[04:26] <maxagaz> hi
[04:26] <maxagaz> In the preseed, this: "d-i partman-auto/choose_recipe select atomic" mean that it will create a swap and a root partition ?
[04:33] <maxagaz> where can i find a manual for the preseed ?
[05:09] <persia> maxagaz, There's an appendix in the installation guide that covers preseeding.
[10:30] <davmor2> Kubuntu is still using grey for the install system :(
[11:30]  * NCommander whacks the installer
[12:01] <CIA-3> tasksel: cjwatson * r1354 hardy-proposed/debian/changelog: releasing version 2.70ubuntu6
[12:05] <CIA-3> tasksel: cjwatson * r1388 intrepid-proposed/debian/changelog: releasing version 2.73ubuntu11.1
[12:31] <CIA-3> wubi: evand * r106 trunk/ (debian/changelog src/wubi/backends/common/backend.py): Removed an erroneous closing parentheses in the use_cd check.
[12:35] <CIA-3> wubi: evand * r107 trunk/src/wubi/backends/common/backend.py: Whoops, fix up that last commit.
[12:37] <davmor2> evand: Kubuntu oem isn't using the tz_map
[12:37] <davmor2> it's using the old one
[12:38] <evand> ah, indeed.  shtylman, do you have any interest or time to look into that?  oem-config is a very similar codebase to ubiquity.
[12:40] <davmor2> evand: is there a bug for it or should I write one?
[12:41] <evand> davmor2: please make one
[12:41] <davmor2> np's
[12:45] <davmor2> evand: oem-config-kde is that the right package?
[12:45] <evand> yes
[12:49] <davmor2> evand: bug 348430
[12:53] <CIA-3> partman-base: cjwatson * r144 ubuntu/ (debian/changelog init.d/parted): Handle filtering of mounted partitions on /dev/mmcblk* (LP: #348411).
[13:01] <CIA-3> grub-installer: cjwatson * r776 ubuntu/ (debian/changelog grub-installer): Handle /dev/mmcblk* in split_device (LP: #348411).
[13:05] <evand> thanks
[13:47] <CIA-3> ubiquity: evand * r3120 ubiquity/ (4 files in 4 dirs):
[13:47] <CIA-3> ubiquity: * Work in bytes rather than percentages in the partition bar code.
[13:47] <CIA-3> ubiquity: * Fix partition bar slider (for resizing) code by adding correct
[13:47] <CIA-3> ubiquity:  calcuations for bounds checking and slider positioning.
[13:48] <evand> finally.  Now to figure out offscreen rendering so I can paint a gtk slider handle instead of a black box.
[13:50] <CIA-3> ubiquity: evand * r3121 ubiquity/ (debian/changelog ubiquity/misc.py): Properly encode data from os-prober (LP: #345573).
[13:51] <CIA-3> ubiquity: evand * r3122 ubiquity/debian/changelog: Add a LP bug reference.
[14:00] <cody-somerville> cjwatson, \o_
[14:00] <cody-somerville> cjwatson, Did Ubuntu patch d-i to fix http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=498143 or something? :)
[14:00] <cjwatson> cody-somerville: yes, but only if you pass cdrom-detect/try-usb on the command line
[14:00] <cjwatson> err
[14:00] <cjwatson> cdrom-detect/try-usb=true
[14:00] <cjwatson> reason being that otherwise it can do the wrong thing with USB hard disks
[14:01] <cody-somerville> does the usb-creator automatically set that?
[14:01] <evand> yes
[14:03] <cjwatson> wow, yeah, Kubuntu's partition bar colour choice is a bit confusing
[14:04] <davmor2> cjwatson: it does make it look like your emptying the drive :)
[14:05] <evand> indeed
[14:06] <cjwatson> is there a bug for it already?
