[10:12] <ev> ScottK, Riddell: http://paste.ubuntu.com/489138/ - partitioning page fixes.  Could either of you let me know how well that works for you?  I still have to add back the "use largest free space" option, but it should otherwise be complete, functional, and hopefully bug free.
[10:25] <Riddell> ev: cool, I'll give it a test shortly
[10:27] <ev> thanks!
[11:44] <CIA-34> ubiquity: evand * r4280 trunk/ (4 files in 3 dirs):
[11:44] <CIA-34> ubiquity: Fetch the translations for the release_notes_only and
[11:44] <CIA-34> ubiquity: update_installer_only strings (LP: #629627).
[11:58] <CIA-34> ubiquity: evand * r4281 trunk/ (3 files in 3 dirs):
[11:58] <CIA-34> ubiquity: Don't use the same name for the error method and GTK label
[11:58] <CIA-34> ubiquity: (LP: #631046).
[12:00] <CIA-34> grub-installer: cjwatson * r864 timeout/ (debian/grub-installer.templates grub-installer): merge lp:~soren/grub-installer/ubuntu-boot-delay
[12:01] <CIA-34> grub-installer: cjwatson * r865 timeout/ (debian/grub-installer.templates grub-installer): boot_delay -> timeout, for consistency
[12:05] <CIA-34> grub-installer: cjwatson * r864 ubuntu/ (3 files in 2 dirs):
[12:05] <CIA-34> grub-installer: Add a preseedable grub-installer/timeout template to adjust the initial
[12:05] <CIA-34> grub-installer: > GRUB timeout (LP: #628418).
[12:07] <CIA-34> grub-installer: cjwatson * r865 ubuntu/grub-installer: consistent behaviour for grub-legacy
[12:08] <CIA-34> grub-installer: cjwatson * r866 ubuntu/debian/changelog: releasing version 1.55ubuntu2
[12:27] <CIA-34> ubiquity: evand * r4282 trunk/ (debian/changelog ubiquity/plugins/ubi-usersetup.py):
[12:27] <CIA-34> ubiquity: Re-use the valid username check from user-setup in the GTK UI
[12:27] <CIA-34> ubiquity: (LP: #631046).
[13:09] <NCommander> cjwatson: hey, not sure if your around, but its been ~2 weeks since I sent mail to d-boot about the changes to libd-i. Any chance we can move towards getting that merged so I can knock that spec out? :-)
[13:10] <NCommander> cjwatson: or how to move this forward in general
[13:15] <cjwatson> I'll look over it
[13:15] <cjwatson> one of our major patch reviewers died recently which hasn't exactly helped
[13:22] <cjwatson> (replied)
[13:35] <davmor2> cjwatson, ev: did you every get the wubi issue resolved?
[13:35] <davmor2> was on hols last week
[13:36] <ev> davmor2: yes
[13:36] <ev> should work now
[13:37] <davmor2> ev: cool I'll give it a spin latter :)
[13:37] <davmor2> ev: what was it for future reference
[13:37] <ev> grub issue
[13:45] <cjwatson>   * wubildr.cfg needs to be run using GRUB's normal parser rather than its
[13:45] <cjwatson>     rescue parser.  Rather than embedding it directly as a configuration
[13:45] <cjwatson>     file, embed it as a memdisk and add wubildr-bootstrap.cfg which invokes
[13:45] <cjwatson>     it using normal mode (LP: #600578).
[13:50] <Riddell> ev: hmm, it was all going well then failure, fault in my VM? http://people.canonical.com/~jriddell/tmp/partitioner-kubuntu-fault.png
[13:51] <ev> could be
[13:52] <ev> Riddell: anything in /var/log/messages of interest?
[13:53] <Riddell> http://muse.19inch.net/~jr/tmp/debug  http://muse.19inch.net/~jr/tmp/syslog http://muse.19inch.net/~jr/tmp/messages
[13:54] <ev> interesting:
[13:54] <ev> Sep  6 12:40:20 ubuntu ubiquity: Resizing the filesystem on /dev/sda1 to 816486 (4k) blocks.
