=== yofel_ is now known as yofel | ||
jbicha | hey, is https://lists.ubuntu.com/archives/quantal-changes/2012-October/011790.html something that needs to land in Edubuntu before release | 18:25 |
---|---|---|
jbicha | I'm a bit confused by the description but it sounds like it won't migrate quantal users who have already logged in to GNOME Classic | 18:27 |
stgraber | jbicha: it sounds like it needs to be in the archive before release, yes | 18:27 |
stgraber | jbicha: it doesn't need to be on a media though, so as long as it's a 0-day SRU and is published in -updates by the time we release, we'll be good | 18:27 |
jbicha | what if somebody installs from the CD and logs in to GNOME Classic before applying updates? | 18:28 |
stgraber | jbicha: I'm not sure and I don't have time to look into this (nor time to respin anyway), but my understanding is that it's only touching the migration code, is that incorrect? | 18:34 |
jbicha | I haven't tested yet | 18:37 |
jbicha | GNOME Classic with effects has been broken in quantal since the gsettings switch and it just sounded from the changelog like the fix doesn't work right for existing quantal users | 18:38 |
stgraber | livecd/ubiquity/target-config/60edubuntu-default-session: chroot /target /usr/lib/lightdm/lightdm-set-defaults --session gnome-fallback || true | 18:39 |
stgraber | so not a problem, we don't use classic, we use fallback | 18:39 |
stgraber | as classic was indeed busted for beta2 and I got bored and hardcoded fallback everywhere | 18:39 |
stgraber | (according to the chngelog ;)) | 18:39 |
jbicha | ok, maybe less of an issue, but with effects is still on your image | 18:39 |
stgraber | yeah, at least users won't be getting it by default, so 0-day SRU is fine | 18:40 |
jbicha | ok, so Edubuntu 12.04 uses Unity by default but 12.10 uses Fallback? | 18:40 |
stgraber | default for desktop is Unity, default for LTSP is fallback | 18:40 |
stgraber | and we have an installer option to use fallback | 18:40 |
jbicha | ok | 18:40 |
stgraber | only way you get classic is by choosing it at login time | 18:41 |
highvoltage | edubuntu 12.04 uses unity 2d by default for ltsp and 12.10 uses fallback by default for ltsp. both use 3d by default | 18:41 |
stgraber | ideally I'd want to switch to classic for 13.04 assuming it's all fixed once the new compiz lands | 18:41 |
jbicha | and are you aware that GNOME plans to kill Fallback, probably for GNOME 3.8? | 18:41 |
stgraber | (but keep unity as the default for desktop, just replace fallback by classic) | 18:41 |
stgraber | yeah, I heard of the plans and they suck but we can't do much about it... | 18:41 |
jbicha | and by killing, they mean removing support in gnome-settings-daemon and other places | 18:42 |
jbicha | I'm thinking Ubuntu will still with gnome-settings-daemon/gnome-control-center 3.6 next cycle so Classic may still be ok for 13.04 | 18:42 |
jbicha | here's the meta bug https://bugzilla.gnome.org/show_bug.cgi?id=682858 | 18:43 |
ubottu | Gnome bug 682858 in general "[meta] Remove fallback support code" [Normal,Unconfirmed] | 18:43 |
stgraber | yeah, at this point I'm trying to ignore what's going on with the desktop stuff and focus on the server side of things where things at least go on the right direction... | 18:44 |
stgraber | it looks like to me that gnome upstream is extremely successful at getting rid of their users, the problem is, I don't feel like shipping a second full desktop environment either, so I'm not sure what we'll end up shipping for 14.04... | 18:45 |
highvoltage | stgraber: is it normal behaviour that a language other than english prompts for folder name changes on first login? | 19:30 |
highvoltage | (seems a bit odd) | 19:30 |
highvoltage | (and I only got it on i386 (so far)) | 19:31 |
stgraber | highvoltage: that part of nautilus/xdg-dirs is racy, we've seen a few reports where you get duplicate entries or badly translated ones, ... | 19:31 |
highvoltage | hmm, and it seems to be telling me that my langpack downloads are incomplete. perhaps that's part of it. | 19:31 |
highvoltage | oh and it popped up 3 more times, eish | 19:31 |
stgraber | oh, showing up multiple times is a first though :) | 19:32 |
highvoltage | is simplified chinese last one or second last item on the isolinux menu? | 19:32 |
stgraber | the one with the least strokes | 19:32 |
highvoltage | ok | 19:33 |
