[02:23] sinzui, so did bug expiry actually get reenabled? [02:23] post bug 333521 [02:23] <_mup_> Bug #333521: Enable bugs expiration for Ubuntu [02:46] poolie: There has been no screaming, so I doubt it. [02:46] The plan fortunately seems to have vanished. [02:47] * thumper needs more coffee [02:49] wgrant: what, you don't like bug expiration? [02:49] lifeless: Not the latest proposal. [02:50] of switching it on without unsetting the flag globally. === Ursinha is now known as Ursinha-afk [08:07] good morning === henninge changed the topic of #launchpad-dev to: Launchpad Development Channel | Week 4 of 10.06 | PQM is open | https://dev.launchpad.net/ | Get the code: https://dev.launchpad.net/Getting | On-call review in irc://irc.freenode.net/#launchpad-reviews | Use http://paste.ubuntu.com/ for pastes [09:08] Morning [10:25] good morning === jtv1 is now known as jtv [11:02] Morning, all. [11:12] morning deryck [11:21] zope.configuration.config.ConfigurationExecutionError: : /var/tmp/launchpad_mailqueue is not a Maildir folder [11:21] wtf [11:33] sigh, why does /var/tmp/launchpad_mailqueue get created with root ownership, yet "make" wants to remove it etc. [11:33] (as my local user) === matsubara-afk is now known as matsubara === thekorn_ is now known as thekorn === Ursinha-afk is now known as Ursinha [14:05] It would be interesting and not too difficult to calculate the actual buildd utilisation at times like this. [14:06] I suspect it's well under half of the 100% that it appears to be. [14:16] It's a bit concerning that one can actually watch buildd-manager's progress through a single cycle over several minutes by refreshing /builders. [14:22] We know. It's getting fixed. [14:37] wgrant: so still 261k where date_started is null and status != 0; [14:37] mmm [14:38] bigjools: Er, really? [14:38] really [14:38] But that is the same as the number you quoted without the filte. [14:38] And there are somewhere around 20k builds pending. [14:38] And probably tens of thousands more suspended. [14:39] So it does not compute. [14:39] well I am using staging's DB [14:39] which is a little out of date but near enough [14:39] oh, meh [14:39] we blitz all pending jobs on staging now [14:39] it has a buildd-manager [14:40] * bigjools dodges the bird that just flew into the office [14:40] What if you say status NOT IN (0, 5)? [14:40] (5 being superseded) [14:40] staging has a build farm now? Handy. [14:41] 120k [14:41] well if you count one machine with two builders on it as a farm ... [14:42] Aha. [14:43] OH~! [14:43] gina. [14:44] 22:14 < wgrant> No date_started but successfully built is fine. [14:44] 22:14 < wgrant> No date_started, but successfully built *and with date_finished or builder* is not. [14:44] I got my issues from that day mixed up :/ [14:44] There were 9 builds in this state at that time. [14:44] It should be the same now. [14:46] gror [14:48] hardware failures....awesome [14:48] Bird-related? :P [14:52] no, the machine I was just running queries on :/ [14:53] Oh, excellent. [14:53] not exactly [15:06] wgrant: can you clarify that clause again? [15:07] oh nm [15:07] little bit distracted today, going to see what breaks when I run the tests with no branch sample data. [15:09] wgrant: http://pastebin.ubuntu.com/456906/ [15:09] gnargh [15:11] oh nm there's only 9, stray clause table [15:11] Ah, good. [15:11] I was WTFing a bit there. [15:12] def just 9 [15:12] ok, now, what shall I poke in the start date [15:12] * bigjools ponders [15:12] The 'builder IS NOT NULL' disjunct was only required when there may have still been incomplete builds. [15:12] You could poke in something just around the rollout time. [15:13] yeah [15:13] Or if you're feeling adventurous actually grab the start time from the logs. [15:13] But that's probably pointless. [15:23] henninge_: did you see my email about caching the message count? [15:24] wgrant: have you got any idea why someone would write that first line of code in IArchive.checkArchivePermission() [15:24] maybe they were on a low IQ day [15:25] bigjools: Ah, that's why it's OOPSing? [15:26] yes [15:26] words fail me [15:26] Well, it's not terribly unreasonable if one assumes it's only used for checking upload permissions. [15:27] it should just return False FFS [15:27] Well, if you're checking whether a source should be accepted into a copy archive, you probably want to crash. [15:27] But this method is used elsewhere, so... === henninge_ is now known as henninge [15:27] no I never want to crash [15:27] returning False is entirely appropriate [15:28] If you get to the stage where you are verifying permissions for a copy archive, you probably want to let someone know. [15:28] Because something is