[10:14] <ara> superm1, hi!
[10:14] <ara> superm1, did bryce fix solved the problem you had with the installation with intel drivers?
[12:48] <CIA-4> pkgsel: cjwatson * r169 ubuntu/debian/ (po/lo.po po/si.po changelog po/se.po): merge from Debian 0.33
[12:50] <CIA-4> pkgsel: cjwatson * r170 ubuntu/debian/changelog: releasing version 0.33ubuntu1
[14:40] <CIA-4> console-setup: cjwatson * r379 ubuntu/debian/ (changelog keyboard-configuration.config):
[14:40] <CIA-4> console-setup: console-setup-udeb (<< 1.53) didn't set the seen flag in the installed
[14:40] <CIA-4> console-setup: system. Cope with this and avoid asking the layout and variant
[14:40] <CIA-4> console-setup: questions again if the default debconf priority is at least high
[14:40] <CIA-4> console-setup: (LP: #705883).
[15:14]  * cjwatson tries to figure out how to detect this console-setup upgrade breakage neatly
[15:15] <cjwatson> it's characterised at least in part by some dialogs being wrongly re-presented and wrongly having the first option as the default, so you tend to end up with keyboard-configuration/modelcode=a4techKB21 and keyboard-configuration/layoutcode=us,af
[15:15] <cjwatson> but I can't really guarantee that some people didn't hit down a bit before getting confused
[15:15] <cjwatson> perhaps I should just do something like:
[15:15] <cjwatson>   if previous-version was one of the broken ones:
[15:16] <cjwatson>     if XKB* in /etc/default/keyboard don't match commented-out XKB* in /etc/default/console-setup:
[15:16] <cjwatson>       reset to values from /etc/default/console-setup and ask questions again
[15:18] <cjwatson> so there'll be some final debconf noise while clearing things up, but it will be confined to people upgrading through natty
[15:18] <cjwatson> or I could s/ and ask questions again//, but that means anyone who's since modified /etc/default/keyboard gets their changes summarily overwritten
[15:18] <cjwatson> maybe if I accompany it with a mail to ubuntu-devel-announce ...
[15:41] <ev> sounds sensible
[16:41] <ev> cjwatson: what are your thoughts on fixing the usb-creator from natty creating usb disks for 10.10 and 10.04 by including maverick's syslinux in usb-creator or building a syslinux-legacy package?
[16:41] <cjwatson> have you tried isohybrid?
[16:42] <cjwatson> if that works it would be so much easier
[16:43] <ev> wouldn't that prevent us from writing a casper-rw file or partition?
[16:45] <cjwatson> I wouldn't rule that out
[16:45] <cjwatson> it might take a bit of fiddling but it should be doable
[16:45] <cjwatson> at least a partition
[16:46] <cjwatson> you *could* import the whole syslinux source and build it again in usb-creator.  it's certainly possible.  it's really ugly though :)
[16:46] <cjwatson> I'm not crazy about syslinux-legacy, it feels like it would tempt people to use it for other things
[16:49] <cjwatson> if isohybrid is unworkable, I'd prefer something that's explicitly private to usb-creator
[16:49] <ev> I'll give it a shot (isohybrid).  I had assumed that this isohybrid stuff worked without a partition table, but I haven't spent any time with it, so further investigation is in order.
