[08:31] <ev> CarlFK: apologies, I had to end on time yesterday as I had evening plans
[08:31] <ev> cjwatson: awesome!
[08:31] <ev> bdmurray: that sounds reasonable, but I agree with cjwatson: a series of steps we can reproduce in a VM would be stellar.
[09:41] <cjwatson> bdmurray: I've never tried to set up Windows in a VM; though one thing I'd say is that quite often Windows partitioning bugs should be reproducible using mkntfs in Linux
[09:46] <CIA-31> ubiquity: evand * r4594 trunk/ (4 files in 3 dirs):
[09:46] <CIA-31> ubiquity: Add a partitioning heading to help clarify what operating systems
[09:46] <CIA-31> ubiquity: may be acted upon by the installer.
[09:48] <ev> trying to reproduce this "everything gets deleted when I select replace Ubuntu" bug now, as there haven't been any follow ups on the bug
[10:58] <ev> bum, https://bugs.launchpad.net/ubuntu/+source/partman-auto/+bug/740881
[10:58] <ubot2> Launchpad bug 740881 in partman-auto "Replace option does not call swapoff for any swap partitions on the disk" [Critical,Confirmed]
[10:59] <ev> (trying to make more of a habit out of filing bugs as I see them, rather than writing them down in notepads as things to do, in the hopes that less falls through the cracks)
[11:35] <CIA-31> ubiquity: evand * r4594 trunk/ (4 files in 3 dirs):
[11:35] <CIA-31> ubiquity: Add a partitioning heading to help clarify what operating systems
[11:35] <CIA-31> ubiquity: may be acted upon by the installer.
[11:35] <CIA-31> ubiquity: evand * r4595 ubiquity/ (debian/changelog ubiquity/misc.py):
[11:35] <CIA-31> ubiquity: * Remove ' (loader)' from the os-prober operating system descriptions on
[11:35] <CIA-31> ubiquity:  the initial partitioning page.
[11:35] <CIA-31> ubiquity: evand * r4596 ubiquity/ (debian/changelog ubiquity/plugins/ubi-partman.py):
[11:35] <CIA-31> ubiquity: * Properly substitute in the non-Ubuntu system when generating descriptions
[11:35] <CIA-31> ubiquity:  for dual boot partitioning.
[11:36] <CIA-31> partman-auto: evand * r603 ubuntu/ (automatically_partition/replace/do_option debian/changelog):
[11:36] <CIA-31> partman-auto: Disable swap partitions on the same disk before deleting the
[11:36] <CIA-31> partman-auto: partition in the replace option (LP: #740881).
[11:39] <CIA-31> partman-auto: evand * r604 ubuntu/debian/changelog: releasing version 93ubuntu12
[11:40] <ev> of course that invoked the oom killer on swapoff for me in testing, but there's not much we can do there right now :-/
[11:40] <ev> (in natty, that is)
[11:42]  * cjwatson tries to figure out the exact algorithm for parsing /proc/self/mountinfo, per bug 738345
[11:42] <ubot2> Launchpad bug 738345 in grub2 "Grub2 entries to boot Wubi incorrect... "Error: file not found, Press any key to continue" at boot" [High,Triaged] https://launchpad.net/bugs/738345
[11:43] <cjwatson> I need to do something with the mnt_id and mnt_parent->mnt_id fields
[11:43] <cjwatson> preferably without actually having to build the full graph structure in GRUB
[11:43] <ev> heh
[11:43]  * cjwatson takes the rare step of getting out some paper
[11:44]  * ogra_ wonders if having no locales on the image at all might cause the breakage of his serial oem-config UI
[11:45] <cjwatson> ogra_: that's possible, you should really have at least en_US.UTF-8
[11:45] <ogra_> i'll generate it and see if something changes
[11:45] <cjwatson>   * Run the debconf frontend in a UTF-8 locale (en_US.UTF-8, since d-i
[11:45] <cjwatson>     guarantees that that always exists) so that newt uses proper window
[11:45] <cjwatson>     borders and can display non-ASCII characters.
[11:45] <cjwatson> (debian/changelog.oem-config)
[11:45] <ogra_> aww
[11:45] <ogra_> k
[11:46]  * ogra_ verifies then
[11:51] <cjwatson> OK, I think I can do this in a single linear pass
[11:55] <ogra_> heh
[11:55] <ogra_> that changed something but still looks broken
[11:59]  * ogra_ tries with TERM=vt100 again
[12:02] <ogra_> hmm, still only a cursor when using vt100
[12:03] <ogra_> and the locale generation (and setting LANG everywhere) didnt really improve much, i have different broken chars for the frames now
[12:03] <cjwatson> your terminal may need to be running in a UTF-8 locale too
[12:04] <ogra_> well, i run it on ubuntu, minicom should pick up my default locale, no ?
