[06:51] <superm1> where are the translation files stored for people who wanted to help translate?
[06:51] <superm1> i was poking around launchpad and not coming to them
[09:16] <cjwatson> superm1: they're under /ubuntu/hardy/+source/debian-installer
[09:53] <xivulon> I took the liberty to add debug boot parameter to /etc/init.d/ubiquity
[09:53] <xivulon> hope it's ok
[09:53] <xivulon> discussed that with evand a couple of days ago'
[10:08] <soren> cjwatson: Can you give me a hint? (I think) I need to add _netdev to the fstab options of any iscsi mount point (and add a bit of code to mountall.sh to match it, but never mind that for now)... Which d-i component should I be looking at?
[10:12] <cjwatson> soren: anything that ships /lib/partman/fstab.d
[10:13] <cjwatson> though I guess you don't want to change all of them individually!
[10:13] <cjwatson> soren: partman-target/finish.d/fstab_hd_entries is probably the place to start
[10:27]  * soren looks
[10:28] <soren> cjwatson: I'm not sure I understand what you mean by changing them individually? I'm not going to involve the user in this, I hope.
[10:28] <soren> cjwatson: Oh.
[10:28] <soren> cjwatson: I get it.
[10:28] <cjwatson> I should possibly have investigated first rather than doing stream-of-consciousness
[10:29] <soren> No worries :)
[10:30]  * soren taps his finger while bzr finishes fetching partman-target
[10:45] <soren> Ah, right, networking is brought up by udev these days! So if I just start open-iscsi before mountall, I'll probably be fine.
[10:45]  * soren tests it
[10:49] <soren> \o/
[11:05] <tjaalton> new kernel is out, perhaps a time to roll out a new d-i as well?
[11:08]  * soren seconds that wish
[11:09] <tjaalton> whizzbang features for the masses! :)
[11:09] <soren> cjwatson: I'm assuming apt-install won't let me install universe packages?
[11:09] <soren> tjaalton: Yeah, virtio ftw!
[11:09] <tjaalton> :P
[11:12] <cjwatson> soren: hmm, I think probably not
[11:18] <CIA-24> debian-installer: cjwatson * r874 ubuntu/ (10 files in 4 dirs): * Move to 2.6.24-7 kernels.
[11:18] <tjaalton> whee
[11:23] <soren> cjwatson: Ok, that's good.
[11:23] <soren> cjwatson: Because then I have an explanation for my final iscsi related problem.
[11:32] <CIA-24> debian-installer: cjwatson * r875 ubuntu/debian/changelog: releasing version 20070308ubuntu28
[11:32] <cjwatson> soren: are you installing from CD?
[11:32] <cjwatson> soren: I think if you install from netboot (possibly netboot mini.iso) it'll let you use universe
[11:32] <soren> cjwatson: I'm using the mini.iso, actually.
[11:33] <cjwatson> soren: or you could add universe to /cdrom/.disk/base_components
[11:33] <cjwatson> that's a little odd then, not sure
[11:33] <soren> I'll check and see what's up.
[11:36] <t105> hello ... has anybody a clue with that one: https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/37527
[11:36] <t105> i'm having that during net install from the mini-iso
[11:36] <ubotu> Launchpad bug 37527 in initramfs-tools "Unable to install initramfs-tools in Dapper 64" [Medium,Confirmed]
[11:36] <t105> yeah thanks, botty ;)
[11:37] <cjwatson> t105: this is usually because something else failed to configure. We need to see the *full* /var/log/syslog (you can extract it using "Save debug logs" from the installer's main menu) to diagnose further
[11:37] <t105> but it seems, the regular installer is broken in some way, too: i cant install from sata-cdrom
[11:38] <t105> ah okay i'll grab the info
[11:38] <cjwatson> the comments at the end of the bug indicate that not using a separate /boot partition may be a workaround
[11:38] <t105> i'm at the point in the installer, where i can start a mini-httpd
[11:38] <cjwatson> problems installing from particular types of devices are typically kernel bugs
[11:38] <t105> yes i read that, but i'm not sure if i can do that. because i'm using a net install
[11:39] <cjwatson> I don't see how a net install is relevant
[11:39] <t105> so as i understand i would have to start up some different system
[11:39] <t105> and copy some files and edit menu.lst
[11:40] <cjwatson> that doesn't make sense to me, I'm afraid
[11:41] <cjwatson> the suggestion is simply to avoid specifying /boot as a separate partition in the installer's partitioner
[11:41] <t105> quote: The workaround suggested by BitJam worked for me as well: I restarted the installation using only /, not telling it about the existing /boot, then declined to install grub (why does it claim that's a fatal error?), then booted into another system to move the new kernel files into the real /boot partition and edited menu.lst to point to them.
