[06:09] Any unetbootin devs around? [11:43] cjwatson, was reviewing initramfs/scripts/local, in relation to bug #226622 and noticed a FIXME (I belive from you) for the mount line [11:43] # FIXME This has no error checking [11:44] Would it be possible to put that errorchecking in (at this stage) so that we also address #226622 [11:44] (and possibly do something similar for casper) [11:44] I don't think that was from me [11:45] the error checking (such as it is) happens later when the initramfs system notices that the root device hasn't really appeared [11:45] but yes, it does produce something unclear; though I'm a little worried that there may be some expected failures at that point [11:47] can we maybe save $? and have a proper panic message within the if [ "LOOP" ] so that it is specific to wubi? Ideally the message should appear also if a user is on usplash [11:48] err, I'm really not sure, I didn't write that originally - it sounds reasonable [11:50] a clear error when ntfs mount fail because of unclean shutdown, would greatly reduce the number of tickets [12:08] +/- http://paste.ubuntu.com/10492/ [12:11] doing that in casper might be slightly trickier though, since the partition is found heuristically [13:13] in lupin-casper we could have an error message if iso-scan/filename is set but no ISO is found and similarly in custom-installation [13:13] we should also check whether the filesystem is mounted ro instead of rw. [13:13] evand ^^ [14:04] indeed, that looks roughly reasonable. [14:10] great, I will do a proper diff later on then, for initramfs-tools and upload some new code into lupin/hardy.proposed as well [14:11] ps any idea when we will have new dailys with the latest fixes? [14:12] cjwatson: ^ I've been wondering this myself. Is that known yet, or is it still in the planning stage? [14:15] error codes for ntfs are listed in man ntfs-3g.probe(8) [14:17] do we have error-remount-ro type of behaviour to worry about? [14:18] evand: for hardy you mean? [14:18] cjwatson: sorry for not being specific, yes [14:18] evand: poke slangasek about that - he has all the equipment now to do it [14:18] will do [14:38] evand, at the end of autopartition-loop we run `apt-install lupin-suport`, when is lupin-support actually installed? in particular is it ok to call update-initramfs in .postinst and is /etc/default/locale available at that stage? [14:39] this is re #136682 [14:41] I would assume it is fine, but would like confirmation [14:46] checking [14:52] thx [14:52] at least in the case of ubiquity, that happens sufficiently late for that to work. And yes, that should be fine. [14:53] good, I'll let Tormod know [15:11] xivulon,evand: fancy doing a demo of wubi at UDS? [15:12] xivulon: I gather you're only there part of the week? [15:13] sure, I'd love to help with that [15:13] cjwatson: absolutely [15:13] I will be there the last two days, so we can do that thu or fri! [15:14] ok, cool [15:14] I've asked Scott et al to pencil that in [15:14] not with my laptop though because it does have ACPI issues :( (#146692) [15:14] cjwatson: speaking of which, did you get the emails on two wubi sessions? [15:15] I can furnish a laptop for the task. [15:15] great [15:15] oh, I did, but my agenda is actually oversubscribed right now - if I get it down to the point where they'd fit, that would be fine, otherwise maybe put them into a whiteboard-scheduled slot? [15:16] we can squeeze in 1 hour I guess if required (as mentioned in my reply) [15:17] works for me. If we run out of time, we can always find vacant space in a hallway. [15:17] sound good to me too! [15:18] only requirement is for cking or some other kernel dev to be around when we discusso kernel stuff [15:23] xivulon: I'll check with Ben in case he has vacant space on his track - doesn't all need to be on the platform track of course [15:23] cjwatson: thanks [15:25] as mentioned by evand, might help to szaka on voip when that happens [15:25] the other one's actually already on my track [15:25] I've replied by mail [15:29] souds good, is the migration topic going to be on a separate session? [15:29] yes [15:30] can we have that too thu-fri pls? [15:30] by the way I was thinking of providing a preliminary script to migrate people over since LVPM has not been update [15:30] d [15:31] should be a matter of copying over files + edit menu.lst/fstab + grub-install [15:31] that is assuming that a target partition is already available [15:34] talking of which... [15:36] I've made a note about your attendance times [15:36] it might also be possible to migrate to a real partition without installing grub... I.E. reusing C:\ubuntu\disks\boot and only changing kopt root [15:36] target partition> this is why it needs to be integrated with ubiquity so that it gets a proper partitioner, boot loader installation, etc. [15:37] otherwise you're going to find yourself gradually reinventing ubiquity, even if it seems like a small job at first [15:37] as mentioned in boston I am in favor of ubiquity integration :) [15:37] I think it's better to reinstall grub - otherwise a user might quite reasonably think they could delete Windows and then bang goes Ubuntu too [15:38] makes sense [15:38] the script mentioned above is only to "fill the hole" while LVPM and or Ubiquity are ready [15:56] I would assume that the Ubiquity migration functionality is supposed to be used via CD, correct? [15:57] any chance of making that also available as an app (possibly using unpartitioned space and/or partitions with no fs/files)? [15:57] I don't think it's fair to make assumptions about something that hasn't really been designed yet ... :-) [15:58] there's no reason you couldn't run ubiquity from a regular system as long as it had something stable to copy from [15:58] while it needs to know about CDs, it's not intrinsically tied to them [15:59] Ah I was thinking more in terms of feature freeze [15:59] since that would require a new ubiquity to be installed in 8.04 [16:04] xivulon: I'm confused. What does this have to do with 8.04? [16:10] evand, say users of wubi in 8.04 want to migrate to a real partition, can we still help them out once the migration facility is in ubiquity? [16:10] one way is to wait for ubiquity 8.10 on CD, the other is to let such users install a version of ubiquity to 8.04 with migration facility and use that [16:11] sorry if I wasn't clear [16:14] I guess that they can use LVPM in the meantime, no biggie [16:28] well, we can discuss this at length at UDS, but regardless of the technical merits, I don't think a brand new feature like this is going to make it into hardy-proposed. [16:30] yeah was expecting that. [16:30] I'm sure it would be possible to build something unofficial [16:31] since the migration tool will be unofficial anyway for 8.04 users (whethere my own scripts or LVPM) we might also consider some early PPP ubiquity build maybe [16:31] cjwatson you beat me :) [16:32] s/PPP/PPA [16:43] evand I think that Tormod branch is almost ready for hardy.proposed (#136682 [16:44] I would like to do some changes to the panic messages in lupin-casper re #226622 but I guess it is better to keep things separately [16:46] xivulon: ok, I'll have a look at it and merge it in. [16:46] can we also have ubuntu-testing to play with http://launchpadlibrarian.net/14240142/lupin-support_0.16%2Blocale110_all.deb [16:49] evand I already have a hardy.proposed branch https://code.launchpad.net/~ubuntu-installer/lupin/hardy.proposed (which is now same as final) [16:54] ok [16:54] evand, i saw that there is a udeb for gptsync. is that actually used anywhere right now/ [16:54] not sure if overriding LANG in /etc/default/locale is desirable [16:56] otherwise code looks good to me :) [16:56] either rdepends doesn't appear to work on udebs, or i don't believe it's in use at all [17:00] mario_limonciell: as far as I can tell it's done inside partman now: [17:01] partman-efi (11ubuntu3) feisty; urgency=low [17:01] * Remove gptsync code; parted does this itself now as of version [17:01] 1.7.1-3ubuntu3. [17:01] parted (1.7.1-3ubuntu3) feisty; urgency=low [17:01] [ Matthew Garrett ] [17:01] * gptsync.dpatch: Perform automatic GPT/MBR partition table [17:01] synchronisation (LP: #46853). [17:01] http://launchpadlibrarian.net/6603773/gpt.diff [17:01] ah hm. i suppose that doesn't handle with recovery partitions well then [17:02] per scott davilla's indication to the mailing list, i started to dig a little bit into it [17:03] i wonder if that patch is literally lifted from gptsync [17:03] or what not [17:03] I have seen a report of problems with recovery partitions, yes [17:03] curious, those mails landed in my inbox, but I didn't get them in the ML folder. [17:03] I understand it was pretty much from gptsync, but there may be some divergence [17:04] I'd appreciate somebody who can conveniently test it looking into it [17:04] mario_limonciell: BTW, I've heard reports that Dell is shipping some systems with four primary partitions, which is obviously very inconvenient if you want to install Linux later. Do you know anything about this? [17:05] cjwatson, this i've not heard. [17:05] could you point me to some indications/ i can poke around [17:06] cjwatson, re gptsync, why was the code integrated into partman rather than just living in the udeb and calling that binary in the first place, do you know/ [17:06] [sorry, my shift is broken right now due to vmware messing a few things up] [17:07] mario_limonciell: unfortunately it was just on IRC [17:07] mario_limonciell: (in theory) much more robust to do it inline in parted, required fewer scary partman hacks, etc. [17:07] ensures that people can't screw up their systems by running parted and forgetting to run gptsync [17:07] but less maintainable in the sense of newer gptsync versions etc [17:07] it's a trade-off [17:08] ah yeah i see. [17:08] gptsync surely doesn't change *that* often [17:08] and libparted ought to know everything about how to handle partitions, really [17:08] well unfortunately the version that is in debian and ubuntu doesn't handle the recovery partitions either [17:08] how are they supposed to be handled? [17:08] i sent a ping to the debian maintainer about updating it [17:08] well it's fixed in the 0.11 release [17:08] he just hasn't touched the package since 2006 [17:10] i'm a bit new to the whole thing, and was looking at it out of curiosity after