[00:04] cjwatson, Riddell: Why's it so slow (the script is not public, AFAICT)? [00:05] sbeattie: is it needed for a paticular reason? if not it' my archive day tomorrow and I'll probably get to it. nobody else is doing regular New processing so it's not unusual for things to be stuck for a week [00:06] wgrant: because the soyuz queue command is slow and it runs it once for each binary package [00:06] Well, it seems like kernel-overrides could be optimised to call it fewer times. [00:07] Making it quicker to start is unfortunately not easy :/ [00:07] Yay Zope. [00:07] Riddell: it's not strictly necessary, but the security team has some tracking scripts that whinge when package sources disappear from the archive. [00:07] Riddell: tomorrow is fine. [00:09] wgrant: on my archive admin days a significant amount of my time is spent waiting for the queue command, it's probably one reason why I'm the only active archive admin left [00:10] Riddell: Do you normally need to do per-binary overrides, or would the current web UI suffice? [00:10] Making that quicker isn't terribly difficult. [00:10] never used the web UI, didn't know one existed [00:10] Riddell: That's what I have to use. [00:11] Riddell: https://launchpad.net/ubuntu/maverick/+queue [00:11] Doesn't let you do per-binary overrides, but is otherwise not bad. [00:12] It is however, unreliable. [00:13] It's not at all rare to get timeouts on it, particularly when doing multiple packages. [00:13] it would help a bit if binarie were put into the component their source is in by default [00:14] If you have complaints, tell us (or maybe just me)! [00:15] It's not deliberately painful, and a lot of things are easily fixed. [00:18] wgrant: When it fails, I retry a few times and if it's consistent, I complain. [00:24] ScottK: Right, we know about the timeouts. [00:24] It's the other issues that seem to also be causing real problems that I'm worried about. [00:24] Functionally, the lack of per binary overrides is the main thing I miss. [00:29] wgrant: Speaking of timeout issues, Error ID: OOPS-1675E1545 is not rare on https://launchpad.net/ubuntu/+source/kde4libs/+publishinghistory [00:29] https://lp-oops.canonical.com/oops.py/?oopsid=1675E1545 [00:29] Retry worked (as it often does). [01:09] wgrant: the problem is that if you call queue on a bunch of binary packages then it takes O(N^2) time [01:09] which is stupid [01:10] so you have to trade off the cost of repeatedly executing queue against the cost of having it override N times as many times as it needs to [01:11] bug 241129 [01:11] Launchpad bug 241129 in soyuz "'queue override' command scales at O(n^2) with the number of packages on the commandline (affects: 1) (dups: 1) (heat: 4)" [Medium,Triaged] https://launchpad.net/bugs/241129 [01:28] cjwatson: Hm, that's pretty awesome. [07:14] Good morning [10:33] slangasek: the server ISO is under size now, with hplip still in there, so we could revert the hplip seed change [10:34] amd64 is at 699 Mb though [10:34] I'd still want to try out at alpha3 how usable is print-server without hplip [11:51] finally! compiz installable again; now waiting for the ecore/edje/elementary stuff to become installable, then we can trigger some images [11:52] in fact, after this publisher run finishes we can trigger amd64/i386 images [12:12] alternates and desktops triggered for all flavours [12:26] first images posted to tracker [12:26] ara: ^ FYI [12:33] pitti, great, thanks! [12:46] pitti, do i get in your way if i fire off an armel build ? [13:01] smoke testing today's kubuntu daily live, seems in reasonable shape, installs fine [13:25] ogra: doesn't work yet [13:25] ogra: we still need some time to sort out the giant FTBFS mess of the new e* stack [13:26] ogra: but if you have an image without netbook-launcher, go ahead of course [13:26] * pitti binNEWs edje [13:27] ogra: FYI, in 1.5 hours I can retry "elementary", so around 18:00 CEST we hopefully have installable arm again [13:30] ara, Riddell, all: all alternate builds posted [13:30] ttx: ^ server, too [13:30] smoser: can you please add the UEC images to the tracker, too? [13:30] pitti: ack [13:32] ubuntu/kubuntu-netbook posted, too [13:32] Riddell: ^ FYI [13:33] pitti, i've never added anything to the tracker myself. [13:33] smoser: ah, who does that usually, for the URC images? [13:34] slangasek, will usually do the UEC images, ara has done them before though. slangasek has a script which takes something like http://uec-images.ubuntu.com/server/maverick/20100803.1/published-ec2-daily.txt as input [13:34] Riddell: oops, just saw the kubuntu-meta upload; I guess you want a rebuild for that? [13:37] pitti: note that server will need to be respinned to get the last eucalyptus in, should land eod today. [13:37] ttx: sure [13:38] ttx: I'm just happy that after all this wrangling with soyuz and cd space, we even have buildable CDs now, so I wanted to build them now before everything breaks all over again :) [13:38] but no problem to do respins [13:38] ttx: first smoketest appreciated, though [13:38] of course [13:39] pitti: kubuntu netbook? that doesn't exist any more [13:39] Riddell: hm, why do we still build them then? [13:39] I thought we