[14:07] <evand> should be
[14:07]  * evand digs
[14:08] <davmor2> cjwatson: I added it to abug that looked the same from mpt but I'm now not sure it is
[14:08] <davmor2> https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/336755
[14:08] <davmor2> it is however very similar in look etc
[14:09] <cjwatson> davmor2: no, it's a different bug
[14:10] <davmor2> cjwatson: I can write a new one although I that was one I was linked to
[14:11] <cjwatson> well, the test is whether fixing one thing would fix the other
[14:11] <cjwatson> in this case that isn't true as far as I can see
[14:11] <cjwatson> so tagging along is a good way for one or the other bug to get lost
[14:11] <davmor2> cjwatson: I'll write a new one if fixing one fixes the other you can always dupe them :)
[14:12] <cody-somerville> evand, did you see my comment on the usb-creator bug I made yesterday?
[14:12] <davmor2> infact evand that might of been the bug you linked me too
[14:14] <evand> cody-somerville: looking now.  Sorry, was caught up in other work this morning.
[14:14] <evand> Hrm, I thought I fixed this previously
[14:14]  * evand tests
[14:15] <evand> arrr, must have broken things with a recent commit
[14:19] <davmor2> cjwatson: bug 348461
[14:19] <davmor2> I've swapped them over on the tracker too
[14:20] <cjwatson> ok, thanks
[14:24] <evand> hrm, actually I cannot reproduce your bug, cody-somerville.  I forgot that mkfs.vfat wont tell the kernel to re-read the partition table, but after calling blockdev --rereadpt, usb-creator tells me that the disk needs to be formatted when I have a single vfat partition spanning the entire disk
[14:24] <evand> that is the set up you had, right?
[14:25] <cody-somerville> evand, I'm not sure.
[14:27] <evand> hrm
[14:38] <shtylman> evand: will look at, np
[14:38] <evand> shtylman: very much appreciated
[14:54] <shtylman> evand: do I need the dvd to do an open install? or can I do it from the live cd?
[14:55] <evand> open install?
[14:55] <shtylman> oem*
[14:55] <shtylman> :)
[14:55] <shtylman> (its still early for me)
[14:55] <davmor2> shtylman: just hit F$
[14:55] <davmor2> F4
[14:55] <davmor2> even
[14:55] <davmor2> oem is listed there
[14:56] <shtylman> k
[14:56] <davmor2> shtylman: so pick a language hit F4 select oem from the list and then the install is pretty much as normal-ish
[14:57] <shtylman> davmor2: doing that now...and then once it is installed, you said the old tz map shows up, instead of new?
[14:58] <davmor2> shtylman: yes.  Basically you run through the "normal" install as oem.  You hit the setup for enduser and reboot.  Then you get 4 pages of end user setup and it's the map in there that is wrong.
[14:59] <evand> shtylman: you can install oem-config from the desktop, and just run it there
[15:00] <evand> mind you, the live CD desktop
[15:00] <evand> not your own
[15:24] <shtylman> evand: does it have to be in a live cd? I have a scrap install and I installed oem-config there, tried to run it...segfaulted
[15:38] <evand> erm, no, it should work fine
[15:38] <shtylman> well, now it doesn't segfault, but also doesn't do anything :)
[15:39] <shtylman> I just run oem-config, right? ... not missing anything there am I?
[15:39] <shtylman> evand: ^
[15:40] <evand> make sure you have oem-config and oem-config-kde installed.
[15:41] <evand> what does the log say?
[15:41] <shtylman> was missing oem-config-kde
[15:44] <shtylman> well...still doesn't do anything :/ does it need certain parameters?
[15:44] <shtylman> tried kde_ui like ubiquity...but nothin
[15:45] <evand> what does the log say?
[15:45] <evand>  /var/log/oem-config.log
[15:45] <shtylman> the installer log?
[15:45] <shtylman> k
[15:45] <shtylman> nothing
[15:45] <shtylman> blank
[15:47] <evand> hrm
[15:47] <evand> sudo oem-config kde_ui ?
[15:48] <shtylman> nope :)
[15:48] <shtylman> whats interesting is I can't remove the oem-config package now...
