[01:37] <cjwatson> doing a bit of hacking in order to update ubiquity, since archive.ubuntu.com mirroring seems busted (notified #is)
[01:45] <CIA-7> ubiquity: cjwatson * r4662 trunk/ (d-i/manifest debian/changelog):
[01:45] <CIA-7> ubiquity: Automatic update of included source packages: localechooser 2.34ubuntu4,
[01:45] <CIA-7> ubiquity: partman-auto 93ubuntu14.
[02:13] <CIA-7> ubiquity: cjwatson * r4663 trunk/debian/ (68 files in 2 dirs): Update translations from Launchpad.
[02:19] <CIA-7> ubiquity: cjwatson * r4664 trunk/debian/ (changelog control):
[02:19] <CIA-7> ubiquity: Add ${misc:Depends} to dependencies of ubiquity-ubuntu-artwork and
[02:19] <CIA-7> ubiquity: oem-config-remaster.
[02:29] <CIA-7> ubiquity: cjwatson * r4665 trunk/debian/changelog: releasing version 2.6.1
[09:12] <ev> thanks for that
[09:59] <CIA-7> ubiquity-slideshow-ubuntu: evand * r348 ubiquity-slideshow-ubuntu/ (453 files in 7 dirs): Update translations from Launchpad.
[10:13] <CIA-7> ubiquity-slideshow-ubuntu: evand * r349 ubiquity-slideshow-ubuntu/debian/changelog: releasing version 39
[11:42] <ev> weird.
[11:43] <ev> so nspluginwrapper can't find libGL.so.1
[11:43] <ev> when flashplugin-installer is run
[11:43] <ev> if I run the same command in the chroot after it crashes and burns, it works fine
[11:45] <ev> and ia32-libs seems to be doing the right thing with respect to creating the symlinks
[11:45] <ev> ugh, I hate bugs that take part of an install to reproduce, eats up too much time.
[11:45]  * ev stops rambling
[11:46] <cjwatson> that sounds like the /etc/ld.so.conf.d problem
[11:47]  * cjwatson hunts for a reference
[11:48] <cjwatson> http://irclogs.ubuntu.com/2011/02/16/%23ubuntu-devel.txt
[11:48]  * ev waits with bated breath
[11:48] <cjwatson> search for the conversation between me and seb128 about libGL there, and try to ignore the false starts
[11:48] <cjwatson> should at least give you context
[11:49] <ev> wow, I think you just saved me a few hours
[11:49] <ev> thanks a bunch
[11:49] <cjwatson> [14:00] <cjwatson> you get to choose between:
[11:49] <cjwatson> [14:00] <cjwatson>  (1) artificial Depends: libgl1, to force the ldconfig cache to be brought up to date
[11:49] <cjwatson> [14:01] <cjwatson>  (2) artificial 'LDCONFIG_NOTRIGGER=y ldconfig' in libgdk-pixbuf2.0-0.postinst, to cope with the fact that the ldconfig cache may not be up to date
[11:49] <cjwatson> might be especially relevant but I haven't looked at nspluginwrapper/ia32-libs ...
[11:53] <ev> right-o
[11:53] <cjwatson> bugger, something is still broken with oem-config and language packs
[11:53] <cjwatson> well, maybe
[11:54] <cjwatson> oh, I should try oem-config/install-language-support=true shouldn't I
[11:57] <ev> so possibly a heads up
[11:57] <ev> something is gobbling up memory in the desktop session
[11:57] <ev> I keep thinking ubiquity is broken only to find the OOM killer did its thing
[11:58] <ev> cjwatson: by the way, did you get to the bottom of the isolinux/localechooser issues?
[11:59] <cjwatson> yeah, it turned out to be a kvm bug
[12:00] <cjwatson> bug 747090
[12:06] <ev> oh wow
[12:06] <ev> reading through that now
[12:10] <cjwatson> not the expected outcome, indeed
[12:14] <ev> wow, great find though
[12:29] <genec> partitioning tool in ubuntu-11.04-beta1-desktop-i386.iso allows me to select to install grub2 to a btrfs single file system (sda1) and fails.  attempting to then select the parent (sda; MBR/reserved sectors) also fails (grub rescue).  Selecting sda first succeeds.
