: Great, thanks. [11:49] wgrant, it's still going... [11:49] jml: Damn. [11:58] * jml stops [11:58] g'night all [11:58] wgrant, I'll let you know how the run goes [11:58] jml: Thanks. [11:58] jml: I guess ec2test won't really work for non-Canonicalites without some hacking... [11:58] It seems to use devpad a bit. === henninge is now known as henninge-lunch [12:16] RuntimeError: You don't have access to a test-runner image. [12:16] Sad. === henninge-lunch is now known as henninge [12:36] I have branches of meta-lp-deps ready for pushing. Any thoughts on the preferred bzr format for new stuff? [12:36] maxb: 2a [12:37] I guess the chances of anyone who isn't already branching launchpad itself wanting to hack on meta-lp-deps are vanishingly remote, so makes sense [12:37] Right. [12:37] And 2.0 is coming RSN. [12:37] * maxb runs some bzr upgrades === matsubara_ is now known as matsubara [13:04] Who owns mailing lists? Both https://edge.launchpad.net/launchpad-foundations and https://edge.launchpad.net/launchpad-registry claim to. [13:07] wgrant, -registry [13:08] salgado: That's what I suspected. I see that Foundations lays claim to Registry itself, too. [13:23] wgrant: The registry team is like a subteam of the foundations team. [13:24] abentley: I see. [13:25] salgado: You are lord of the Apache log parser, aren't you? [13:25] wgrant, I wrote the log parser that counts downloads of LibraryFileAliases [13:27] salgado: Right. I've refactored and split it so it can be used to count PPA downloads too. Do you want to look at that, or will anybody do? [13:27] wgrant, I'd be happy to have a look [13:29] wgrant, how are you going to count PPA downloads? are PPAs served by the librarian now? [13:30] salgado: No. There will be (well, I sort of have it mostly working already) an alternate script which works out the archive and package. It uses the same infrastructure as the librarian one, but uses different database stuff and path format. [13:32] I see [13:32] salgado: It would actually be much harder if PPAs were served by the librarian. [13:33] salgado: So, https://code.edge.launchpad.net/~wgrant/launchpad/refactor-librarian-log-parser and https://code.edge.launchpad.net/~wgrant/launchpad/refactor-librarian-log-parser-part-2 are what I have now. Can you advise me if I'm on the right track or have done something utterly stupid, as seems likely? === al-maisan_ is now known as al-maisan [13:38] wgrant, can you give me an overview of the changes in each branch? [13:40] salgado: Sure. Give me a sec. [13:42] salgado: So, basically, all the non-LFA-specific code has moved into lp.services.apachelogparser. The first branch does everything except stuff from cronscripts/parse-librarian-apache-logs.py, while the second branch includes that. [13:43] salgado: This leaves a fairly tiny amount of code in canonical.launchpad.scripts.librarian_apache_log_parser and cronscripts/parse-librarian-apache-logs.py, and a similarly small amount for PPAs later on. [13:44] salgado: can you mark your bugs on https://edge.launchpad.net/launchpad-registry/+milestone/2.2.8 as FIX COMMITTED or move them to 3.0? I want to close the milestone [13:46] sinzui, done [13:46] thanks === spiv_ is now known as spiv [14:04] wgrant, the refactorings in the first branch look good. for the second I'll need more time [14:04] wgrant, why do you think you've done something stupid there? are you seeing any test failures or unexpected behaviour? [14:05] salgado: The second one has just one unique revision, with a 200 line diff. [14:05] salgado: Oh, no, just guessing I've done something horribly wrong. [14:06] yeah, I'm looking at the 200-line diff here, but it seems to touch all the code of ParseLibrarianApacheLogs.main(), which I don't remember anything about [14:07] salgado: Ah, right, that one is a bit big. [14:09] wgrant, anyway, I'll have a closer look at that. will get back to you soon [14:09] salgado: Thanks. The first one is sane enough to propose a merge? [14:09] wgrant, I think so [14:10] salgado: Lovely. Thanks [14:55] barry: bac: I'm going to convert the list of simple page conversions to bugs that the we can assign to ourselves at our own pace. [15:02] sinzui: grat [15:02] sinzui: er, great! [15:28] danilos: when you have a chance, please set up the gnu mailman