[11:42] <t105> well, i'm confused about "booted into another system..."
[11:42] <t105> while... there is none =)
[11:42] <cjwatson> that's somebody who is really desperate to use a separate /boot
[11:42] <cjwatson> doing so is not usually required, except on old systems
[11:43] <cjwatson> you don't need to mess around with any of that if you're just doing a standalone install on a reasonably current machine
[11:43] <t105> ah ok. but theres one thing... my root will be on a raid1
[11:43] <cjwatson> and I assume that something with a SATA CD-ROM is fairly current
[11:43] <cjwatson> ah
[11:43] <t105> yes it's a dell t105
[11:43] <t105> it is sata-only
[11:44] <t105> theregular cd starts fine, but the installer is dumb on mounting the cd it is loaded from ;)
[11:44] <cjwatson> it is my understanding that /boot can be placed on RAID1 as long as you use LILO
[11:45] <cjwatson> that's actually unlikely to be installer dumbness
[11:45] <cjwatson> the CD booting just indicates that your BIOS can find it
[11:45] <t105> yes lilo was the possibility... some helpful guy on #xubuntu suggested that lilo is ill
[11:45] <cjwatson> whereas the installer needs the Linux kernel to be able to find it too
[11:45] <cjwatson> hence, probably a kernel bug
[11:45] <cjwatson> a newer release of Ubuntu, or even the 6.06.2-to-be CDs that are in testing at the moment, might help
[11:46] <t105> hm. what i read regulary that there is a problem with enumeration and sata-controllers...
[11:46] <cjwatson> (http://cdimage.ubuntu.com/ubuntu-server/dapper/daily/current/)
[11:46] <t105> i've downloaded xubuntu installer yesterday, that seems pretty up-to-date?
[11:46] <cjwatson> in any case, if you post your syslog I can have a look at the initramfs-tools problem
[11:47] <cjwatson> Xubuntu what version?
[11:47] <t105> 7.10
[11:47] <cjwatson> you pointed to a dapper bug, I assumed you meant that
[11:47] <cjwatson> ok
[11:47] <t105> err... no
[11:47] <t105> sorry for confusing
[11:47] <cjwatson> please post syslog then
[11:48] <t105> ok lets see (it takes some time to go through all these steps again)
[11:50] <t105> p.s. i'm using the net install atm
[11:53] <t105> { in the regular install (alternate cd) i can't seem to do anything, it just says "insert cdrom". when i press "help", there are some messages unable to mount the cd dir }
[12:05] <xivulon> any issue if I move wubi to gpl-v3? main reason is to exchange code with win32-loader
[12:07] <t105> the first error occurs (i switch to the log console): debootstrap: cannot find /lib/modules/2.6.22-14-generic, then update-initramfs: failed for /boot/initrd.img-2.6.22-14-generic
[12:07] <t105> dpkg: subprocess post-installation script returne error exit status 1
[12:10] <t105> p.s.: using the amd64 7.10. minimal cd
[12:11] <t105> cjwatson: how can i post syslog to you? it's pretty long, i guess irc is flood protected
[12:12] <cjwatson> t105: attach it to the bug; certainly don't put it on IRC
[12:12] <t105> the area where initramfs fails is:
[12:13] <t105> ah ok i need an account hm
[12:13] <cjwatson> you could also use paste.ubuntu.com, but it should really be attached to the bug otherwise it's liable to get lost
[12:14] <cjwatson> the log posted at the start of the bug sort of looks like the kernel didn't get installed, but it isn't entirely clear
[12:15] <cjwatson> soren: https://wiki.ubuntu.com/KvmVirtManagerEtc doesn't indicate it, but gfxboot just displays a blank screen for me in hardy too; is that known?
