=== bigjools-afk is now known as bigjools [00:50] -> monthly allergy shot. Urgent things you can SMS/ring. [03:28] wtf [03:28] how can bug 951365 be a regression [03:28] <_mup_> Bug #951365: Accessing milestone on project group via webservice api returns has_bugs object

< https://launchpad.net/bugs/951365 > [03:29] wgrant: so i think that happens because a project groups milestones are IProjectGroupMilestones, not IMilestones [03:29] mwhudson: The only thing I can think of is milestone tags. [03:30] That may have changed things. [03:30] ah [03:30] i suppose that might have resulted in some shuffling [03:34] milestone tags was only a 3000 line change... [03:34] Pfft [03:35] wgrant: good guess [03:35] What happened? [03:36] wgrant: well, that was the change that made IProjectGroupMilestone not inherit from IMilestone [03:37] Ah [03:37] That would do it indeed. [03:38] i _guess_ the milestone collections in iprojectgroup need to be overridden to declare returning IProjectGroupMilestone [03:43] ARGH [03:43] I hate lazr.config [03:44] Particularly some of the ways we use it. [03:45] config.error_reports.error_dir can be overridden by error_dir in a script-specified config section. [03:45] Except that the config schema has some dev paths as the codehosting sections' error_dirs. [03:45] Which means that dropping the codehosting error_dirs from the prod configs won't fall back to the configured production default, but the dev default. [03:54] hm, i wonder if i've run rocketfuel-setup on this machine [03:55] mwhudson: If /etc/hosts is 600,000 lines long, yep. [03:56] appears i have not [03:56] LXC! LXC! [03:57] ah good point [03:57] It's nice and easy on precise [03:57] I cleaned up the howto on whatever day was earlier this week [04:00] Speaking of precise, I'm trying to figure out how to upgrade my fileserver. [04:01] It's Ethernet module isn't in Lucid's kernel, so it's dkms'd, but I don't want it to be dkms'd on upgrade, since the module is in Precise's kernel. [04:01] ''' [04:05] wgrant: this howto? https://dev.launchpad.net/Running/LXC [04:05] That one [04:06] cool [04:07] wgrant: fix the schema ? I don't see any reason for it to have dev data in it, we have a dev config for that [04:07] lifeless: Yeah, fixed in devel a few minutes back. [04:07] It's still full of crap, but it's slightly less offensive and stupid crap. [04:19] wgrant: launchpad-dependencies: Depends: python-apt (>= 0.7.94.2ubuntu6.4) but 0.7.94.2ubuntu6 is to be installed [04:19] (in the lxc) [04:20] mwhudson: Sure you ran rocketfuel-setup? That should be in the launchpad PPA [04:20] mwhudson: No -updates? [04:20] Or -updates [04:20] StevenK wins [04:20] Hmm [04:20] I'm pretty sure my fresh container had that by default [04:20] mwhudson: Has this machine run LXC before? [04:20] bad precise lxc cache probably [04:20] wgrant: yes, i lxc-destroy-ed the lxc i had [04:20] sources.list was 1 line [04:21] mwhudson: there is a cache as well [04:21] oh excellent [04:21] Ah [04:21] I'd suggest restarting [04:21] oh ok [04:21] sudo rm -r /var/cache/lxc/lucid [04:21] Then lxc-create again [04:21] Vast improvements were made in the last couple of months [04:21] Such that it's no longer torture to get LXC working. [04:21] It'll redebootstrap an environment that actually works [04:22] heh [04:22] ok [04:22] maybe a note along these lines could be added to the wiki [04:22] * wgrant adds === jtv1 is now known as jtv [04:39] mwhudson, Are you fixing LP #951365? :) [04:39] <_mup_> Bug #951365: Accessing milestone on project group via webservice api returns has_bugs object

< https://launchpad.net/bugs/951365 > [04:39] cody-somerville: its a possibility [04:41] come on intertubes, go faster [04:58] mwhudson: Going OK? [04:58] wgrant: yes [04:58] Great. [04:58] I haven't actually completely tested the latest version of the instructions from scratch :) [04:58] heh [04:58] well rocketfuel-setup --no-workspace is grinding away [04:59] Great. [05:04] wgrant, wallyworld: Either of you want to look at https://code.launchpad.net/~stevenk/launchpad/bugs-information_type-mail/+merge/104483 ? [05:05] StevenK: Underscores? [05:05] Really? [05:06] I don't think we have underscores anywhere else in the UI [05:06] Mail or Web. [05:06] wgrant: 