=== compbrai1 is now known as CompBrain === joejaxx_ is now known as joejaxx [18:01] evand ping [18:24] cjwatson ping [18:26] xivulon: pong [18:29] hey evand couple of quick questions [18:29] a few users reported errors with m-a and wubi [18:30] Unable to open the registry file at '/mnt/migrationassistant/WINDOWS/system32/config/software' [18:30] I think the path case is the issue [18:30] but that make the installer stop [18:31] how can I go around that for the time being? Do you know the relevant file on top of your head? [18:31] Issue 2: I'd like to silence the dialog that complains about security.ubuntu.com being unreachable, can I preseed that? [18:32] apt-setup/security_host [18:32] thanks chwatson [18:32] actually, no [18:33] apt-setup/service-failed [18:33] except that's in hardy :) [18:33] use apt-setup/security-updates-failed in gutsy [18:33] what value do I preseed to make it skip? [18:33] xivulon: that error alone will not make the installer stop. I'd need to see the full logs. [18:34] anything, it just needs to be seen. the empty string would be traditional [18:34] d-i apt-setup/security-updates-failed note [18:34] evand see http://ubuntuforums.org/showpost.php?p=3638010&postcount=101 [18:35] don't have the syslog but can ask the user that reported that [18:35] m-a works well for me [18:36] Issue 3 is that when ntfsresize fails, users are shown posix_filesystem_required [18:37] that message is quite confusing, since what they should see is "Please run chkdsk /r from windows and reboot again into Ubuntu" [18:37] ntfsresize --info [18:39] Issue 4: not sure how to get right the grub-drive, the drive order of what is seen by grub4dos and what is seen from linux is not always the same [18:39] that's it!! [18:42] cjwatson, can I show a custom message in ubiquity from autopartition-loop and force ubiquity to quit? [18:42] that is to go around #3 [18:44] So, Two Questions: 1) is there a python parser or generator for preseed configuration files 2) Is there a python parser for template files, and whats the easiest way to amass a collection of them [18:45] CompBrain: I am unaware of any tool to generate preseed files. The closest thing is the kickstart GUI to generate kickstart files that the installer can handle. [18:45] Ok, I borrowed pseudo code from debconf-(get,set)-selections and build a module for that purpose [18:47] please keep in mind that debconf doesn't have to be flat files forever, so it's best to interface directly with the existing tools so that you're well insulated from any changes to the underlying database format. [18:47] I'm mostly interested here in building a preseed configuration widget [18:52] evand is the msg above enough for you? [18:53] xivulon: what message? [18:57] http://ubuntuforums.org/showpost.php?p=3638010&postcount=101 [18:58] xivulon: http://ubuntuforums.org/showpost.php?p=3655006&postcount=113 [18:59] good [19:00] By the way more than a failure I guess the installer would just sit there when in automatic-mode [19:01] If I wanted to display a message box in Ubiquity, from within autopartition-loop what would be a good route? [19:01] db_input? [19:02] And to make ubiquity exit? [19:06] CompBrain: you'd have to develop it [19:06] I doubt it would be in python; that's the implementation language of neither debconf nor cdebconf [19:06] you can use debconf-copydb to help [19:30] cjwatson I notice that grub-installer uses an existing device.map, how is that generated? [19:31] xivulon: grub-install (from grub) generates it [19:32] having a look [19:36] hmm cannot find where that is actually generated [19:36] I can see line 413: echo "(hd0) $bootremovable" > "$device_map" [19:36] but that does not look like it... [19:40] in turn, it gets grub to do it [19:40] $grub_shell --batch $no_floppy --device-map=$device_map <$log_file [19:40] quit [19:43] hmm that's similar to what I am using [19:43] echo quit | grub --batch --device-map=/boot/grub/device.map [19:43] but I end up with disks that do not match what is seen by grub4dos (at least accordingly to user reports) [19:44] hd0 