[12:15] <cjwatson> (I forget if I've asked you before)
[12:15] <soren> cjwatson: Your kernel is not up-to-date, I'm guessing?
[12:16] <soren> cjwatson: Ah.. bootstrap-base, which calls install_extra, which installs the queued packages from apt-install, has installer-menu-item 6500. apt-setup has 7000... So no universe for queued apt-install stuff.
[12:16] <cjwatson> -5 I think
[12:16] <soren> Is there a Contents file for all the installer stuff?
[12:17] <soren> cjwatson: You need -7.
[12:17] <cjwatson> base-installer will use universe if you do the base_components thing above, though
[12:17] <cjwatson> ah, ok, thanks
[12:18] <cjwatson> d-i Contents> I don't think so unfortunately
[12:18] <soren> cjwatson: Ah, yes, sure. I'm not trying to force it for local testing. I just wanted to identify if moving open-iscsi to main would fix my problem, and that seems to be the case.
[12:18] <cjwatson> nod
[12:18]  * soren -> lunch
[12:32] <t105> cjwatson: ok, here we go: https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/37527
[12:32] <ubotu> Launchpad bug 37527 in initramfs-tools "Unable to install initramfs-tools in Dapper 64" [Medium,Confirmed]
[12:50]  * t105 gets some food
[12:55] <cjwatson> t105: so were you installing onto a partition that already had some data on it?
[12:57] <CIA-24> apt-setup: cjwatson * r119 apt-setup/ (17 files in 8 dirs):
[12:57] <CIA-24> apt-setup: * Pre-populate apt's lists directory with signed Release files for
[12:57] <CIA-24> apt-setup:  archive.ubuntu.com (and mirrors) and security.ubuntu.com, to protect
[12:57] <CIA-24> apt-setup:  against downgrade attacks right from initial installation.
[13:03] <xivulon> evand, on winfoss replacement, had a preliminary look and nsis does not support onmouseover/onmouseclick.
[13:05] <xivulon> so the alternatives are, wrapping a dll plugin for the control (nlikely for friday), have a simpler dialog with 3 radio buttons + next + cancel (difficult to deuglify), have a go with pywin (~2MB)
[13:07] <xivulon> second option seems more plausible
[13:08] <xivulon> it will look a bit like our old interface: http://omattos.co.uk/ubuntuscreens/4.JPG
[13:09] <xivulon> except that the 3 radio will be for "Demo and Full Installation", "Install within Windows",  "More Info" and icons to match
[13:11] <CIA-24> apt-setup: cjwatson * r120 apt-setup/debian/changelog: releasing version 1:0.31ubuntu3
[13:13] <xivulon> Icon 1 = Tango CD, Icon 2 = ???, Icon 3 = Tango help, Header =???
[14:01] <CIA-24> ubiquity: cjwatson * r2442 ubiquity/ (5 files in 4 dirs):
[14:01] <CIA-24> ubiquity: * Fetch http_proxy and no_proxy from gconf (if possible) immediately
[14:01] <CIA-24> ubiquity:  before configuring apt.
[14:16] <evand> xivulon: given those options, perhaps heno would be more willing to use buttons for the option titles.
[14:17] <evand> imho, a clicking a radio box, then a button, is far too much work :)
[14:30]  * cjwatson wonders what the self.popcon is not None business in on_advanced_button_clicked is all about
[14:30] <cjwatson> trying to see how it could ever be other than None without getting past that ...
[14:30] <evand> hrm
[14:31] <cjwatson> oh, I see, ubiquity.components.summary does it
[14:31] <cjwatson> (I'm adding a proxy configuration bit to the advanced dialog, per networkless-installation-fixes
[14:31] <cjwatson> )
[14:33] <evand> fantastic
[14:36] <tjaalton> yeah, noshell-hack works
[14:40] <xivulon> evand, sure we can have buttons as well, but those are basic buttons no pixmap in there
[14:43] <evand> that shouldn't be a problem, but I'm curious to see what heno says in his reply given the options.
