[00:06] cjwatson: what's the rootdelay equivalent for the hardy livecd? [00:42] hrm, break=mount seems to be the only way to extend the live media recognization that I can see. On the upside, this behavior on this hardware doesn't appear to be a regression over 8.04.2 [00:43] (where this behavior==get dropped into busybox by the initramfs) [04:36] how can I hack the netbook installer to install less packages so i can fit the netbook remix into 2GB? [04:38] can I just edit dists/*/*/binary-i386/Packages.gz? === astronouth7303 is now known as astro73|0000 === astro73|0000 is now known as astronouth7303 [05:15] astronouth7303: Unfortunately its not that easy, and it depends on what packages you want to remove. [05:16] it's for a netbook. I *need* firefox, pidgin, skype, and gedit [05:16] and a terminal [05:16] Well what don't you need? [05:17] openoffice, evolution [05:17] rhythmbox [05:17] java, maybe [05:17] astronouth7303: Unfortunately the people who really know how to do this are not currently around, however if you are around in about 5 to 6 hours, the installer maintainers may be able tgo help you. [05:17] ok [09:31] cjwatson: in reviewing the oem-config merge, I noticed the comment at the top of http://bazaar.launchpad.net/~mterry/ubiquity/oem-config-merge/annotate/head%3A/bin/oem-config-firstboot . Couldn't we fix that by using an upstart job with "start on stopped last-good-boot"? [09:40] sbeattie: yeah, I don't know if there's any alternative ... [09:40] evand: actually we can probably just ditch it, as last-good-boot's been abandoned [09:40] which is why I haven't been worrying about that comment recently :) [09:41] oh cool, last I had heard we were looking for someone to get it actually working [09:42] it'd be interesting to attempt it in grub2 [09:42] so I take it then that this was a resource problem, not that the problem has magically been solved elsewhere [09:58] well, it was sort of not the best place or way to do it anyway [09:58] using debconf in the boot process => Keybuk sadface [09:58] so I think any replacement for it is not going to have the same problem [09:59] gotcha [10:29] finish-install: cjwatson * r830 ubuntu/ (debian/changelog finish-install.d/90console): [10:29] finish-install: Adjust serial console handling to cope with upstart 0.6 configuration [10:29] finish-install: file paths. [10:31] finish-install: cjwatson * r831 ubuntu/debian/changelog: releasing version 2.23ubuntu2 [10:34] finish-install: cjwatson * r832 ubuntu/ (debian/changelog finish-install.d/90console): Fix minor mistakes in upstart 0.6 handling. [10:35] finish-install: cjwatson * r833 ubuntu/debian/changelog: releasing version 2.23ubuntu3 [10:35] (whoops) [11:31] uh, interesting: http://pastebin.ubuntu.com/214541 . From what I've gleaned by digging through the partman/parted code, that shouldn't ever happen. [11:36] versions of libparted1.8-12 and partman-base? [11:44] 1.8.8.git.2009.06.03-1ubuntu2 and 131ubuntu1 [11:45] odd then, as you say that shouldn't happen ... [11:45] happy debugging ;-) [11:45] heh, indeed [14:07] cjwatson, I'd like to patch cdrom-detect so something like cdrom-detect/try-hdd=true can work as expected. [14:07] Hi All. Can i ask a question related remastersys here? [14:07] cjwatson, I figure since someone hasn't already maybe theres a reason [14:07] ubuntu - remastersys - kde [14:07] c0nfus3d: you can ask, but I'm not sure anyone here has relevant expertise [14:08] cody-somerville: mostly because it's starting to look like spaghetti code and I kind of think it should be reorganised, but haven't worked out how yet [14:08] cody-somerville: (and a number of Debian people reckon you should use hd-media and iso-scan in those circumstances instead) [14:08] cody-somerville: ... well, why not use hd-media and iso-scan? [14:09] cody-somerville: the problem with booting off the hard disk, of course (and it applies to either of those approaches) is that it tends to completely screw the partitioner [14:09] cody-somerville: which is why it isn't something we generally encourage [14:11] i am trying to customize ubuntu and installed using ubiquity. When i use remastersys to make a live CD, everything works fine, except my desktop settings (wallpaper does not change) - i am using KDE 4 [14:11] after some search i could find that we have to copy the files from the home folder to the /etc/skel. But its not working. Any idea or workaround? [14:12] this is the list of files - http://imagebin.org/55424 [14:12] /etc/skel will affect *all* created users [14:12] yes - but do you know which file i need to change to change my wallpaper? [14:12] absolutely no idea :) [14:12] that'd be a KDE thing [14:12] is /etc/skel being copied at all? it should be [14:13] it is getting copied - i mean do i have to copy all the files in my home folder to /etc/skel? [14:13] surely not, but you'll have to ask KDE folks to find out where the wallpaper preference is stored [14:13] ok fine, let me try my luck there - thank you so much cjwatson [14:14] best to keep /etc/skel as minimal as possible, generally [14:14] ok i understand - thanks [14:14] let me try my luck with KDE guys [14:14] thanks - :) [14:14] good luck, sorry I couldn't help more [14:14] no problem [14:19] cjwatson, Are hd-media and iso-scan udebs or something? [14:20] cjwatson, The reason I Want to do it in cdrom-detect is because I want to dd the cdrom to a partition to become a recovery partition [14:20] cjwatson, that way things Just Work (TM) [14:23] except for the partitioner, which doesn't work [14:23] but feel free to try ;-) [14:24] hd-media is an installation method, with its own debian-installer build - it's mostly used for USB sticks but basically you put the kernel and initrd on the filesystem and then dump the .iso straight onto the filesystem as a single file [14:24] iso-scan is the main udeb that powers it [14:25] other people have done the recovery partition technique, but only with casper-based live CDs, to my knowledge [14:25] that's the reason casper has the UUID stuff in it [14:26] I exaggerate slightly when I say the partitioner doesn't work. Actually I think it mostly works as long as you make absolutely sure not to touch anything before or including the recovery partition. It has no checking for this, though, so it's very easy to shoot yourself in the foot [14:31] cjwatson, doesn't the installer get loaded into the ramdisk? [14:32] the disk is still mounted [14:32] for example it still needs it to fetch .debs off; loading all those into memory would be absurd [14:32] cjwatson, wouldn't a casper-based live cd still have the same partitioning issue? [14:32] yes. read what I wrote above :) [14:32] the people who have done it are just rather cautiously stepping around the problems [14:39] patching cdrom-detect should be straightforward - just grep for try-usb in its source and do the same kinds of things [14:39] list-devices is in the debian-installer-utils source package [14:39] it should already have the support you need but you'll probably want to look at its interface [14:40] * cody-somerville nods. [14:41] evand... huuuuuuuullllllllllooooooooo :) [14:41] evand: got good news? [14:42] rgreening: I've separated the main binary again [14:42] hehe. :) [14:42] still working on the windows and devkit backends [14:42] help welcome there [14:42] fighting parted to the death at the moment [14:42] do you have any pointers to docs for devicekit-disks? I'm not familiar with it. [14:42] trying to find the source of a bug that increasingly looks like a local issue [14:43] heh [14:43] there's a link to the API in the devicekit-disks backend [14:43] build a dep and push to a PPA so we can test [14:43] thats' why I got packaging fixed so I could test on other systems [14:43] :) [14:44] http://hal.freedesktop.org/docs/DeviceKit-disks/ [14:44] evand: what was the branch again? cleanup? [14:44] yup [14:44] kk [14:44] evand: let me read up on it... and get back to you... [14:46] sure thing [15:02] Can anyone here help me with a preseed install? [15:02] I can't manage to preseed the one question: "unmount partitions that are in use?" [15:03] Hi, I'm trying to preseed an installation with extra debs. Unfortunately one of the debs require an interactive step (Do you wanna build kernel module? Yes/No) [15:04] Anyway around this ? [15:05] On an installed system, you can run: [15:05] debconf-get-selections --installer > file; debconf-get-selections >> file [15:05] To get a list of all possible questions, from what I understand [15:05] I found "partman-base partman/unmount_active boolean false", but it doesn't seem to fix my problem. [15:06] I'm trying to preseed an installation from a usb key, and I can't get around it asking me if I want to unmount the usb key. Obviously I don't, because I'm installing from it. [15:07] cudev: I'm no expert .. but booting with DEBCONF_DEBUG=5 and see'ing the ALT+F4 screen might help you understand which variable needs to be preseeded [15:08] kim0: thanks, I'll give it a try to start [15:09] cudev: any idea how to suppress an interactive deb installation :D [15:09] Can you try running those commands I listed? I don't know how, but it might be in there [15:10] The output is not that long, and they are all commented as to what they do [15:10] I wouldn't actually recommend using debconf-get-selections --installer as a source [15:10] it's better to use the installation guide [15:10] Where is that? [15:10] help.ubuntu.com, fairly clearly linked I think [15:10] debconf-get-selections --installer includes a *lot* of stuff that shouldn't be preseeded, and at the moment unfortunately there's no easy way for non-experts to distinguish [15:11] cudev: what version of Ubuntu are you using? [15:11] OK, but the list at that site doesn't have the option to remove the prompt I am getting [15:11] https://help.ubuntu.com/community/InstallCDCustomization ? [15:11] that said, in kim0's case, debconf-get-selections --installer is probably one reasonable way to find the question that's being asked [15:11] kim0: https://help.ubuntu.com/9.04/installation-guide/amd64/appendix-preseed.html [15:11] Or change to whatever system you are trying to install [15:12] alternatively, you can boot the installer with DEBCONF_DEBUG=developer, and you'll get a big pile of output in syslog that should include the question being asked [15:12] cjwatson: the question is virtualbox package asking if it should compile kernel modules .. would I expect that to show up in debconf-get-selections output ?? [15:12] it ought to, but I can probably just answer that one - one moment [15:12] oh cool ! [15:12] * kim0 anxiously waiting :D [15:13] cjwatson: entirely my fault and ultimately an utter waste of time. I built the package on 9.04, it linked against libparted1.8-10. Building on Karmic fixes it. [15:13] kim0: what version of Ubuntu? [15:13] cjwatson: 8.04 [15:13] evand: heh :) ok [15:14] cjwatson: and the latest virtualbox-3.0 deb [15:14] btw, I was installing on Ubuntu 9.04 64-bit server [15:15] kim0: ... where does that package come from? [15:15] cjwatson: sun repos .. it's the closed source one [15:15] it's not in our archive as far as I can see [15:15] URL? [15:15] ok moment [15:15] I need to look at it to answer this question, you see [15:15] cjwatson: that's the apt repo [15:15] deb http://download.virtualbox.org/virtualbox/debian jaunty non-free [15:16] in my case hardy not jaunty [15:17] cudev: a workaround is 'd-i partman/filter_mounted boolean false' (and debconf-get-selections --installer wouldn't have told you this, I don't think). However, IMO it's a bug that you're seeing this question in the first place. Is your installer image modified at all from standard Ubuntu? [15:18] Yes, I am installing from a flash drive [15:19] cjwatson: Is that it? => db_get virtualbox/module-compilation-allowed [15:19] And I ripped open the initrd file so that I can place the preseed script right in it [15:19] Thus, I can bypass the language and keyboard questions [15:19] At least, that's as I recall [15:20] I made it about a year ago for 8.04, and now I am doing it again in the same way to get it working for 9.04 [15:25] I did it from a tutorial I found, and it worked fine for 8.04. The USB key contains: [15:25] install folder, syslinux folder, recipe file, the iso, and a ldlinux.sys executable [15:25] The install folder contains the linuz file and a modified initrd.gz file, and the syslinux folder is basically a copy of the iso's isolinux file [15:25] kim0: still looking, one moment [15:25] cudev: you could of course just preseed those questions on the kernel command line :) [15:26] how can I hack the netbook