[11:58] <saispo> hi
[11:58] <saispo> how can i force to include ide-modules-2.6.17-10-generic-di on my custom cd ?
[11:59] <saispo> i have an error on ide-modules in syslog when try to format :/
[11:59] <cjwatson_> make sure it's in the installer seed
[12:01] <saispo> hi cjwatson :)
[12:03] <saispo>  * /^.*-modules-.*-386-di/
[12:03] <saispo> i have this...
[12:04] <cjwatson> which does not match anything-generic-di.
[12:05] <saispo> yep
[12:05] <saispo> readd this
[12:05] <cjwatson> in edgy we used the 386 kernel in the installer though
[12:05] <cjwatson> if you've changed that then you'll have to make corresponding seed changes
[12:05] <saispo> yep
[12:06] <saispo> but i don't understand why i have an ide error module...
[12:06] <cjwatson> nor do I; you're much better-placed to debug it locally
[12:07] <saispo> hw-detect: Missing modules 'ide-mod (Linux IDE driver),
[12:07] <saispo> ide-probe-mod (Linux IDE probe driver), ide-detect (Linux IDE detection),
[12:07] <saispo> ide-floppy (Linux IDE floppy)
[12:07] <saispo> i have this exactly
[12:07] <saispo> and when i use a scenario for partitioning, i have an error with "no system file founds"
[12:07] <saispo> (i backport partman-auto)
[12:08] <saispo> partman:   No matching physical volumes found
[12:08] <saispo> i have this too
[12:11] <cjwatson> ignore those warnings. You've misdiagnosed them.
[12:12] <saispo> ok, you think, the error is not in this track ?
[12:12] <cjwatson> the hw-detect ones, I mean
[12:12] <cjwatson> try manual partitioning and see if it presents any disks
[12:12] <saispo> manual partitioning work
[12:13] <cjwatson> then it is not a matter of missing modules, obviously
[12:13] <saispo> hw-detect must be backported ?
[12:13] <cjwatson> if you think about it
[12:13] <cjwatson> NO
[12:13] <cjwatson> stop flailing around randomly :-)
[12:14] <cjwatson> "No matching physical volumes found" is irrelevant as well
[12:14] <cjwatson> what's the exact error you get from autopartitioning?
[12:15] <saispo> no system files have been choosed for the root (/) partition
[12:15] <cjwatson> is that really the exact error? (it contains a grammatical mistake)
[12:19] <cjwatson> http://www.chiark.greenend.org.uk/~sgtatham/bugs.html
[12:19] <cjwatson> specifically the "Works for me. So what goes wrong?" section
[01:00] <CIA-4> ubiquity: jriddell * r1875 trunk/ubiquity/frontend/ (5 files): merge in kde-advanced-partitioner
[01:01] <Riddell> that CIA stuff is hassle to set up
[01:01] <cjwatson> yeah, is just an initial cost though
[01:03] <CIA-4> ubiquity: jriddell * r1876 trunk/debian/ (changelog ubiquity-frontend-kde.install): Add a changelog entry and add new .ui files to debian/ubiquity-frontend-kde.install
[01:05] <CIA-4> ubiquity: cjwatson * r1877 ubiquity/ubiquity/frontend/gtkui.py: revert spurious debugging output
[01:08] <CIA-4> ubiquity: cjwatson * r1878 ubiquity/ubiquity/frontend/ (kde-ui.py partition_create_dialog.ui): spell dialog consistently (I know, it's the US spelling, but it will make things easier if it's consistent)
[01:11] <cjwatson> otherwise looking pretty good; thanks for that!
[01:11] <cjwatson> I'll upload later today
[01:12] <CIA-4> ubiquity: cjwatson * r1879 ubiquity/debian/changelog: stray tab
[01:15] <CIA-4> ubiquity: cjwatson * r1880 ubiquity/ (configure configure.ac): bump to 1.3.22
[01:16] <Riddell> oops, I only checked kde-ui for debugging to remove
[01:19] <cjwatson> no problem
[01:20] <Riddell> cjwatson: I take it you plan to add some buttons at some point, so people don't have to discover the right click menu?
[01:20] <cjwatson> yeah
[01:20] <cjwatson> was planning to do that pretty soon actually
[01:21] <cjwatson> they'll be needed once the disk bar UI gets in
[01:21] <Riddell> yep
[01:22] <Riddell> cjwatson: still planning on ubiquity slideshow for feisty?

