[00:37] cjwatson ping === mpt_ is now known as mpt [09:42] not hugely important, but trying to netboot hardy fails since it can't find a suitable kernel [09:44] "couldn't find package installation-report; info: Found kernels '' " [09:51] odd [09:51] stick the full syslog somewhere? [09:54] sure, a sec [09:58] cjwatson: http://users.tkk.fi/~tjaalton/foo/syslog [10:01] the installation-report bit is odd but not relevant [10:02] although possibly indicative ... it should be able to find that since it's in the archive [10:02] ah [10:02] echo "deb $APTSOURCE $DISTRIBUTION $COMPONENTS" > /target/etc/apt/sources.list [10:02] echo "deb $APTSOURCE $DISTRIBUTION-updates $COMPONENTS" > /target/etc/apt/sources.list [10:02] spot the mistake [10:03] Heh. [10:07] tjaalton: fixed in base-installer 1.85ubuntu3, thanks [10:10] wow great [10:11] ah, I didn't spot the error :P [10:12] yes, sources.list is quite minimal on the system [10:12] heh [10:34] heh. amazing what damage missing a single > would do :) [12:23] cjwatson: works great [12:25] excellent [12:28] now if aptitude could be a bit more forgiving, since it bails out if a single package can't be found :) [12:29] I preseed pkgsel/include with a list of packages, and if one of those is missing (which is common no a development version), it just fails [12:29] so in this case the list for gutsy is no good === cjwatson_ is now known as cjwatson [16:21] How does the installer decide which kernel to install? [16:40] err, the answer is quite long. Which bit do you care about? [16:41] (as in, what logic does it use? what things does it look at? where can I find the code?) [16:47] Good question. :) I want to make the jeos iso install the -virtual kernel flavour. [16:48] cjwatson: A pointer to the code will probably be the the easiest for you and will probably give me all the info I need. [16:49] ah, in that case it's actually easier to just tell you [16:49] soren: preseed 'd-i base-installer/kernel/override-image string linux-virtual' [16:49] That seems a bit harsh. [16:49] it's in data/$DIST/preseed/ in debian-cd [16:50] alternatively; if only the -virtual kernel is on the CD, it should already select it automatically [16:50] see base-installer/kernel/i386.sh [16:50] if linux-generic is there, it'll use that in preference, so you need to make sure linux-generic isn't there [16:50] Yes, ISTR something about that, and I thought it would be the only kernel, but the boot seed gave me the -generic one as well. [16:51] Ok, I'll figure something out. [16:51] you probably shouldn't be including the boot seed ...? [16:51] No, that's my conclusion as well. [16:51] jeos should substitute for boot [16:51] jeos: boot required minimal [16:51] ubuntu-seeds/ubuntu.hardy/STRUCTURE [16:52] I think that's the error [16:52] I just thought it would be an easy shortcut to certain things, but it's causing me more problems than it's solving :) [16:52] Right. I'll hand pick the packages I need, put them into the jeos seed and remove boot from that line in STRUCTURE. [16:53] there's nothing in boot you need that isn't already in jeos [16:54] Oh. [16:54] You're right. [16:54] I must have looked at germinate's output and got confused. [16:55] easy to do, expanded dependency trees are *big* [16:55] Hm... [16:56] http://people.ubuntu.com/~ubuntu-archive/germinate-output/ubuntu.hardy/boot shows a lot of stuff I want that's not in http://people.ubuntu.com/~ubuntu-archive/germinate-output/ubuntu.hardy/jeos [16:56] hrmm, I'm guessing I should look outside of ubiquity/d-i for the cause of bug 173008 [16:56] and the bot appears to be MIA [16:56] soren: you need to look at the seeds themselves rather than the output [16:57] looking at the output will just be confusing [16:57] Quite. [16:57] :) [16:57] cjwatson: Ok, thanks. I'll work it out. [16:57] soren: boot there is pulling in lots of base-system-y stuff [16:57] because boot is Weird [16:57] cjwatson: And jeos will get that from... required? [16:57] it's basically a nasty hack around confusion caused by the kernel having massive dependencies but being conceptually at the bottom of the stack [16:57] right [16:58] Ok, got it. [16:58] if you want to compare expanded lists, compare the union of required+minimal+boot with the union of required+minimal+jeos [16:58] it'll make a lot more sense [16:58] Alright. Thanks again. [16:59] evand: probably console-setup, but err [16:59] weirdness [16:59] indeed [16:59] probably means that console-setup just isn't working at boot time [16:59] or that X didn't pick it up [16:59] wasn't X failing to generate xorg.conf at some point? [17:00] * cjwatson follows up to the bug ... [17:00] (feel free to answer the same questions if you have an affected system handy) [17:00] I do, but I'll have to run through another install again. Shouldn't take too long though. [17:01] I'll queue it up in the event that he doesn't reply in a timely manner. [17:01] s/in/for/ [17:03] ok, what the HELL [17:04] Ctrl-Alt-F1 tries to switch to a console and FAILS [17:04] Oh, it's not just me? [17:04] first time that's happened to me in nine years of using Linux [17:04] I thought I had knackered something. [17:05] * evand is afraid of that key combination. nvidia does not play nice. [17:07] oops, apparently stracing X isn't a great plan [17:07] cjwatson: 9 years? N00b! :) [17:07] anyway, seems to work fine from gdm's X server, but not once logged in [17:08] soren: hippy [17:08] :p [17:30] yup, that's it [17:31] /etc/default/console-setup is right [17:31] /etc/X11/xorg.conf is non-existant [17:31] that'd do it [17:31] iz xorg bug [17:31] yay, not our problem ;) [18:09] cjwatson: mind if I fix bug 62950 with a s/use CD2 or// ? [18:09] https://bugs.launchpad.net/ubuntu/+source/syslinux/+bug/62950 [18:09] figured I'd ask first as it's been outstanding for a while, and I wasn't sure if there was a reason why it wasn't covered in previous attempts to remove references to CD2 [18:15] evand: yeah, sounds reasonable to me [18:19] I'm assuming I can edit the existing patch, rather than shoving a patch to a patch in there, right? [18:20] I only ask as this is apparently stolen from SuSE, but I imagine trusting whoever updates that to check the changelog would be reasonable. [18:21] and updating anything in debian/ doesn't require patching anyway, right? [18:22] edit the existing patch in debian/patches/ [18:23] probably easiest to use dpatch-edit-patch to do so [18:23] run it with appropriate args, it'll give you an unpacked source tree, you edit the source directly, it fixes up the patch to do what you said [18:23] indeed [18:25] I'll take care of the merge while I'm here [21:03] i know this probably isn't the best place to ask this question, but does anyone know of a quick way to display all the packages installed on my system and the license each one uses? === ceeka1 is now known as seekay [21:06] seekay: /usr/share/doc/*/copyright [21:07] COLUMNS=100 dpkg -l |awk '/^ii/{print $2}' for a list of packages === cjwatson_ is now known as cjwatson [23:10] cjwatson: i sent a mail around last week about problems installing grub using a 7.04 netinstall from a usb disk. i have removed all lines regarding grub from my preseed, but it still always installs grub into /dev/sda (my USB disk at install time, thus overwriting syslinux!) i can reboot, do a grub-install to my hard disk, and it works fine, but it seems wrong that i should have to re-install syslinux every time on my usb key... any thought