[13:54] <ev> Sep  6 12:40:20 ubuntu ubiquity: The filesystem on /dev/sda1 is now 816486 blocks long.
[13:54] <ev> Sep  6 12:40:20 ubuntu ubiquity:
[13:54] <ev> Sep  6 12:40:22 ubuntu ubiquity: resize2fs 1.41.12 (17-May-2010)
[13:54] <ev> Sep  6 12:40:22 ubuntu ubiquity: The filesystem is already 816486 blocks long.  Nothing to do!
[13:55] <Riddell> that doesn't look healthy
[13:55] <ev> Riddell: what steps did you take at the partitioner?
[13:55] <ev> also, can I have /var/log/partman
[13:56] <Riddell> http://muse.19inch.net/~jr/tmp/partman
[13:57] <Riddell> the "Install side by side" step
[14:03] <ev> okay, did you move the slider at all?
[14:06] <Riddell> ev: yes
[14:26] <CIA-34> usb-creator: cjwatson * r320 allow-system-internal/ (9 files in 7 dirs):
[14:26] <CIA-34> usb-creator: Add an --allow-system-internal option (Unix only) to allow installation
[14:26] <CIA-34> usb-creator: to system-internal devices such as hard disks. This is useful when
[14:26] <CIA-34> usb-creator: preparing test USB images in KVM.
[15:09] <CIA-34> ubiquity: cjwatson * r4283 ubiquity/ (debian/changelog scripts/plugininstall.py ubiquity/misc.py):
[15:09] <CIA-34> ubiquity: Factor out common /proc/mounts handling into a new
[15:09] <CIA-34> ubiquity: ubiquity.misc.mount_info function.
[15:41] <CIA-34> ubiquity: evand * r4284 trunk/ (3 files in 3 dirs):
[15:41] <CIA-34> ubiquity: Do not cut off the text of the Try and Install buttons by setting
[15:41] <CIA-34> ubiquity: their width to just the largest initial width of the two
[15:41] <CIA-34> ubiquity: (LP: #629437).
[15:43] <cjwatson> hmm.  the mtime of /casper-rw is getting hashed into apt-cdrom's identification
[15:43] <cjwatson> that's really not going to help matters
[15:48] <cjwatson> I'd sort of like to move casper-rw to .disk or something, but then that would require casper changes and everyone would have to have a new version of usb-creator
[15:48] <cjwatson> and we'd keep on getting bugs about broken OEM installs from people who haven't
[15:51] <ev> assuming btrfs lands in 11.04, perhaps we should consider replacing the casper-rw machinery with something that leverages btrfs
[15:51] <ev> and move things around then
[15:51] <cjwatson> you know, I'm half-tempted to use a dummy apt cdrom method
[15:52] <cjwatson> all the media-change stuff is a pain and it would be good to be able to just say "no, damnit, /cdrom is it"
[15:52] <cjwatson> or I suppose I could use LD_PRELOAD
[15:52] <ev> works for me
[15:53] <cjwatson> LD_PRELOAD might actually be easier; the filename for the method executable comes from the URI scheme, so it's hard to substitute
[15:55] <cjwatson> but LD_PRELOADing C++ code, OMG
[15:55] <ev> haha
[15:58] <cjwatson> and since casper uses apt-cdrom, anything else that uses apt (e.g. jockey in the live session) will have the same problem
[15:59] <cjwatson> moving casper-rw out of the root seems like the only feasible option
[16:01] <cjwatson> does .disk seem like a reasonable place for it?
[16:01] <cjwatson> or the casper subdirectory, I suppose
[16:02] <ev> indeed
[16:15] <CIA-34> usb-creator: cjwatson * r320 trunk/ (debian/changelog debian/control usbcreator/install.py): Use python-debian for Ubuntu release version comparison.