installer to install less packages so i can fit the netbook remix into 2GB? [15:26] can I just edit dists/*/*/binary-i386/Packages.gz? [15:26] cudev: I'd like to see a DEBCONF_DEBUG=developer syslog from your preseeded installation attempt - make sure you set a non-secret password if you're preseeding that, as it'll show up in the log [15:27] astronouth7303: I suppose you could but that's horribly horribly roundabout. why not just tell pkgsel/tasksel to install some different set of packages? [15:27] ... how would i do that? [15:27] astronouth7303: you'd have to get round authentication warnings if you edited Packages.gz and in general it'll be an utter pain to maintain [15:27] Alright, I'm working on getting some debug output now, I'll be back later [15:27] https://help.ubuntu.com/9.04/installation-guide/i386/preseed-contents.html, search for pkgsel [15:27] i just need to do this once [15:28] if you're doing it once, all the more reason not to edit Packages.gz. It'll take you ages to get it right, trust me [15:28] it'd take *me* ages to get it right [15:28] compare with whatever preseed file in /preseed/ on the disk it is that it's using [15:29] /preseed/netbook-remix.seed, I think [15:29] tasksel tasksel/first multiselect mobile-netbook-remix [15:29] so you could take that out and replace it with either some smaller task, or just a giant list of packages in pkgsel/include [15:29] (d-i pkgsel/include string ...) [15:32] cjwatson: Would you by any chance have a degraded raid1 setup recipe for hardy :) [15:33] now i have to figure out what set of packages is installed... [15:34] kim0: not my field, I'm afraidi [15:34] afraid [15:34] kirkland might know [15:35] cjwatson: kim0: howdy [15:35] kirkland: hi there :) [15:35] kim0: what do you mean by "recipe" ? [15:35] um .. in a preseed file [15:35] kim0: so yes, virtualbox/module-compilation-allowed seems to be the right thing to set: 'virtualbox-3.0 virtualbox/module-compilation-allowed boolean false' [15:35] for a degraded raid1 installation [15:35] kim0: i do not, however, mathiaz does [15:35] according to the templates file in that package [15:35] kim0: ask mathiaz in #ubuntu-server [15:36] kim0: i generally test that by hand [15:36] kirkland: thanks a lot [15:36] cjwatson: awesome .. thanks a million [15:37] kirkland: I'm curious about that testing, incidentally, given that I found a "greq" typo in mdadm recently that had been there since pre-hardy ;-) [15:38] hmm, no, can't have been pre-hardy, intrepid maybe [15:39] cjwatson: I've got the syslog file here. This is without trying the workaround you suggested [15:39] but it looked as though it would never ask the degraded question during the partitioner, only (at best) later [15:42] cudev: sorry, I can't do DCC [15:42] cudev: paste.ubuntu.com, perhaps? [15:42] I particularly can't do DCC because you're advertising a private IP address, presumably behind a NAT ;-) [15:42] 192.168.66.53 here is not the same as 192.168.66.53 at your end ... [15:43] Fair enough [15:44] http://paste.ubuntu.com/214825/plain/ is a script I found under a rock somewhere to post files to paste.ubuntu.com, BTW [15:45] (and I've asked our sysadmins to add a file upload widget, so that you don't have to cut-and-paste big chunks [15:48] oh, bah, iso-scan puts its mount on a different mountpoint [15:49] cudev: can I just see /proc/mounts, to make sure of my fix? [15:49] from the running installer [15:49] Oh, I already restarted the installer to attempt your fix [15:50] how necessary is scim? [15:50] I can tell you in a second if it worked [15:51] astronouth7303: not especially unless you need input support for a non-Latin-script language [15:52] ok [15:53] cjwatson: the workaround seems to have worked. Thanks a lot! I've been trying at this for awhile now [15:54] Was it a hardware thing then, is that why it was asking me that when it should not have? [15:54] Also, where did that work around come from? You said that I wouldn't have found it in debconf-get-selections. Is there a more complete list somewhere? [15:54] the source, I'm afraid [15:55] I'm not actually sure yet why it's happening, my initial guess looks as though it may be wrong [15:55] I assume that the workaround you gave