[15:48] <shtylman> dpkg fails
[15:49] <shtylman> (probly unrelated...but I was just gonna try a clean install)
[15:56] <shtylman> trying a clean install now
[16:03] <evand> sorry, in a meeting.  I would see if you have better luck with kvm and a live CD
[16:03] <cjwatson> shtylman: this isn't really how you run oem-config
[16:03] <cjwatson> shtylman: you run oem-config-prepare after rebooting into the installed system
[16:03] <cjwatson> shtylman: then you reboot again
[16:04] <cody-somerville> Does partman support creating sa quashfs + ext3 partition w/unionfs hybrid type thing?
[16:04] <shtylman> cjwatson: ok, thanks...will try that
[16:06] <cjwatson> cody-somerville: not at present
[16:06] <cody-somerville> cjwatson, what would be the sanest way to accomplish that atm?
[16:06] <cjwatson> can I get back to you after this meeting?
[16:07] <cody-somerville> sure
[16:10] <shtylman> evand: ok...I finally got it running :) ... so want me to just migrate my map over to the oem-config package...I assume?
[16:10] <evand> shtylman: please do :)
[16:10] <shtylman> k
[16:17] <CIA-3> ubiquity: evand * r3123 ubiquity/debian/changelog: Add a LP bug reference.
[16:52] <cody-somerville> evand, whats your launchpad id?
[16:53] <evand> ~evand
[16:59] <CIA-3> ubiquity: cjwatson * r3124 ubiquity/ (d-i/manifest debian/changelog debian/rules): merge from lp:~ubuntu-installer/ubiquity/jaunty-beta
[17:05] <cody-somerville> cjwatson, are you available now?
[17:06] <evand> 1.12.1, no?  I would think that the next time a release happens from trunk we'll be in karmic.  Or am I forgetting how we normally do this.
[17:14] <cjwatson> cody-somerville: call, sorry :-/
[17:14] <cjwatson> evand: whoops
[17:14] <cjwatson> evand: no, we normally keep going on trunk after beta
[17:15] <cjwatson> evand: and normally, once we hit beta, we bump to an even minor number for release
[17:15] <cjwatson> so strictly beta should have been 1.12.0 :-)
[17:17] <CIA-3> ubiquity: cjwatson * r3125 ubiquity/debian/changelog: fix version for post-beta
[17:18] <cjwatson> cody-somerville: (still on call, but can manage a small bit of attention) so this would come out to two partitions?
[17:18] <cjwatson> one for the squashfs, one for ext3?
[17:18] <cjwatson> physically, I mean
[17:18] <cody-somerville> MY understanding is that it would be a squashfs file + ext3 partition w/ unionfs
[17:19] <cjwatson> evand: I don't mind if you want to change this, but I was sort of expecting to work as before - I wonder though how unstable trunk really is between beta and final - I would expect "not much"
[17:19] <cjwatson> cody-somerville: squashfs file stored where?
[17:19] <cody-somerville> smagoun, Where will the squashfs file be stored?
[17:20] <smagoun> cody-somerville: Current setup for squashfs-based images looks like this: 2 ext3 partitions. On the first partition we have /boot and a squashfs image. The second partition is a data partition. We use unionfs to merge the squashfs + data together
[17:24] <cjwatson> right, so anything generic will be a pain in the arse, but you could do it as a custom hack
[17:24] <cjwatson> do you care about manual setup, or just automatic?
[17:25] <smagoun> manual setup of disk partitioning? (I came to the conversation late - sorry)
[17:25] <cjwatson> 16:04 <cody-somerville> Does partman support creating sa quashfs + ext3 partition w/unionfs hybrid type thing?
[17:25] <cjwatson> 16:06 <cjwatson> cody-somerville: not at present
[17:25] <cjwatson> 16:06 <cody-somerville> cjwatson, what would be the sanest way to accomplish that atm?
[17:25] <cjwatson> partman being the component of d-i/ubiquity that handles partitioning
[17:26] <smagoun> Manual partitioning is interesting in the general case; for the squashfs-specific case we can do without it if that makes things easier
[17:27] <cjwatson> so I think the easiest approach would be to make this an attribute of ext3 partitions in partman
[17:27] <cjwatson> i.e. an ext3 partition can have "squashfs-ness" or "cow-ness"
[17:28] <cjwatson> where does the installer get the squashfs from?
