[00:20] <stgraber> FWIW, tested the current wubi trunk on Windows 8 and it worked to install Ubuntu from the network, didn't try with a media as I didn't have any around. Now installing Windows 8 in a VM to have a better test environment.
[00:57] <GrueMaster> I would think we should special case Windows 8 as a nuke & pave installation.  But that's just me.
[01:41] <infinity> GrueMaster: Heh.
[01:41] <infinity> GrueMaster: wubi via Windows Update?
[01:42] <GrueMaster> That would be better than the McAffee Antivirus update that made nt.dll a virus.
[01:44] <GrueMaster> Er, svchost.exe.  My bad.  Still was a funny oops.
[14:14] <cyphermox> cjwatson: thanks for the extra review for ifblacklist_migrate.sh. I don't think I'll rewrite it in python or anything else yet though, I'd just like to fix this and move on to something else. definitely something to think about later though, but at this point it's also shipped by NM for the only purpose of providing it to d-i, perhaps that should be folded into the d-i 55NetworkManager finish-install.d script
[14:16] <cjwatson> mm, we didn't do that because it was tied to what NM happened to want at any given point so it made more sense for it to be in NM
[14:16]  * cjwatson goes quietly insane writing a parted gptsync test script
[14:18] <cyphermox> ah
[15:14] <stgraber> cjwatson: Do you see a potential problem with the patch from bug 888281 (except that it's been done the wrong way around so needs reversing :))?
[15:14] <ubot2`> Launchpad bug 888281 in wubi "Boot from FAT32 error and return to grub menu" [Undecided,Confirmed] https://launchpad.net/bugs/888281
[15:22] <cjwatson> stgraber: oh!  I was ignoring it on the basis that hardcoding loop0 was obviously wrong ;-)
[15:22] <cjwatson> wonder why I did that
[15:23] <cjwatson> argh argh context-format diff with no context
[15:23] <cjwatson> I think it's wrong though - I'll follow up
[15:24] <cjwatson> there's no reason why grub's loop device ordering should match that in the booted OS, and it often won't
[15:29] <cjwatson> hm, well
[15:29] <cjwatson> I suppose that it's *telling* grub to name it that way; but it seems like it would go wrong if somebody loop-mounted something while update-grub was running
[15:32] <cjwatson> oh, no it wouldn't, that wouldn't change the existing numbering
[15:32] <cjwatson> OK, I think I've convinced myself that it's fine
[15:33] <stgraber> ok ;) are you planning a grub2 upload soon where you can bundle that patch or should I apply it now?
[15:39] <cjwatson> stgraber: I've applied it to Debian; I expect I'll upload at some point over the next few days
[15:40] <stgraber> cool, thanks
[16:22]  * stgraber starts looking at ubiquity's keyboard indicator now that we have a gi binding for it
[16:46] <kyleN> ev. Hi. how is the list of keyboard layouts in oem-config populated (precise). can we restrict the list to a target set easily?
[16:50] <ev> kyleN: it peers into console-setup's brain via debconf. I don't believe there is a way to restrict the set presented.
[16:51] <kyleN> ev. does the selection default to the best choice for the language the user selected?
[16:51] <ev> yes
[16:51] <ev> the language and timezone
[16:52] <kyleN> ev. ok thanks
[16:52] <ev> sure thing
[16:54] <cjwatson> stgraber: cool; sounds like it will take a while, though, so I think I'll go ahead and upload ubiquity if you don't mind
[16:55] <stgraber> cjwatson: I have it done here, just need testing :)
[16:56] <cjwatson> oh wow
[16:56] <stgraber> apparently the gi port has an identical API to the old python-xklavier, so I just switched to that and changed the build-deps and deps
[16:56] <cjwatson> where was the existing code?