me just doesn't scan for mounted devices. Is that correct? [15:55] it's unlikely to be a hardware thing, but could be related to the particular mode of installation in use here [15:55] that's right [15:56] ugh. Looking at the source was going to be my next step if the awesome people here couldn't help me. === nxvl_ is now known as nxvl [15:56] it's one of those things that I haven't documented because it's strictly an installer bug workaround [15:56] and I'd rather fix the bugs :-) === Nicke_ is now known as Nicke [16:17] odd. even though I changed the packages in preseed, the disk space wanted by the partitioner is the same (2097426432B) [16:18] oh, I didn't know you were talking about that [16:18] what partitioning mode? [16:19] custom [16:19] controlled by a preseed file? [16:19] I just thought that it might be smart enough to calculate how much space was needed based on what was installec [16:19] no, i just do it by hand [16:19] I don't know of anything in the partitioner that hardcodes that size [16:20] although it certainly isn't smart enough to calculate that - it doesn't really have enough information readily available [16:20] can you clarify what you mean by "the disk space wanted by the partitioner", then? [16:21] when I partition, it gives me a warning about root not being big enough (it's 1998MB). After removing openoffice, evolution, scim, and extraneous video drivers, that number didn't change [16:22] oh, I can believe that; it would speed things up if you could give me the exact text of the warning so I can search for it and explain how to turn it off [16:22] that number is not related to packages installed [16:23] i don't care that much; it's just one of those oddities [16:23] I do :- [16:23] :-) [16:23] I care about anything that causes people to ask questions [16:23] next time I run the installer, i'll let you know [16:23] * cjwatson looks around to try to find it [16:25] oh, that'll be ubiquity, won't it [16:25] astronouth7303: graphical installer? [16:25] yup. off the netbook remix [16:25] damn, sorry, in that case I was leading you up the garden path a bit [16:25] that warning is based on the size of the live filesystem; pkgsel/tasksel is irrelevant [16:26] ah [16:26] you just need to take the live filesystem and remove anything you don't want from it [16:26] sorry, I was thinking that UNR still used the text-based installer, which is of course wrong [16:26] (it used to, I think) [16:26] ... that requires remastering the disc, doesn't it? [16:26] right [16:26] but that's unavoidable - the graphical installer operates by copying the live filesystem to disk [16:27] oh [16:27] https://help.ubuntu.com/community/LiveCDCustomization [16:27] yeah, #ubuntu linked me yesterday [16:28] actually, in current versions there might be a bit of a get-out [16:28] hm? [16:28] except that there's no way to preseed away the partitions-too-small warning [16:28] i can hit ok [16:29] but the installer will remove any packages that are listed in /cdrom/casper/filesystem.manifest but not in /cdrom/casper/filesystem.manifest-desktop, and in recent versions of ubiquity it actually skips copying their files to disk [16:29] so if you're very very careful you can just strip out files from /cdrom/casper/filesystem.manifest-desktop - that's still a remastering but a much simpler one [16:30] it won't necessarily help you out much if you get dependencies wrong, though, so you'll have to be careful [16:30] i.e. it expects the list in manifest-desktop to include all the dependencies [16:30] well, this is off a flash drive, so if I don't have to rebuild the filesystem, that's good [16:30] i'm only removing stuff, not adding things [16:30] evand: really we ought to make ubiquity/partition-too-small sensitive to file blacklisting, not that that's easy :) [16:31] and if i screw it up, i'll have aptitude fix it after the fact [16:31] we'd have to calculate the blacklist *even earlier*, and I suspect there are some apt-install commands that happen after that point ... [16:31] right - I think it'll probably tolerate minor screwups, basically by refusing to remove those packages [16:35] just in case, is there an installer that uses the preseed package