[17:30] <smagoun> Our current build tool spits out a squashfs image by default (our USB + ISO-based installers work from a squashfs image; at install time we either copy the image file directly to the target or unpack the squashfs and copy the contents to the target, depending on whether the target is squashfs+ext3 or straight ext3)
[17:32] <cjwatson> so it's somewhere on the installation image, with a known path
[17:33] <smagoun> correct
[17:38] <cjwatson> cody-somerville,smagoun: so I'd be thinking of something like this:
[17:38] <cjwatson> http://paste.ubuntu.com/137635/
[17:38] <cjwatson> that's in partman-target
[17:38] <cjwatson> obviously finish.d/copy_squashfs needs to be executable
[17:38] <lool> cjwatson, evand: I didn't merge the change by persia to have flash-kernel-installer only on armel and in the bootloader section; I prefer if you review it as well
[17:39] <cjwatson> cody-somerville,smagoun: that doesn't add any UI, but you can do something like squashfs_source{ /path/to/squashfs } squashfs_target{ /path/where/you/want/it/copied } cow{ any data you need for that } in a partman recipe
[17:39] <cjwatson> get the idea?
[17:40] <smagoun> cjwatson: alright, thanks! I think that makes sense.
[17:40] <cjwatson> lool: which change is that?
[17:41] <lool> https://code.launchpad.net/~persia/ubuntu-seeds/platform.jaunty+flash-installer-
[17:41] <lool> is-an-armel-boot-loader-installer/+merge/4875
[17:41] <lool> arg, /me slaps mutt
[17:41] <lool> https://code.launchpad.net/~persia/ubuntu-seeds/platform.jaunty+flash-installer-is-an-armel-boot-loader-installer/+merge/4875
[17:42] <cjwatson> oh, a seed branch
[17:42] <lool> right; I'm sorry I'm half sick and probably not being very clear
[17:42] <persia> I asked lool to review because I was modifying his last change to the seed.
[17:42] <cjwatson> lool: that's fine and can be merged. do you want me to do it?
[17:42] <lool> cjwatson: i have it merged here and will just push
[17:43] <cjwatson> go ahead then
[17:43] <cjwatson> p.s. 'bzr di | vipe | ubuntu-paste' really handy
[17:44] <cjwatson> cody-somerville,smagoun: I'm going to throw this away from my local tree, and paste.ubuntu.com probably expires after a while, so do you have a copy?
[17:44] <lool> Err StevenK pushed it already
[17:44] <smagoun> cjwatson: yes - thanks
[17:57]  * ogra waves
[17:58] <BlackLukes> is there anyone who's active in ubiquity development?
[17:58] <BlackLukes> yesterday I was asking about what code is used to display the partition bar in ubiquity as seen here: http://www.askdavetaylor.com/2-blog-pics/ubuntu-install-pic6.png
[17:59] <robbiew> BlackLukes: evand should be able to help
[18:00] <BlackLukes> thanks
[18:02] <cjwatson> BlackLukes: you'll find it easier to get an answer if you don't drop off IRC all the time - I tried to answer you a couple of times yesterday but you'd vanished
[18:02] <BlackLukes> sorry I had some connection problems later
[18:03] <cjwatson> anyway, the code is in ubiquity/segmented_bar.py and called from ubiquity/frontend/gtk_ui.py
[18:03] <ogra> so i'm setting ubiquity/configure_bootloader=false on cmdline ... i see it fine in /proc/cmdline, but echo get ubiquity/install_bootloader | sudo debconf-communicate disagrees about the setting
[18:04] <BlackLukes> I found it, thanks!
[18:04] <cjwatson> ogra: well, if you're setting ubiquity/configure_bootloader=false on the command line, you can hardly expect ubiquity/install_bootloader to be set :-)
[18:05] <ogra> well "echo get ubiquity/install_bootloader | sudo debconf-communicate" returns 0 true
[18:05] <ogra> cat /proc/cmdline tells me ubiquity/configure_bootloader=false
[18:05] <ogra> so it gets lost somewhere
[18:06] <ogra> could it be because i dont define a preseed file on cmdline ?