[01:22] <Riddell> I see
[01:22] <cjwatson> not sure, it's getting late and I still need to sit down with Ken and talk about artwork for it
[01:23] <Riddell> evand: what are the chances of migration assistant for Kubuntu Feisty?
[01:24] <cjwatson> Riddell: I think you missed it in the KDE .ui, but I included a stub partition_list_buttonbox widget in the GTK one
[01:30] <evand> Riddell: that would be difficult as I have not had the time to write the functions for importing into the KDE equivs, but when I finish bug fixing I'll start working on it.  I think it's too late for Feisty though.
[01:35] <Riddell> evand: ok, a shame but entirely understandable
[01:37] <evand> Riddell: Indeed, but don't let this be an indication of my commitment to getting m-a working for both DEs.  I will definitely start coding for it as soon as I get the chance.
[01:39] <evand> hah, thanks
[01:40] <cjwatson> cr3: gtk-on-next is a bug, yes, and I could believe that
[01:40] <saispo> cjwatson: excuse me, my boss want to see me
[01:40] <cjwatson> cr3: copying the desktop CD to a partition and booting from it is probably not a great test, I'm afraid
[01:41] <saispo> cjwatson: it's my translated message of the french message ;)
[01:41] <cjwatson> saispo: I would rather have the untranslated message. I can look it up myself.
[01:41] <saispo> ok
[01:41] <cr3> cjwatson: darn, how is booting from a partition different?
[01:41] <saispo> will give you
[01:41] <cjwatson> saispo: when I ask for an exact error message, it's because I want to grep for it. Home-made translations don't help me at all.
[01:43] <saispo> i understand
[01:43] <cjwatson> cr3: actually, looking at the relevant casper code, it might be ok
[01:43] <cjwatson> cr3: just make sure you copy the whole CD and not just the squashfs, as ubiquity uses files outside the squashfs
[01:44] <cjwatson> cr3: /preseed/ubuntu.seed can be used to preseed the desktop installation, but the extent to which the desktop installer is preseedable is still quite limited; see the ubiquity-automation spec
[01:46] <cr3> cjwatson: I copy the integrality of the desktop cd to its own partition. unfortunately, I have to cp, I haven't managed to dd and then boot off the iso partition.
[01:48] <cjwatson> cr3: if you can get it to boot at all, I suppose it will be fine
[01:56] <saispo> cjwatson: the exact message in french is : "Aucun systme de fichier n'a t choisi comme racine"
[01:56] <saispo> "Veuillez corriger cela a partir du menu de partitionement"
[01:57] <cjwatson> ah, partman-target/no_root
[01:57] <cjwatson> saispo: could you put your preseed file somewhere I can see it, with any passwords removed?
[01:58] <saispo> cjwatson: ok, i can
[01:58] <saispo> thks
[02:01] <saispo> cjwatson: http://www.zarb.org/~saispo/saispo.seed you can see it here
[02:03] <cjwatson> saispo: looks ok, I think; can I see /var/log/syslog and /var/log/partman from the installer too?
[02:03] <saispo> ok
[02:04] <cjwatson> saispo: err, you still have d-i    partman-auto/disk string /dev/discs/disc0/disc set there
[02:04] <cjwatson> saispo: you should remove that with the partman-auto you've backportede
[02:04] <cjwatson> backported
[02:05] <saispo> ok
[02:05] <saispo> will try without it :)
[02:09] <saispo> cjwatson: same error
[02:10] <saispo> it's not my scenario ?
[02:10] <saispo> because in partman log, i see sda
[02:10] <saispo> you want the log ?
[02:15] <cr3> I'm attempting to preseed edgy and the installer is prompting on Partition disks. wasn't there some 'manage' or somesuch configuration to define for partman?
[02:15] <saispo> http://www.zarb.org/~saispo/log.partman, http://www.zarb.org/~saispo/log.syslog
[02:15] <saispo> cr3: i try to preseed edgy too and have some error with partman
[02:16] <cr3> aha! I remember what I was thinking of, it was partman-auto/method, but I specified that.
[02:17] <cr3> saispo: I remember getting preseeding to work with edgy when partitioning the whole drive, but now I'm trying partman-auto/expert_recipe
[02:17] <saispo> cr3: same as me
[02:17] <cr3> saispo: interesting!
[02:18] <saispo> cr3: http://www.zarb.org/~saispo/saispo.seed it's my preseed but don't work
[02:18] <saispo> cr3: have you an example of a preseed file for the whole drive ?