as seen by grub4dos becomes hd1 by the time I generate device.map [19:45] nothing I can reproduce on my machines... [19:46] am I missing anything? [22:06] Hello all. I've got a quetion related to installing ubuntu(specifically the gutsy ubuntustudio.iso) From a usb-stick. I've got syslinux working and I'm using the initrd and vmlinuz from hd-media however the iso scan is failing to load the iso file on the stick. I set the ramdisk_size to 850MB in my syslinux.cfg, [22:06] Anyone had any issues getting this to work before? [22:08] I'm wondering if there the debian-installer doesn't recognize the non-standard ubuntu iso? Or mayber the iso should be renamed in 8.3 (dos) format... [22:08] Yes, the file name does in .iso. :) [22:18] Hello? [22:21] what iso are you using? is it the alternate one? [22:22] Yeah, the alternate dvd iso. [22:22] It seems to be the only one to download too... [22:23] Its the ubuntustudio-7.10-alternate-i386.iso [22:24] don't remember on top of my head what checks are done, one sec [22:24] Cool, thanks! [22:27] it checks for .disk/info (which should be there) [22:28] and for the existence of /cdrom/dists/$distlink/Release [22:28] I can mount the .iso real quick and double check...just a sec [22:29] Release is there. [22:29] allowed distros are determined by the line "for distlink in stable testing unstable ; do" in iso-scan.postinst [22:29] What did you mean by .disk/info though? [22:30] it's a file that should be in the folder .disk within the cdrom [22:31] but the reason might be a mismatch between the distlink in the initrd and the one in /cdrom/dists [22:31] Top-level of the iso contains the folders: dists, doc, install, isolinux, pics, pool, pressed and the files: cdromupgrade, md5sum.txt README.diskdefines [22:32] ls -al [22:32] cat .disk/info [22:33] Yeah, .disk is there. cat output: Ubuntu-Studio 7.10 "Gutsy Gibbon" - Release i386 (20071015) [22:34] as mentioned, you have to check distlink in iso-scan.postinst vs folders in /cdrom/dists [22:34] Anyway I can make a modified initrd to match the distlink in the cdrom? Or vice-versa? [22:34] Ah [22:35] probably easier to do an initrd... certainly much faster [22:36] Ok. How do I check the distlink in the initrd? [22:36] Find the source I'm guessing? [22:38] or extract the initrd files [22:40] Is there a quick way to do that from the term? [22:41] Konqueror doesn't seem to want to do it. [22:45] gunzip -c -9 $initrd | cpio -i -d -H newc --no-absolute-filenames [22:47] Ok, cpio gave me two Operation not permitted errors for dev/console and dev/null but that shouldn't matter. [22:57] I can't seem to find iso-scan.postinst [23:01] If my memory serves me well should be in /var/cache/lib/apt/ [23:04] Heh, /var/cache/anna contains nothing, but /var/lib/dpkg/info has it. [23:04] yeah that's the one [23:04] I'm lessing it now. [23:05] under the register_cd() funtion I've found the following line: [23:05] for distlink in stable testing unstable ; do [23:06] relfile=/cdrom/dists/$distlink/Release [23:06] if [ --e $relfile ] && [23:07] egrep -q 'Architectures:.* '$(udpkg --print-architecture)'( |$)' $relfile [23:07] then [23:17] Comment are as follows: [23:17] # Make sure that the iso is usable for the architecture. If so, [23:17] # set the suite and codename to the suite/codename that is on the CD. [23:17] # This assumes that there will be no more than one distribution on [23:17] # the CD, and that one of the testing, stable, or unstable links will [23:17] # point to it. Since the CDs currently have many links, parse the [23:17] # Release file to get the actual suite name to use. [23:23] Well from what I can tell they match. [23:25] It seems that it reads the Release file on the cdrom to see and then Checks to make sure the given iso file matches the Ubuntu release naming convention. [23:26] And the Release file reports it as simply a standard release of Ubuntu 7.10. [23:26] Since that's the case I'm going to shorten the name of the iso file on the usb-stick and see if that turns the trick. [23:27] Thanks for all the help!!