[12:29] <cjwatson> genec: try a daily build - we've been fixing btrfs problems in grub since beta-1
[12:30] <genec> I know the 110410 fails properly and brings you back to the typical Debian installer stage menu (uncertain on proper name)
[12:31] <cjwatson> oh, I guess there might not be a current enough daily build with the last round of grub fixes
[12:31] <cjwatson> bug 732149 is the main one I'm referring to
[12:31] <genec> saw it
[12:31] <cjwatson> I'll give that scenario a try (8th item on to-do list)
[12:31] <genec> as far as I can tell, grub2 refuses to install directly to a btrfs with an absurd reason.
[12:32] <genec> reading the source out of the bazaar tree
[12:32] <cjwatson> please keep the conversation strictly technical
[12:33] <cjwatson> is it the "isn't known to reserve space for DOS-style boot" message?
[12:33] <genec> "btrfs filesystem which isn't known to reserve space for DOS-style boot" when the first 64kiB are reserved for boot loaders (more than MBR/reserved sectors)
[12:34] <cjwatson> yeah, that should be easy to fix, it's just a missing flag in a struct
[12:34] <cjwatson> can you please file a bug report on the Ubuntu grub2 package about this?
[12:34] <genec> absolutely.
[12:34] <cjwatson> thanks
[12:35] <cjwatson> bumped up higher in to-do, next to a few other btrfs backports I want to do
[12:36] <cjwatson> (I'm working on bug 711926 at the moment)
[12:44] <arand> (following a discussion in +1) Does natty support btrfs+full disk encryption? Maverick seemed to do so, did maverick set up subvolumes in a similar manner that natty does, and if not, could this be what is causing issues?
[12:47] <genec> cjwatson: is there any tags you want me to add?  is there a tag that indicates it should go upstream to grub2?
[12:48] <cjwatson> genec: don't worry, I'll take care of that
[12:48] <genec> cjwatson: bug 757446
[12:48] <cjwatson> arand: maverick didn't do the subvolumes thing.  I don't know about btrfs + full-disk encryption - would need to see logs
[12:48] <cjwatson> genec: thanks
[12:50] <cjwatson> oh, full-disk encryption might confuse automatic addition of rootflags=subvol=@
[12:51] <cjwatson> I can see how that might happen.  Probably not sanely fixable until such time as GRUB has its own LUKS support
[12:51] <arand> So would it be possible that just adding that manually might boot a system?
[12:51] <cjwatson> it's possible; I'm not certain
[12:51] <genec> cjwatson: if you need logs, I have the full /var/log directory as a tarball and can furnish select logs if needed.
[12:51] <cjwatson> genec: I don't need logs in your case, thanks
[12:53] <genec> it appears grub2's btrfs support is mostly there.  I've been meaning to test this since alpha-2 was released
[13:00] <ev> I'm amazed there's even an Isle of Man mirror... https://bugs.launchpad.net/ubuntu/natty/+source/ubiquity/+bug/751666
[13:02] <cjwatson> wildcard alias to archive.ubuntu.com :)
[13:02] <cjwatson> (zzzzz.archive.ubuntu.com exists, in the same way)
[13:04] <ev> ahh
[13:16] <genec> In the interim, I created a new section at https://help.ubuntu.com/community/btrfs#Fresh%20Install%20on%2011.04-beta1%20Natty with a link to the bug
[13:56] <CIA-7> ubiquity: cjwatson * r4666 trunk/ (debian/changelog ubiquity/plugins/ubi-console-setup.py): Handle keyboard layouts with only one variant (LP: #711926).
[14:12] <CIA-7> ubiquity: evand * r4667 trunk/partman/finish.d/apt_clone_save: Remove set -x from apt_clone_save.