[02:18] <cjwatson> cr3: saispo has an unusual scenario because he's backporting partman-auto
[02:19] <saispo> cjwatson: i can remove it :)
[02:19] <cjwatson> any chance of me actually having time to analyse your log files? :-P
[02:19] <saispo> cjwatson: i can test withtout a backported partman if you want
[02:20] <cjwatson> saispo: that log is from an installation run where you did manual partitioning
[02:20] <cjwatson> saispo: I want logs from an autopartitioning run
[02:20] <saispo> k
[02:20] <saispo> cjwatson: i have :)
[02:20] <cr3> saispo: the only difference I can notice between our partman configuration is that I use: d-i partman-auto/expert_recipe string boot-root :: ...
[02:20] <cjwatson> cr3: partman-auto/method is only relevant in feisty
[02:21] <cr3> whereas you use: d-i partman-auto/expert_recipe     string root ::
[02:21] <cjwatson> (both are wrong, incidentally, although not importantly so - you should use : not ::)
[02:21] <cr3> saispo: I don't know the difference between boot-root and root though, perhaps I found it on feisty by accident :(
[02:21] <cjwatson> (but this is probably not the cause of this problem)
[02:21] <cjwatson> cr3: they're arbitrary identifiers; makes zero difference
[02:22] <saispo> cjwatson: http://www.zarb.org/~saispo/log.partman, http://www.zarb.org/~saispo/log.syslog
[02:22] <saispo> new log
[02:22] <saispo> you can see no partition detection
[02:23] <cjwatson> saispo: which simply means the disk is blank
[02:23] <cjwatson> it detected the disk fine
[02:23] <saispo> yep
[02:23] <saispo> but why he don't create the partion ?
[02:23] <cjwatson> saispo: could you update the preseed file you're using?
[02:24] <cjwatson> http://www.zarb.org/~saispo/saispo.seed
[02:24] <saispo> yes i can update it
[02:24] <saispo> with what ?
[02:24] <cjwatson> your current version
[02:24] <cjwatson> 13:05 <saispo> will try without it :)
[02:24] <saispo> i try with this
[02:24] <saispo> but not work
[02:24] <cjwatson> urgh, hang on
[02:24] <cjwatson> you backported partman-auto. How much of it?
[02:24] <saispo> i take partman-auto from feisty
[02:25] <saispo> i put it on my local repository, he ovveride the partman-auto from edgy
[02:26] <cjwatson> saispo: I suspect this simply won't work without a local installer expert. Sorry.
[02:27] <cjwatson> I think you'll just have to use 'd-i partman-auto/disk string /dev/sda' for now and deal with having to have different preseed files for different machines.
[02:27] <saispo> ok :(
[02:27] <cr3> I copied syslog, partman and my preseed file here: http://people.ubuntu.com/~cr3/preseed/edgy/
[02:27] <cjwatson> I did warn you that this wasn't a guaranteed option
[02:28] <saispo> i say...
[02:28] <cjwatson> cr3: your disk is hda, but your preseed file says sda
[02:28] <cjwatson> cr3: fix that and it should work
[02:28] <saispo> cjwatson: it's possible to have a hook wich detect if it's an hda or an sda and load a template for partition ?
[02:29] <cjwatson> saispo: yes, kickstart does that, but I'm afraid I don't have time to run you through it, sorry
[02:30] <saispo> ok, i understand, no problem
[02:30] <saispo> if [ "$(cat /sys/block/hda/devices/media)" = cdrom ] ; then echo part-auto-sda.cfg; else echo part-auto-hda.cfg; fi
[02:30] <saispo> this things may work ?
[02:35] <cjwatson> maybe
[02:35] <saispo> cr3: you use DEBCONF_DEBUG=3 ?
[02:35] <cjwatson> I really don't want to have to debug that if it goes wrong though
[02:35] <cjwatson> DEBCONF_DEBUG=5 surely
[02:41] <saispo> k
[02:41] <saispo> thks
[02:43] <cr3> yes, DEBCONF_DEBUG=5
[02:44] <cr3> cjwatson: thanks, that's the second time I get bitten by that problem, I need to make it a variable interpolated per machine :(
[02:58] <cr3> I'm using the same partman-auto/expert_recipe as on http://people.ubuntu.com/~cr3/preseed/edgy for another machine and I'm getting: Can't have a partition outside the disk.
