[00:41] <CarlFK>  subprocess.Popen(inst_composite, stdin=subprocess.PIPE)   is that where you want subprocess.stderr ?
[00:42] <CarlFK> subprocess.Popen( [...], stdout=subprocess.PIPE, stderr=subprocess.PIPE )
[00:43] <CarlFK> out,err = p.communicate()
[00:43] <cjwatson> we've dealt with it
[00:44] <cjwatson> and no, that's not the right fix
[00:44] <cjwatson> it's better to have stdout go straight to stderr (stdout=sys.stderr) rather than having to handle all the output explicitly in ubiquity
[00:46] <CarlFK> ah - I thought... never rmind :)
[01:28] <ev> *grumbles about having hardcoded an efi question despite having constructed the facilities to find the correct platform-specific ones*
[01:28] <ev> tomorrow though
[09:29] <StevenK> Whee, grub2 *and* GPT on this new machine, two things I have no clue about.
[09:30] <StevenK> I'm getting "This GPT partition lbale has no BIOS Boot Partition; embedding won't be possible"
[09:30] <StevenK> I suspect I have screwed something up, but I have no clue what.
[09:41] <cjwatson> StevenK: http://grub.enbug.org/BIOS_Boot_Partition
[09:41] <StevenK> cjwatson: Which, helpfully, is a 404. :-)
[09:50] <cjwatson> urgh
[09:51] <StevenK> cjwatson: Sorry.
[09:51] <cjwatson> unfortunately none of the current GRUB developers have admin access to the wiki
[09:51] <cjwatson> http://replay.waybackmachine.org/20090613031004/http://grub.enbug.org/BIOS_Boot_Partition is sort of OK
[09:52] <cjwatson> guess I need to write some proper documentation on that for the texinfo manual
[09:53] <StevenK> cjwatson: So I've created /dev/sda[12] on the disks, but given that page I set the flag on the disk itself or the partition?
[09:53] <cjwatson> oh, wait, I kind of did
[09:53] <cjwatson> the partition
[09:53] <cjwatson> http://www.gnu.org/software/grub/manual/grub.html#Installing-GRUB-using-grub_002dinstall
[09:53] <StevenK> cjwatson: Why didn't d-i do this for me?
[09:54] <cjwatson> http://www.gnu.org/software/grub/manual/grub.html#Images
[09:54] <cjwatson> it should have done.  if it didn't I want a bug with logs
[09:54] <StevenK> Bleh
[09:54] <StevenK> :-)
[09:54] <cjwatson> it definitely has code for it
[09:55] <StevenK> And parted /dev/sda1 bios_grub errors with "expecting a partition number"
[09:56] <cjwatson> oh, yeah, that old version of the page is wrong
[09:56] <cjwatson> http://www.gnu.org/software/grub/manual/grub.html#Images has the right rune
[09:56] <cjwatson> I should probably create a new section for that though
[09:57] <cjwatson> or http://webcache.googleusercontent.com/search?q=cache:http://grub.enbug.org/BIOS_Boot_Partition
[09:59] <StevenK> Heh
[09:59] <StevenK> Right, Grub installed, let's see if it boots
[10:00] <StevenK> ... and it reboots constantly. Awesome
[10:08] <cjwatson> I'm not going to be able to look right now, I'm afraid :-/
[10:08] <cjwatson> if you can rescue-disk it and at least extract logs, that would be great
[10:09] <cjwatson> please tell me this isn't a Mac
[10:25] <CIA-31> ubiquity: cjwatson * r4623 trunk/ (debian/changelog tests/test_ubi_partman.py): Fix test failure on non-x86 architectures.
[10:35] <CIA-31> ubiquity: cjwatson * r4624 trunk/debian/changelog: releasing version 2.5.32
[10:42] <StevenK> cjwatson: Heh, it's fine. Looking at the partition manager after a reboot, it's my fault. And no, it isn't a Mac.
[10:42] <cjwatson> ok, what was wrong?
[10:43] <StevenK> Partition 1 is the 100MB RAID1 /boot, which is helpfully marked as biosgrub, not raid
[10:43] <StevenK> So, uh, I fail :-)
[10:45] <cjwatson> ah, heh
[11:51] <CIA-31> ubiquity: cjwatson * r4625 trunk/ (bin/ubiquity-dm debian/changelog):
[11:51] <CIA-31> ubiquity: Shut down the debconf-communicator instance started in DM.run. This
[11:51] <CIA-31> ubiquity: should reduce the number of spurious crash reports we get.
[15:40] <aliguori> Hi, I'm trying to automate a guest installation in QEMU.  In an ideal world, it would work with just the standard ISO and no special network infrastructure
[15:40] <aliguori> I can get a guest booting based on the kernel/initrd in the ISO, I'm then trying to feed it a preseed
[15:41] <aliguori> I see two very strange things: 1) If I use a preseed/early_command to create a preseed file in the rootfs, and then use preseed/file to point to it, the installer still prompts for questions
[15:41] <aliguori> but curiously, if i pass a wrong file name, it throws an error, so it's doing somethign with it
[15:41] <cjwatson> can you extract a syslog from it?
[15:42] <aliguori> any other preseed action command doesn't do anything unless i have a preseed/file or preseed/url specified, so just adding "preseed/run" to the kernel command line, the run command never gets executed
[15:42] <aliguori> cjwatson, once interactive install kicks off, how can i get to a shell?