translation group [15:30] barry: done, https://translations.edge.launchpad.net/+groups/mailman-translators [15:31] barry: I've made you the owner (you can reassign to a team or whatever) so you can assign each per-language team [15:31] barry: bac: You can see the UI 3.0 update bugs at https://edge.launchpad.net/launchpad-registry/+milestone/3.0 [15:31] barry: we usually have people register teams like mailman-l10n-XX [15:31] (XX is the language code) [15:31] sinzui: great, thanks! [15:31] danilos: fantastic, thanks [15:32] sinzui: looking [15:32] sinzui: so i will just claim some bugs and start working on them [15:33] barry: yep. I think the forms are just change the layout and add the page_title to the view. [15:35] +1 [15:49] abentley: say i use bzr co -r '2009-07-27' to create a working tree at a particular point in the past, how do I update the working tree to say its state one week later? [15:49] bzr co - r '2009-08-03' fails with .bzr already exists [15:49] I want to give ownership of some branches to ~launchpad. If I post to launchpad-dev@, I guess there are some people on the launchpad dev team with the relevant super admin powers to take the branches? [15:50] maxb: yep, you can't give it away yourself? [15:50] You can't give to a team you're not a member of [15:50] Which is fair enough, if inconvenient in this instance [15:50] maxb: ok, we need to ask a LOSA to do it then, what branches? [15:51] I was thinking that perhaps being a ~bazaar-expert would be enough [15:51] abentley could tell us :-) [15:52] flacoste: there isn't a way to do that. update -r is a missing feature. [15:52] great [15:52] abentley: revert -r? [15:52] not quite the same [15:52] that would change the file text, but not what revision the tree thought it was related to [15:52] bac: I see you changes a bug to Medium. What does that mean? [15:52] True. [15:53] revert -r seems to work [15:53] flacoste: but now try bzr status [15:53] well, for my purpose, i simply care about the content of the tree [15:53] sinzui: that it's higher than a low. sorry, i forgot our ban on mediums [15:54] I recall having a long discussion on #bzr about why update -r was missing, and it being something to do with the confusion about what it should mean for bound branches [15:54] bac: you you are committing us to fixing the bug by a certain time, it is High. otherwise it is low [15:54] sinzui: i can change it to high if you want, but it's in review now and will be fix committed soon [15:54] high it is [15:54] revert will work fine for my purpose here [15:55] abentley: does being a ~bazaar-expert give you the powers to reassign some branches away from me to ~launchpad? [15:55] bac: Bugs uses Medium to mean we want to fix it bug we wont commit to it. I think that stinks because it implies Low is "Someone wants my to fix their bug but I do not wan to." [15:56] cprov: hey. is poppy soyuz-related? If so, am I right in assuming we still use it? [15:56] gary_poster: I thought it was the upload ftp server? [15:56] gary_poster: yes, it is. [15:56] Yes, it's... rather critical. [15:56] :-) ok [15:56] ok thanks all [15:57] maxb: I am not sure. [15:57] gary_poster: Porting it off ThreadedAsync? [15:57] yeah [15:57] in spare cycles ;-) [16:23] What's wrong with buildbot? The page won't load ... [16:43] danilos, ping. [16:43] deryck: hi === matsubara is now known as matsubara-lunch [16:56] sinzui: on the product index page, the in-line programming languages picker is broken. [16:56] bac: I saw [16:56] ok [16:57] danilos: ping [16:58] barry: pong [16:58] bac: there's a bug open on lazr-js about that [16:58] it's a lazr-js bug [16:58] barry: ok. [16:58] danilos: hi. will you be around for a little while to talk about your page titles email? i am tasked with trying to figure it out and would like to pair with you on this [16:59] barry: not for a lot longer, and I believe I should have a call with flacoste in a few minutes [17:00] barry: let me see if flacoste is around for a call (I know he was in the UI call), and if we are not having a call, we can do it instead [17:00] danilos: okay, it can fill you in with the basics. i'll figure out what i can and we can chat tomorrow. [17:00] s/it/he [17:00] barry: ok, cool [17:01] danilos: i'm free! [17:01] flacoste: cool :) [17:01] danilos: