=== Beret- is now known as Beret [06:15] hrm, I didn't know sinzui was a Gnome/gtk person as well. [06:16] Ah,gedit developer plugins. [06:25] * StevenK grumbles at Jenkins four persistant test failures [06:28] ... which don't fail on db-devel. Bleh. [06:33] StevenK: I will gladly grumble more about buildbot [06:33] :) [06:34] No one else seems to [06:35] My merge is pending because of buildbot :( [06:37] No, your merge was rejected due to buildbot-poller and testfix mode [06:37] oh. [06:41] buildbot just handles builds of devel and db-devel. buildbot-poller runs via cron, and checks to see how the last builds finished and changes the config of PQM. [06:55] Ahhh. [07:21] nigelb: A build has been forced, so we are likely out of testfix if you want your branch landed [07:24] StevenK: \o/ [07:24] * nigelb heads for lunch and offline-ish things [07:29] Which requires me to know which branch ... :-) [08:02] * G can't get my dev instance of launchpad working with e-mail :S [08:20] G: What's not sending mail? [08:20] G: Appserver mail will go to root@localhost. Script mail is not sent at all. [08:21] wgrant: yeah, I think I just realised my problem... bug mail is scripted [08:21] send-bug-notifications.py -vv or so [08:22] I've got postfix setup as a localhost mail server, with @canonical.com @ubuntu.com going to my user's mbox [08:23] Dev appservers redirect mail to root@localhost, and scripts (well, all those that use zopeless, which is just about all) send to nowhere. Most scripts will output the mail to stderr when given a -v or two. [08:24] I run LP is a lucid LXC container... I have script mail turned on there, because its postfix relays through my host, which delivers everything to my user. [08:25] wgrant: how did you enable the script-mail? [08:25] I may tell you once you ensure everything goes to a local user, not just @canonical.com and @ubuntu.com :) [08:25] wgrant: oh I have :) [08:26] Tested various non-Canonicalish domains, and none leave your machine? [08:26] Anyway, for bug notifications just give send-bug-notifications.py a couple of -vs, and it will show you what it "sends"/. [08:27] http://pastebin.ubuntu.com/681695/ [08:29] The second one resulted in an NDR... [08:29] wgrant: yep, but so does my real @gmail.com [08:31] I've got inet_interfaces set to loopback_only & default_transport to error [08:32] Ah. I use default_transport = local, luser_relay = wgrant [08:32] and I have virtual_alias_maps w/ @canonical.com & @ubuntu.com njones [08:33] don't worry, it was the first thing I checked, last thing I wanted to do, was spam Mark Shuttleworth :) [08:33] It would be amusing. [08:33] For those of us that aren't you, that is. [08:34] hahah yeah ): [08:35] only reason why I want to play w/ bug e-mails is because I think I spotted a bug where e-mails are sent out in the wrong order [08:35] I only know of one case. [08:35] Subscribed/assigned emails can be sent from the appserver. [08:35] During the request. [08:36] So they're not queued like the rest. [08:36] It can be a bit strange. [08:36] I forget exact details. [08:36] Anyway, in configs/development/launchpad-lazr.conf, right down the bottom, there should be a sinzui comment about zopeless mail. Flip that setting. [08:38] well take this for example: http://dev.nigelj.com/buge-mail.png logically speaking, the [NEW] should arrive before the non [NEW] [08:39] but I've also seen it recently when doing: assign to me, + comment at the same time [09:06] G: That's probably more to do with the Message-IDs and References headers. Have you checked the timestamps? [09:09] wgrant: yeah, 2 seconds in it [09:09] wgrant: which makes me think it's the order it's put in the queue [09:10] hmmm looks like watching TV + LP 'make run' and the python send bug mail script is too much for my laptop [09:10] G: [NEW] is probably sent direct from the appserver. [09:10] Yeah, LP likes its RAM. [09:10] Particularly on amd64. [09:10] Yay Python. [09:10] wgrant: I had a poke round, and it seems they are both sent the same way [09:11] I think it's a case of wrong order, which is why I wanted the e-mails via postfix just to test the theory and have everything in one place [09:11] LP likes it's RAM, and TV likes it GPU/CPU for the h264 [09:12] G: If you have an example from production, the headers will tell you where they are sent. loganberry/ackee run send-bug-notifications and other scripts, while chaenomeles/gac/soybean/wampee run appservers. [09:12] wgrant: ahhh thanks [09:12] in that cause you are right [09:13] [NEW] comes from the appservers, the others are from send-bug... [09:13] * G takes it servers are named after fruits or something there? [09:13] There have been a few naming schemes. [09:14] First was penguins. Which is why there is gentoo.ubuntu.com. [09:14] Then there were Antarctic bases. [09:14] oh I thought that was a big joke [09:14] Then elements (you'll see lots of these on https://launchpad.net/builders) [09:14] I'd seen it mentioned somewhere [09:14] Then