stgraber | I never remember the order, I just look for the "simple" one :) | 19:33 |
highvoltage | (it's the second-last one, ftr) | 19:34 |
highvoltage | stgraber: do those test cases get parsed by anything (I notice you added my name and date, I'll continue to do the same) | 19:36 |
stgraber | highvoltage: they get parsed by me when I'm tired, so properly aligned and following the format defined in the header is a good way of reducing my reading time ;) | 19:40 |
stgraber | 19:40 < stgraber> highvoltage: they get parsed by me when I'm tired, so properly aligned and following the format defined in the header is a good way of reducing my | 19:41 |
stgraber | reading time ;) | 19:41 |
highvoltage | stgraber: I got your message before my container died, at least :) | 19:42 |
stgraber | ok :) | 19:42 |
highvoltage | (I keep running into this strange bug where if my lxc package updates, half of my containers die, and then around 2 minutes later, the other half) | 19:42 |
highvoltage | (so I kind of knew it was coming as my irc session was in the second batch :p) | 19:42 |
alkisg | highvoltage: when creating the greek live CDs I noticed I had to run `locale-gen --purge` somewhere at the bottom of the process so that the chinese locales were removed and didn't cause the "langpack incomplete" prompts later on | 19:43 |
alkisg | Removing the pt or es locales works fine, but there's a bug somewhere in removing the chinese ones | 19:43 |
alkisg | And it's not cleanly removed, and locale-gen --purge is required etc etgc | 19:43 |
alkisg | `locale -a` after a non-chinese installation should demonstrate the problem.... | 19:44 |
highvoltage | alkisg: maybe that has something to do with there being an ubuntu chinese edition where there's some specific stuff for that | 19:47 |
alkisg | Yeah it's special-cased in some areas of the installation process | 19:47 |
stgraber | highvoltage: doing greek amd64 in oem with encrypted home now | 19:48 |
highvoltage | I'm doing simplified chinese but just a straight install since I don't want to risk clicking on something and not knowing what it means | 19:49 |
highvoltage | (although chinese isn't as well translated as french) | 19:49 |
* alkisg won't even ask if the greek keyboard is working in 12.10, he's certain that it's still broken... :-/ | 19:49 | |
stgraber | highvoltage: can you borrow a machine for dvd testing on your side? | 20:08 |
stgraber | highvoltage: I can do usb testing but I don't have any drive or blank media here | 20:08 |
highvoltage | yep | 20:11 |
highvoltage | I'll write it to DVD here and then install at home on mac mini / lg laptop | 20:13 |
stgraber | highvoltage: are you testing 20121017.1 at the moment? if so, can you also post results to iso.qa.ubuntu.com (no need for details, I just want to get the count roughly right) | 20:14 |
highvoltage | stgraber: yes, I synced right before testing | 20:15 |
highvoltage | will do | 20:15 |
stgraber | wow, it's always so weird to install in arabic... | 20:15 |
highvoltage | iso tracker seems a bit slow atm. | 20:17 |
stgraber | yeah, that usually happens around release... some results took over 30s to post here | 20:18 |
stgraber | I'll need to look at better indexing for a few fields | 20:18 |
highvoltage | what happens? does the db get hit real hard? | 20:18 |
highvoltage | or is it on a host with other stuff? | 20:18 |
stgraber | it's running on its own server (quadcore xeon). Seems like mostly IO bound and postgres not being properly tweaked | 20:20 |
stgraber | I have 5GB of free memory so it's a bit weird seeing the DB being so slow | 20:20 |
stgraber | I'd need to do some tweaks to postgres/apache but I first need to replicate the load issues on staging and then ask IS to update the config, so it takes time... :( | 20:21 |
stgraber | my current guess is that the query looking for bug history is taking quite long to parse the LP bug table and that I need some better indexes on that stuff | 20:23 |
stgraber | doh, yeah, found the problem, the LP bug table isn't indexed by bugnumber... | 20:23 |
stgraber | highvoltage: added the index, you should see a huge speed difference | 20:33 |
highvoltage | stgraber: cool | 20:35 |
stgraber | just testing wubi on i386 then will mark that one as ready too | 22:11 |
stgraber | and then will go to bed :) | 22:12 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!