[10:25] bug 725408 is getting odder by the minute, it seems [12:38] phew, partman crashed while resizing and I was worried there was an undiscovered bug here [12:38] fortunately it was just the OOM killer putting my in my place [12:38] I should probably have a look at memory usage in O though. I'm sure we can trim some fat from ubiquity. [12:48] reimplement in C [12:51] hahaha [12:51] you do realize you're only trolling yourself, being the original author and all [12:55] not quite original :) [12:58] original required some spanish skills iirc :) [12:59] ah yes [12:59] I forgot about them [13:33] Is the default and intended behaviour in ubuntu to always mount the top-level of the btrfs, and use rootflags=subvol=@ to mount root for example? [13:34] ...or rather, have the top-level as the "default root" when it is mounted with no options? [13:37] yes [13:38] well, ideally there would be a way to set @ as the default subvolume without making it hideously painful to mount other subvolumes, but AFAIK that isn't straightforward in btrfs at the moment [13:38] on the whole I'd not recommend relying on any particular behaviour of mounting an Ubuntu btrfs installation with no options at the moment [13:40] Ok, there was a user which used a grub.cfg with no rootflags=subvol=@ option which claimed to be able to boot, and had not touched the set-default, so thus I was confused. [13:42] that's due to a grub2 bug [13:42] it sometimes fails to parse the btrfs filesystem properly and so omits subvol=@ [13:42] I keep saying this in different bug reports but apparently there are enough of them that people don't notice [13:43] Yea, I've been trying to figure this out in Bug #752506 (this is basically the issue you are talking about right?) [13:45] So this is an issue in grub-probe, rather? [13:46] there's already a triaged bug for it somewhere [13:47] bug 732149 [13:47] it's very likely the same thing, I'll get to it as soon as I can manage to reproduce it [13:52] The way I've been able to do it is basically repeatedly reinstalling the linux-image and linux-headers packages, but it is not by far a sure way to do it. [13:54] ok, thanks [13:55] I don't need a 100% method, it's OK if it takes a while to tweak the filesystem into the right (wrong?) state [14:37] ev, how ya doing? [14:38] mterry: pretty good, and yourself? [14:38] ev, klattimer and I think we've got localized geonames results [14:38] hooray! [14:38] two branches: https://code.launchpad.net/~mterry/ubuntu-geonames/search-alt/+merge/56647 and https://code.launchpad.net/~karl-qdh/ubuntu-geonames/with-charset/+merge/56753 [14:38] first allows one to search over the altnames [14:38] second allows you to use funky characters [14:38] we swiped it from a forum post [14:39] seems to allow results for chinese characters for example, though we don't know how accurate it is, since we don't speak chinese [14:39] but we figured we can wait for bug reports... :-/ [14:39] hah [14:39] awesome [14:39] odd, I just saw his merge request but yours hasn't come through yet [14:40] I'll have a look now [14:41] ah, nevermind, there it is [14:47] ev, any idea what the default timeout is supposed to be on the geonames stuff in ubi-timezone? it doesn't seem to be explicitly set [14:51] oo, yikes that's quite the oversight [14:51] from what I can tell reading the code there is no timeout [14:52] as that's determined by socket.setdefaulttimeout(), which has a default value of None, indicating no timeout. [15:05] netcfg: cjwatson * r1249 ubuntu/ (96release-dhcp-lease debian/changelog debian/netcfg.install): [15:05] netcfg: Move release-dhcp-lease finish-install script after unmounting [15:05] netcfg: filesystems, to avoid interfering with iSCSI. [15:08] netcfg: cjwatson * r1250 ubuntu/debian/changelog: releasing version 1.60ubuntu3 [15:15] ev, oh fun. that explains my weirdness where I can't get out of hangs behind a proxy [15:24] superm1: not sure why I went with urllib2 in the first place, given that I presume asynchronous callbacks are available in gio for this sort of thing. [15:24] with a timeout still set, of course [15:43] mterry: looks good, about to merge. [15:43] ev, awesome. hopefully we get some non-English feedback [15:43] well I just tried Shanghai in Chinese and it worked fine, but indeed [15:45] good [15:45] oh interesting [15:46] I was actually trying without the branches merged by accident and the search worked (though returned English results) [15:46] no matter [15:52] ev, well... I'm surprised it matched the chinese [15:52] ev, but you could get altname results before, but they would return the canonical English name, yah [15:53] indeed [15:55] we should make unit tests for this. Perhaps a fun thing for me to play with over the weekend if I find time. [16:02] mterry: does http://localhost/?query=上海 work for you? Wondering if I just haven't recached something. [16:03] ev, ('Shanghai - \xe4\xb8\x8a\xe6\xb5\xb7', 'Shanghai Shi', 'China', 121.45806, 31.22222) [16:03] ('\xe4\xb8\x8a\xe6\xb5\xb7', 'Shanghai Shi', 'China', 121.45806, 31.22222) [16:03] ('\xe4\xb8\x8a\xe6\xb5\xb7', 'Shanghai Shi', 'China', 121.45806, 31.22222) [16:03] ('\xe4\xb8\x8a\xe6\xb5\xb7', 'Shanghai Shi', 'China', 121.45806, 31.22222) [16:03] ('\xe4\xb8\x8a\xe6\xb5\xb7', 'Zhejiang Sheng', 'China', 121.05804, 29.32955) [16:03] we should probably do something about those duplicate results. they were there before this round of patches and I've dealt with it client-side [16:03] but just seems inefficient [16:08] ah, the problem was chrome [16:08] doing it by hand in python worked fine [16:08] okay merging for reals now === robbiew1 is now known as robbiew [16:16] contacting IS now [16:21] and done [16:45] ev, something I realized... the admin and country names aren't translated... does the data provide that? [17:54] ev, asynchronous would be better for the whole page if possible. there's a hang before the page starts when rdate is trying to contact ntp.ubuntu.com too [18:38] ev: can you look over and maybe triage bug 728710 [18:38] https://bugs.launchpad.net/ubuntu/+source/debian-installer/+bug/728710 "d-i apt-setup/local0/repository not working" [19:09] any clue how long it is? [19:09] whops, wrong chan [22:50] rootskel: cjwatson * r630 ubuntu/ (9 files in 6 dirs): merge from Debian 1.94 [22:53] rootskel: cjwatson * r631 ubuntu/debian/changelog: releasing version 1.94ubuntu1 [23:13] debian-installer: cjwatson * r1435 ubuntu/ (57 files in 2 dirs): Update help text translations from Launchpad. [23:16] debian-installer: cjwatson * r1436 ubuntu/debian/changelog: releasing version 20101020ubuntu27 [23:16] (actually arranging to have that uploaded in two hours' time via at(1), in order to include the new rootskel)