[03:05] <saispo> cr3: you thing a variable work for switching between sda or hda ?
[03:07] <cjwatson> I seriously recommend you live with two preseed files until feisty's out
[03:08] <cjwatson> cr3: can I get the syslog for that?
[03:11] <saispo> cjwatson: yep
[03:11] <saispo> i will begin my job under feisty when i finish to fix bug of my edgy support
[03:11] <saispo> :)
[03:11] <saispo> ls
[03:21] <CIA-4> ubiquity: cjwatson * r1881 ubiquity/ubiquity/ (components/partman.py frontend/gtkui.py frontend/kde-ui.py): move allowed-actions logic into Partman component
[03:31] <CIA-4> ubiquity: cjwatson * r1882 ubiquity/ubiquity/frontend/ (gtkui.py kde-ui.py): more informative partman menu entries
[03:34] <cr3> cjwatson: http://people.ubuntu.com/~cr3/preseed/feisty/syslog
[03:34] <cjwatson> cr3: partman would be good too
[03:36] <cr3> cjwatson: uploaded
[03:37] <CIA-4> ubiquity: cjwatson * r1883 ubiquity/ubiquity/frontend/ (gtkui.py kde-ui.py): fix get_actions arguments
[03:39] <cjwatson> cr3: do you know if the disk has one of those Host Protected Area things?
[03:39] <cjwatson> 'cos it looks like the kernel may be lying to partman about the size of the disk ...
[03:40] <cjwatson> cr3: I'd try to reproduce it with parted from the command line, and if you can, take it to a kernel bug
[03:48] <cr3> cjwatson: I'll try a few more machines with the same preseed file
[03:54] <cr3> cjwatson: you were right! what's ironic is that I had actually tried it on two machines beforehand and the two returned the same error. I then tried it on a few other machines after your diagnosis and they're working perfectly.
[03:55] <cr3> I'll log a bug about the other machines, did you notice anything in either syslog or partman that might be useful to highlight in my bug report?
[03:55] <cjwatson> no, better to just attach the whole thing
[03:56] <CIA-4> ubiquity: cjwatson * r1884 ubiquity/ (debian/changelog ubiquity/frontend/gtkui.py): - Add contextual buttons to GTK frontend (part of LP: #83166).
[03:56] <saispo> cjwatson: i decided to wait for feisty for auto partitioning
[03:56] <saispo> i let the user to partition the disk with the feisty release
[03:56] <saispo> it will be a new feature for the next release ;)
[04:02] <cjwatson> cr3: btw, your gtk-go-forward bug is 86047
[04:05] <CIA-4> ubiquity: cjwatson * r1885 ubiquity/ (3 files in 2 dirs):
[04:05] <CIA-4> ubiquity: * Retranslate Forward button when going back from summary page
[04:05] <CIA-4> ubiquity:  (LP: #86047).
[04:09] <CIA-4> ubiquity: cjwatson * r1886 ubiquity/debian/changelog: LP: #85642 coincidentally fixed
[04:18] <CIA-4> ubiquity: cjwatson * r1887 ubiquity/ (debian/changelog ubiquity/frontend/gtkui.py): * Fix up step numbers if migration-assistant is disabled (LP: #85711).
[04:20] <CIA-4> ubiquity: cjwatson * r1888 ubiquity/debian/changelog: reorg
[04:29] <cr3> I'm getting:  The following packages have unmet dependencies: ubuntu-desktop: Depends: gdm but it is not going to be installed, Recommends: ttf-thai-tlwg but it is not going to be installed
[04:30] <cr3> that's probably a problem with my mirror or perhaps in the way I'm defining mirrors...