didn't [13:39] Riddell: ok, I can take them off the tracker and remove them from the cronjb [13:40] they are commented out in the crontab [13:40] pitti: we're still waiting for the last KDE 4.5.0 packages to get published, I'd also like the packagekit update and of indeed the meta packages on there [13:41] Riddell: ok; can I ask you to trigger a respin once that's done? my afternoon is full of meetings, and I have a presentation to give at 1400 UTC, too [13:42] yes I can do that [13:43] Riddell: cheers [13:43] Riddell: ok, so I'll ignore the current kubuntu desktop build [13:43] i. e. not add it to tracker [13:48] libmail-dkim-perl depends on libdigest-sha-perl, which is in Universe. It has an approved MIR from some time ago (Bug #243306). I'm not sure why that doesn't show up on the problems page, but would someone please promote libdigest-sha-perl. [13:48] Launchpad bug 243306 in libdigest-sha-perl (Ubuntu) "MIR for libdigest-sha-perl (heat: 4)" [Undecided,Fix released] https://launchpad.net/bugs/243306 [14:01] pitti, sigh, asac promised to be careful :P [14:02] ogra: well, he had the entire stack synced, and it has very sloppy build deps [14:02] but we are almost through now, just two more hours :) [14:02] pitti, i know, i wouldnt have acked the sync right before A3 if i had known it needs the whole stack new [14:03] let's just hope that it'll actually work :) [14:03] we're far from being ready with the settings and i had planned to adjust them until A3 [14:03] oh, hmm [14:03] netbook-launcher-efl needs a rebuild, too [14:03] against the -06 ABI [14:03] it currently depends on *-05 [14:03] ogra: ^ mind uploading that? [14:03] i thought asac pulled a new upstream version in [14:04] https://edge.launchpad.net/ubuntu/+source/netbook-launcher-efl/+changelog [14:04] last upload is from April [14:04] afaik that was the whole reason for doing the syncs of the efl stack [14:04] maybe he's planning to [14:04] -> #u-desktop [14:04] thats what i understood [14:41] pitti: wubi is broken still. Drops straight into grub shell on first boot after installing on windows side [15:06] ttx: revert the seed change> your call :) === jjohansen is now known as jj-afk [15:07] slangasek: I'll keep it that way. We can revert for beta :) === bjf[afk] is now known as bjf [15:37] slangasek: If you have a moment, would you please promote libdigest-sha-perl? See channel scrollback from ~2 hours ago for details. [16:08] * Riddell builds Kubuntu CDs === jj-afk is now known as jjohansen [17:45] slangasek, when you get a minute, can you populate iso tracker with http://uec-images.ubuntu.com/server/maverick/20100803.1/published-ec2-daily.txt [18:12] * Riddell builds kubuntu CDs again [18:16] smoser: added please check I got that right http://iso.qa.ubuntu.com/qatracker/build/ubuntuserver/all [18:16] did you do it by hand ? [18:16] how else do you do it? [18:17] slangasek, has a script. and the serial numbers are supposed to be the ami ids :-( [18:17] slangasek, how about providing me with that script, so next time i ask someone else i can say "use this, its easier" [18:18] ScottK: libdigest-sha-perl moved to main [19:00] smoser: the script isn't done yet, it still has to be edited for each new milestone; I'm working on fixing that today [19:01] slangasek, thanks. i didn't mean to pester. only to save other people time and heartache [19:01] smoser: it's not pestering, this is something I've needed to get done for a couple of months now to eliminate myself as SPOF :) [19:01] new kubuntu CDs up, added to tracker [19:08] Can we have the xubuntu Desktop images added to the tracker? [19:22] Riddell: yippie [19:22] charlie-tca: looking [19:22] Thanks [19:22] charlie-tca: sorry, had an afternoon full of meetings, got a bit sidetracked from the release stuff [19:22] no problem [19:22] I ran some anyway [19:23] done [19:29] Thank you [19:37] is there such a thing as SRU freeze for 10.04.1? [19:38] Riddell: yes, there is [19:38] Riddell: we had to inflict it so that we can test CDs built from -proposed for a while [19:40] pitti: so that's still in place? [19:40] yes, it is [19:40] pitti: where was that announced? [19:40] Riddell: shoudl have been during the sprint [19:41] it was not announced === bjf is now known as bjf[food] [19:45] that seems rather a problem [19:45] Riddell, right, thanks to the lack of r-m I guess but that should be fixed rsn [19:53] smoser: AMIs posted; script finished; lp:~ubuntu-archive/ubuntu-archive-tools/trunk/post-amis-to-iso-tracker.py [19:53] slangasek, thanks. [19:53] pitti: ^^ [19:53] slangasek: thanks [19:53] regarding the amis... well, we're re-spinning :) [19:53] but the script is great. [19:53] pitti: sure! sorry it took so long [19:53] smoser: 'sok, really easy to post the new ones when they're ready ;) [19:53] slangasek: did you meet Kate yet? say hello from me [19:53] pitti: I did - you know Kate? [19:53] slangasek: not yet :) [19:53] heh :) [19:54] but let's say, I'm not at all unhappy about getting back a RM [19:55] right :) === bjf[food] is now known as bjf [21:16] good night everyone [21:16] no -efl tody, so no arm images === apachelogger is now known as gnomelogger === bjf is now known as bjf[afk] [23:40] Riddell: Thanks.