i am going to get some lunch, flacoste and you can fill me in on anything you decide and i'll read the scrollback when i return [17:01] barry: ok, though it'll mostly be phone call :) [17:03] cool [17:03] noodles775: ping === salgado is now known as salgado-afk === jtv1 is now known as jtv === beuno is now known as beuno-lunch [17:29] sinzui, jtv, deryck - Could I get your 3.0 conversion progress reports by email? [17:29] deryck, I picked you because intellectronica is on leave (and I'm sending the email out for him) [17:29] rockstar, np. [17:29] rockstar: mwhmyumsoonmhwmsI'mthroughdinnermyom [17:31] jtv, :) [17:42] abentley, rockstar: as members of ~bazaar-experts can you reassign ownership of branches to arbitrary team? [17:42] flacoste: I don't know. [17:43] ok, i'll ask a LOSA anyway [17:45] flacoste: they can, I checked in a dev launchpad instance [17:54] rockstar: yhm === matsubara-lunch is now known as matsubara [18:16] abentley, hey - does hitchhiker have a way of killing and recreating the .bzr directory of a branch? [18:16] abentley, I need to restart an import branch, but the user has made it a development focus, so I can't delete and try again. I was thinking maybe I'd use hitchhiker to be clever. [18:18] rockstar: It can delete the .bzr directory, it can't create a new one. [18:18] rockstar: but bzr can. [18:21] abentley, hm, as a bzr-expert, can I just init the path then? [18:21] rockstar: Yes. [18:22] abentley, hm, when I try to get into the branch, it tells me the dir doesn't exist... wtf? [18:23] Oh wait, I think it's trying to get into the mirrored area. [18:23] Nope that wasn't it. [18:24] rockstar: This is an import branch? I don't think you have access to those. [18:24] abentley, yeah, that might be it. [18:25] abentley, although that seems kinda odd. I can mess with people's owned branches, but I can't mess with the ones that are owned by a team I'm a member of. :) [18:26] Anyway, I'll just have the user remove the series link, and then do the change through the UI. === salgado-afk is now known as salgado [18:40] sinzui, I'm thinking of doing the person-edit* pages now. is that ok or has someone already taken them? [18:40] salgado: take them all [18:41] salgado: I broke it into smaller bugs: bug 421975, bug 421976 [18:41] Bug #421975: Update identity/location pages to UI 3.0 [18:41] Bug #421976: Update keys/wikiname pages to UI 3.0 [18:41] sinzui, cool, I'll assign them to me [18:42] salgado: we can avoid creating links between the edit pages if we add the edit icon to the information on the +index page. [18:43] salgado: beuno really wants us to do that. It might be awkward if we cannot add it to the 2.0 +index page [18:44] sinzui, ok, I'll give that a try, but '2.0'? [18:44] salgado: the user page is still 2.0 [18:45] salgado: I see we have links to the edit pages in now...maybe we should keep them. When we update the +index to 3.0, we can remove them from the edit pages. [18:45] * sinzui thinks that is a fastest way to update the pages [18:47] that sounds like a plan === beuno-lunch is now known as beuno [18:50] * barry is doing the /people/+new{person,team} pages [18:51] sinzui, currently it's only the +edit page that has links to other pages. should I add them to +edit[irc/jabber/etc]? [18:53] salgado: we could create a menu, but I think that is work we should avoid. We use a related pages menu on an edit page if it is not possible to get to the form from the +index page [18:53] salgado: branding is only form that we cannot make a link to from an +index page. [18:54] salgado: that is why is suggested that we may want to update the 2.0 +index page [18:54] salgado: You could pick something else from https://edge.launchpad.net/launchpad-registry/+milestone/3.0 and we take take up the person edit pages /after/ the user page is 3.0 [18:55] that might be a better idea [18:56] barry, I think you'll want to assign bug 421972 for you, then ;) [18:56] Bug #421972: Update merge pages to UI 3.0 [18:57] sinzui, I'll take bug 421966, then [18:57] Bug #421966: Update contact, annoucement, driver pages to UI 3.0 [18:57] fab, you'll be done is 2 hours [18:57] maybe less [18:58] salgado: i grabbed bug 421974 [18:58] Bug #421974: Update new person/team pages to UI 3.0 [18:59] barry, duh. I misunderstood you and were thinking you'd be doing the /people/+*merge pages. nevermind me [19:01] :) [19:08] sinzui, are all form pages supposed to have the