[15:10] <t105> cjwatson: err... was asleep some time =] not really existing data. it was just the 4th or so try, i just kept the partitions. reformatted them. dunno about /boot, but i'd suppose its overwritten anyway?
[15:14] <cjwatson> t105: problem is that if there's already an initramfs on /boot then update-initramfs will try to update it, and will break because the kernel isn't installed at that point
[15:14] <cjwatson> t105: are you sure you didn't skip past a warning about installing to an unclean partition?
[15:14] <cjwatson> something like this:
[15:14] <cjwatson> _Description: Proceed with installation to unclean target?
[15:14] <cjwatson>  The target file system contains files from a past installation. These files
[15:14] <cjwatson>  could cause problems with the installation process, and if you proceed,
[15:14] <cjwatson>  some of the existing files may be overwritten.
[15:15] <t105> yes there was a warning like that
[15:15] <cjwatson> it meant it
[15:15] <cjwatson> "could cause problems with the installation process" :-)
[15:15] <t105> but i could only proceed... or not?
[15:15] <cjwatson> it's a boolean question, you can answer no
[15:15] <t105> becaus it happened the first time, too
[15:16] <cjwatson> you need to format that partition
[15:16] <t105> it appeared the first time when everything was virgin, and all other times didn't change
[15:17] <t105> so i thought it has no effect
[15:17] <cjwatson> depending on the status of the previous installation, you may be lucky
[15:17] <cjwatson> the breakage is not guaranteed
[15:18] <t105> hehe, not guaranteed?
[15:19] <cjwatson> I'm not sure I believe you that everything was virgin when it appeared the first time, TBH
[15:19] <cjwatson> the check is very simple
[15:19] <cjwatson> "does /target/bin/sh exist?"
[15:19] <CIA-24> base-installer: cjwatson * r323 ubuntu/ (debian/changelog library.sh):
[15:19] <CIA-24> base-installer: * If base-installer/use_unclean_target is asked, emit a warning to the
[15:19] <CIA-24> base-installer:  logs so that we know about it when diagnosing problems.
[15:20] <t105> well i just said "reformat everything", but installer keeps on moaning that there is data.. hm
[15:20] <cjwatson> I think you must have missed something
[15:20] <cjwatson> somehow or other
[15:21] <t105> well i said "no" to the unclean install question, went back to the partitioning screen and changed everything to reformat
[15:21] <cjwatson> I would recommend rebooting at that point to make sure
[15:22] <cjwatson> it is possible to do it without that but there are ways to get it wrong
[15:22] <t105> sounds wise
[15:22] <cjwatson> soren: shouldn't there be a bug closure for that AMD family 16 thing in base-installer?
[15:24] <soren> cjwatson: Hm... Yes, I guess there should.
[15:26] <soren> cjwatson: It was marked as private, but I guess that shouldn't matter. Don't know why I thought it should.
[15:27] <cjwatson> oh, maybe best to upload without then
[15:27] <cjwatson> it can be useful for tracking though
[15:27] <cjwatson> let me know, I'll do an upload shortly
[15:27] <CIA-24> net-retriever: cjwatson * r342 ubuntu/debian/changelog: releasing version 1.18ubuntu2
[15:29] <t105> cjwatson: there was an extra option for the partition: delete all data. formatting alone did nothing. well it takes ages to delete, but maybe that's needed. i guess rebooting makes no sense, since partitioning is in the workflow and doesn't ask to reboot
[15:30] <cjwatson> oh, I have to admit I have no idea whether that's been tested recently
[15:30] <cjwatson> reboot and tell the partitioner to format the partitions right from the start
[15:30] <t105> there was an option "leave existing data" or "format  partition". the latter had no effect, i think he wasn't evven formatting anything
[15:30] <cjwatson> the latter sets a flag and causes the partition to be formatted later
[15:31] <t105> hm. :confused:
[15:31] <t105> is there an option for that? i started boot: cli
[15:32] <cjwatson> an option for what?
[15:32] <t105> for the installer
[15:32] <cjwatson> to do what?