[15:42] <aliguori> (early_command does get run no matter if file or url is specified FWIW)
[15:43] <aliguori> fyi, the way i'm feeding in the preseed file is using early_command to read it from a the serial port
[15:44] <aliguori> ah, another VT
[15:44] <cjwatson> that's the easiest way yes
[15:47] <aliguori> cjwatson, http://pastebin.ubuntu.com/586894/
[16:09] <ev> cjwatson: mpt and I were sitting down to discuss redesigning the advanced partitioning page in ubiquity for O. I recalled you wrote this https://wiki.ubuntu.com/Ubiquity/SoftwareRaid , but I wonder if the focus going forward should be on btrfs instead (as it supports raid like configurations and and seemingly all of the use case for LVM)
[16:10] <ev> what are your thoughts?
[16:10] <cjwatson> btrfs is too early to commit to it
[16:10] <cjwatson> there'd be significant risk of it just plain not working
[16:11] <cjwatson> bear it in mind, but it shouldn't be the design focus at the moment
[16:17] <ev> cjwatson: sure, I didn't mean we should commit to implementing that part of the design now
[16:17] <ev> but that the visual design for anything beyond what we already support should be of functionality provided in btrfs
[16:18] <ev> I guess it's of no matter as the interface for creating logical volumes would look the same no matter which implementation you choose
[16:18] <cjwatson> I'm not even sure anything like that in btrfs is stable
[16:21] <ev> indeed, but given our history with respect to implementing any advanced partitioning options, I suspect all of btrfs will be rock solid by the time we get around to adding support for logical volumes
[16:21] <ev> but as mentioned, I was wrong
[16:21] <ev> as presumably a logical volume is a logical volume, regardless if it's LVM or btrfs underneath
[16:22] <ev> as far as the visual design is concerned
[16:22] <cjwatson> a successful design would hopefully make it look that way for logical volumes, yes
[16:22] <cjwatson> I'm less sure you can do that for RAID
[16:23] <ev> mirroring and striping are terms in the btrfs vocabulary as well
[16:23] <ev> though as you suggest, they're still very much in development
[16:26] <ev> cjwatson: at any rate, are there specific things you'd like us to address in the redesign?
[16:26] <seb128> hi
[16:27] <seb128> bug #727726 is still an issue, should it be assigned to someone?
[16:27] <ubot2> Launchpad bug 727726 in ubiquity "ubiquity panel is about 4px instead of 30 on install" [Low,Confirmed] https://launchpad.net/bugs/727726
[16:27] <cjwatson> ev: as long as it's vaguely implementable in terms of partman-{lvm,raid} ...
[16:27] <seb128> it means you don't see the icons in there by default if you don't know they are there
[16:27] <ev> cjwatson: but of course
[16:28] <ev> seb128: huh, I wonder what we're doing wrong there.
[16:28] <ev> I'll take it as it's my poor copy and paste job that got us here
[16:29] <ev> assigned
[16:29] <seb128> ev, thanks
[16:29] <cjwatson> ev: I think pretty much all my thoughts on it should be in the old spec
[16:30] <seb128> ev, btw I still get http://people.canonical.com/~seb128/ubi.png as well
[16:30] <ev> cjwatson: https://wiki.ubuntu.com/Ubiquity/AdvancedPartitionerRewrite ?
[16:30] <seb128> is that a bug or not?
[16:30] <seb128> ie used to have text on the right iirc?
[16:30] <ev> seb128: only if you have an internet connection (the release notes link)
[16:31] <cjwatson> ev: yeah, sounds right - syncing an image so my network is very very slow
[16:31] <seb128> ev, ok, that's why then, so notabug, thanks ;-)
[16:31] <cjwatson> ev: looks like I meant https://wiki.ubuntu.com/Ubiquity/SoftwareRaid, that's more current
[16:31] <ev> sure thing
[16:32] <ev> okay
[16:32] <ev> mpt: ^
[16:33] <mpt> ok
[16:33] <ev> huh, why can't I target this bug for natty I wonder https://bugs.launchpad.net/ubuntu/natty/+source/ubiquity/+bug/727726/+nominate
[16:33] <ubot2> Launchpad bug 727726 in ubiquity "ubiquity panel is about 4px instead of 30 on install" [Low,Confirmed]
[16:37] <cjwatson> 'cos it's already targeted for natty
[16:44] <ev> oh
[16:44] <ev> heh
[17:42] <seb128> ev, cjwatson: is the keyboard selector dialog in ubiquity supposed to have a layout selected on the right column? or does "no selection" means "default for the selected country"?
[17:46] <cjwatson> it is supposed to have a variant selected on the right; I think there's a buglet there
[17:49] <seb128> cjwatson, ok, it doesn't in french, I will check for open bugs and file one if there is none
[17:49] <cjwatson> I mean "it ought to have, but it currently does not" - English too
[17:49] <seb128> cjwatson, btw I confirm that the odd xkb dialog being shown issue is fixed
[17:50] <cjwatson> good
[17:50] <seb128> cjwatson, do you need a bug for the right list selection thing?
[17:50] <cjwatson> yes please
[17:50] <seb128> or is it known as "being worked, no need of a bug"
[17:50] <seb128> ok
[18:49] <seb128> cjwatson, bug #745137
[18:49] <ubot2> Launchpad bug 745137 in ubiquity "no layout variant displayed as selected in the right side of the screen" [Undecided,New] https://launchpad.net/bugs/745137
[22:41] <juzzy_> hi
[22:41] <juzzy_> i recently installed ubuntu, and encrypted my drive
[22:41] <juzzy_> but how can i find my encryption key via the shell again?
[22:41] <juzzy_> it told me the command, but now i forget