'information type' doesn't work for the command, 'information_type' does. :-( [05:06] informationtype does as well :) [05:07] Does this follow value parsing rules similar to status? [05:09] wgrant: informationtype works, I guess [05:10] information-type [05:10] * mwhudson bikesheds [05:10] wgrant: Status just calls transitionToStatus(), no checking [05:11] I didn't think CreateBugParams supported that method, but ICBW. [05:11] StevenK: But how does it parse the value?" [05:11] wgrant: It's done via setAttributeValue, so the parsing is in the guts of the mail handler, I guess. [05:12] StevenK: The behaviour of status and informationtype parsing has no reason to differ. [05:12] It sets a dbschema = BugTaskStatus [05:13] wgrant: Except for 'Private' and the forbidding of 'Proprietary'. [05:14] True. [05:14] The the latter is unrelated. [05:14] It's a temporarily post-parse filter operation [05:14] Bah [05:14] It is too cold to type sensibly [05:14] Oh? [05:14] It's 24degC in my study [05:18] wgrant: I think after we drop Private, we can change informationtype to be like Status [05:21] Or we could just declare we won't support Private on the mail interface, and I can change it now. [05:21] * StevenK laughs [05:31] * mwhudson wonders if rocketfuel-setup has asked for his root password for the last time... [05:32] Bug #992184 seems rather special [05:32] <_mup_> Bug #992184: lib/lp/services/database/doc/textsearching.txt fails intermittently/rarely on parallel tests < https://launchpad.net/bugs/992184 > [05:33] yes! [05:33] * mwhudson disappears [05:36] StevenK: i was doing school pickup, i assume you don't need me to look at your branch? [05:37] wallyworld: I was having a discussion with wgrant until his Internet broke due to no power. [05:37] lol [05:38] * wallyworld remembers a time when there wasn't any internet even with power [05:39] wallyworld: and where 9600 baud seemed like an impossible speed dream? [05:39] my first modem was 300/75 [05:39] so yes :-) [05:39] Igot lucky. started on 1200 [05:39] LUXURY [05:40] took ages to download all that ascii 'porn' [05:40] wallyworld: Download? Feel sorry for the poor bugger that had to type it in. [05:40] rotflmao [05:41] i still laugh every time an 80's movie shows a modem instantly connecting without the audible handshake [05:41] i used to love the handshake [05:42] I had a friend at uni would could whistle 2400 baud [05:42] s/would/who/ [05:42] so he could connect to the CIA and order a nuclear strike? [05:42] Haha [05:44] spm: bah, 300/75? thats speedy [05:44] lifeless: itym wallyworld [05:45] ohbahdee [05:45] what was slower than 300/74? [05:45] yesIdoo [05:45] 75 [05:48] wallyworld: a) I was trolling, b) 110 baud [05:48] i was genuinely curious :-) [05:48] 110 baud, wow that's slow [05:49] http://en.wikipedia.org/wiki/Bell_101 [05:49] but since baud refers to state changes per second, if each state change encoded 16 bits it wouldn't be soooo bad [05:49] My first modem was a 300/75 [05:50] * wallyworld is now reminiscing about the Hayes command set [05:50] those were the days [05:50] http://en.wikipedia.org/wiki/List_of_device_bandwidths is pretty good for reminiscing [05:50] they left off the 2 cans either end of a taught piece of string [06:09] I'm rather pleased to say that I've compeltely forgotten all modem commands beyond AT. and I'm proud that I'm unsure if even that's correct. [06:09] I can remember ATDT [06:22] +++ [06:23] * wgrant stabs u-boot [06:23] Or flash-kernel [06:23] One of those [06:23] Haha [06:27] I started with a 56kbps modem, you old people. [06:30] luxury [06:38] wgrant: modern luxury [07:38] good morning [08:10] There we go [08:11] Running/LXC tested and minimised and prettified === StevenK_ is now known as StevenK === cjwatson_ is now known as cjwatson [09:32] https://dev.launchpad.net/Running/RemoteAccess is a bit sad. [09:34] With SNI widely supported it seems like we don't really need the two IP addresses any more. [09:36] widely supported enough, really? [09:37] for dev use [09:37] yes [09:37] The only thing in the config that needs it is private loggerhead, which will only be accessed by a web browser, and everything except those using Windows XP's crypto stack supports SNI. [09:37] long as we don't