seeing that there has been these successes with getting ubuntu on the apple tv [17:13] I suspect the answer is simply to forward-port patches from gptsync one by one [17:14] that's very attainable then. i'll make some notes of this conversation, scott may be convincing me to get one of these atv's in the near future so i may be able to help with that. [17:16] cjwatson, re the 4 partitions, i'll send some e-mails around. if you get any more indications of models or locales this is happening in, could you let me know/ [17:16] ok, I'll keep my ears open, thanks [17:16] seems like it should be a general policy to use logical partitions wherever possible [17:18] i agree, but unfortunately some of the overseas factory teams sometimes aren't very cognoscente of the possibility of people needing more than 4 partitions for any reason [17:18] that's why it needs to be policy rather than left up to individual initiative ;-) [17:18] ;] [17:19] Ubuntu only creates a primary partition now if there isn't one already, so it should be as well-behaved as you can get in this regard [17:39] cjwatson, initial responses back would like to know which locale at least. do you know where the person(s) who indicated this were located? [17:41] mario_limonciell: I looked at my logs; it was astronut on #ubuntu-devel, who I believe is in the US [17:41] 17:40 [Freenode] -!- astronut [n=astronut@sfnc-162-39-87-194.sandhills.us] [17:41] cjwatson, okay. i'll pass that on. thanks [17:41] he wasn't very clear, but he did claim that it came with four primaries [17:41] 20:39 cjwatson: i'm not sure... thisone has a a tiny fat16, two large ntfs (C: and D:, d has label backup and had some empty folders and some data) and then a hidden fat32 w/ restore data [17:43] both for the tiny fat16 utility partition, the recovery partition, the factory process doesn't allow for non primary (internal limitations), but i'm not sure why a backup partition would have been created, especially being primary [17:49] can Windows cope with being booted off a logical partition? [17:50] I'm not sure [17:51] even if it couldn't, that backup partition if really necessary should be able to be logical [17:52] yeah, you'd have thought [17:55] I don't suppose windows can be booted off of a logical partition, because a DOS MBR can't address a logical partition in an extended partition [18:01] lowmem: cjwatson * r74 ubuntu/ (3 files in 2 dirs): merge from Debian 1.28 [18:02] lowmem: cjwatson * r75 ubuntu/debian-installer-startup.d/S15lowmem: revert to Debian's version [18:03] lowmem: cjwatson * r76 ubuntu/debian/changelog: releasing version 1.28ubuntu1 [20:18] Is there any reason why we don't currently disallow /boot (or by extension /) on XFS in ubiquity or include lilo to use with it in such circumstances? [20:39] partman-target: evand * r723 ubuntu/ (debian/changelog finish.d/clear_partitions): [20:39] partman-target: * Ensure that if we clear the root partition, / is owned by root:root [20:39] partman-target: (LP: #224446). [22:09] Hi tormod, I am Ago [22:09] thanks a lot for the locale patch :) [22:12] xivulon: hi! it's an honour - wubi is the greatest thing [22:12] ...after sliced bread [22:12] ...and d-i [22:12] ...and ubiquity [22:13] :) [22:13] for linux adoption... since the live cd [22:13] or vfat support :) [22:15] you seem capable with c also, and have windows [22:15] there is one bug I cannot understand and have looked at the 200 times [22:15] maybe a pair of fresh eyes would help [22:16] bug #207137 [22:16] there is a stand-alone app (see hampus comments) [22:16] I only have a USB drive with an (unbootable) ntfs partition, but I can boot into NT with vfat. [22:17] basically this is a small app to extract a CD to an ISO [22:17] code is very simple: gets a handle for the device, keeps reading it in a loop and pushes the bytes into a file [22:19] yet for some users it does not work (usually at the end) and you get 2 or 3 (different) errors [22:20] code is here: http://launchpadlibrarian.net/14044790/CD2ISO.zip [22:21] compiled with http://www.codeblocks.org/ [22:21] of course the above is an open invitation for anyone else since myself and hampus are out of idea [22:22] s [22:22] evand, you too ^ [22:23] this is the last bug I am tracking for point release (have a few open ones but those should be manageable) [22:27] tormod also added a small comment for #136682 [22:28] ehm c++ [22:59] evand: we should definitely make it possible to install lilo. GRUB+XFS has a race condition; it may work for some people but it can't currently be made to work across the board [23:17] cjwatson: are you around? [23:17] Zelut: for a short while [23:18] it's late for me and I have an early meeting tomorrow [23:18] feel free to leave a message or send mail [23:18] I was going to do some preseed work to automate an ubuntu-server install. [23:18] anything you might be able to point me to? [23:21] start with https://help.ubuntu.com/8.04/installation-guide/, and you probably also want to make sure you incorporate each of the things in /preseed/ubuntu-server.seed from an Ubuntu server CD unless you explicitly choose otherwise [23:36] evand panic text re 226622 http://paste.ubuntu.com/10631/