fruits. [09:14] And now stars, although there aren't many of them yet. [09:15] heh, Native New Zealand Birds for my network + VPS' [09:15] unfortunately some aren't overly suitable, because although they are Maori, they could be read the wrong way :) [09:16] There have been a number of names vetoed, sadly :( [09:17] I bet :) [09:23] come to think of it, amazing that only 4 machines are needed to run the main part of launchpad.net [09:25] There are lots more behind the scenes, but yes, those four serve all the production appserver requests. [09:26] oh yeah, I saw some comment the other day about the DB servers and that [09:27] and I'd hate to think how much storage is backing the PPAs and that but still [09:27] A bit :) [09:29] (For what it's worth I used to be a Fedora guy, but I kept running into brick walls, and PPAs was actually something that converted me [09:29] hated having to roll my own RPMs and keeping track of them somehow [09:29] Arch has AUR, openSUSE has OBS... does Fedora have anything? [09:30] not unless they added something to Koji in the last year [09:30] (Koji been the buildsystem) [09:32] I've never understood that. Everyone else has a TLA... even Debian. But noooo, Fedora just has to go and use a word instead :P [09:32] haha [09:33] but yeah, one of the things I like about Launchpad, is that it's all interconnected like it should be [09:33] (at least imo) [09:33] Well, a bit :) [09:34] Fedora has about 5 or 6 core 'applications' and then they still have to use RH's Bugzilla ;) [09:34] There is great benefit in having them all linked together. But also a significant cost with the current way things are done. [09:34] ie. a 650KLOC monolithic Zope webapp. [09:35] But things are moving in the right direction now. [09:35] tbh, I thought Zope had died out [09:36] I'm not sure that Zope 3 was ever really alive. [09:46] hmmm can't seem to confirm the bug I was looking for but I think I found a completely different one [09:47] It wouldn't be an issue for launchpad.net, but for instance, my machine is +1200, script mail, comes through with the correct date/time, but e-mail from the app layer, comes through at NZST, but saying it's -0000 [09:47] Yeah, there's a bug for that. [09:48] Until a year or so ago the appservers ran in BST, so some mail was hour off for a few months a year. [09:48] I'm pretty sure I filed it years ago... [09:49] Bug #262040 [09:49] <_mup_> Bug #262040: Bogus timestamps on some unbatched bugmail

< https://launchpad.net/bugs/262040 > [10:57] Project devel build #1,030: STILL FAILING in 4 hr 38 min: https://lpci.wedontsleep.org/job/devel/1030/ [13:10] wgrant: btw, thanks for the help before, managed to get at least one of the issues I wanted sorted [14:32] staging is down? [18:05] I love who StevenK's branches are always deleting or killing or burning :D [18:05] *how [18:05] GAH [21:57] qa-ing merge proposals on qastating seems to be fun [22:55] mwhudson: All the files are restricted, so yeah, you basically have to create a new one. [22:55] yeah, it's not so bad i guess, just a bit tedious [22:59] wgrant: i have a packaging question for you. i upgraded to oneiric beta over the w/e and it hosed my system so i reinstalled from scratch [22:59] wgrant: and now lp devel dependencies won't install [22:59] because postgressql-8.4-debversion can't be installed [22:59] What's the error? [22:59] Hmm. [22:59] Ah. [22:59] It's only here for 9.1 [23:00] Grab it from Natty for now. [23:00] because libapt-pkg.4.11 is instgalled with oneiric and it needs libapt-pkg4.10 [23:00] Ah. [23:00] i forced libapt-pkg-4.10 to install [23:00] but synaptic keeps removing it [23:00] because it says it's broken [23:00] It is. [23:01] GIve me a few minutes and I'll have an oneiric postgresql-8.4-debversion in the PPA... [23:01] cool! thanks [23:01] i don't know much about this stuff [23:01] But you should use a lucid LXC :) [23:02] wgrant: yeah, haven't reinstalled lxc yet [23:02] but lp does run ok with my current broken set up [23:04] Hmm. We really should upgrade to 9.1 soon. [23:05] yes, an upgrade to 9.1 would be good [23:05] scary though given the pain with 8.3->8.4 [23:15] Hmm, not sure if the new version will actually build with 8.4. May have to rename the old one and build that as well. [23:22] wgrant: i think this happened because there was a recent apt upgrade in oneiric. so is your system broken yet? or are you yet to see this issue locally? [23:22] wallyworld_: I use LXC :) [23:23] ah of course [23:23] i'll reinstall lxc but the memory saving didn't seem worth the hassle. but this on the other hand... [23:43] wallyworld_: Could you test the postgresql-8.4-debversion from ppa:wgrant/experimental? [23:43] It finally built. [23:43] wgrant: sure will. thanks very much [23:44] * wallyworld_ opens synaptic [23:48] wgrant: it works \o/ [23:48] Yay. [23:48] * wgrant copies. [23:48] now all i need to do is disable zeitgeist because it keeps easting all my cpu and oneiric is looking good