[16:28] <cjwatson> ev: is the str(self.target) stuff in usbcreator/install.py still necessary?  it seems to be cast to str in some contexts but not in others
[16:29] <ev> cjwatson: not sure, that's purely PyQt being god-awful.  GTK knows how to handle unicode in a sane fashion.
[16:31] <cjwatson> right, I know, but what I don't get is why the KDE frontend works at all since there are some places where self.target isn't cast
[16:33] <cjwatson> self.target is only assigned to in one place, so maybe cast it there?
[16:37] <ev> sure
[16:54] <ScottK> ev: I can do some testing later today if you still need it.
[16:55] <ev> ScottK: please do.  I haven't had a chance to look at Riddell's bug, but I'm keen to see if it works for you.
[16:56] <ScottK> ev: OK.  Should be able to give it a whack in about 90 minutes.
[16:57] <ev> cool, thanks!
[17:16] <NCommander> cjwatson: thanks for your review. I agree with your suggestions and working on rewqorking the patch to incorperate it.
[17:16] <ev> weird, indicator-session off the custom panel can't seem to talk to ConsoleKit over dbus, despite the session daemon running.
[17:24] <cjwatson> ev: should I land that usb-creator patch?
[17:25] <ev> cjwatson: please do
[17:27] <CIA-34> usb-creator: cjwatson * r321 trunk/ (9 files in 7 dirs): merge lp:~cjwatson/usb-creator/allow-system-internal
[17:51] <CIA-34> ubiquity: evand * r4285 trunk/ (6 files in 6 dirs):
[17:51] <CIA-34> ubiquity: Update the panel to use the new location for the panel background,
[17:51] <CIA-34> ubiquity: and force a redraw when setting it.
[17:51] <ev> damn
[17:51] <CIA-34> ubiquity: evand * r4285 trunk/ (debian/changelog src/panel/panel.c):
[17:51] <CIA-34> ubiquity: Update the panel to use the new location for the panel background,
[17:51] <CIA-34> ubiquity: and force a redraw when setting it.
[17:51]  * ev whistles, hopes no one noticed that
[18:33] <ev> ah, ck-launch-session sets XDG_SESSION_COOKIE, which is only going to picked up by metacity.  Rats.
[18:33] <ev> right, home time
[18:33] <ScottK> Riddell: When you're testing ubiquity changes, how do you usually go about doing it?  Did you make a .deb with the patch that's somewhere reachable?
[18:34] <ev> ScottK: you can patch in the live environment, or in casper (setting break=bottom on the kernel command line: https://wiki.ubuntu.com/Installer/Development/Tips)
[18:34] <ev> right, out
[18:34] <ScottK> Thanks.
[18:36] <Riddell> ScottK: in this case I patched the files, and just did sudo cp to move them into the right place http://muse.19inch.net/~jr/tmp/ubiquity/
[18:37] <ScottK> Thanks.
[18:37] <Riddell> ev, ScottK: i tried the change again on real hardware and it worked fine
[18:37]  * ScottK will try to squeeze testing in around the other chores for the afternoon.
[18:37] <Riddell> and I've been getting write errors while upgrade testing today
[18:38] <Riddell> so i think my VM (or my laptop harddisk) is having problems rather than the patch
[18:41] <ScottK> OK.
[20:27] <ScottK> Meh.  No ubiquity testing for me today.  My wife is back from her trip earlier than I'd expected.
[21:06] <NCommander> cjwatson: so I'm working on reworking the libd-i patch, and after some googling, I think it makes sense to implement the new guess function as a weak symbol (especially because this leaves Debian free to not implement it until its needed)
[21:07] <NCommander> is that going to cause a problem, or should I rework my WIP patch to use strong symbols only
[21:37] <cjwatson> NCommander: um, I'm not sure, will probably be easier to judge based on a patch which I can try out, since I always have to look up linker details
[21:40] <NCommander> cjwatson: ah :-). I don't think we need to touch the linker voodoo in libd-i if its a weak symbol, but once I have a new patch ready (probably by your tomorrow)