[12:04] <ogra_> from gnome-terminal i mean
[12:05] <ogra_> ogra@isis:~$ env|grep ^LANG
[12:05] <ogra_> LANG=de_DE.UTF-8
[12:11] <ogra_> aha ... having en_US.UTF-8 generated locally too gets me a cursor (still broken frames though) even with TERM=xterm
[12:12] <ogra_> ah, crap, no, i had exported vt100 before
[12:17] <ev> wow, I'm awful
[12:17] <ev> patch incoming for the replace issue
[12:17] <ev> also, is there a downside to using umount -f, or should we be using it everywhere?
[12:18] <cjwatson> are you using NFS a lot?
[12:19] <mterry> ev, cjwatson: how do I open a bug against geoname-lookup.ubuntu.com?
[12:19] <ev> cjwatson: no, but every now and again partman seems to get caught up around the umount calls in replace/reuse
[12:19] <ev> mterry: uhm, that's a good question
[12:19] <ev> there isn't a project for it
[12:20] <ev> creating now
[12:20] <mterry> ev, I got a bug against indicator-datetime about searching for "san francisco"
[12:20] <mterry> ev, and it seems to be a server issue
[12:20] <cjwatson> mterry: I think I'd use RT
[12:20] <ev> or that
[12:20] <cjwatson> ev: umount -f is only documented as helping with NFS.  Maybe umount -l instead?
[12:20] <ev> actually
[12:20] <ev> I control the code, I think
[12:21] <mterry> ev, so I come bug you on IRC then?  :)
[12:21] <mterry> ev, https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/740874
[12:21] <ubot2> Launchpad bug 740874 in indicator-datetime ""Location" auto-complete menu doesn't know about San Francisco, USA" [Undecided,New]
[12:22] <ev> cjwatson: sure, lazy would work, presumably
[12:24] <ev> mterry: digging
[12:28] <ev> https://rt.admin.canonical.com/Ticket/Display.html?id=39986
[12:28] <ev> https://rt.admin.canonical.com/Ticket/Display.html?id=42610
[12:28] <ev> https://code.launchpad.net/~ev/+junk/geoname
[12:28] <ev> I'll make a proper project for that now
[12:31] <mterry> ev, ok, I'll assign a couple bugs to it
[12:34] <ev> mterry: cool
[12:34] <ev> https://code.launchpad.net/ubuntu-geonames
[12:34] <ev> we'll still have to bug IS to do pulls
[12:35] <ev> as I don't believe I have access to that server
[12:36] <mterry> ev, OK, targetted a couple to the project.  thanks
[12:40] <ev> okay, I created a team and added you, so you should be able to commit directly
[12:40] <ev> if you need to and you can't, let me know and I'll chase it down
[12:41] <mterry> k, thanks
[12:46] <ev> I've also updated those two RTs to point at it
[12:51] <CIA-31> partman-auto: evand * r605 partman-auto.ubuntu/ (automatically_partition/replace/do_option debian/changelog):
[12:51] <CIA-31> partman-auto: * Make sure we have a partition selected for autopartitioning, not the
[12:51] <CIA-31> partman-auto:  disk, when using the replace option (LP: #738366).
[12:51] <ev> ugh
[13:26] <ogra_> oh, screen works a lot better than minicom, intresting
[13:26] <ogra_> UTF-8 is still broken though, but i have colors and a cursor
[13:59] <ev> nice, agy already pointed at the new branch location for geonames
[14:00] <ev> someday we'll get a real layout engine in our UI toolkit... :-/
[14:05] <mterry> ev, is geonames localized in any fashion?  always seems to give back english answers?
[14:06] <ev> if memory serves, not exactly
[14:06] <ev> we're at the mercy of the data dump from geonames.org
[14:06] <cjwatson> that's what we want in the installer case, I think - localisation happens on the client side
[14:08] <ev> localizing on the client side for the results of a search in the timezone box would be quite the endeavor as it returns more than olsendb locations
[14:09] <mterry> It would be nice if we could just add the language in the query, right?  Like &lang=fr
[14:10] <mterry> then get the same set of results, but with french names from the database
[14:15] <ev> yeah, that would be nice, though we'd have to add tables or construct the database differently for that
[14:16] <ev> I think
[14:16]  * ev looks a bit harder rather than just guessing :-P
[14:20] <mterry> It's definitely in the source data, but yeah, we might not be replicating it in the database
[14:20] <ev> actually, one way of doing it would be to just to try to match against alternatenames and return that instead of name
[14:21] <ev> it wouldn't ensure all the text was localized
[14:21] <mterry> well, if we don't have a french translation for something, better to show english than nothing i'd say
[14:21] <ev> but if you searched for espana, you wouldn't get spain back
[14:21] <ev> indeed
[14:22] <mterry> I think that would be fine, as long as you were in spanish locale
[14:23] <ev> I don't think they're mapped to locales in the database
[14:23] <ev> its a comma separated varchar(500)
[14:23] <ev> alternatenames, that is
[14:24] <ev> I know, lets just create a script that pipes it all through google translate once and tries to guess the language table to place it in ;)
[14:26] <mterry> ev, the 3rd column in alternatenames seems to be a locale
[14:27] <mterry> or maybe that's just the country of the location?
