[15:05] <ogra_> cjwatson, i just filed bug 890261
[15:05] <ubot2> Launchpad bug 890261 in ubiquity "can not execute oem-config in a chrooted environment" [Medium,New] https://launchpad.net/bugs/890261
[15:05] <ogra_> FYI
[15:06] <cjwatson> tell me when it has debugging information :)
[15:06] <cjwatson> I honestly have real trouble believing that chrooting is the problem
[15:06] <cjwatson> in and of itself
[15:07] <ogra_> yeah, its likely the way i call it and dont specify a frontend
[15:07] <ogra_> fyi, linaor wants to use it exactly the wayx i described in the bug, that should solve a lot of differences with the linaro images :)
[15:30] <ogra_> cjwatson, btw, for proper execution we will need a switch for the final reboot call ... chrooted reboot here causes a host reboot for me
[15:30] <ogra_> i guess we dont want that on image builders :)
[15:32] <ogra_> (i.e. when it drops me into the rootshell after a failed attempt and i ctrl-d out of it)
[15:37] <cjwatson> probably a matter of moving the handling of ubiquity/reboot to somewhere the installer will see it
[15:37] <cjwatson> s/the installer/oem-config/
[15:38] <ogra_> yeah, well, its only happening if i get into the fallback shell anyway
[15:38] <ogra_> and i dont plan to actually run ubiquity chrooted, only oem-config
[15:38] <ogra_> so if all works fine it should never show up
[15:43] <CarlFK> bug 889656
[15:43] <ubot2> Launchpad bug 889656 in debian-installer "installer stops using proxy " [Undecided,New] https://launchpad.net/bugs/889656
[15:44] <CarlFK> is that the right package, and can I can help debug code?
[15:46] <cjwatson> CarlFK: what is in /target/etc/apt/apt.conf at the point of the failure?
[15:47] <CarlFK> cjwatson: ill have that for you in 20 min
[15:47] <cjwatson> also /etc/resolv.conf and /target/etc/resolv.conf
[15:51] <cjwatson> I suspect that one of the packages installed during the pkgsel phase has clobbered the configuration installed by apt-setup
[15:51] <cjwatson> tracking down which may be a nightmare
[15:55] <cjwatson> CarlFK: (BTW, don't bother carefully selecting snipped-out bits of syslog to put in the bug description; I always skip over that and go straight for the full log anyway)
[15:56] <CarlFK> cjwatson: yeah, I some how thought it would only be a screen full ... so much for my estimates
[16:17] <CarlFK> cjwatson: http://dpaste.de/h1Ess/
[16:18] <CarlFK> it hasn't thrown the error yet, but it seems to be stuck in the "Retrieving file 1 of 2" part
[16:18] <cjwatson> Do you recognise "pc8" as a proxy?
[16:18] <CarlFK> yes
[16:18] <CarlFK> that's also 192.168.0.1
[16:19] <cjwatson> It's not your preseeded one, but it's somehow ended up in /target/etc/apt/apt.cnf
[16:19] <cjwatson> .conf
[16:19] <cjwatson> (BTW please attach this to the bug, I'm doing other things and I definitely can't guarantee to finish this today)
[16:19] <CarlFK> ah, sorry, it is the preseeded one now
[16:20] <cjwatson> then I need a new matching syslog
[16:20] <CarlFK> understood. no prob
[16:20] <CarlFK> do you need my preseed files too?
[16:20] <cjwatson> never hurts
[16:20] <CarlFK> k
[16:20] <cjwatson> not seeing anything obvious here unfortunately
[16:21] <cjwatson> will probably need to attempt to set up a matching environment to debug it
[16:22] <CarlFK> I'll attach a tar that has all the configs and a script to build it
[16:23] <cjwatson> realistically, I'm on dev release maintenance this month ... maybe a good opportunity for somebody else to get into debugging the installer
[16:26] <CarlFK> how likely is this effecting debian?  I know a dd that might be up for it, but he might get grumpy if it's just ubuntu
[16:47] <cjwatson> hard to say
[16:47] <cjwatson> but if he's negative about Ubuntu then he might not be the best person to investigate something that starts as an Ubuntu bug :-)
[16:51] <CarlFK> he has meet you: Holger Levsen <holger at debian.org>  h01ger
[16:53] <CarlFK> He helped me with pycon, so had to use my ubuntu laptops, "I'm surrounded by Ubuntu... woa."
