[07:24] <superm1> in its current state, should partman be able to understand if it is preseeded a symlink to a disk rather than the actual disk itself? (I haven't tried yet)
[07:25] <superm1> eg one of the items living in /dev/disk/by-id
[10:25] <mib_p87ibcq2> i need some help
[10:26] <mib_p87ibcq2> ...
[10:26] <mib_p87ibcq2> what do you type to get the gutsy packages onto your computer?
[10:27] <mib_p87ibcq2> apt-get install "???????"
[10:27] <mib_p87ibcq2> what is the name of the packages?
[10:27] <cjwatson> which gutsy packages?
[10:28] <cjwatson> there are several thousand of them
[10:28] <mib_p87ibcq2> for ubuntu 7.10 server edition
[10:28] <cjwatson> what do you currently have installed?
[10:28] <mib_p87ibcq2> it's textbased atm, i need to get
[10:28] <mib_p87ibcq2> the gutsy packages on there so i can use it
[10:29] <cjwatson> if you already have Ubuntu installed, then you're done with this channel; I suggest #ubuntu-server
[10:29] <mib_p87ibcq2> where ca i get info on the packages?
[10:29] <mib_p87ibcq2> *can
[10:31] <mib_p87ibcq2> ???
[10:32] <mib_p87ibcq2> nvm then
[10:32] <mib_p87ibcq2> cya
[10:33] <cjwatson> #ubuntu-server, please
[10:42] <CIA-24> debian-installer: cjwatson * r899 ubuntu/debian/changelog: releasing version 20070308ubuntu34
[11:05] <TheMuso> evand: I'll be around from 20:00UTC tomorrow if you can spare me some time to work through ubiquity a11y stuff.
[11:05] <TheMuso> evand: Sorry, today.
[11:44] <CIA-24> debian-installer: cjwatson * r900 ubuntu/ (33 files in 2 dirs): * Update help text translations from Launchpad.
[11:48] <CIA-24> debian-installer: cjwatson * r901 ubuntu/debian/changelog: releasing version 20070308ubuntu35
[11:52] <simosx> A few weeks ago I updated the translation for gfx-boot-theme. With Ubuntu 8.04 Beta I do not see the updated translations in the boot screen.
[11:52] <simosx> How can I verify that my translations made it to the package?
[11:52] <simosx> https://translations.edge.launchpad.net/ubuntu/hardy/+source/gfxboot-theme-ubuntu/+pots/bootloader/el/+translate
[11:56] <cjwatson> simosx: I updated them just an hour or so ago
[11:56] <cjwatson> if they're up-to-date in Launchpad and don't need review or anything, they'll be there
[11:56] <cjwatson>  msgid "^Install Ubuntu"
[11:56] <cjwatson> -msgstr ""
[11:56] <cjwatson> +msgstr "^Εγκατάσταση Ubuntu"
[11:57] <cjwatson> that kind of thing yours?
[11:57] <simosx> yep, that's mine.
[11:57] <cjwatson> ok, should be up-to-date in tomorrow's CD then
[11:58] <simosx> is tomorrow's CD something like Beta2? Are these daily builds?
[11:59] <cjwatson> daily builds, I mean
[11:59] <cjwatson> http://cdimage.ubuntu.com/daily-live/current/ or http://cdimage.ubuntu.com/daily/current/
[11:59] <cjwatson> we won't be doing a beta2, I expect
[12:00] <simosx> okay, cool. will test tomorrow's CD image as well. thanks.
[12:01] <grrrreg> matin
[12:01] <grrrreg> oops
[13:59] <_MMA_> Well after typing in the wrong window, has anyone see the "Failed to mount /" issue on the resent Alt disks after the partitioning step?
[14:01] <cjwatson> I think we need a bit more information than that; logs would be good
[14:01] <cjwatson> (it could easily depend on your partitioning choices or the existing contents of your disks)
[14:03] <_MMA_> cjwatson: This is using the Alt disk on install after picking any of the partitioning options so far. I'll also mention this is using VirtualBox. Im gonna try on HW now.
[14:04] <cjwatson> if you can get me logs first, that would be good
[14:05] <_MMA_> Im unsure where to find logs on an Alt disk install. Sorry.