[18:06] <ogra> i would have assumed that ubiquity (or debconf) picks it up anyway from cmdline
[18:06] <cjwatson> ogra: casper is what's responsible for copying command-line parameters into debconf
[18:07] <cjwatson> ogra: see scripts/casper-bottom/24preseed
[18:07] <ogra> hmm, casper is definately used, else i wouldnt have a live session
[18:07] <cjwatson>                 */*=*)
[18:07] <cjwatson>                         question="${x%%=*}"
[18:07] <cjwatson>                         value="${x#*=}"
[18:07] <cjwatson>                         casper-preseed /root "$question" "$value"
[18:07] <cjwatson>                         ;;
[18:08] <cjwatson> ogra: um, so you keep saying you're putting ubiquity/configure_bootloader=false on the command line
[18:08] <shtylman> evand: have you updated the oem-config code to have the raw_utc_offset in the tz.py file?
[18:08] <cjwatson> ogra: that needs to be ubiquity/install_bootloader=false
[18:08] <shtylman> evand: I need that for the timezone map
[18:09] <ogra> cjwatson, err, sorry, typo
[18:09] <ogra> indeed its ubiquity/install_bootloader=false
[18:09] <ogra> no trace of the word ubiquity in casper.log
[18:12] <ogra> i see "Loading preseed file..." in casper.log
[18:13] <ogra> so the script seems to be executed
[18:16] <cjwatson> ogra: I think if I were you I'd use break=top and stick a set -x on the second line of 24preseed
[18:16] <ogra> hmm
[18:16] <ogra> k
[18:19] <cody-somerville> cjwatson, how does the Ubuntu cds boot into the install only mode?
[18:19] <cody-somerville> (for the desktop CDs that is)
[18:20] <ogra> only-ubiquity iirc
[18:20] <ogra> or flipped ... ubiquity-only .... one of them shoudl work
[18:20] <cody-somerville> yea, but how does it work exactly?
[18:21] <superm1> its put in /proc/cmdline and keyed off in the ubiquity init script
[18:21] <superm1> which starts ubiquity-dm
[18:21] <ogra> take a look at casper and look for the string
[18:21] <cody-somerville> ok
[18:24] <ogra> cjwatson, hmm, so i canged my cmdline ... from "console=ttymxc0,115200 console=tty boot=casper LIVEMEDIA=/dev/mmcblk0p1 ubiquity/install_bootloader=false" to "console=ttymxc0,115200 console=tty file=/cdrom/preseed/ubuntu.seed ubiquity/install_bootloader=false boot=casper LIVEMEDIA=/dev/mmcblk0p1 --" and guess what ... debconf has the setting
[18:24] <ogra> i dont get why ... but it seems to work now
[18:26] <CIA-3> partman-target: cjwatson * r755 ubuntu/ (3 files in 2 dirs):
[18:26] <CIA-3> partman-target: Introduce partman/mount_style (choices: traditional, label, uuid) to
[18:26] <CIA-3> partman-target: allow controlling how filesystems are mounted. Default this to uuid, and
[18:26] <CIA-3> partman-target: stop using labels by default since they have unavoidable problems with
[18:26] <CIA-3> partman-target: removable disks (LP: #347817).
[18:27]  * ogra fires up another install test
[18:27] <persia> Oh good.  It's always nice when you have two problems, and you fix one, and the other goes away.
[18:27] <ogra> yeah
[18:27] <ogra> though its totally beyond me why it works now
[18:27] <cjwatson> ogra: maybe a command line length limit?
[18:27] <cjwatson> seems a bit short, but
[18:28] <ogra> err, i made it a lot longer now
[18:28] <persia> cjwatson, A minimum limit?  Making it longer made it work.