before the page's main heading, like in http://people.canonical.com/~salgado/double-heading.png ? [19:09] salgado: That is a header issue that barry is going to work on, and it related to bug 417089 [19:09] Bug #417089: the base-layout heading-slot should not render any heading for IPrimaryContext [19:10] salgado: ignore it since we have more than 50 pages that do exactly what you are seeing [19:11] ok, cool [19:11] right. i'm going to work on that as soon as i get this other bug into review (should be within the hour) [19:11] danilos: bug 422150 describes the other problem i mentioned in our skype [19:11] Bug #422150: Not all series are displayed when setting up translation branches [19:11] barry: cool, thanks! [19:12] barry: don't forget the "john is the uploader" bug as well :) [19:13] danilos: that one's in the email i just sent, following up to that thread [19:13] hey is anybody having trouble resolving the dns for launchpad.net? [19:14] barry: excellent, always a step ahead, thanks! === abentley is now known as abentley-lunch [19:14] :) [19:14] hi sinzui. can i have a mid-imp chat with you re: bug 422128? [19:14] Bug #422128: Making a private team the owner of a project fails if the project has a ProductRelease [19:14] hmm. it seems happy again. [19:14] sinzui: i guess i should steal the assignment of bug 417089 [19:14] Bug #417089: the base-layout heading-slot should not render any heading for IPrimaryContext [19:15] barry: I'll paste a conversation into the bug [19:17] barry: I updated the bug [19:17] sinzui: thanks [19:34] bac: sorry, I got distracted. Yes talk about the bug now [19:54] hi sinzui [19:55] bac: skype? [19:55] ok [20:13] cprov: ping [20:13] EdwinGrubbs: pong [20:14] cprov: I'm trying to make a team the maintainer of a sourcepackagerelease, so that I can test the showing/hiding of the Maintained Packages link in the portlet. Is there a test helper function for that? [20:16] EdwinGrubbs: not specifically, but you can create new SPRs using SoyuzTestPublisher.getPubSource(maintainer=a_person) [20:17] thanks [20:19] EdwinGrubbs: np === abentley-lunch is now known as abentley [20:48] EdwinGrubbs, has the team page landed yet? [20:49] beuno: no, I'm working on tests for it. [20:49] EdwinGrubbs, super [20:54] flacoste: someone looked at lp this weekend and was a bit horrified to discover that not only did our JS tests not pass, they didn't even run because of a trivial bug (http://python.pastebin.com/m34a9f134). This highlighted to me that mars' absence is kind of a big deal, especially with 3.0's focus. [20:54] AFAIU, we don't have automated tests, and we don't have any chance of getting a non-red buildbot for this. I don't think I have anybody on foundations that can work on this, but it feels a bit urgent. Do you have any thoughts? [20:55] (we don't have automated tests that are run, I mean, apparently) [20:55] gary_poster: i don't unfortunately [20:57] flacoste: ok. are we getting further and further in debt because of that right now, or is custom JS halted ATM because of the template changes? [20:57] gary_poster: kind of the latter [20:57] everyone is working on template conversion [20:58] so we shouldn't be adding JS much at this point [20:59] flacoste: ok. I guess I'll bring up with team leads that problem, and ask that if any JS-aware dev starts to move from templates to new JS things, they move instead to fixing the tests. I think that should be the next JS thing to happen, with or wothout mars [20:59] i agree [20:59] ok cool [21:06] flacoste, what are you using to create the burndown chart for the UI conversion? I was thinking about doing a per-week checkout and creating a report per week [21:06] thought about all kinds of fancy things, but considering we're 3 weeks to go... [21:06] beuno: it's all taken care of [21:06] beuno: i have the data [21:06] flacoste, super [21:07] and updating the script and HTML template to show it [21:07] beuno: i'll send you an email with update instructions [21:07] flacoste, perfect, looking forward to seeing what it looks like :) [21:07] beuno: actually, if you are not going away soon, i'll ask you for feedback on