[14:05] <cjwatson> /var/log#
[14:05] <cjwatson> er, without the #
[14:05] <_MMA_> k
[14:05] <cjwatson> 'anna-install openssh-client-udeb' and you can copy them out. I need syslog and partman
[14:08] <evand> TheMuso: ok.  I think I've tracked down the problem, but I don't see an easy fix.  The GTK module in ubiquity needs to be imported as the ubuntu user for it to work, but we cannot drop privileges via the EUID as GTK doesn't like that.
[14:18] <CIA-24> ubiquity: cjwatson * r2599 ubiquity/ (debian/changelog ubiquity/components/partman.py): * Fix crash with the "don't use" partitioning option (LP: #132611).
[14:28] <superm1> cjwatson, would you expect partman to work with EDD at this point?
[14:28] <superm1> if you passed a udev symlink to a EDD drive in the preseed?
[14:35] <cjwatson> superm1: er, not very sure, never tried it - what does 'mapdevfs <edd device name>' say?
[14:37] <superm1> cjwatson, it maps it to /dev/sda
[14:37] <superm1> as expected at least for the disk I tried
[14:39] <cjwatson> in that case I think it should work
[14:39] <cjwatson> since you're asking I assume it doesn't ;)
[14:39] <superm1> well I haven't tried yet, but it will require a new factory image since EDD was fixed in udev yesterday
[14:40] <superm1> so I was jumping the gun to see if I should further investigate before trying
[14:40] <superm1> and also you fixed an item related to it in grub this morning I see
[14:51] <cjwatson> right, testing of that would be fantastic; I've got some unit testing but an end-to-end test, ideally of something that used to fail, would be brilliant
[14:52] <superm1> cjwatson, well as soon as it shows up in a DVD image i'll make the changes to our seeds and bootstrap scripts to see how it works out
[14:53] <cjwatson> the algorithm inside grub was a little more complex than I would have liked
[14:53] <superm1> did it look to be portable to grub2 still?
[14:53] <cjwatson> though I'm quite pleased with the elegance of the result
[14:53] <superm1> i havent looked over the diff yet
[14:53] <cjwatson> I've never looked at grub2's internals much, but I gather that its device.map generation is quite similar to grub's
[14:54] <superm1> for one locale we need to ship grub2 for showing localized strings in the grub menu, so i'll take a look through
[14:54] <cjwatson> http://bazaar.launchpad.net/~ubuntu-core-dev/grub/ubuntu/annotate/824?file_id=edddevicemap.diff-20080402124634-7yrytbi62n8zuyz0-1
[14:56] <superm1> looks like its not horrid at least
[15:28] <CIA-24> ubiquity: cjwatson * r2600 ubiquity/ (4 files in 3 dirs): * Use localised, human-readable names for partitioning methods.
[16:15] <CIA-24> ubiquity: cjwatson * r2601 ubiquity/ (bin/ubiquity-wrapper debian/changelog):
[16:15] <CIA-24> ubiquity: * Reorganise the wrapper script to cope with kdesu's broken argument
[16:15] <CIA-24> ubiquity:  handling.
[16:59] <mario_limonciell> cjwatson, well an easy port to grub2 was just a dream unfortunately:
[16:59] <mario_limonciell> 14 out of 14 hunks FAILED -- saving rejects to file device.c.rej
[17:08] <cjwatson> heh, I imagine it won't be *that* easy
[17:08] <cjwatson> but the intent of the changes should be clear
[17:09] <cjwatson> be careful you catch any copies of the fprintf blocks I removed
[17:09] <cjwatson> the number of those may differ
[17:25] <mario_limonciell> well it will be a bit low on the todo for now - after giving EDD a shot in the installer, partman still doesn't like EDD stuff.  I've got a small suspicion of the cause, but if its not it, i'll summarize it all in a bug
[17:29] <mario_limonciell> cjwatson, before trying it though, the way partman-auto/disk is used, it might eventually have support for preseeding multiple disks.  would it be a fair assumption to currently ignore that?
[17:37] <cjwatson> well, it doesn't have such support at the moment, but please don't make it more difficult
[17:37] <mario_limonciell> okay
[19:54] <evand> xivulon: taking this here in case anyone wants to disagree with what I'm about to say
[19:55] <evand> note that ubiquity in automatic-ubiquity mode requires 256MB, not 384MB
[21:13] <TheMuso> evand: Hrm right.
[21:13] <mario_limonciell> cjwatson, I added the support to partman-auto to handle EDD for the case that we are doing.