use python to access anything [09:37] wgrant: ^ [09:37] Yes [09:38] As long as we have the main vhost first, python will be happy [09:38] oh, is that the degraded behaviour ? [09:39] I'm pretty sure it degrades the same as missing Host with HTTP [09:39] It picks the first matching vhost [09:40] I'm thinking I'll add a make target which tweaks the config to listen on * and allow a user-specified subnet. [09:51] how can I set up an 'ec2 demo' server so that it's possible to log on? [09:52] jml: What breaks when you try? [09:52] SNI - only ten years late [09:52] wgrant: I get told that the openid provider is unavailable [09:52] OpenID Provider Is Unavailable at This Time [09:52] The openid provider was unavailable. Please try again in a moment. [09:52] You can also join the #launchpad IRC support channel on irc.freenode.net to ask for further assistance. [09:52] jml: That's slightly unpleasant of it. [09:53] wgrant: yes. the capitalization is egregious. [09:53] Both the fact that there's an error and the text content of the error [09:53] Yes [09:53] At least it doesn't say FreeNode [09:53] Let's see. [09:56] Slow ec2... [10:04] wgrant: can you run a query on dogfood for me? [10:04] wallyworld: Sure [10:04] https://pastebin.canonical.com/65435/ [10:05] it's for a job [10:05] so not 100% critical is blazingly fast [10:05] replace the policy ids [10:05] Ubuntu? [10:05] sure, why not :-) [10:06] it's to select all bugs a person is subscribed to but doesn't have a grant for (including via a team) [10:08] I expect it will be very slow, but let's see. [10:09] i may need to use a CTE [10:09] i have a bunch of tests passing, just need to optimise :-) [10:10] wow [10:10] it's so nice to have so many Australians around [10:10] hi jml :-) [10:11] lifeless: Do you have time to talk in about 10 or 11 hours, about MAAS architecture? [10:11] wallyworld: hi [10:11] hey jml! Ah bugger, I'm not really Australian [10:11] jml: how are they hanin', mate? [10:11] bigjools: bloody immigrants. [10:11] bigjools: fuck off [10:12] wallyworld: all good, thanks [10:12] wallyworld: I'd be quiet if I were you or I'd tell everyone you were listening to Britney today. Oops, too late :) [10:12] ow ow, it hurts to laugh [10:12] jml: bastards eh :) [10:13] bigjools: lies, all lies [10:13] and also you had Lady Gaga on [10:13] I bet you wear a dress when nobody's looking [10:14] no, just panties [10:14] with the peep hole [10:14] that way i can wear them when people are looking [10:14] wallyworld: I'd spell it http://paste.ubuntu.com/964307/ [10:14] wallyworld: Which is fast [10:14] * wallyworld looks [10:15] It satisfies my subscriptions to Ubuntu in 30ms [10:15] ubuntu-crashes-universe in 1010ms hot [10:15] Which is not bad. [10:16] wgrant: i didn't realise btf has access_grants [10:16] That doesn't have the restriction to Ubuntu because i am forgetful [10:16] But it's easy enough to add into the query [10:16] wallyworld: Yeah, it has all the access info so searches can be fast. [10:16] And this is a search :) [10:17] wgrant: does && mean AND? i've not seen thhat before [10:17] wallyworld: && is array intersection [10:18] ok. is that postgres specific? or standard sql? [10:18] The array stuff is a postgres extension [10:18] AIUI [10:18] i bet storm doesn;t support it, so i will have to hand code the query :-( [10:18] It's 3 lines to define && [10:18] suck it up [10:18] And I think we already have one in the codebase [10:18] Bug subscriptions use it [10:18] pointer? [10:19] lib/lp/bugs/model/structuralsubscription.py: oper = "&&" [10:19] bigjools: i've said it before and i'll say it again: fuck off [10:19] I'd move that into lp.services.database.stormexpr with the rest [10:19] wallyworld: However [10:19] wallyworld: muahaha :) [10:19] wallyworld: Those access checks are the same as the ones in bugtasksearch [10:19] wallyworld: You can probably reuse get_bug_privacy_filter [10:19] See how the access checks are done in there [10:19] It's not pretty yet, but eventually. [10:20] night all [10:20] Fuck off bigjools :) [10:20] wgrant: wallyworld gets away with it but not you [10:20] Aw [10:20] get back to your basement [10:20] wgrant: i'll have a look. i may need to refactor what i've done. i've defined an indirect_bugs_filter