[18:28] <cjwatson> oh, hmm
[18:28] <persia> I wonder if it is related to the final "--"
[18:28] <cjwatson> I doubt it
[18:28] <ogra> and /proc/cmdline has all it should
[18:28] <cjwatson> the purpose of -- is that everything after it is copied to the installed system
[18:29] <ogra> yeah, and the casper script doesnt seem to make any distinction here
[18:29] <cjwatson> I'm sure a set -x trace would make it clear
[18:29] <ogra> it just processes the words
[18:29] <ogra> heh
[18:30] <ogra> it did match LIVEMEDIA=/dev/mmcblk0p1 .... now i've put the preseed options before that
[18:30] <cjwatson> none of the cases in that script would match LIVEMEDIA=/dev/mmcblk0p1
[18:30] <ogra> would be my guess ...
[18:30] <cjwatson> stop guessing
[18:30] <cjwatson> get data
[18:30] <ogra> well, i'll test it later
[18:31] <ogra> more important is that it is set now, i need my install to run through
[18:31] <ogra> but wont forget to investigate ... for now i need to find the proper cmdline for my beta image first
[18:32] <ogra> sad that ubiquity still doesnt fully fit in 600px - 2 panels .... i have about 1px to reach the buttons
[18:33] <ogra> but its a lot better than before already
[18:37] <persia> ogra, It fit for me, as long as I ran it from the menu.  If I run it manually, it sizes differently.
[18:37] <ogra> ah
[18:37] <ogra> i never ran it from the menu in my life :)
[18:38] <ogra> there is that convenient desktop icon :)
[18:38] <persia> That ought be the same as running it from the menu.
[18:39] <persia> It's calling it from the command line that I consider manual (but do often anyway for debug purposes).
[18:39] <ogra> well, the buttons hide behind the bottom panel and i see exactly a 1px line of each button
[18:39] <charlie-tca> The panels are different sizes from the menu and from the desktop install icon
[18:39] <ogra> ??
[18:39] <ogra> the panels are 24px everywhere
[18:40] <ogra> in a live session at lest
[18:40] <ogra> *least
[18:40] <charlie-tca> Maybe the screen is different sizes?
[18:40] <ogra> 800x600
[18:40] <charlie-tca> I get all the buttons from the desktop, but can not see back or forward from the menu
[18:43] <persia> charlie-tca, Do you have any idea about https://bugs.launchpad.net/ubuntu/+source/xfwm4/+bug/348585
[18:44] <persia> I presume that's supposed to be a Xubuntu bug, but I hope it's not the common case.
[18:45] <charlie-tca> Possibly duplicate of bug 325958
[18:46] <charlie-tca> Which I don't think is large fonts, but just resolution issues that I don't know how to fix
[18:47] <charlie-tca> I'm gonna guess he installed the live cd to usb instead of the alternate cd
[18:48] <persia> Thanks.  I'll point the reporter at 325958.
[18:48] <CIA-3> oem-config: cjwatson * r638 trunk/ (3 files in 2 dirs):
[18:48] <CIA-3> oem-config: Set passwd/auto-login=true rather than implementing auto-login ourselves
[18:48] <CIA-3> oem-config: (LP: #347900).
[19:42] <shtylman> evand: ~shtylman/oem-config/timezone <- updated to new timezone map
[19:50] <shtylman> cjwatson: how come ubuquity doesn't build oem-config? seems like there is alot of overlap?
[20:01] <maco> hi, just wanted to say you guys are doing an awesome job. one of my friends just told me that her 5 year old daughter installed ubuntu by asking her 7 year old brother to read the words to her (she cant read yet), in order to prove the neighbor wrong regarding "linux is too hard to install"
[20:02] <maco> (not devel related but figured youd appreciate hearing it)
[20:08] <ogra> GRRR, ubioquity still dies in the bootloader step
[20:09] <ogra> so obviously something unsets the preseed value during install
[20:10] <ogra> cjwatson, at the end of the ubiquity run "echo get ubiquity/install_bootloader | sudo debconf-communicate" returns true again ... while it returned false when i started
[20:11] <shtylman> maco: :)
[20:56] <cjwatson> shtylman: yes, there is quite a lot of overlap and we've been meaning to merge them for ages now
[20:56] <cjwatson> oem-config actually predates ubiquity and was in some respects a prototype for some of the things ubiquity does
[20:57] <cjwatson> maco: cool!