[14:30] <CIA-7> ubiquity: cjwatson * r4668 trunk/ (debian/changelog ubiquity/plugins/ubi-console-setup.py):
[14:30] <CIA-7> ubiquity: Make sure to set a new variant default when changing keyboard layout
[14:30] <CIA-7> ubiquity: (LP: #745137).
[16:02] <arand> cjwatson: Reported https://bugs.edge.launchpad.net/ubuntu/+source/grub2/+bug/757631 It seems like it is indeed only grub-probe not managing to detect btrfs correctly (again :) that is the sole issue..
[16:03] <cjwatson> arand: my explanation above probably still holds.  We aren't likely to fix this for natty
[16:05] <cjwatson> I'll add a release-notes task
[16:06] <arand> cjwatson: Ok, fair enough. As long as there a warning or so, although it can be argued that it shouldn't really be expected to work, but then again, if it did work in 10.10 it's worth noting.
[16:07] <cjwatson> grub couldn't boot off btrfs in 10.10
[16:08] <cjwatson> though, OK, the subvolume thing is different
[16:08] <cjwatson> I'll add a partman-btrfs task for the warning, but I can't promise to get to it for release - it seems like a fairly niche configuration TBH
[16:10] <cjwatson> but thanks for the report
[16:14] <CIA-7> ubiquity: evand * r4669 trunk/ (debian/changelog scripts/plugininstall.py):
[16:14] <CIA-7> ubiquity: Deal with the confirmation key being translated in jockey-text (LP:
[16:14] <CIA-7> ubiquity: #757208).
[16:14] <CIA-7> ubiquity: evand * r4670 trunk/debian/changelog: fix changelog
[16:36] <bdmurray> cjwatson: have you seen bug 683904?
[16:37] <cjwatson> bdmurray: yes, but I was leaving it to mvo
[16:38] <bdmurray> cjwatson: okay, from his last comment it wasn't clear to me if he was still working on it
[16:38] <cjwatson> I'm asking him
[16:44] <cjwatson> I don't think I really have bandwidth to take it on, though
[17:39] <cjwatson> ev: are you working on anything in ubiquity at the moment?  we should upload for tomorrow's builds
[17:43] <ev> nothing in particular.  The migration-assistant reuse option interaction is going to require a bit more thought, I've found a bug exposing an issue with the bootloader drop down, but that's not super-critical
[17:43] <ev> so yeah, I'm happy to sort an upload
[17:43] <ev> shall I?
[17:43] <ev> or do you have things in the pipeline?
[17:43] <cjwatson> nope
[17:43] <cjwatson> go ahead
[17:43] <ev> cool, on it now
[17:49] <CIA-7> ubiquity: evand * r4671 trunk/ (134 files in 3 dirs): Update translations from Launchpad.
[17:55] <CIA-7> ubiquity: evand * r4672 trunk/debian/changelog: releasing version 2.6.2
[18:01] <CIA-7> debian-installer: cjwatson * r1437 ubuntu/ (3 files in 2 dirs): Update help text translations from Launchpad.
[18:02] <CIA-7> debian-installer: cjwatson * r1438 ubuntu/debian/changelog: releasing version 20101020ubuntu28
[18:37] <ev> Apr 6 22:20:17 ubuntu install.py: self.copy_all()
[18:37] <ev> Apr 6 22:20:17 ubuntu install.py: File "/usr/share/ubiquity/install.py", line 357, in copy_all
[18:37] <ev> Apr 6 22:20:17 ubuntu install.py: os.unlink(targetpath)
[18:37] <ev> Apr 6 22:20:17 ubuntu install.py: OSError: [Errno 21] Is a directory: '/target/usr/local/man'
[18:37] <ev> ...I feel like we've been here before
[18:37] <ev> bug 753024
[18:38] <ev> I've assigned it to myself and will have a ponder
[18:51] <superm1> ev, a side effect of the apt clone?
[18:52] <ev> but the code path has that under an explicit S_ISLINK check
[18:52] <ev> unless it's a broken link
[18:52] <ev> actually we even check for that
[18:58] <superm1> OK yeah that's pretty odd then