helper method in accesspolicy.py [10:21] wallyworld: get_bug_privacy_filter also handles public bugs, which I omitted from that query [10:21] Mostly because I forgot [10:21] ok. thanks. will have a look [10:27] :) [10:30] jml: Ah [10:30] jml: You still have that ec2 instance? [10:31] wgrant: yes. [10:31] jml: I suspect it will unbreak if you s/$INSTANCE_IP/*/ in /etc/apache2/sites-available/local-launchpad [10:32] hmm. [10:32] jml: ec2 demo forces it to listen on the external interface, when testopenid.dev will point to localhost [10:32] So the local appserver can't talk to the provider. [10:32] wgrant: ok, will try that. [10:33] launchpad-developer-dependencies only suggests libapache2-mod-wsgi. I guess it should depend now. [10:34] Ah [10:34] No [10:34] It also only suggests apache2, so I guess ec2 installs that manually. [10:34] yeah, haven't looked sorry. [10:36] wgrant: ok. Now I get "unknown email address" when I try to log in / register. [10:36] jml: admin@canonicalcom? [10:36] admin@canonical.com [10:37] wgrant: that gives me the following: http://pastebin.ubuntu.com/964360/ [10:38] jml: Try clicking Log In / Register again [10:38] I suspect your session might have died. [10:38] wgrant: ooh, I only just noticed that when I got that error, the log in / register link was gone and replaced with Foo Bar (~name16) [10:38] Aha [10:39] which is weird but workaroundable. [10:43] Retrying with my fixes now. [10:54] wgrant: thanks. === al-maisan is now known as almaisan-away [11:17] Fourth time lucky... [11:24] wgrant: are your fixes to lp-dev-utils or to launchpad? [11:27] wgrant: I'm going to leave now – I have a lunch appointment. My IRC proxy will be connected. [11:31] jml: lp-dev-utils [11:32] wgrant: ah, thanks. I was going to do a patch for the '*' thing but will await yours. [11:32] * jml really off [11:32] Ah [11:32] Neds an image rebuild to pick up the new package [11:32] That's why it's not working. [11:33] * wgrant will commit and publish a new image. [11:33] Thanks for letting us know. [11:33] Hmm, then poolie can drop 523. Which might be difficult to organise. === almaisan-away is now known as al-maisan === danhg_ is now known as danhg === dobey_ is now known as dobey [13:02] jml: lp-dev-utils is fixed, and there's an updated image [13:02] so demo should work properly now [13:02] Also, buildout is slow [13:06] News at 11 [13:18] wgrant: thanks. === jcsackett_ changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: jcsackett | Firefighting: - | Critical bugs: 3.47*10^2 === jcsackett_ is now known as jcsackett [14:15] can someone please land lp:~jml/lp-dev-utils/public-ip [14:15] abentley approved it [14:18] jml: done. [14:18] abentley: thanks! [14:31] running a dev instance via 'ec2 demo', looks like the private xmlrpc server isn't running [14:31] nothing is listening on 8097 [14:31] how do I start that up? === al-maisan is now known as almaisan-away [14:38] hi [14:38] g'morning :) [14:38] sorry jml, not sure there. Looking at the makefile not sure which bits start up the xmlrpc [14:41] fwiw, we're getting https://pastebin.canonical.com/65462/ when we try to do an xmlrpc request [14:41] achuni: is that right? [14:41] and are being redirected from http to https. [14:42] jml: that's just hitting the xmlrpc endpoint in a browser, I can try an xmlrpc request in a bit [14:42] achuni: ah ok. [14:42] http->https redirect, yep. afaict we should be using https anyway [14:43] *ahem* [14:43] 8087 [14:44] never mind [14:44] it is running, and I'm a fool. [14:47] achuni: I get http://pastebin.ubuntu.com/964783/ [14:47] jml: looks good [14:47] achuni: good? is that what you expect? [14:47] (does launchpadlib support proxies? Can I ask it to use one?) [14:48] jml: well, up until the 403 response, yep [14:48] achuni: if it didn't hurt so much, I would laugh bitterly [14:49] :D [14:49] nothing on the help wiki [14:49] jml: you think that's authorizing by IP just? we don't do any kind of authentication for that call afaict [14:49] it just uses httplib2 though [14:49] so it's probably whatever that uses [14:50] right [14:51] googling doesn't reveal much [14:52] :( [14:53] httplib2 supports proxies via the socks module [14:53] as in, by writing code [14:53] yeah, checking