[02:27] c [08:29] lfaraone: ok, I'll take a look at it today [10:15] partman-auto-lvm: cjwatson * r209 ubuntu/ (66 files in 3 dirs): [10:15] partman-auto-lvm: Ask how much of the VG should be used for logical volumes, rather than [10:15] partman-auto-lvm: unconditionally using it all (LP: #160156). [10:32] Hi. Does anyone know if its possible to enable dm-multipath during an ubuntu install? Are there command line options to pass which enables it? [10:59] there's been some work on that in jaunty, but probably doesn't work in prior releases [10:59] I think it's disk-detect/multipath/enable=true in jaunty [11:00] partman-auto-lvm: cjwatson * r210 ubuntu/debian/changelog: releasing version 32ubuntu2 [11:23] Ok. thanks [11:24] I'll give that a try [12:19] ubiquity: evand * r3026 ubiquity/ (debian/changelog ubiquity/timezone_map.py): [12:19] ubiquity: Only select cities that are in the same UTC offset as where the mouse [12:19] ubiquity: clicked. [12:20] Short of the poor lining up of the map against the time zone cities, the time zone map should be usable now. [12:51] * cjwatson perpetrates some entertaining ncurses abuse [12:57] I think I may have cracked manual package selection [13:21] except for some reason it's exiting right after displaying the terminal [13:24] oh, I forgot to override SIGCHLD handling, bah [13:43] evand: great. [13:58] user-setup: cjwatson * r153 ubuntu/debian/ (changelog user-setup-udeb.install): [13:58] user-setup: Disable installation of pre-pkgsel.d/10kdesudo; it does nothing for [13:58] user-setup: Ubuntu, and causes a confusing message that worries some people. [14:38] sudo bterm -f /usr/share/oem-config/unifont.bgf -l en_GB.UTF-8 -- sh -c 'env DEBCONF_USE_CDEBCONF=1 DEBCONF_DEBUG=20 LD_LIBRARY_PATH=/usr/lib/debug NCURSES_TRACE=8191 ./progress.config 2>err' [14:39] ridiculous test-harness command lines of our time [14:50] Anybody report issues with Ubuntu Jaunty alternate cd failing to install? [14:55] frequently ;-) you're going to have to be a *little* bit more specific than that! [14:57] failing at 6%, missing packages at lib-c6? [14:58] The following packages have unmet dependencies: Feb 13 14:28:07 in-target: libart2.24-cil: Conflicts: libart2.0-cil but 2.20.1-1ubuntu1 is to be installed [14:59] also libgconf2.24-cil [14:59] libgnome-vfs2.24-cil [15:00] Then three missing pacages [15:00] that's not an installer bug, it just means the system isn't actually installable at the moment [15:01] the installer is just the messenger [15:01] Who needs to know, or is it bug report time? [15:01] we have automatic reports for this so it usually sorts itself out [15:01] Okay. Thanks [15:01] it's good to report this when a milestone is coming up, but when it's just a random daily build it tends to be unnecessary [15:01] but thanks :) [15:48] welcome :) [15:48] thank you :) [15:49] my current branch is ~shtylman/+junk/ubiquity but I will rename it at some point, I was just following a quick bzr guide as I had never used it before until like 3 days ago [15:50] ~shtylman/ubiquity/arbitrary-name-for-your-branch would be neater and would make it show up on the ubiquity project page [15:51] ahh..thanks for the heads up, I will do that... any advice on how to do that if I already have that branch? is there a move I can use? [15:52] there's a rename feature in LP [15:52] excellent [15:52] https://code.launchpad.net/~shtylman/+junk/ubiquity/+edit [15:52] the naming scheme is OWNER/PROJECT/NAME by which you should be able to figure out which bits to edit [15:56] ok, done...it is now ~shtylman/ubiquity/kdeui [16:00] shtylman_: if you follow the IRC notification part of this: https://wiki.ubuntu.com/InstallerDevelopment , your commits will show up in here and will generally make it easier to follow the changes you're making [16:01] not a requirement though :) [16:02] what type of changes does it notify? and would this be recommended? I don't have a problem setting it up if its helpful [16:04] (side questions about bzr) I previously made a change to the timezone_map.py file for testing, but now I just want to replace mine with the one from the master branch, I know I can merge and commit, but I don't want to merge, just replace...help? [16:05] merge, copy the .OTHER file to yours, resolve [16:05] thanks === Nicke_ is now known as Nicke [16:06] or copy the file from the other branch after merging but before committing, if the merge doesn't leave a .OTHER file [16:07] you can also do stuff with bzr diff -rbranch:../wherever | patch but TBH this way is easier [16:08] yea...the .OTHER approach worked, thanks ... I am finding bzr to be more like svn (in commands and behavior) than git, even though it is also distributed [16:08] it notifies the channel of every commit, it's generally helpful as it's much easier for us to keep track of your branch [16:09] every commit? or every push? [16:11] every commit [16:14] most of us use "bound branches" so that commit==push [16:14] although I unbind if I'm going to be without network access for a while [16:14] (bzr help bind, bzr help unbind, bzr help checkout) [16:14] interesting...I will look into that [16:14] or maybe bzr help checkouts, more usefully [16:15] bzr is more like cvs/svn command-wise by design, yes [16:15] whereas git is explicitly unlike them by design; philosophical difference [16:16] heh [16:18] what is really cool to me is how well it integrates