the actual charet [21:07] flacoste, I am not [21:07] have TONS of things to do [21:07] *TONS* [21:07] caps and bold [21:08] which i think means you are locked to your computer for a while... [21:27] So, now we have branches for launchpad-dependencies, I want to submit some changes - what's the right thing to do with debian/changelog here - completely update it ready to build the new version? Leave it saying UNRELEASED? [21:44] maxb: are you asking for policy or advice? [21:44] i'm pretty sure there's no policy yet... [21:44] either? both? === abentley1 is now known as abentley === salgado is now known as salgado-afk === thumper-zzz is now known as thumper [22:01] beuno: I really need to talk to you about breadcrumbs soon [22:01] beuno: I have a call with abentley now [22:01] beuno: perhaps in an hour and a half maybe? [22:01] thumper, sure. One hour sounds better, but we can see what happens :) [22:03] beuno: http://people.canonical.com/~flacoste/conversion.html [22:04] flacoste: nice! [22:04] thumper: well, arguably... [22:04] kind of show in our face that we aren't going to make it [22:05] flacoste: I've submitted a CP request for the branch listing timeouts in production [22:05] ok, i'll look at it in an instant [22:05] flacoste: ta [22:06] * thumper waits for abentley to call [22:09] maxb: i'd vote for "ready to build" [22:14] beuno: i sent you an email with the update instructions [22:14] flacoste: ok, and "bzr tag" the new revision too? [22:16] I guess I could even put the package to my PPA, but then if the changes are not accepted as-is, I have a version in my PPA which I can't delete and revise [22:16] flacoste, your page is great. I fear we have too many people setting up this graph. It might be nice to have one canonical page (no pun intended) [22:18] rockstar: beuno is the canonical source now [22:19] flacoste, he doesn't have your super sexy graph. [22:19] rockstar: he will, my copy is just a demo, it's not updated [22:20] maxb: about bzr tag, i have no idea how this works, so whatever you feel makes sense [22:20] maxb: regarding you publishing some version, i don't think it's a big deal. if changes are not accepted, we'll see them in the changelog [22:20] flacoste, okay. We should maybe think about putting it somewhere official looking. [22:20] maxb: or we can use UNRELEASED until it's merged approved [22:23] jml: please make --headless go faster :( [22:23] That might be best ... I put the branch up as UNRELEASED, file a merge proposal, wait for approval, then do one more commit changing UNRELEASED->karmic, do the bzr tag, upload the source package to my PPA, and then a ~launchpad member just has to pull/push the changes and copy the packages between PPAs [22:24] flacoste, cool, thanks [22:25] I will update now [22:25] maxb: that sounds good [22:25] sinzui: ec2 --headless is too slow so i'm crossing my fingers and jfdi [22:26] barry, I think you meant to direct your statement about --headless to mwhudson [22:26] mwhudson is not headless! [22:27] barry: your changes should have only affected the doc/browser/stories tests. if something fails, I think you have justification to remove the offender [22:28] i thought i saw a bug from jml on --headless, but maybe it was mwhudson (or maybe it's assigned to him) [22:28] sinzui: yep. worst case, backing out the button change is easy [22:28] yes, --headless should go faster [22:28] unfucking buildbot seems higher priority right now though === matsubara is now known as matsubara-afk [22:32] flacoste, looks like we're doing a great job at deleting templates [22:33] beuno: well, not enough i think, we still have more than 60% to go and 3 weeks left [22:33] we need to make a serious dent in that curve this week [22:34] :) [22:35] nice burndown chart. it really helps visualize the progress. [22:40] mars: whoa! hey man! how goes it? [22:40] gary_poster, busy :) [22:40] heh, I bet [22:41] nothing much else to say, really. Just normal newborn stuff, but going through 30 nappy changes a day, instead of the usual 12. [22:41] lol [22:42] cool, well, congratulations. hope you send some pics to warthogs eventually [22:42] names first, then pictures :) [22:42] heh, ok, I'll agree to that ordering [22:42] flacoste, http://people.canonical.com/~beuno/conversions.html [22:42] everyone