[15:32] <t105> to partition first and what you said
[15:32] <t105> cjwatson	reboot and tell the partitioner to format the partitions right from the start
[15:33] <cjwatson> yes, of course
[15:33] <t105> yes but thats what i did... i manually configured the partitions and so it went
[15:34] <cjwatson> you can either use the guided partitioning options (erase whole disk, etc.) or use manual partitioning and make sure it says "Format the partition: yes" in the detailed view for each of the partitions you're going to mount
[15:34] <t105> yes i did enter everything manually
[15:34] <t105> because root will be on a raid partition
[15:35] <t105> just boot has an own small partition for grub
[15:35] <cjwatson> I guess it's possible there's some craziness going on with RAID
[15:35] <cjwatson> it might be worth clearing everything out first with mdadm
[15:35] <cjwatson> though that shouldn't affect /boot
[15:35] <t105> because someone suggested lilo is somewhat ill whilst can boot from raid
[15:35] <cjwatson> there are no errors in the log that might indicate trouble formatting disks, though
[15:36] <cjwatson> "ill"?
[15:36] <t105> nah, i don't remember the exact word but... so i said ok, let's use grub
[15:37] <t105> maybe it is easier to recover if /boot is a regular partition?
[15:37] <t105> and not on raid1
[15:37] <cjwatson> I thought you said you did have /boot on a regular partition
[15:37] <cjwatson> that would be likely to simplify things, yes
[15:38] <t105> so i did, /boot is a small ext3 partition
[15:39] <soren> cjwatson: I think we should put the bug nr. in the changelog. If the bug ever gets unprivated, it'll be useful. If not, no harm done.
[15:39] <cjwatson> soren: ok, go for it
[15:40] <t105> but as i said, in manual partition config i told him to reformat /boot. i thought all data would be lost? but obviously it didn't format it
[15:40] <cjwatson> this is getting to the point of being extremely hard to figure out what's going on without sitting in front of the machine and watching you operate it :-/
[15:41] <t105> heell yes ;-)
[15:41] <cjwatson> to get past it, I'd be inclined to boot in rescue mode, mount /boot, and rm everything on it
[15:41] <t105> i will wait till "delete all data" on root is ready, then i'll delete all data on /boot, and let's see what happens
[15:41] <CIA-24> base-installer: soren * r324 base-installer/debian/changelog: Add reference to bug 187869 in debian/changelog.
[15:41] <ubotu> Bug 187869 on http://launchpad.net/bugs/187869 is private
[15:42] <t105> yes kinda like that. but.. deleting 240 gig takes ages...
[15:42] <t105> i don't want to stop him right now ;)
[15:42] <t105> who knows... *magic*
[15:43] <t105> i'll go to the drugstore meanwhile *g*
[15:43] <cjwatson> 240GB? hope you have backups :)
[15:44] <cjwatson> do you mean "Delete the partition" rather than "Delete all data"?
[15:45] <t105> no, data
[15:46] <cjwatson> and this is 7.10?
[15:46] <t105> there is none, btw ;)
[15:46] <t105> well its german btw, so i translate back on the fly ;)
[15:46] <cjwatson> what is the German text? I can look it up
[15:46] <cjwatson> it always helps to know exactly what you're doing
[15:46] <t105> it sais.. dfelete all data on partition
[15:47] <t105> Lösche Daten auf RAID1 Gerät #0
[15:47] <t105> so my daughter pulls me... brb ;)
[15:47] <cjwatson> OH
[15:47] <cjwatson> you're using the crypto erase thing
[15:48] <cjwatson> you might as well let it go now I guess, but it isn't what you want
[15:48] <cjwatson> the way to erase the contents of a partition is to format it
[15:49] <cjwatson> the crypto erase tool might well not even work properly on RAID; I'm honestly not sure
[15:49] <cjwatson> I wonder how you ended up seeing that; that option is only presented if you tell the installer to make that partition encrypted
[15:57] <soren> cjwatson: Hmm... preseed is not in bzr?
[15:57] <soren> cjwatson: Is that on purpose?
[16:00] <cjwatson> it's not modified in Ubuntu
[16:00] <cjwatson> and looks like we've never imported it
[16:00] <cjwatson> you can either go through the dance to register and import it, or else just upload a modified source package
[16:02] <CIA-24> hw-detect: soren * r77 hw-detect/ (debian/changelog debian/disk-detect.templates disk-detect.sh): Add preliminary iSCSI support.
