[01:22] <TheMuso> o/c
[09:32] <lool> evand: Hola!
[09:32] <evand> Hallo
[09:32] <lool> evand: We have an issue with the dh 7 move of ubiquity
[09:33] <lool> I just sent a merge proposal which I think will help resolve it
[09:33] <evand> indeed, I took a brief look at that the other day, but didn't come up with anything
[09:33] <lool> I didnt do a test build though
[09:33] <evand> oh, awesome!
[09:33] <lool> Would be cool if we could have an ubiquity upload not too far in the future: we'd like to do some livefs images on armel, we'd like to confirm they are fixed
[09:34] <evand> speaking of which, is there a best practices document for building for ARM?  Something with qemu and pbuilder, that is.  I ran into qemubuilder, but I don't know if that's used much inside the mobile team.
[09:34] <lool> (never got a working livefs image in karmic on armel so far!)
[09:34] <evand> lool: I'll sort it out now
[09:34] <lool> evand: I think qemubuilder should work albeit I dont use it; I suspect it's very slow; you have access to the ports machine on request to IS if you like
[09:34] <lool> It's rimu.canonical.com
[09:35] <evand> oh cool, I'll fill out an RT to get access, just in case something crops up again.
[09:35] <lool> evand: Another way is pushing to the ~canonical-arm-dev PPA; the reason it's limited to canonical is that the builders are virtualized like the i386/amd64/lpia ones under xen so IS prefers giving access sparingly to such non-virtual PPAs
[09:35] <lool> We can add you to the team if that's useful to you
[09:36] <evand> ideally I'd like to be able to build in rapid succession, without filling your PPA with lots of test builds
[09:36] <lool> evand: The porter machine will probably change soon, but you'll stick in the porters group for the next one: we're moving from armv5 to armv7 buildds so it will be a new host in a couple of weeks I expect (hope :)
[09:37] <evand> cool
[09:37] <lool> evand: Worst case, I can give you ssh access to real hardware; I have a bunch of boards running with remote SSH for some folks
[09:37] <evand> very cool, thanks.  I'll let you know if it gets to that point.
[09:39] <lool> evand: Hmm sorry I didn't include a changelog entry
[09:40] <lool> Usually I checked whether the projects are debcommit style or not but I forgot this time around
[09:41] <evand> no worries, I can craft one with the appropriate credits.
[09:56] <CIA-33> ubiquity: evand * r3388 ubiquity/debian/ (ubiquity.dirs-any changelog rules):
[09:56] <CIA-33> ubiquity: Fix code around the debhelper 7 migration to build on ARM again.
[09:56] <CIA-33> ubiquity: Thanks Loïc Minier!
[10:22] <evand> just trying to sort out why localechooser/debian/templates.tmp is ending up in the delta with this change
[10:41]  * StevenK laments the lack of wubi fixes for UNR
[10:46] <CIA-33> ubiquity: evand * r3389 ubiquity/debian/changelog: releasing version 1.99.10
[10:52] <evand> StevenK: looking into it now
[10:53] <StevenK> evand: Thanks :-)
[11:02] <CIA-33> wubi: evand * r144 trunk/ (debian/changelog src/wubi/frontends/win32/cd_menu_page.py):
[11:02] <CIA-33> wubi: Factor in free space on the source device when calculating whether
[11:02] <CIA-33> wubi: it's too big to use Wubi with. This fixes using Wubi on
[11:02] <CIA-33> wubi: sufficiently large USB disks.
[11:06] <evand> StevenK: mind giving http://people.canonical.com/~evand/wubi/karmic/wubi-r144.exe a try?  New CD builds will pick it up as well.
[11:10] <evand> bah, ubiquity still broken.  Investigating.
[11:13] <StevenK> evand: Certainly -- might take me a few
[11:13] <evand> no worries, take your time
[11:15] <evand> ah, spotted it
[11:20] <CIA-33> ubiquity: evand * r3390 ubiquity/debian/ (changelog rules): Correct test for ubiquity.dirs.$DEB_HOST_ARCH.
[11:21] <lool> evand: Argh I forgot to rename the debian/ubiquity.dirs.foo to .dirs-foo
[11:22] <lool> Sorry about that
[11:22] <davmor2> evand: will the wubi fix work for kne too?
[11:23] <evand> lool: why rename them?  Isn't okay to just modify the test and copy to have a period instead of a dash?  (As that's what I just did).