[21:02] <shtylman> cjwatson: is oem config originally from debian?
[21:02] <cjwatson> no, I wrote it
[21:02] <cjwatson> check the changelog :)
[21:03] <cjwatson> though Jonathan Riddell and Anirudh Ramesh wrote the KDE frontend, I think
[21:07] <shtylman> cjwatson: :), gotcha
[21:56] <shtylman> cjwatson: question, commonly, when you make a branch and lets say your changes get merged in, do you delete that branch? and branch again if you do more work? or do you keep the branch around and keep merging in from master?
[22:01] <cjwatson> shtylman: you can do either, it depends on the purpose of the branch
[22:02] <cjwatson> shtylman: if it's a specific branch for a particular topic, you might well throw it away once it's been merged; if it's a general "my main branch of this package" branch, you'd probably keep it
[22:03] <shtylman> I see, what about merging in changes? is bzr merge or pull preferred?
[22:06] <cjwatson> doesn't really matter, but the only way you'll be able to use pull is with --overwrite and that throws away the old branch history
[22:07] <cjwatson> bzr will handle either, just depends what you want the result to look like :-)
[22:07] <shtylman> gotcha
[22:12] <cody-somerville> cjwatson, Where did you say ubiquity looks for the preseed file?
[22:15] <cjwatson> cody-somerville: ubiquity doesn't :-) casper handles preseeding
[22:15] <cody-somerville> ah
[22:15] <cjwatson> cody-somerville: and it looks wherever it's told to look, with the file= or url= boot parameter
[22:15]  * cody-somerville nods.
[22:16] <cjwatson> oh, it'll also load /preseed.cfg from the initrd if it exists
[22:23] <cody-somerville> Would this work? http://pastebin.ubuntu.com/137811/
[22:24] <cjwatson> well, some of it's made up
[22:24] <cjwatson> 11 d-i     apt-setup/uri_type      select d-i
[22:25] <cjwatson> dunno where that comes from but it doesn't exist
[22:25] <cjwatson> 35 d-i     partman-auto/disk       string /dev/hda
[22:25] <cjwatson> you're sure it'll be hda not sda?
[22:25] <cjwatson> if there's only one disk on the system, it's better to leave partman-auto/disk unpreseeded and then partman will pick the first one
[22:26] <cjwatson> much of netcfg is not used in ubiquity - only netcfg/get_hostname, netcfg/get_domain, netcfg/dhcp_ntp_servers
[22:27] <cjwatson> at the moment anyway
[22:27] <cjwatson> 51 d-i     console-tools/archs string skip-config
[22:27] <cjwatson> 53 d-i     console-keymaps-at/keymap select us
[22:27] <cjwatson> ah, I get it, you're reading from Debian preseeding documentation. Don't do that, use Ubuntu documentation instead. They differ in ways that will confuse you later.
[22:27] <cjwatson> 52 d-i     debian-installer/locale string en_US
[22:27] <cjwatson> 55 d-i     languagechooser/language-name-fb    select English
[22:27] <cjwatson> 56 d-i     debian-installer/locale             select en_US.UTF-8
[22:28] <cjwatson> duplicated debian-installer/locale, and you don't need to set languagechooser/language-name-fb
[22:28] <cjwatson> all that tzconfig stuff can go, time/zone is sufficient
[22:28] <cjwatson> kbd-chooser/method can go (maybe from some ancient Ubuntu documentation, pre-edgy?)
[22:29] <cjwatson> xserver-xorg/autodetect_monitor is no longer used and can go
[22:30] <cjwatson> you don't need to preseed mirror/suite
[22:30] <cjwatson> the rest is probably OK although give it a run through :)
[22:31] <cody-somerville> Will "d-i	passwd/user-password-crypted    passwd " work?
[22:32] <cjwatson> I'm not sure
[22:32] <cjwatson> what are you trying to do?