[21:13] <mario_limonciell> its linked in the branch on bug 186686
[21:13] <ubotu> Launchpad bug 186686 in dell "[Hardy Feature Request] Add EDD support" [High,Fix released] https://launchpad.net/bugs/186686
[21:13] <TheMuso> evand: Would it be better to simply run everything as root?
[21:13] <mario_limonciell> just verified that it works for us
[21:14] <cjwatson> meh, it's a bug-fix ... ish ;-)
[21:16] <cjwatson> mario_limonciell: hmm, I'd recommend not calling it "seed" - "seed" in Ubuntu terms means something you feed to germinate. I'll rename that preseed, I think
[21:16] <cjwatson> +                               found = 1
[21:16] <mario_limonciell> okay
[21:16] <cjwatson> mario_limonciell: legal shell would be found=1
[21:16] <cjwatson> and local found=0 above
[21:17] <mario_limonciell> appeared to work for me, didn't realize it wasn't legal :)
[21:17] <cjwatson> $ found = 1
[21:17] <cjwatson> sh: found: not found
[21:17] <cjwatson> $ foo () { local found = 0; }
[21:17] <cjwatson> $ foo
[21:17] <cjwatson> sh: local: 0: bad variable name
[21:17] <cjwatson> sh: : bad variable name
[21:17] <cjwatson> so, not sure how it worked for you :) that's busybox sh
[21:18] <evand> TheMuso: I think we're getting close to having to resort to that.  I cannot think of a way around this, though I'm willing to give it some more thought and poke around the GTK code.
[21:18] <cjwatson> bash and dash fail similarly
[21:18] <mario_limonciell> cjwatson, well maybe the errors were obfuscated then in a log in /var/log at least
[21:18] <cjwatson> mario_limonciell: given that, could you make the changes I mentioned and retest?
[21:18] <mario_limonciell> yes will do.
[21:18] <TheMuso> evand: Ok, I'm happy to test whatever.
[21:19] <cjwatson> mario_limonciell: actually, uh, this can't be right
[21:20] <cjwatson> mario_limonciell: partman_disk is a device node (e.g. /dev/sda), but dev_to_partman returns a partman device directory (e.g. /var/lib/partman/devices/=dev=sda)
[21:20] <cjwatson> you can't compare those two
[21:20] <cjwatson> if you just want to canonicalise a device name, use mapdevfs
[21:21] <cjwatson> mario_limonciell: I think it worked for you due to the shell bug - found would never have been equal to 0
[21:21] <mario_limonciell> oh hmm.  could my results be skewed then from multiple ubiquity runs with partman storing information in debconf?
[21:21] <cjwatson> that's possible too, but I think a conjunction of bugs (sort of) cancelling each other out is more likely
[21:22] <cjwatson> my reading of this code is that it will always use the first disk
[21:22] <mario_limonciell> okay i'll sort it out
[21:22] <cjwatson> thanks
[21:22] <cjwatson> happy to merge it once that's sorted
[21:37] <cjwatson> mario_limonciell: oh, also, as a general rule, it would help avoid confusion if you used UNRELEASED as an upload target in changelogs rather than hardy until it's actually being uploaded
[21:38] <mario_limonciell> cjwatson, ah yeah that's my normal behavior, I forgot this time.
[21:39] <cjwatson> ok
[22:46] <xivulon> cjwatson, I would really need the metalink url (140458), as mentioned in the bug last comment, I would encourage the use of a url just for wubi
[22:49] <xivulon> with static metalink files to be copied off wubi-installer.org/metalinks/8.04
[22:56] <xivulon> evand the code in rev 92 looks good to me! I like it better!
[22:56] <xivulon> did not test it though
[23:01] <cjwatson> xivulon: Steve was blocked on getting mirror code from webmaster for a while, but that's done; he's now hoping to get it sorted out early next week
[23:02]  * xivulon checking release schedule
[23:02] <xivulon> ok so we have time for the release candidate
[23:03] <xivulon> and hopefully a couple of days before that
[23:03] <xivulon> good enough for me
[23:03] <xivulon> thanks
[23:05] <xivulon> In the keyring now I have my own key (which should be removed) and 1024D/FBB75451
[23:06] <xivulon> that is for verifying the md5s of the metalinks, so I am assume the latter will be used