[16:49] <ev> noted
[16:49] <ev> thanks bunches for the advice, as always
[16:50] <cjwatson> but if isohybrid is unworkable then I think we need to figure that out, because whenever we finally move to grub it will bear a lot of resemblance to isohybrid
[16:50] <cjwatson> isohybrid does use a partition table (possibly optionally, I'm not sure)0
[16:51] <cjwatson> doc/isolinux.txt says:
[16:51] <cjwatson> The ISO 9660 filesystem is encapsulated in a partition (which starts
[16:51] <cjwatson> at offset zero, which may confuse some systems.)  This makes it
[16:51] <cjwatson> possible for the operating system, once booted, to use the remainder
[16:52] <cjwatson> one thing to establish would be whether there are any other changes you need to make to the filesystem
[16:52] <cjwatson> if there are then you'd need to pack them back into an iso before running isohybrid, which might be slow/fiddly
[16:52] <ev> ah yes, there would be
[16:53] <cjwatson> or you could use xorriso to create a hybrid image in one shot
[16:53] <ev> (the kernel command line parameters in the configuration, assuming that's done on the filesystem)
[16:53] <cjwatson> yeah
[16:53] <cjwatson> xorriso's in main as of maverick, since we needed it for grub2
[16:54] <cjwatson> but the general approach of keeping it as an iso9660 filesystem as well would dodge the whole bullet of having to install a different boot loader, which seems worth trying
[16:54] <ev> indeed
[16:55] <cjwatson> you can look at grub-mkrescue if you want examples of xorriso usage
[16:55] <ev> will do, thanks
[16:55] <cjwatson> 16:51 [Freenode] !kornbluth.freenode.net *** Message to #ubuntu-installer throttled due to flooding
[16:55] <cjwatson> boo @freenode
[16:56] <ev> wow
[16:56] <cjwatson> I think it objected to me pasting four lines from isolinux.txt
[16:56] <ev> that was hardly anything approaching a flood
[16:56] <ev> ah
[16:56] <ev> still, quite agressive
[16:56] <cjwatson> yeah
[17:00] <CIA-4> ubiquity: superm1 * r4491 ubiquity/debian/changelog: add bug number
[17:09] <cjwatson> http://paste.ubuntu.com/559096/ vomit
[17:09]  * cjwatson builds to see what happens
[18:42] <bladernr> Has the policy changed for Natty Server where X is going to be installed??
[18:42] <bladernr> http://pastebin.ubuntu.com/559136/
[18:43] <bladernr> mlegris just pointed out that gtkperf is trying to run on servers, and turns out that x11-utils is being installed for some reason (along with other X packages, see pastebin)
[18:44] <cjwatson> no policy change although console-setup has always pulled in a certain amount via xkb-data
[18:45] <cjwatson> can I have an installer syslog?
[18:46] <cjwatson> hm, should topic that
[18:46] <bladernr> cjwatson:  let me see if I can grab it for you
[18:47] <bladernr> cjwatson:  just installer/syslog or would you prefer a tar of installer/
[18:47] <cjwatson> just installer/syslog
[18:48] <cjwatson> this looks like what happens if you install the uec-live task
[18:48] <cjwatson> gtkperf certainly isn't installed by default, it's in universe
[18:50] <cjwatson> (I mean the ubuntu-uec-live task, btw)
[18:51] <bladernr> right...  I'm starting to learn about using tasksel to find these things out, so sorry for bugging you about some of this stuff
[18:51] <cjwatson> was this an install you did, or somebody else?
[18:53] <bladernr> it's one of the preseeds on the enablement servers in Boston (for hardware cert)
[18:54] <bladernr> AFAIK, our preseed uses defaults and doesn't install any options for server
[18:54] <bladernr> I'm going to do a local install to make sure it's not something our preseed is pulling in though
[18:54] <cjwatson> then ideally I'd get a log with the DEBCONF_DEBUG=developer boot parameter, to try to determine why ubuntu-uec-live is being pulled in
[18:54] <bladernr> and the log is in your e-mail
[18:54] <cjwatson> thanks
[21:35] <cody-somerville> Whats the value for no password?
[21:35] <cody-somerville> (for use with preseeding)
[21:36] <cody-somerville> chpasswd no longer accepts no password for using unencrypted version no longer works
[23:02] <bladernr> cjwatson:  you can safely ignore my x11-utils question from before.  Turns out (I didn't find this out until well after I'd become invested in this) that the person who brought it to my attention initially had installed desktop on a server, which explains everything.