seriously broken. === Chex_ is now known as Chex [15:28] there's too much of that paranoia in Soyuz [15:28] EdwinGrubbs: yes I saw it but have not read it yet. [15:28] sorry ;) [15:29] bigjools: I think it's a handy defensive tactic, given how badly things can go wrong. [15:29] Just... not when applied incorrectly. [15:29] the latter is my beef [15:30] That line has been there for a while. [15:30] i wonder why it hasn't been problematic before. [15:31] I guess maybe canUpload wasn't actually used for that before :/ [15:32] It is a reasonably novel method. [15:38] it's used when ascertaining retry availability [15:38] if you can upload you can retry [15:38] I know why it's used now. [15:39] I'm just wondering why it never appeared before, when that code has been there for ages. [15:39] And those pages worked a few months ago. === Ursinha_ is now known as Ursinha [15:46] you have to hit the build in a certain state [15:47] if you can upload you can retry <--- I can upload to ~launchpad/+archive/ppa but I can't retry [15:47] maxb: it only applies to the main archive :) [15:47] consistency? What's that? :-) [15:47] maybe it should not [15:48] I've been meaning to fix that for copy archives for a while. [15:48] I didn't realise it affected PPAs too. [15:49] well for copy archives it should only let the owner and admins [15:49] Or custom uploaders, if they are added. [15:49] No reason to forbid it, and it's handy. [15:49] custom uploaders are sorta evil at the moment [15:49] eg. MOTU should be able to retry builds in the rebuild archive, without having to poke doko. [15:50] I want to move all upload permissions to explicit ArchivePermissions rather than relying on the ownership of the PPA [15:50] Yes yes yes. [15:50] All this implicit permission stuff is silly. [15:51] Although I guess we won't see those ACLs for a while now. [15:51] Does the launchpad api have a sleep type thing built in, or am I going to get banned from the site for using the api too much? [15:52] cakofony: not unless we can see it grossly affecting performance [15:53] bigjools: I'm working on an opensource project that collects as much data as possible from forge sites -- we pull a lot of data :-s [15:53] cakofony: is that a one-off or ongoing? [15:54] bigjools: we do a run every month or so (it probably takes a day or two) [15:54] bigjools: before I found the api, I was downloading about 100,000 html pages per run, this should speed things up quite nicely [15:55] okay - I might put you on to someone else like, errr, mrevell perhaps [15:55] and jml might be interested too [15:55] what [15:56] cakofony, there's no built in sleep [15:56] cakofony, if you become a problem, we'll let you know :) [15:56] cakofony, what's the website? [15:56] jml: flossmole.org [15:56] cakofony, thanks. [15:57] jml: thank you! I'm calling the client in my program "python-flossmole" [15:58] cakofony, are you syncing bug data, or just projects & developers? [15:58] jml: only projects teams and developers === matsubara is now known as matsubara-lunch === matsubara-lunch is now known as matsubara === Ursinha is now known as Ursinha-lunch === salgado is now known as salgado-lunch === salgado-lunch is now known as salgado [19:27] after amateur tries to update MLT to 0.5.6 i have left without ffmpeg modules and even ffpmeg is installed, kdenlive says that some not installed at all. also it says that some sound module is not installed. i spent all day to make "lines and dots" bug dissappear (white lines and dots - was promised to be fixed in MLT 0.5.5) and i couldn't make it, even worse - now modules "avformat module", "Quimage module", "Title module" are missing and reinstallin [19:27] g of the program and ffmpeg does not helping. [19:27] help me please to make this thing work correctly. my skype is "woanerges", or write me here. please, bro's, come on, i need some support here! [19:27] white dots and lines examples: [19:27] http://kdenlive.org/sites/default/files/shot1_0.png [19:28] http://www.youtube.com/watch?v=nrFXr_bx2a0 === EdwinGrubbs is now known as Edwin-lunch [21:26] abentley, would you object if I factored out the sourcepackagerecipebuild views in browser sourcepackagerecipebuild.py ? [21:27] rockstar, I don't understand. [21:28] abentley, so lp.code.browser.sourcepackagerecipe has views for sourcepackagerecipebuilds in them. I want to move them out. [21:28] rockstar, alright. [21:29] abentley, cool, thanks. === zyga_ is now known as zyga === matsubara is now known as matsubara-afk [22:14] morning === salgado is now known as salgado-afk [23:02] hiya thumper [23:02] hi lifeless === beuno_ is now known as beuno === Edwin-lunch is now known as EdwinGrubbs === krkhan_ is now known as krkhan