[16:57] <cjwatson> ah, ubiquity/misc.py
[16:57] <stgraber> tests/test_misc.py and tests/test_misc.py both were simply disabled with a warning pending the port to gi
[16:58] <stgraber> so far it looks like http://paste.ubuntu.com/874822/ but it hasn't been tested yet, waiting for tests/run to finish building all the d-i components :)
[16:59] <cjwatson> if I'd known it was that easy ;-)
[17:02] <stgraber> well, then we still have all the keyboard indicator bugs, the most annoying one being bug 800561
[17:02] <ubot2`> Launchpad bug 800561 in libxklavier "No way to add other keymap than english on Live CD" [Wishlist,Confirmed] https://launchpad.net/bugs/800561
[17:04] <stgraber> >>> engine = Xkl.Engine()
[17:04] <stgraber> Segmentation fault (core dumped)
[17:04] <stgraber> stgraber@arkose-tmpMYJ4gW:~/data/code/ubiquity/ubiquity$
[17:04] <stgraber> ...
[17:05] <infinity> stgraber: That looks promising.
[17:05] <stgraber> :)
[17:06] <stgraber> always interesting to see python segfault...
[17:06] <infinity> Well, I assume the segv is in a module, not python.
[17:06] <infinity> (Or in a library linked by a module)
[17:07] <stgraber> yeah, it's just that with the old bindings we'd get a nice python stacktrace, now we get a segfault and need to debug it ... :)
[17:15] <bdmurray> cjwatson: what else would be useful in bug 949992?
[17:15] <ubot2`> Launchpad bug 949992 in update-manager "Upgrade from Lucid to Precise results in broken grub configuration" [Undecided,New] https://launchpad.net/bugs/949992
[17:17] <cjwatson> the "Removing update-grub hooks" bit is fine
[17:17] <cjwatson> anyway it's busted debconf configuration
[17:17] <stgraber> cjwatson: I'll push the gi change but keep it disabled with the old warnings, then upload ubiquity, anything else you want to include before uploading?
[17:18] <cjwatson> stgraber: nothing from me
[17:20] <CIA-32> ubiquity: stgraber * r5237 ubiquity/.bzrignore: Add po/Makefile.in to bzr ignore list
[17:20] <cjwatson> bdmurray: I've followed up to the bug
[17:22] <CIA-32> ubiquity: stgraber * r5238 ubiquity/ (4 files in 3 dirs): Switch from python-xklavier to gir1.2-xkl-1.0 though keep the code disabled as the binding segfaults
[17:24] <CIA-32> ubiquity: stgraber * r5239 ubiquity/ (d-i/manifest debian/changelog debian/control): releasing version 2.9.25
[17:26] <stgraber> doh, forgot to run pyflakes and should have ...
[17:26] <stgraber> doing some ugly branch fixing now, will take a sec
[17:32] <stgraber> ok, all fixed, hopefully nobody will have pulled during these 5 minutes...
[17:36] <stgraber> uploaded
[18:14] <stgraber> cjwatson: bug 950125
[18:14] <ubot2`> Launchpad bug 950125 in ubiquity "[ftbfs] Fails to build due to test failing" [Undecided,New] https://launchpad.net/bugs/950125
[18:14] <stgraber> and confirmed with the .25 upload (failed on the builder)
[18:15] <stgraber> that's odd because the test passed fine here on my machine
[18:15]  * stgraber checks if there's some gtk/theming upgrades available
[18:18] <stgraber> ok, no updates though I opened that unity session a few days ago, let's try with a clean one ...
[18:18] <cjwatson> #ubuntu-devel
[18:18] <cjwatson> shnatsel mentioned it there
[18:18] <cjwatson> I hate that test, it's not like it's actually accurate, but we need something like it ...
[18:19] <cjwatson> it might be sensitive to some default theme we don't build-dep on
[18:21] <stgraber> I'm also usually running without xvfb, not that it should cause a big difference
[18:21] <cjwatson> running through sbuild now
[18:22] <stgraber> ok
[18:22] <cjwatson> not that that will do much other than agree with the builder, I expect
[18:23] <stgraber> I guess we could compare the .24 and .25 build log to figure out what changed
[18:28] <stgraber> confirmed that the test passes fine on my laptop with a fully up to date Precise, so the missing build-dep sounds very likely
[18:30] <stgraber> and passes fine in xvfb too
[18:30]  * stgraber starts disecting build logs
[18:34] <stgraber> nothing obvious in the list of packages getting installed on the buildd: http://paste.ubuntu.com/874958/
[18:36] <stgraber> I'll try a PPA upload adding light-themes as build-dep, see if the helps
[18:36] <cjwatson> I wonder about locales
[18:36] <cjwatson> though you'd think we'd have noticed
[18:37] <cjwatson> I think I've tried light-themes before in an attempt to make that test more reliable and it wasn't obvious it helped
[18:37] <cjwatson> but feel free
[18:37] <cjwatson> do you have a version diff?