[11:23] <evand> davmor2: remind me what the issue with kne was?
[11:23] <evand> incidentally, I'll look at the cd builds after I'm done with ubiquity
[11:23] <lool> evand: dh_installdirs will use the . file
[11:23] <lool> evand: That is on armel it will prefer ubiquity.dirs.armel instead of ubiquity
[11:24] <lool> Which is why a dash is used so that dh doesn't see the arch specific files for its arch override logic and we can use them for our arch complement logic
[11:25] <davmor2> evand: it shows wubi but after you type in your name and pass and hit next it  does nothing.  It can't find the cd (there isn't one) and caouldn't download anything as networking was broken in kne alpha 4
[11:26]  * lool hope he is clear enough
[11:26]  * evand processes
[11:35] <evand> lool: Is this correct then: http://pastebin.ubuntu.com/256821/ ?
[11:36] <evand> davmor2: I'm a bit confused by that statement.  Can you elaborate on what you mean by "networking was broken in kne alpha 4"?  Wubi's networking code?
[11:40] <davmor2> evand: sorry me going mad and getting myself confused.
[11:41] <davmor2> evand: ignore that bit
[11:41] <evand> but it still does not download the CD?
[11:42] <davmor2> evand: I got a strange firewall on my xp install that was causing network issues so even if wubi tried it couldn't download at the time.
[11:43] <davmor2> evand: however I spent entire day installing slowly xp onto my netbook so I could test this on real hw properly now :)
[11:43] <evand> I'm assuming this was before we put in the fix to ignore dashes in the distro name.  Should be fixed now.
[11:43] <evand> If the next set of CDs aren't working for you, please file a bug and poke me
[11:46] <davmor2> evand: I might try and squeeze in a kne this afternoon with a wubi test we'll see
[11:46] <davmor2> if not I'll give it a try over the weekend
[11:52] <lool> evand: Yes
[11:53] <evand> good deal
[11:53] <lool> evand: And it makes it really identical ot .install which was my intent
[11:53] <lool> I should really have done a test build, but bzr branching is so slow with the ubiquity branch that I didn't want to do it on an armel machine
[11:53] <CIA-33> ubiquity: evand * r3391 ubiquity/debian/ (5 files):
[11:53] <CIA-33> ubiquity: Rename ubiquity.dirs.* to ubiquity.dirs-* to avoid dh_installdirs using
[11:53] <CIA-33> ubiquity: the files as overrides.
[11:53] <lool> I wonder whether the branch could be made faster by some bzr upgrade
[11:57] <evand> we've done it once before, though I have no idea what it's at now (bzr info lp:~ubuntu-installer/ubiquity/trunk says the format is "unnamed").  I suppose we should go to 2a at some point.
[11:58] <StevenK> Heh, unnamed
[12:10] <CIA-33> ubiquity: evand * r3392 ubiquity/debian/changelog: releasing version 1.99.11
[12:42] <evand> success!  ubiquity for ARM is built.  Thanks for the help, lool.
[12:48] <lool> Thanks to you too; sorry for the initial issue
[12:50] <evand> no worries
[14:11] <davmor2> evand: did those cd's get re-spun yet?
[15:00] <StevenK> evand: Huzzah, that gives me an Install inside Windows button, which when clicked says "'NoneType' object has no attribute 'name'"
[15:17] <evand> StevenK: ok, I'll ook into it
[15:17] <evand> davmor2: respinning now
[15:18] <davmor2> evand: Cool :)
[16:13] <superm1> hm so maybe the reason standard desktop CDs aren't seeing this (if it's caused by rsyslog) is that the gnome session interface is what is getting used to reboot the machine
[16:13] <superm1> kubuntu uses a KSMServerInterface to do it
[16:23] <davmor2> evand: out of curiosity what are you actually respining?  cr3 wants to check that alternate is one of them for the automated tests
[16:27] <cr3> davmor2: I've got automated desktop testing too :)
[19:39] <CIA-33> ubiquity: mterry * r3353 plugins/ (9 files in 4 dirs): support simple plugins that don't have an ask script but want to talk to debconf
[20:06] <CIA-33> ubiquity: mterry * r3354 plugins/ubiquity/frontend/ (gtk_ui.py kde_ui.py): let plugin translation prefixes actually work in gtk/kde
[20:42] <CIA-33> ubiquity: mterry * r3355 plugins/scripts/install.py: let InstallPlugins access debconf