[09:33] <CIA-52> ubiquity: cjwatson * r3747 ubiquity/ (debian/changelog ubiquity/frontend/kde_ui.py):
[09:33] <CIA-52> ubiquity: * KDE frontend:
[09:33] <CIA-52> ubiquity:  - Don't add partition bar frames for devices we can't find for whatever
[09:33] <CIA-52> ubiquity:  reason (LP: #485450).
[09:52] <CIA-52> ubiquity: cjwatson * r3748 ubiquity/ (debian/changelog ubiquity/components/ubi-console-setup.py):
[09:52] <CIA-52> ubiquity: Pass model, layout, variant, and options to rewrite_xorg_conf
[09:52] <CIA-52> ubiquity: (LP: #514646).
[09:55] <CIA-52> pkgsel: cjwatson * r155 ubuntu/debian/ (changelog postinst):
[09:55] <CIA-52> pkgsel: Install zh-hans language packs for zh_CN, and zh-hant language packs for
[09:55] <CIA-52> pkgsel: zh_TW (LP: #517621).
[09:57] <CIA-52> pkgsel: cjwatson * r156 ubuntu/debian/changelog: releasing version 0.25ubuntu5
[09:58] <CIA-52> ubiquity: cjwatson * r3749 ubiquity/ (debian/changelog scripts/install.py):
[09:58] <CIA-52> ubiquity: Install zh-hans language packs for zh_CN, and zh-hant language packs for
[09:58] <CIA-52> ubiquity: zh_TW (LP: #517621).
[10:18] <CIA-52> ubiquity: cjwatson * r3750 ubiquity/ (debian/changelog ubiquity/components/ubi-console-setup.py): Fix crash on keyboard variant selection (LP: #512953).
[10:20] <dmarkey> is it possible to disable specific entries in tasksel
[10:33] <cjwatson> dmarkey: yes, preseed tasksel/skip-tasks
[10:33] <cjwatson> ev: don't suppose you could have a look at bug 507343?
[10:34] <dmarkey> cjwatson: genious, as always
[10:34] <cjwatson> there's a report from karmic as well
[10:34] <ev> cjwatson: will do
[10:34] <cjwatson> thanks
[10:43] <ev> hrm, the persistent debconf changes to casper appear to have the effect that debconf isn't writing to its database
[10:44] <ev> just a heads up, I'm looking into it
[10:45] <cjwatson> ugh
[11:02] <dmarkey> cjwatson: can one customise the "installation is finished" message
[11:03] <ogra> cjwatson, so in my log from last night you think the locking was the issue ? (i wouldnt know where that comes from, nothing uses debconf here apart from ubiquity)
[11:04] <cjwatson> dmarkey: no
[11:04] <ogra> and yes, i think libc6-vfp is gone since we have vfp by default
[11:04] <cjwatson> dmarkey: not without editing source and rebuilding anyway (but that should be obvious!)
[11:04] <cjwatson> ogra: it looked like it.  you may think you aren't using debconf but some packages are being reconfigured under the covers and they apparently use it.
[11:05] <ogra> when could that happen ? i mean i boot directly into ubiquity-dm ...
[11:06] <cjwatson> it's being done by the configure_hardware function in /usr/share/ubiquity/install.py
[11:06] <CIA-52> hw-detect: cjwatson * r138 ubuntu/ (debian/changelog hw-detect.sh):
[11:06] <CIA-52> hw-detect: Don't try to install libc6-vfp any more; it hasn't been built since
[11:06] <CIA-52> hw-detect: eglibc 2.10.1-0ubuntu10 in karmic.
[11:06] <cjwatson> which some people do want to happen in oem-config, so it isn't as simple as just dropping it
[11:06] <ogra> or rephrasing that, could it be ubiquity or could it be the bootstrapping
[11:06] <cjwatson> it is a bug in ubiquity
[11:06] <ogra> ok
[11:06] <cjwatson> or at least it appears to be
[11:07] <ogra> well, i run debootstrap and the subsequent apt-get under DEBIAN_FRONTEND=noninteractive ... so i was fearing it could be a leftover from the build run
[11:07] <cjwatson> nothing to do with thta
[11:07] <cjwatson> that
[11:08] <ogra> if you think its clearly ubiquity then i'm fine for the script side at least :)
[11:08] <cjwatson> the problem is that ubiquity uses debconf itself, but then tries to reconfigure some packages that also use debconf, and in oem-config mode they end up trying to use the same database
[11:08] <ogra> right
[11:08] <cjwatson> and it appears not to be passing down the necessary fds or whatever
[11:08] <cjwatson> this is just a top-level analysis, I haven't seriously looked into the bug
[11:08] <cjwatson> please do file it separately from your earlier one
[11:09] <ogra> will do, i guess you'D like syslog with it ?
[11:09] <ev> it'd be fantastic if initramfs-tools supported break=top,bottom.  Might have to hack that in at some point.
[11:09] <ogra> (the full one)
[11:10] <cjwatson> ogra: yes
[11:10] <cjwatson> ev: yes :)
[11:23] <ev> http://paste.ubuntu.com/373137/ - does that look reasonable?