with launchpad [16:19] makes starting development and keeping up really easy [16:19] indeed, launchpad is fantastic for such things [16:25] to do the bound branch, can I just bzr checkout my branch? I did that but now I want to merge changes from the master branch and it is complaining, is there more I need to do to inform it of the master branch? like to I have to re bind to the master then merge and bind back? [16:25] to convert a normal branch to a bound branch, 'bzr bind lp:~...' [16:25] you should not bind to the master branch; that will create confusion [16:26] what is the complaint? [16:26] k [16:26] no location specified...I think cause I never bound after the checkout [16:26] um [16:26] more likely that you didn't give a branch to bzr merge? [16:26] yea :) [16:27] normal would be 'bzr merge lp:~ubuntu-installer/ubiquity/trunk' or 'bzr merge /path/to/local/checkout/of/same' [16:27] ahhh [16:27] you only have to do that the first time; it will remember thereafter, as long as you're only ever merging from a single branch [16:28] ok [16:30] ok, cool...all is well, thanks [16:36] evand: Hello why is encrypted home being dropped? [16:36] or is it just from ubiquity? [16:36] davmor2: security concerns. It's not really encrypted if swap is not encrypted. [16:37] evand: so will it be dropped from d-i too then for that reason? [16:37] yes [16:37] Okay :) [16:40] evand, are you dropping it entirely from the GUI or just hiding widgets and such until a more secure solution is developed? [16:40] I'd ideally like to do the latter [16:42] well if the former happens, can you comment out the line or two related to it in mythbuntu_ui in your commit? [16:51] absolutely [16:51] evand: I know this will sound daft but if you've got to go to the extreme of encrypting home and swap then why not just promote encrypted lvm, or am I missing something else? [16:52] davmor2: a conversation to have with kirkland :) [16:52] kirkland: ^ [16:52] davmor2: several reasons ... [16:52] davmor2: encrypted lvm requires a password on boot [16:53] davmor2: and it encrypts everything, including stuff that doesn't necessarily need to be encrypted [16:53] davmor2: like /lib, /usr, and so on [16:54] also encrypted LVM is mounted even when the user in question is logged out; encrypted home is finer-grained [16:54] davmor2: also, the way we're doing encrypted home allows each user to encrypt (or not encrypt) their home dir with different keys [16:54] davmor2: which, as cjwatson points out, is unmounted after the user logs out [16:54] davmor2: finally, using the approach we've taken with encrypted home, it's possible (although not yet trivial) to make incremental backups of your encrypted home dir data [16:55] davmor2: it's not really possible to do an incremental backup of an entire lvm [16:55] kirkland: right now I get it a bit more thanks :) [16:56] davmor2: sure [16:56] davmor2: some will prefer encrypted lvm, no doubt, and there's no interest in removing that [16:57] kirkland: So why is it so hard to encrypt swap, is it because it is accessed by more than just 1 user? [16:57] davmor2: well, it's actually very easy to encrypt swap [16:58] davmor2: but doing so the "easy" way breaks hibernation on multi-user computers [16:58] davmor2: http://ubuntumagnet.com/2007/11/creating-encrypted-swap-file-ubuntu-using-cryptsetup [16:58] davmor2: that's a 4-step process to encrypting swap, and exactly what i use [16:58] davmor2: but i can't (and don't want to) hibernate my machine [16:59] davmor2: though i use suspend all the time [16:59] davmor2: with that mechanism, swap is encrypted with a different random key on boot, every time [16:59] davmor2: to enable hibernation to work, you would need a second, wrapping key [17:00] davmor2: and you'd need to enter that wrapping passphrase on boot [17:00] davmor2: and you'd need to share that passphrase with any other users who might want to resume that machine from a hibernation state [17:01] davmor2: i consider that a relative corner case, but I'm in the minority in that vote, i think :-) [17:01] kirkland: Yes I think anyone with a job and laptop might take issue with you :D [17:04] kirkland: So are we likely to this implemented and working full and stable by the next release? [17:04] 9.10 [17:04] rather than 9.04 [17:05] davmor2: i actually have a couple of ideas how to solve the encrypted swap issue [17:05] davmor2: i'm hoping for those to land in 9.10 [17:05] davmor2: kees had an interesting suggestion [17:05] davmor2: i'm going to add a little script to ecryptfs-utils that will do the steps in that article url i pasted above [17:15] evand: hey, so encrypted-home is going to be dropped from the desktop installer, but i think it should remain in the server/alternate installers [17:15] evand: and i'd recommend just commenting it out [17:15] evand: i think we can address the concerns by 9.10 for swap and others [17:16] kirkland: why can it stay in one and not the other? Can you reply in the email thread if you haven't already, I have to run to do some pre-vday errands and I suspect I wont be back until Monday [17:19] evand: sure. basically, it comes down to the fact that the user would need to setup encrypted swap themselves postinstall using a script that i provide [17:23] evand: i don't think that's too much to ask of users installing via server/alternate installer [21:31] is keyboard_names.py autogenerated?