[18:39] <stgraber> cjwatson: http://paste.ubuntu.com/874972/
[18:39] <cjwatson> odds on gtk I reckon :)
[18:40] <stgraber> gnome-icon-theme might also be interesting :)
[18:40] <stgraber> but yeah gtk is 1st on the blame list as usual ;)
[18:48] <stgraber> test build running with a build-dep on ubuntu-artwork
[18:48] <stgraber> I suspect build-dep on light-themes wasn't enough as the gsettings override file is shipped by ubuntu-artwork
[19:12] <stgraber> cjwatson: yay! https://launchpad.net/~stgraber/+archive/experimental/+build/3270148
[19:13] <stgraber> so apparently build-dep on ubuntu-artwork does the trick
[19:13] <cjwatson> interesting!
[19:13] <stgraber> should I just commit that to trunk and release 2.9.26?
[19:14] <cjwatson> yeah, please
[19:14] <cjwatson> thanks :-)
[19:14] <stgraber> np
[19:15] <CIA-32> ubiquity: stgraber * r5240 ubiquity/debian/ (changelog control): Add a build-depend on ubuntu-artwork to force the right theme to be used in the unit tests
[19:16] <CIA-32> ubiquity: stgraber * r5241 ubiquity/debian/changelog: releasing version 2.9.26
[22:26] <stgraber> right, so looks like ubiquity will now also contain a bluetooth agent ;) (bug 644198)
[22:26] <ubot2`> Launchpad bug 644198 in ubiquity "Ubuntu LIve Cd does not enable bluetooth before choice menu in live-cd" [High,Triaged] https://launchpad.net/bugs/644198
[22:35] <stgraber> cjwatson: hmmm https://launchpadlibrarian.net/95997096/buildlog_ubuntu-precise-amd64.ubiquity_2.9.26_FAILEDTOBUILD.txt.gz now ubuntu-artwork fails to install on amd64 for some reason ... I guess I'll just retry the build as it worked in my PPA
[23:16] <stgraber> infinity: any clue as to what's happening there https://launchpadlibrarian.net/96008343/buildlog_ubuntu-precise-amd64.ubiquity_2.9.26_FAILEDTOBUILD.txt.gz ?
[23:17] <stgraber> infinity: it built fine on all the other architectures and worked in my amd64 PPA (though a bit earlier)
[23:17] <infinity> Archive skew, perhaps?
[23:17] <stgraber> I downloaded the LP chroot and ran the same apt-get in it and couldn't reproduce
[23:17] <infinity> amd64 is a bit backlogged.
[23:17] <infinity> Though, if you can't reproduce, that's odd.
[23:17] <stgraber> yeah and I'd have expected to see the skew on my laptop too (I'm running amd64) and I definitely don't
[23:18]  * infinity has a look.
[23:18] <stgraber> well, I'm affected by an archive skew but not that particular one apparently ;)
[23:18] <stgraber> (gwibber is held back here)
[23:19] <infinity> I might have to push for zram-config by default on all installs next cycle.
[23:19] <infinity> Installing it on my laptop has made it much happier.
[23:20] <stgraber> yeah, it sounds interesting and much cleaner than compcache
[23:20] <infinity> It kinda just works.  And my ghetto upstart job seems to do the trick.
[23:22] <infinity> Can reproduce here just fine.
[23:22] <infinity> It's the gconf skew.
[23:23] <infinity> (Which I already scored up for other reasons)
[23:23] <infinity> So, retry the build in ~40 minutes, should be fine then.
[23:23] <stgraber> hmm, ok, wondering if it's not my squid being a bit too clever and making me avoid that specific skew, that'd explain it I guess