[14:27] <ev> ohhh
[14:27] <ev> I was looking at the column in geonames, not the table
[14:27] <ev> neat
[14:28] <mterry> this data is whack.
[14:28] <mterry> 120903  165131          Tell Sabaïne
[14:28] <mterry> 120905  165131          Tall Şabayn
[14:28] <mterry> 120904  165131          Tell Sanaïne
[14:28] <mterry> those are presumably all the same place?  but different names?
[14:28] <ev> oh yeah, it's all sorts of weird in places
[14:28] <mterry> they have different locations though and no 3rd column
[14:43] <cjwatson> ev: oh, that kind of answer, OK
[14:43] <cjwatson> I was thinking of timezone IDs
[16:05] <cjwatson> argh, I fix one bit of Wubi and another mysteriously falls over
[16:06]  * cjwatson plays whack-a-mole
[16:08] <cjwatson> double argh, now kvm is being mysteriously silent
[16:09] <ev> wuh oh
[16:13] <cjwatson> ah, it was bug 736351
[16:13] <ubot2> Launchpad bug 736351 in vgabios "natty vgabios installs links into /usr/share/kvm" [Medium,Fix released] https://launchpad.net/bugs/736351
[16:14] <cjwatson> I think.  Upgrading cleared it up, anyway
[16:35] <CIA-31> grub-installer: cjwatson * r1200 ubuntu/debian/changelog: releasing version 1.60ubuntu2
[16:37] <CIA-31> localechooser: cjwatson * r160 ubuntu/debian/changelog: releasing version 2.34ubuntu3
[16:38] <CIA-31> partman-auto: evand * r607 ubuntu/debian/changelog: releasing version 93ubuntu13
[16:42] <CIA-31> partman-auto-loop: cjwatson * r56 ubuntu/debian/ (changelog partman-auto-loop.templates po/templates.pot): Improve wording of partman-auto-loop/unclean_ntfs.
[16:49] <CIA-31> partman-auto-loop: cjwatson * r57 ubuntu/debian/po/templates.pot: fix Report-Msgid-Bugs-To
[16:50] <CIA-31> partman-auto-loop: cjwatson * r58 ubuntu/debian/changelog: releasing version 0ubuntu19
[17:16] <bdmurray> ev: is there a reason bug 740903 doesn't have a natty task?  I'm pretty sure that kate keeps an eye on the natty tasks so that wouldn't show up.
[17:16] <ubot2> Launchpad bug 740903 in ubiquity "return_to_partitioning fails when the replace option fails" [High,Confirmed] https://launchpad.net/bugs/740903
[17:17] <ev> bdmurray: done, thanks.
[17:18] <bdmurray> ev: thanks, I'd have been happy to do it - just wanted to make sure
[17:18] <ev> right-o
[17:20] <ev> any objections to a new ubiquity?
[17:22] <bdmurray> that'd be good for the bug day fwiw
[17:36] <CIA-31> ubiquity: evand * r4598 trunk/ (139 files in 3 dirs): Update translations from Launchpad.
[17:38] <ev> ah, looks like I'm still blocked on builds/publisher anyway
[19:24] <Jemt> Hello. The ubuntu user on the Ubuntu Live CD (remastered) does not honor settings in /usr/share/polkit-1/actions/org.freedesktop.upower.policy (I disabled Suspend and Hibernation). However, it works for new users created on the Live CD. How can I solve the problem for the default ubuntu user ?
[19:25] <Jemt> (I ask in here since this seems to be the place were the actual Ubuntu developers/geeks are hanging out)
[20:15] <CIA-31> ubiquity: evand * r4599 trunk/ (d-i/manifest debian/changelog):
[20:15] <CIA-31> ubiquity: Automatic update of included source packages: console-setup
[20:15] <CIA-31> ubiquity: 1.57ubuntu17, grub-installer 1.60ubuntu2, localechooser 2.34ubuntu3,
[20:15] <CIA-31> ubiquity: partman-auto 93ubuntu13, partman-auto-loop 0ubuntu19.
[20:54] <CIA-31> ubiquity: evand * r4600 trunk/debian/changelog: releasing version 2.5.28
[21:14] <CIA-31> ubiquity: evand * r4601 trunk/ (debian/changelog po/LINGUAS): Remove cy from LINGUAS. Fixes build failure.
[21:17]  * ev grumbles
[21:22] <CIA-31> ubiquity: evand * r4602 trunk/debian/changelog: releasing version 2.5.29
[22:09] <ev> hm, it would be nice if the debconf protocol had a means to get the list of variables used for substitution
[23:14] <ev> well, worked around it for now by manually parsing templates.dat
[23:15] <ev> (want to be able to assert question_only_has_variables(question, variables=[]) in tests, in case the template changes once the tests are written)
[23:16] <ev> err rather, I now do