information? [16:36] (also, that installation attempt failed - not enough disk space) [16:36] I don't think UNR does d-i-based builds any more, sorry [16:37] cjwatson: indeed, quite the chicken and egg situation [16:37] we could move those earlier, maybe to commit.d scripts or something [16:37] oh, check.d runs before commit.d doesn't it [16:37] bah [16:40] cudev: I think I sort of see what might be going on here; looks like the mount-filtering code isn't smart enough to deal with loop mounts [16:40] cudev: could you please file a bug on partman-base about this? [16:44] is there anything core in UNR that uses mono? [16:45] shouldn't think so, nothing core in Ubuntu uses mono, basically peripheral apps [16:47] http://people.ubuntu.com/~ubuntu-archive/germinate-output/mobile.jaunty/rdepends/ALL/libmono0 (useful but a bit hard to read without practice) says that it's just f-spot and tomboy [16:50] cool [16:55] (take 5 on the installer) === dpm is now known as dpm-afk [17:12] wow, i remove a few packages and suddenly it takes a while to calculate files [17:18] i could be wrong, but i think it hung at 15% "Calculating files to skip copying..." [17:19] that step can be a bit slow; I optimised it as far as I could in 9.04 [17:20] it has a fundamentally fairly hard job to do [17:20] ok, so it's ok if it doesn't budge from 15? [17:20] basically it's testing whether each of the packages in turn are safely removable, in order to try to construct the maximal set of things to remove [17:20] well, it should budge eventually [17:20] ok [17:20] but if you've removed a lot of stuff, it gets quadratically more expensive :-( [17:20] oh [17:21] O(n**2) is not good [17:21] better than O(2^n) :-) [17:21] cjwatson: I'm back. You still want me to file that bug? [17:22] cudev: yes please [17:23] k [17:24] At https://bugs.launchpad.net/, correct? [17:24] https://bugs.launchpad.net/ubuntu/+source/partman-base/+filebug [17:24] debian-installer: cjwatson * r1120 ubuntu/ (9 files in 6 dirs): [17:24] debian-installer: Disable GTK frontend for now; the GDK directfb backend needs some work [17:24] debian-installer: in the current development series. [17:27] cjwatson: should this be related to bug #276656, or start a new bug? [17:27] Launchpad bug 276656 in partman-base "It's possible to install to the device the installation files reside on." [High,Fix released] https://launchpad.net/bugs/276656 [17:27] As I was saying :) [17:27] That's marked as fixed though .. reopen, or new? [17:29] new bug, please [17:29] this is a special case [17:29] it's specifically the case where /cdrom is a loop-mount from a file on a certain filesystem, rather than being a mount of that filesystem directly [17:30] in fact you can pretty much just say that in the bug if you want :) it's just good if you're the submitter rather than me [17:31] gotcha [17:35] Done. Bug #397901: [17:35] Launchpad bug 397901 in partman-base "Asked to unmount the installing device" [Undecided,New] https://launchpad.net/bugs/397901 [17:35] Ubuntu bug 397901 in partman-base "Asked to unmount the installing device" [Undecided,New] [17:46] cjwatson: i think take 5 worked. It got past copying files w/o running out of space [17:48] cool [17:48] cudev: thanks [18:03] cjwatson, superm1 mentioned that he thought that there were some installer changes required to support grub 2. I was wondering if you could shed any light on that. [18:04] the changes we made ages ago? [18:04] if you're talking about hardy, I'm sure it needs all kinds of things, I wouldn't recommend it myself [18:04] grub-installer is fairly tightly integrated [18:05] cjwatson, So grub-installer for jaunty should be fine to handle grub2? [18:06] sort of, it has known bugs [18:06] grub2 in jaunty was for experimental purposes only [18:06] I seriously wouldn't recommend shipping it on production systems yet [18:07] I have a spec assigned to me to sort it all out for karmic [18:08] if you insist on doing it in jaunty, you should probably scan the grub-installer and grub2 changelogs for interesting fixes [18:17] cjwatson: thank you, ubuntu is now installed in <2GB. Need to tweak it down more. [18:18] Why does stuff depend on openoffice? [18:19] language-support-* you mean? that's being