[16:03] <soren> I'll just upload it for now, and add the import thing to my todo list. I'm under a bit of time pressure.
[16:06] <cjwatson> sure
[16:06] <cjwatson> (that's what I'd probably do too, though I have got a bunch of stuff imported in the past)
[16:07] <soren> ISTR that you did special magic to import certain bits? Something about you having history that wasn't available elsewhere?
[16:14] <CIA-24> hw-detect: soren * r78 hw-detect/debian/changelog: Releasing 1.58ubuntu2
[16:17] <soren> cjwatson: It seems more appropriate if it's the installer team that imports preseed, though. </lameexcuse>
[16:17] <soren> :)
[16:17] <t105> cjwatson: back. ehm.. no, i didn't select any encryption stuff. as the base package failed, i went back to the oartition menu, and voila there was that option. but not mentioned about encryption
[16:20] <t105> btw, "erase" is still 84%. i suppose: the data, if any, is gone. ;D
[16:28] <evand> Is it reasonable to make the free space slider work over the entire partition, stopping at the min and max sizes, and converting that value to the percentage of the maximum percentage of the free space behind the scenes?
[16:28] <evand> Or would it be more clear to continue working in percentage of available space.
[16:29] <evand> cjwatson: ^
[16:40] <cjwatson> soren: you can be in the installer team if you like
[16:40] <cjwatson> soren: as long as you're prepared with mail filters, since ubuntu-installer is subscribed to all ubiquity bugs
[16:41] <cjwatson> soren: I did by-hand imports for ones that had historically had lots of Ubuntu uploads
[16:41] <cjwatson> (based on Launchpad imports) but that's not the case here
[16:41] <cjwatson> evand: my gut reaction is that the former gives you a better idea of scale
[16:41] <evand> yay, I was hoping you'd say that
[16:42] <evand> now to just figure out how to do it
[16:42] <soren> cjwatson: Re becoming member of installer team: I was afraid you were going to say that :)
[16:42] <evand> (getting the real size of the partition that is, as the question itself is of no help)
[16:42] <soren> cjwatson: I'll think about it.
[16:42] <evand> I've already got the min and max stopping done, so that's not an issue.
[16:48] <t105> *whooot*? i'm trying to repartition my hds, but i can't delete a new raid md1... but i cant't: "probably in use"
[16:49] <t105> the partition tool seems to get confused... hm
[16:51] <t105> well ok it is deleted, but i can't remove the first raid1 entry.
[16:54] <t105> let's see, if it is still there after reboot . i deleted ALL partitions and i hope he git it
[17:06] <t105> q: is there a way to start partitioning from the installer in a way i can delete all partitions? because i have no chance to do that from the regular install routine, as it keeps asking for root before writing partition tables
[17:09] <t105> kinda boot option boot: partman
[17:17] <seb128> hi
[17:17] <seb128> is there a known bug about having no LANG or LANGUAGE set to /etc/environment on new installs?
[17:25] <cjwatson> t105: not as a boot option; normally you would just delete all partitions, create the new ones, and *then* say "Finished with partitioning"
[17:25] <cjwatson> t105: it only issues the warning you're talking about if you try to proceed past partitioning without creating some partitions that the rest of the installer can work with
[17:28] <cjwatson> t105: that said, if that isn't working, then a possible workaround would be to delete all the partitions, then select "Configure the Logical Volume Manager"; that ought to cause it to prompt you whether you want to write changes to disk, at which point you'd say yes
[17:28] <cjwatson> then you can create the partitions you want
[17:29] <cjwatson> seb128: that was intentionally removed, and it only writes to /etc/default/locale now
[17:29] <cjwatson> seb128: the changelog is:
[17:29] <cjwatson>   * No longer write locale settings to /etc/environment after #389466 was
[17:29] <cjwatson>     fixed.
[17:29] <evand> localechooser sets /etc/default/locale presumably rather than /etc/environment.  Is this an error?
[17:29] <cjwatson> seb128: is that bug fixed in Ubuntu's gdm?