[11:24] <cjwatson> LGTM
[11:26] <ev> cool, uploaded
[11:37] <ogra> ev, how does that work from a user perspective, i ctrl-d out of one breakpoint and it moves to the other ?
[11:38] <ev> ogra: indeed
[11:38] <ogra> sweet !
[11:39] <CIA-52> tzsetup: cjwatson * r513 ubuntu/ (debian/changelog tzsetup): Only set http_proxy if mirror/http/proxy is non-empty (LP: #519241).
[11:45] <CIA-52> tzsetup: cjwatson * r514 ubuntu/debian/changelog: releasing version 1:0.26ubuntu5
[12:25] <ogra> cjwatson, bug 429146 wrt libc6-vfp ... feel free to wipe the patch
[12:26] <ogra> (apparently it was gone in karmic already, sorry that nobody notified you)
[12:26] <cjwatson> ogra: already done, look at scrollback about 1h20m ago ;-)
[12:26] <cjwatson> suppose I could upload it
[12:26] <ogra> heh, missed that
[12:27] <ogra> i'd love to test the recent changes ... at least for the debconf frontend
[12:27] <ogra> oh, thats only in hw-detect ... not oem-config
[12:28] <CIA-52> hw-detect: cjwatson * r139 ubuntu/debian/changelog: releasing version 1.73ubuntu2
[12:38] <ara> cjwatson, ping
[12:48] <cjwatson> ara: hi
[12:49] <cjwatson> ev: https://bugzilla.gnome.org/show_bug.cgi?id=605137 making progress ...
[12:49] <ara> cjwatson, hello, I am working on setting up the testcases for iscsi root installations
[12:49] <ara> cjwatson, (ubuntu-server lucid up to date iso)
[12:50] <ara> and, although I can login to the target, it does not appear in the list of available disks
[12:50] <ev> ah, awesome
[12:51] <cjwatson> ara: odd, can I see logs?  what kind of target are you using?
[12:52] <cjwatson> ara: if a software target, do you have a Lun entry in the appropriate Target block in /etc/ietd.conf?
[12:52] <cjwatson> mine looks like this:
[12:52] <cjwatson> Target iqn.2009-07.org.ucam.vpn.pelham.wl.sarantium:test
[12:52] <cjwatson>         Lun 0 Path=/srv/iscsitest,Type=fileio
[12:52] <ara> cjwatson, yes, I have
[12:52] <ara> cjwatson, http://imagebin.org/84186
[12:53] <cjwatson> ok, that does look good
[12:54] <ara> but, it does not show the new disk
[12:54] <ara> the one that appears is the "physical" one
[12:56] <cjwatson> oh, you mean that fdisk shows the physical one?
[12:58] <ara> yes
[12:58] <ara> but not the iSCSI target
[13:03] <cjwatson> ara: can I see logs, please?
[13:04] <ara> sure, let me try to post them somewhere
[13:22] <ara> cjwatson, I found the typo in the ietd.conf file
[13:23] <ara> cjwatson, strange thing is that it does not report it when you start the daemon
[13:23] <ara> it is working now, thanks
[13:23] <cjwatson> yeah, it's all a bit delicate :-/
[13:23] <cjwatson> great
[13:35] <superm1> cjwatson, did you see whether preseeding was working during Jamie Bennett's merge (bug 518272)?
[13:41] <cjwatson> superm1: the core of it did work, but ev's in the process of fixing a bug whereby the debconf database doesn't get saved.
[13:41] <superm1> cjwatson, ah okay :)
[13:45] <ev> I suspect we're not poking the right environment variables.  strace indicates that it's just not writing to config.dat-new like it normally would.
[13:45] <ev> for debconf, that is
[13:58] <cjwatson> we're killing debconf-communicate with SIGTERM, which it may not consider to be a clean shutdown
[13:58] <cjwatson> it would be worth trying to close its file descriptors first, and then waiting for it to shut down by itself
[13:58] <cjwatson> indeed maybe just removing the kill would do it ...
[14:09] <ev> nope, it doesn't shut down
[14:09] <ev> SIGINT doesn't help either
[14:10] <cjwatson> that suggests to me that some other process still has an fd held open to it
[14:10] <cjwatson> look through /proc/*/fd/
[14:12] <cjwatson> if that's the case, then some judicious 3>&- 4<&- redirections ought to help
[14:28] <dmarkey> is there a way to get grub-pc to generate menu.lst
[14:28] <dmarkey> in tandem to grub.conf
[14:34] <cjwatson> no, menu.lst is exclusively for grub-legacy.
[14:34] <cjwatson> (and grub.conf is Fedora's name for menu.lst; I assume you actually mean grub.cfg)
[14:34] <dmarkey> yep
[15:38] <dmarkey> is it possible to get partman to start the 1st partition on LBA 128 instead of LBA 63
[15:38] <dmarkey> strange question i know
[16:00] <cjwatson> not really.  why?