[16:54] <CarlFK> I think I'll bug him - might be amusing :)
[17:01] <cjwatson> I know Holger, yes
[17:01] <cjwatson> fixture of debconf :-)
[17:01] <cjwatson> (er, the conference, not the program)
[17:02] <cjwatson> CarlFK: maybe try reproducing it in Debian first?
[17:02] <CarlFK> that's worth a shot
[21:19] <CarlFK> cjwatson: looks like it is caused by one of my pkgsel/include packages: http://pastie.org/2863857
[21:19] <CarlFK> I pull that line, no problem.  put it back, proxy problem...
[21:21] <CarlFK> squid-deb-proxy-client  seems a likely clupret
[21:35] <CarlFK> http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/oneiric/squid-deb-proxy/oneiric/view/head:/debian/squid-deb-proxy.postinst
[21:35] <CarlFK> that gets run as part of the install, right?
[21:44] <superm1> cjwatson: oh i forgot to ask you at UDS, is there any liklihood that the switchover isolinux->grub2 is going to happen for precise, or is it still put off blocked on the same stuff as previous?
[21:46] <cjwatson> superm1: if it happens in the same cycle as I have a baby, it won't be by my efforts ...
[21:46] <cjwatson> (at this point I really don't object to somebody else attacking it; I seem to have failed so far)
[21:46] <superm1> cjwatson: ah I see :)  what were the major blockers that were needed for it?
[21:48] <cjwatson> there are several things that need to be done in gfxmenu to implement the look of gfxboot, and upstream's guidance on the subject has been that they'd like it to be done in a way that allows the same menu file to be usable without gfxmenu; i.e. it needs to degrade gracefully
[21:48] <cjwatson> the basic UI elements are written up as work items, but I never really got very far with them, unfortunately
[21:49] <cjwatson> the UEFI menu is proof that we *could* use grub if we didn't care about the look of the thing, so it's all about UI at this point
[21:49] <superm1> okay that's good to know
[21:49] <cjwatson> (well, not quite proof I suppose, but I'm pretty sure there are no blockers at that level anyway)
[21:50] <cjwatson> I would be happy to support and assist respins that wanted to use grub and didn't mind a basic UI
[21:50] <cjwatson> I don't know if that helps
[21:50] <superm1> well it would be for OEM stuff, but i was hoping to not jump ahead of distro
[21:50] <superm1> in case there are other parts that break unexpectedly
[21:52] <superm1> ideally i'd like to help with some of those work items, but i'm not sure how time commitment will look for me on the matter
[21:53] <superm1> er better worded, i'm not sure i'll have the time to commit seeing them all through
[21:53] <cjwatson> basically this cycle I'm getting this +1 maintenance team going, doing some performance work on germinate/launchpad so that we can run the publisher more frequently, possibly squeezing in a tool to generate a preseed file based on an install, and anything else is a bonus
[21:53] <cjwatson> because as of February my attention span is going to drop through the floor
[21:54] <superm1> right
[21:55] <superm1> Ok.  well if I do get the time to help out then, i'll look to you to help review what I come up with hopefully before you disappear for a while
[21:55] <cjwatson> right, I do feel bad about not getting this done, I don't like getting stuck on things
[21:56] <cjwatson> but that boils no potatoes
[22:07] <cjwatson> CarlFK: I can imagine that not helping matters, yes
[22:07] <cjwatson> yes, the postinst of anything you install will be run in /target during pkgsel
[22:08] <CarlFK> cjwatson: im still digging.. the URL i posted is for the squd server config, not the client