[17:29] <evand> ah, nevermind
[17:29] <cjwatson>    * Parse /etc/default/locale in gdm pam files, for now (closes: #389466)
[17:29] <cjwatson> seems to be fixed in our gdm too
[17:30] <cjwatson> auth    required        pam_env.so readenv=1 envfile=/etc/default/locale
[17:30] <seb128> cjwatson: no it's not, which means currently english is used for everybody on login
[17:30] <cjwatson> seb128: so what's the line above doing? :-)
[17:30] <cjwatson> /etc/default/locale is becoming standard, and is neater - I think we should honour it
[17:31] <seb128> cjwatson: well, /etc/init.d/gdm has
[17:31] <seb128> if [ -r /etc/environment ]; then
[17:31] <seb128>   if LANG=$(pam_getenv -l LANG); then
[17:31] <seb128>     export LANG
[17:31] <seb128>   fi
[17:31] <seb128>   if LANGUAGE=$(pam_getenv -l LANGUAGE); then
[17:31] <seb128>     export LANGUAGE
[17:31] <seb128>   fi
[17:31] <seb128> fi
[17:31] <seb128> which was how we used to set the LANG and LANGUAGE to use
[17:32] <seb128> cjwatson: is pam_getenv supposed to use /etc/default/locale?
[17:33] <seb128> because pam_getenv -l LANG gives nothing
[17:33] <seb128> and there is no LANGUAGE set /etc/default/locale neither
[17:34] <cjwatson> hmm, might be worth bugging slangasek about this; I think pam_getenv -l should nowadays also check /etc/default/locale
[17:34] <cjwatson> its manual page hints at that
[17:35] <cjwatson>         # We set LANGUAGE only if the languagelist is a list of
[17:35] <cjwatson>         # languages with alternatives. Otherwise, setting it is useless
[17:35] <t105> cjwatson: the problem was, i deleted all partitions, but somehow an old raid device survived and i couldn't get rid of it. i made a simple root partition, started, rebooted, and then i was able to delte partitions and NO old raid was there.
[17:35] <cjwatson>         if echo "$LANGLIST" | grep -q ":"; then
[17:35] <cjwatson>                 sed -i 's/^# LANGUAGE=$/LANGUAGE=\"'"$LANGLIST"'\"/' $DESTFILE
[17:35] <cjwatson>         fi
[17:35] <cjwatson> seb128: in other words it's not an error if LANGUAGE isn't set there
[17:36] <t105> cjwatson:  btw, i'm doing a fresh install. i selected another mirror (ireland) this time, so let's see if it works
[17:36] <cjwatson> t105: ok, I'm sorry but at the moment you're outside my experience
[17:37] <seb128> cjwatson: ok
[17:37] <t105> i was just wondering if there might have been some inconsistent mirror last time my install broke
[17:37] <t105> does that happen?
[17:38] <cjwatson> it is possible, but would not match your symptoms
[17:39] <t105> because... i did everything as the first time but used other mirror, now it sais "install ok"
[17:40] <t105> and voila, kernel is up
[17:42] <t105> *phew* that was one hard thingy
[17:43] <t105> now it's installing xubuntu ;-)
[17:43] <t105> cjwatson: thank you very much!
[17:45] <cjwatson> interesting
[17:45] <cjwatson> glad to help, even though I'm still not entirely sure what went wrong ...
[17:48] <t105> yes, me too
[17:49] <t105> but as i read the syslog i had the feeling, that some packages where installed in wrong order - or wrong place
[17:49] <t105> whatever that had caused
[17:50] <cjwatson> no, I understand where that impression came from but it's not the case
[17:51] <cjwatson> initramfs-tools is installed before the kernel, correctly so - however it is a bug in initramfs-tools that it fails to handle being configured on a system that has /boot/initrd.img-BLAH but not /lib/modules/BLAH
[17:51] <t105> ;-)
[17:51] <cjwatson> evand: this grub-installer stuff in summary.py is breaking for me
[17:52] <cjwatson> evand: I can see one obvious fix (it's handling subp.communicate output wrongly), but it seems weird for it to be using os-prober to get the list of possible targets for grub at all?
[17:52] <t105> hm. looking at the problems those other people had, maybe there is a bug in initramfs-tools?