if I can ask launchpadlib to pass in the right args [14:53] nothing in the stack seems to have environment-based http proxy support. [14:54] * achuni hacks /usr/lib/python2.7/dist-packages/httplib2/__init__.py [14:55] :( [14:55] achuni: just double checking, you're not waiting on me for any questions, right? [14:56] jml: nope, thanks [15:04] jml: kind of success. It seems to have worked, I now get Permission denied: '/var/tmp/launchpad_mailqueue/tmp/1336057390.27818.domU-12-31-39-08-04-1D.1557865906 [15:04] achuni: hmm. that might be an fs thing. [15:05] odd. [15:05] achuni: I've made those writeable. Try again? [15:06] weird, let me pastebin [15:07] jml: https://pastebin.canonical.com/65469/ [15:07] jml: still a Perm denied failure, but now on an unspecified location [15:08] effing Python errors [15:08] how many hours have been lost because the 'os' module doesn't raise errors with function arguments [15:09] * jml takes a deep, slow breath [15:10] achuni: oh, my bad. [15:10] achuni: there were more directories that needed permission changing [15:10] (my criticism of Python is still valid though) [15:11] achuni: try again? [15:16] jml: I haz creds :) [15:16] achuni: \o/ [15:16] achuni: what's next? [15:17] jml: standup, but I'll try those creds out next, in a bit === mpt_ is now known as mpt [15:18] achuni: heh, thanks. [15:24] bac: all clear === matsubara is now known as matsubara-lunch === salgado is now known as salgado-lunch === deryck is now known as deryck[lunch] [16:25] hi there, was wondering if anyone could help me figure out why a few mails from the openstack list are occasionally silently vanishing into a bitbucket instead of reaching me? [16:25] I receive most of them [16:25] and have checked my spam folder [16:29] adam_: are they showing up on the archive? [16:29] czajkowski: yes [16:29] e.g. https://lists.launchpad.net/openstack/msg11012.html [16:36] adam_: any issue with your email as nobody else has said there are any LP mail issues [16:37] czajkowski: not aware of any other issues or dropped mail from other sources, no [16:38] hi guys, I'm writing an ubuntu accomplishment that should test if a person had personally verified an SRU bug to earn the tester trophy. I'm using python api for launchpad, any idea how it could be done? [16:38] czajkowski: don't suppose you'd be able to pinpoint that mail in the MTA logs and see if it was accepted by the MTA my end? [16:39] adam_: I can ask and see [16:39] if it was then I'll know I have to fight battles with our corporate email team ... [16:39] thanks! [16:40] adam_: and your email address it would have gone to is ? [16:40] aspiers@suse.com [16:43] adam_: checking [16:44] awesome, thanks :) [16:55] sinzui: thanks for helping out with the review queue. :-) [16:59] czajkowski: gotta go now, but I'll stay logged in to catch anything you might say ... thanks again! [16:59] adam_: I'll be gone shortly and no update yet [16:59] czajkowski: ok, I'll be around tomorrow too [17:00] ok [17:00] of course I'm reachable on that email address too ... well, hopefully ;-) [17:00] jcsackett, I forgot it was Thursday [17:00] sinzui: it's all good. that was a sincere thank you. that branch was the largest in my queue. :-P [17:01] czajkowski: UK timezone, that is [17:01] jcsackett, :) [17:02] czajkowski: hopefully we'll overlap by at least one hour but you never know on this channel ;-) [17:05] adam_: look like it did make it to that end === deryck[lunch] is now known as deryck [18:09] allenap: I'd be delighted to === lifeless_ is now known as lifeless [18:13] allenap: I'm going back to bed for a bit (0600 here) but will ping when I'm on deck [18:23] * deryck reboots, will brb [18:34] does anyone know why launchpad_mailqueue is unwriteable on ec2test? [18:35] ec2test/testrunner.py:426 [19:16] jcsackett: if you have time for a small review, I have a bite sized one for you: https://code.launchpad.net/~benji/launchpad/bug-992692/+merge/104599 [19:41] how do I create new users on a demo instance? [19:41] jml: there is a script [19:42] though we should really get demo talking to real openid [19:42] oh wait, 'make-lp-user'? [19:43] I believe so [19:43] I think I wrote that. [19:43] Anyway. [19:44] achuni: you've got shell access, right? [19:44]