[22:32] <cody-somerville> Set an empty password
[22:33] <cjwatson> casper does:
[22:33] <cjwatson> d-i passwd/user-password-crypted password U6aMy0wojraho
[22:33] <cjwatson> since U6aMy0wojraho happens to be the hash of a blank password
[22:33] <cody-somerville> I'll steal that then
[22:33] <cjwatson> BTW, the type field should be "password" not "passwd", not that it matters very much
[22:34] <cjwatson> put a comment above the U6aMy0wojraho bit if you do :-)
[22:34] <cjwatson> as it happens, I think that this would also work as of jaunty:
[22:34] <cjwatson> d-i passwd/user-password password
[22:34] <cjwatson> d-i user-setup/allow-password-empty boolean true
[22:35] <cjwatson> which might be clearer assuming I'm not just talking out of my arse
[22:35] <cody-somerville> Okay, I'll try that :)
[22:39] <cody-somerville> This is what I have now: http://pastebin.ubuntu.com/137829/
[22:44] <cody-somerville> should the preseed file be on the cd or in the live rootfs?
[22:47] <cody-somerville> Also, if I preseeded during the live chroot creation, would I get the same affect?
[22:52] <cjwatson> cody-somerville: easier to put it on the CD and use file=/cdrom/...
[22:52] <cjwatson> cody-somerville: preseeding during the live chroot creation is unwise. That means that the preseeding will be active in the installed system too.
[22:52] <cjwatson> # Post install APT setup
[22:52] <cjwatson> d-i     apt-setup/hostname      string archive.ubuntu.com
[22:52] <cjwatson> d-i     apt-setup/directory     string /ubuntu/
[22:52] <cjwatson> where are you getting this stuff from?
[22:53] <cjwatson> d-i     netcfg/get_hostname     string  localhost
[22:53] <cjwatson> d-i     netcfg/get_domain       string  localdomain
[22:53] <cjwatson> I think explicitly setting localdomain is unwise
[22:54] <cjwatson> see e.g. http://lists.debian.org/debian-devel/2005/10/msg00559.html
[22:54] <cody-somerville> I used http://d-i.alioth.debian.org/manual/en.i386/apb.html
[22:54] <cjwatson> don't use Debian documentation for preseeding
[22:54] <cjwatson> https://wiki.ubuntu.com/Installer/FAQ
[22:55] <cjwatson> even so, apt-setup/hostname and apt-setup/directory aren't in the Debian documentation either
[22:57] <cody-somerville> I've been looking at different examples on the web as well
[22:58] <cjwatson> use the installation guide on help.ubuntu.com
[22:58] <CIA-3> ubiquity: cjwatson * r3126 ubiquity/ (debian/changelog ubiquity/frontend/base.py): Always set auto-login for OEM installations.
[22:58] <cjwatson> lots of preseeding documentation floating around the web is wrong
[22:58] <cjwatson> or specific to a particular version without realising it
[22:59] <cjwatson> hah, I know where you got apt-setup/hostname and apt-setup/directory from
[22:59] <cjwatson> those used to be implemented by base-config, which we removed in dapper
[23:01]  * cody-somerville grins.
[23:02] <cjwatson> bad documentation on the web never dies ...
[23:06] <CIA-3> ubiquity: cjwatson * r3127 ubiquity/ (debian/changelog ubiquity/frontend/kde_ui.py):
[23:06] <CIA-3> ubiquity: * KDE frontend:
[23:06] <CIA-3> ubiquity:  - Hide auto-login and require-password radio buttons for OEM
[23:06] <CIA-3> ubiquity:  installations.
[23:09] <CIA-3> ubiquity: cjwatson * r3128 ubiquity/ (debian/changelog scripts/install.py):
[23:09] <CIA-3> ubiquity: Remove reimplementation of gdm and kdm auto-login support for OEM
[23:09] <CIA-3> ubiquity: installations, since user-setup handles this now and the duplication
[23:09] <CIA-3> ubiquity: causes oem-config to be unable to undo autologin (LP: #347900).
[23:27] <cody-somerville> cjwatson, Is it safe to use the same bootstrap on a stable release for image builds?