[17:52] <cjwatson> t105: yes, there is
[17:52] <cjwatson> t105: that is why there is a bug open on initramfs-tools ;-)
[17:53] <evand> cjwatson: Ok, I'll take a look at it.  The use of os-prober is to get human readable names.
[17:53] <cjwatson> evand: right, but it only works if there's something there already
[17:53] <t105> well... then this is a monday bug. or a bug depending on nationality (server) {insert wild guessing here}
[17:53] <cjwatson> oh
[17:53] <cjwatson> I think I understand
[17:54] <cjwatson> t105: I think I explained it adequately above, at least to my own satisfaction ;-)
[17:54] <evand> right, so if a partition isn't going to be there in the future it doesn't include it in the list.
[17:55] <CIA-24> ubiquity: cjwatson * r2443 ubiquity/ubiquity/components/summary.py: handle the output of subp.communicate correctly; [0] is stdout, then splitlines to get a list of items one per line
[17:55] <cjwatson> evand: how does that look? beforehand it crashed for me (os-prober's output was empty)
[17:57] <evand> ah, much better.  Sorry about that, I thought I tested os-prober being empty, but apparently not.
[17:57] <cjwatson> looks better now. slightly odd that it still defaults to (hd0) which isn't in the list though
[17:57] <cjwatson> maybe default to the first disk instead? but otherwise it looks great, an excellent improvement
[17:58] <evand> that's still somewhat I work in progress.  I have to finish off the KDE bits and obviously do some more testing.  But I've been stock piling ubiquity branches as of late :/.
[17:58] <evand> ah, will do
[17:59] <evand> and thanks
[18:52] <CIA-24> ubiquity: cjwatson * r2444 ubiquity/debian/ (changelog ubiquity.install-any): * Update installed file list for apt-setup 1:0.31ubuntu3.
[19:27] <cjwatson> hmm, apt-setup's signed Release handling is a little bit broken wrt pockets; will fix when I get a chance
[19:28] <cjwatson> in fact, very broken. argh.
[20:06] <CIA-24> ubiquity: cjwatson * r2445 ubiquity/ (87 files in 6 dirs):
[20:06] <CIA-24> ubiquity: * GTK frontend:
[20:06] <CIA-24> ubiquity:  - Add proxy configuration to the advanced dialog (LP: #131421).
[20:19] <CIA-24> ubiquity: cjwatson * r2446 ubiquity/ (debian/changelog ubiquity/components/summary.py):
[20:19] <CIA-24> ubiquity: * Import DebconfError correctly in summary component (thanks, Jesus L.
[20:19] <CIA-24> ubiquity:  Alvaro; LP: #177442).
[21:56] <evand> cjwatson: do you think this is reasonable? http://pastebin.ubuntu.com/4359/
[22:47] <t105> cjwatson: i found some strange entries in my debug.0 : "... CFI: Found no ck804xrom @ffc00000 device at location zero
[22:47] <t105> JEDEC: Found no ck804xrom @ffc00000 device at location zero"
[22:48] <t105> googles has some resources about that... nvidia... driver... >=4 gb ram .... kernel-patch?... well that kind of stuff no one wants to run into ;)
[22:49] <t105> i guess there is no chance including nvidia-stuff in the installer
[23:14] <t105> hm, sata-nv IS in ubuntu... strange
[23:52] <cjwatson> evand: I think you should be able to use PARTITION_INFO $oldid rather than iterating over PARTITIONS
[23:53] <cjwatson> evand: I'm having a hard time figuring out what totsize actually is - it doesn't seem to be the total of anything. Is it the original size of the partition being resized?
[23:53] <cjwatson> t105: can't help you here, try #ubuntu-kernel
[23:53] <cjwatson> the installer is strictly userspace :)
[23:54] <cjwatson> evand: maybe use $hcursize instead?
[23:55] <cjwatson> in fact - looks like you could just use the value of new_size, unless I'm misunderstanding something; the current size returned by GET_RESIZE_RANGE should be the same as what you get back from iterating over PARTITIONS
[23:56] <cjwatson> but I haven't seen the ubiquity code so am not entirely sure