refactored in karmic [18:19] not just that [18:19] up until now we haven't had a better way to say "these packages are needed if you have these other packages installed and are using this language" [18:19] well I imagine it's a different reason in each case then ... [18:19] there are relations between thunderbird, openoffice, and evolution [18:19] specifically with language and docs [18:20] yes, there are [18:20] well, actually, I rather doubt that anything thunderbirdish depends on openoffice.org [18:20] I think you might have to back that one up ;-) [18:21] i'd have to dig out the dependencies again [18:21] I can't find any evidence of that. language-support-en (and friends) depend on all of the things you mentioned [18:21] so in your case, you probably just want to rip out language-support-* in practice [18:22] true [18:22] i can probably pull out thunderbird, too [18:22] evolution can't be completely removed without screwing with the desktop env, as it turns out [18:23] compiz will be interesting [18:25] I thought UNR defaulted to maximus [18:25] it does [18:25] but compiz is still installed [18:25] maximus replaces compiz? [18:26] they both implement window management [18:26] neither replaces the other as such [18:26] well, all i know is that my computer can't handle all these fades [18:40] cjwatson, What generally generates /etc/initramfs-tools/conf.d/resume? === rgreening_ is now known as rgreening [20:10] Hi .. Is there some way to list the loaded udebs inside the installer ? [20:39] messed up my partitons...http://pastebin.com/d6bfdffdc [21:09] cody-somerville: base-installer [21:09] kim0: they're in /var/lib/dpkg/status === robbiew is now known as robbiew-afk [21:27] cjwatson: thanks .. it does show that the raid module is installed .. but it's still erroring out [21:28] cjwatson: syslog mentions the multiraid line .. as the last line before it gave an error [21:29] err, I'll need a bit more exact detail than that if you expect me to debug it :) [21:29] developers tend to really hate being given paraphrased error messages ;-) [21:29] cjwatson: anyway I can copy that syslog out of the box ? [21:29] anna-install openssh-client-udeb [21:29] then you'll have scp [21:29] ok [21:32] cjwatson: I pasted here: http://paste.ubuntu.com/215068/ ... Thanks for taking a look [21:41] kim0: I need to see your preseed file as well [21:41] please try to put it up in a way that preserves things like spacing [21:42] cjwatson: paste.ubuntu.com as text preserves right [21:42] not sure about spaces at the ends of lines, but give it a try [21:42] if I can't work it out from that, I'll tell you [21:43] cjwatson: http://paste.ubuntu.com/215072/ [21:44] kim0: so, doesn't look as if I can tell from that; I think you have a space at the end of a line or something and it's confusing it [21:44] cjwatson: http://paste.ubuntu.com/215076/ [21:44] cjwatson: vim shows end of lines as $ now [21:45] that should help eh [21:45] kim0: ok, line 13 [21:45] get rid of those spaces [21:45] you're on 8.04, right? this is bug 245256, fixed in later releases [21:45] Launchpad bug 245256 in preseed "Lines containing only spaces cause installer to error" [Undecided,Fix released] https://launchpad.net/bugs/245256 [21:46] cjwatson: OMG :) [21:46] * kim0 testing [21:49] cjwatson: You are da man :D [21:49] cjwatson: this consumed me almost 2 full days of head banging [21:49] cjwatson: thanks a zillion man [21:50] cool, sorry about that bug [21:50] I agree that whitespace sensitivity can get a bit confusing ... [21:50] It's great that I know about that now [21:50] cjwatson: Yeah [21:51] one wouldn't expect this really [21:51] perfect [22:41] cjwatson: do you know if umenu in hardy is version sensitive too? Only I still get the invalid cd but now if I browse it I can run wubi which I couldn't yesterday [22:42] it= the cd [22:51] sorry, I don't [22:52] I grepped all of wubi for 8.04.2, if there's anything else then it's something more complicated that (a) I'm not familiar with (b) I probably won't know how to fix [22:52] needs somebody more familiar with wubi than I :) [22:52] cjwatson: wubi is working fine now when you fire it up [22:53] I'll send a mail to xivulon and evand then [22:54]