is happy and healthy, btw. [22:42] excellent [22:47] :( [22:47] my network connection went down while running tests on an instance. [22:48] maxb: btw, I have a branch that cleans out the last traces of the old zope branch, ~gary/launchpad/poppyasync. We've had more than the usual share of spurious failures today, so I'm retrying an ec2test run. Should be in either within the next few hours, or the next ~16 or so. I tried to make some progress on the karmic/buildout problem but getting things running on karmic is still not done, and I couldn't focus on it exclusively. [22:48] But anyway, I'm working on these things. [22:49] bye all [22:52] beuno, WOOT! === flacoste is now known as flacoste_afk [23:01] rockstar: skype? [23:01] thumper, here [23:01] rockstar: skype doesn't think so [23:02] Skype == retarded [23:03] lol [23:07] looks like we're in testfix [23:07] barry: shouldn't be [23:07] We never SHOULD be in testfix [23:07] mwhudson: hmm. i just had a branch fail because it didn't match the testfix regexp [23:11] barry: hmm [23:12] hello [23:12] mwhudson: otoh, r9280 seems happy [23:12] (in buildbot) [23:12] barry: jtv's branch seems to be being processed [23:13] barry: regexps are a notoriously wonderful user interface component, are you sure pqm is telling you it's in testfix? [23:14] mwhudson: i just deleted the email :( but i resubmitted :/ the regexp had no ui= or r= match that i could tell [23:14] mwhudson: let's see what happens with jtv's and my branches [23:15] The size of the diff (15093 lines) is larger than your specified limit of 1000 lines [23:15] oof [23:20] barry: can you say [ui=edwin, sinzui] yet? [23:20] I do not think you can [23:20] ui= [23:21] really? istr doing that in a previous branch [23:21] it would be crazy if we can't since we have ui mentats [23:22] barry: without a request to a losa to update the *provided* RE, it wont happen [23:22] are there any losas around right now? [23:22] barry: no we're all hiding. sorry. :-P [23:22] barry: first we need to get the current RE [23:22] barry: none, none whatsoever! [23:22] spm: can you give us the current regexp for pqm? [23:22] you guys... :) [23:23] arrrgghhh!!! stop typing 'bzr' when you want 'barry' [23:23] barry: sure. db-devel? devel? [23:23] spm: devel [23:24] barry: commit_re=(?is)^\s*(:?\[testfix\])?\[(?:release-critical|rs?=[^\]]+)\]\[ui=(?:.+)\] [23:25] mwhudson, sinzui ^^ see? ui=.+ [23:25] * jml tries running the tests for this branch again. [23:25] barry: cool [23:25] yay! well, let's see what happens [23:26] jtv's change landed [23:27] cool [23:29] * jml files bug 422274 [23:29] Bug #422274: PQM regex rejection emails are confusing [23:29] beuno: ping [23:30] thumper, pong [23:30] beuno: call? [23:30] thumper, in 10 [23:30] ok [23:31] flacoste_afk: how close are we to python 2.5 now? [23:34] thumper, ready [23:34] thumper: http://dev.launchpad.net/LaunchpadOnKarmic has a list of known test failures [23:36] success this time [23:46] thumper, https://wiki.canonical.com/Launchpad/UI/Navigation [23:46] barry: oh, before, it's possible that the buildbot-poller hadn't noticed that we were out of testfix yet [23:47] spm: how often does the buildbot-poll.py script run? [23:48] mwhudson: that's what i was thinking [23:48] mwhudson: every 5 [23:48] spm: thanks [23:48] MOAR EVENTS PLS [23:49] mwhudson: so [23:49] mwhudson: there is a fix branch script; that I've been nagging you to run :) [23:50] mwhudson: and as Andrew points out, users cannot fix their mirrored branches themselves - we overlooked that in the initial work. [23:50] lifeless: on which branches? [23:50] bug 354036 [23:50] Bug #354036: ErrorFromSmartServer - AbsentContentFactory (unfixable by users) error when pulling a branch from the mirrored area [23:50] lifeless: it sounds like it's going to take a LOSA to run this script [23:50] lifeless: is there much value in going through me? [23:51] * mwhudson puts his build engineer hat on so hard it cover his eyes [23:51] you could add the ability for people to say 'please remirror' [23:51] which would be generally useful as a in-extremis knob. [23:55] yes, there's a bug about that already [23:55] which i am also not going to work on this month [23:55] ok [23:55] then its an rt tissue [23:55] yeah [23:56] barry, still around?