[16:01] <cjwatson> if you're asking about this for the sake of SSD alignment, that's a somewhat bigger question than just the position of the first partition, and I'm hoping parted 2.1 will help
[16:43] <dmarkey> cjwatson: its for block alignment with xen, between the guests partition table and the backend storage device
[16:43] <dmarkey> probably applied for kvm too
[16:44] <dmarkey> applies*
[17:03] <cjwatson> dmarkey: right, alignment is really a bigger problem, I'd rather try to solve it in one block rather than cherry-picking bits around the corners
[17:04] <cjwatson> the way to get libparted to do this kind of thing is to apply what it calls "constraints"
[17:05] <dmarkey> cjwatson: can that be preseeded?
[17:05] <dmarkey> btw i'm working on 9.10 here
[17:05] <cjwatson> dmarkey: no, you're probably stuck on this, sorry
[17:06] <cjwatson> the installer just doesn't have this feature right now
[17:06] <dmarkey> cjwatson: how about a format the disk before the installer kicks in
[17:06] <dmarkey> in an early command or something
[17:07] <cjwatson> sure, you can hack around with that, but it's horrendously difficult
[17:07] <cjwatson> you have to duplicate chunks of partman and convince it not to run somehow
[17:07] <cjwatson> I'm afraid you're kind of on your own there
[17:08] <dmarkey> so, fdisk wont do what i want?
[17:09] <cjwatson> fdisk might
[17:09] <cjwatson> like I say, you'll have to work this out for yourself I'm afraid, you're well outside normal installer territory
[17:10] <cjwatson> in future I'd like to have the installer do this itself, but I don't think I am able to help with doing it on 9.10
[17:14] <dmarkey> cjwatson: but if i set up something early, the partitioner will probably overwrite it
[17:14] <cjwatson> hence
[17:14] <cjwatson> 17:07 <cjwatson> you have to duplicate chunks of partman and convince it not to run somehow
[17:15] <cjwatson> i.e. you may find you have to set up fstab for yourself and stuff like that
[17:15] <cjwatson> although I suppose you could try creating a small dummy partition at LBA 128, and then tell the partitioner to use the remaining free space
[17:15] <cjwatson> that might work, with care
[17:17] <dmarkey> thanks alot
[17:18] <cjwatson> let me know if it does, in case somebody else asks? :-)
[17:54] <fragadelic> cjwatson - hi, this is fragadelic the guy who made remastersys.  been getting some reports from folks that ubiquity in karmic is having some issues due to unresolved deps - are you guys in the process of updating ubiquity for karmic?
[18:07] <cjwatson> fragadelic: wasn't aware of that, do you have any detail?
[18:08] <cjwatson> we don't normally update ubiquity in point releases, but in any event I'd guess this would be a problem with individual packages, not with the installer itself?
[18:08] <cjwatson> given that it worked at release time :)
[18:08] <cjwatson> err - I meant to say "we don't normally update ubiquity in non-LTS releases" (since they don't get point releases, so no CDs etc.)
[18:09] <cjwatson> can make exceptions for really bad stuff
[18:10] <ogra> are the gtk/qt UIs of oem-config supposed to have a tasksel run as well or is that debconf_ui only ?
[18:13] <cjwatson> debconf_ui only
[18:13] <cjwatson> built-in tasksel was just for the server mode.  you can always run tasksel manually
[18:14] <ogra> indeed
[18:14] <ogra> i have a tasksel gui for the rootstok-gui app i was just wondering if i could leave that part completely to oem-config
[18:15] <ogra> but then i'll go with my current solution
[18:16] <cjwatson> you could try adding a UI page for it to ubiquity
[18:16] <cjwatson> wouldn't have to be used as a mandatory thing
[18:16] <cjwatson> there's probably some complex debconf plumbing involved, though
[18:16] <ogra> yeah
[18:17] <ogra> my ui just parses the desc file and hands over ^$task for each selected task to apt ...
[18:17] <ogra> its really primitive but does the job
[18:18] <ogra> and i think i need to do the run in advance actually, since i select the oem-config ui based on the fact if x/gtk/qt is installed
[18:19] <ogra> hmm, yeah, that would actually not work
[18:52] <tbrijeski> cjwatson - are you there
[21:00] <tbrijeski> cjwatson or anyone else - my question is about the version of ubiquity in karmic - is it changing or has it changed? some users of remastersys on karmic are reporting they can't install as ubiquity won't install due to deps. I have to go soon but will check the irc logs if anyone answers. thanks in advance
[21:32] <cjwatson> tbrijeski: you left *just* before I could answer
[21:33] <cjwatson> I already answered you under a different name
[21:33] <cjwatson> 18:07 <cjwatson> fragadelic: wasn't aware of that, do you have any detail?
[21:33] <cjwatson> 18:08 <cjwatson> we don't normally update ubiquity in point releases, but in any event I'd guess this would be a problem with individual packages, not with the installer itself?
[21:33] <cjwatson> 18:08 <cjwatson> given that it worked at release time :)
[21:33] <cjwatson> 18:08 <cjwatson> err - I meant to say "we don't normally update ubiquity in non-LTS releases" (since they don't get point releases, so no CDs etc.)
[21:33] <cjwatson> 18:09 <cjwatson> can make exceptions for really bad stuff
[21:34]  * cjwatson will try mail
[21:34] <ogra> heh