[04:32] <cr3> weird, the package is indeed on the mirror: pool/main/t/thaifonts-scalable/ttf-thai-tlwg_0.4.4-0ubuntu1_all.deb
[04:49] <cjwatson> that's generally a consequence of other breakage; look for the first error rather than the last, and/or retry with -o Debug::pkgProblemResolver=true to the relevant apt tool
[04:49] <cjwatson> unlikely to be an installer problem
[04:56] <cr3> cjwatson: I'm not sure where I could specify that option. I suspect the problem is in the multiselect, so would I define something like this in my preseed: tasksel tasksel/first -o Debug::pkgProblemResolver=true multiselect ubuntu-standard, ubuntu-desktop
[04:56] <cjwatson> no no no
[04:56] <cjwatson> not talking about the preseed file
[04:57] <cjwatson> I mean chrooting in by hand after it all goes pear-shaped
[05:05] <cr3> there seems to be a conflict between gdm and gnome-session, which might be caused by my multiselect which specified ubuntu-standard and ubuntu-desktop. I'll try with just ubuntu-desktop as specified in the ubuntu.seed file on the alternate cd
[05:08] <cr3> hm, that really doesn't make sense considering both gdm and gnome-session are in ubuntu-desktop metapackage :(
[05:10] <cr3> gdm: Breaks: gnome-session (<= 2.17.91-0ubuntu1) but 2.17.91-0ubuntu1 is to be installed, so there's probably a problem with my ordering of mirrors
[05:14] <cjwatson> oh, if it's feisty, then this sort of thing happens transiently
[05:15] <cjwatson> generally the answer (if you're not in core-dev) is "wait and try again later"
[05:15] <cr3> cjwatson: crap, I realy on the alternate install a lot, maybe current is fixed...
[05:19] <cr3> or maybe I can fix it temporarily by mounting a unionfs over the iso and copying another gdm or gnome-session package over it, but that probably means I'll have to regenerate the Release files
[05:19] <cjwatson> cr3: the milestone releases shouldn't have this problem. Dailies may be arbitrarily broken
[05:19] <cjwatson> noooooo
[05:19] <cjwatson> use a milestone :)
[05:19] <cr3> I am using the milestone
[05:19] <cr3> I will test the alternate from CD just in case the problem is with my preseeding
[05:20] <cjwatson> hmm, could be, that sort of problem shouldn't have got past basic testing
[05:20] <cr3> cjwatson: I have had alternate install failures on herd 2, so I installed everything manually
[05:21] <cr3> I mean, I installed the desktop manually, I couldn't test the alternate at all
[05:38] <cr3> the alternate install from CD worked, so the problem is indeed with my preseed file
[05:40] <cr3> I wonder when mirror/http/hostname is being used and when apt-setup/local0/repository is being used...
[05:44] <cr3> so, I added the mirror to the local0 repository and incremented the number of the other repositories, just in case that made a difference
[06:13] <cr3> aha! so the problem was that the installation seems to be alternating randomly between repositories when they specify the same target
[06:36] <cr3> damnation! now the problem is that the alternate CD only has a partial repository on it! I would need to mirror main and restricted for a milestone which I don't think is possible
[10:55] <stephans> anyone awake here?
[10:59] <cr3> stephans: yep, what's up?
[10:59] <stephans> hey
[11:00] <stephans> are the repositories broken on the cds?
[11:00] <stephans> I have loop mounted the cd and also the dvd to a location on a web server...
[11:00] <cr3> stephans: they're partial, but not broken as far as I know
[11:00] <stephans> ah
[11:01] <cr3> stephans: the dvd might be complete though, what seems to be missing or conflicting or somesuch?
[11:01] <stephans> when i try to do an install off of the loopmounted cd the installer can not find the file:
[11:01] <stephans> cdrom-core-modules-2.6.17-10-386-di_2.6.17.1-10.34_i386.ude
[11:01] <stephans> b
[11:02] <stephans> This is a proble both with the cd and the dvd.
[11:02] <stephans> if I remove the instruction ks=... and let the install go off the inet... then it is ok
[11:03] <stephans> but takes for ever...
[11:03] <cr3> stephans: you can specify alternate repositories in your preseed file, don't know about kickstart
[11:03] <stephans> the ks.cfg contains the source path to the web server with the loop mounted iso of the dvd
[11:04] <stephans> kickstart is supposed to be the same thing...
[11:04] <stephans> it is also easy to use...
[11:04] <cr3> stephans: ok, so specify the loop mounted iso as primary mirror and another mirror as secondary
[11:04] <stephans> hmmm...
[11:05] <stephans> ok you stumped me...
[11:05] <stephans> I do not know how to do that..
[11:05] <cr3> I don't know how to do that with kickstart either
[11:05] <stephans> I have not found a good how to for the preseed stuff...
[11:05] <cr3> stephans: the installation-guide-i386 package contains a useful example file
[11:05] <stephans> ok
[11:05] <cr3> dpkg -L installation-guide-i386 | grep example
[11:07] <stephans> ok have to install it....
[11:10] <stephans> what is the supported or official way to do unattended installs with ubuntu?
[11:22] <cr3> stephans: kickstart and preseed are supported and official