[07:50] <xnox> cjwatson: true. I've wrote a python script with python-pam help, to start / authenticate / open_session and that does not create logind session for me. (well at least loginctl list-sessions doesn't who it). Alternatively executing dbus calls to logind API to CreateSession (similar to how CK worked) ends up killing the master pid for the session and everything below it.
[07:56] <xnox> well, I think I see what was wrong. Switching to pam service to "su" makes it work, and getenvlist is now giving me XDG_* / logindish variables.
[10:45] <xnox> looks like ubuntuone plugin doesn't like to be offline =/
[13:48] <brendand> xnox, is there any known issue with creating precise usb sticks using usb-creator-gtk in saucy?
[13:48] <brendand> xnox, mine aren't working
[13:48] <xnox> brendand: use "dd"
[13:49] <xnox> brendand: in general they should work & boot in legacy bios mode.
[13:49] <brendand> xnox, any instructions for that?
[13:49] <brendand> xnox, i do notice a suspicious 'Installing EFI bootloader' message when creating the image
[13:49] <xnox> brendand: "dd if=path/to/precise-desktop.iso of=/dev/sdX" where /dev/sdX is the usb-stick device.
[21:26] <GrueMaster> Anyone seeing a bad keyboard mapping when using oem-config on an installed server running 12.04?
[21:43] <GrueMaster> On boot after running oem-config-prepare on a 12.04.2 installed server, oem-config starts and asks for launguage.  Problem is the arrow keys don't work properly, and you can't select the proper language (english in this case), nor can you use the tab key or the enter key.  Pressing F5 gives me the english selection, but thats as far as I have been able to get.  We have tried 5 different brands of keyboards (all USB), and even remove
[21:43] <GrueMaster>  console to the server (Intel RMM4).
[21:44] <GrueMaster> The real odd part is that <ctrl><alt>-F2 will give me a login prompt, and I can login just fine with the account I created initially.  No keyboard issues.
[21:46] <infinity> GrueMaster: Can you file a bug?  We're mid-beta here, and not many people available to do real-time debugging.
[21:47] <GrueMaster> I'm looking through launchpad at the moment to see if there may be one already.  Lots of open, untouched bugs.
[21:47] <GrueMaster> But yes, I plan on filing one.
[21:47] <cjwatson> ubiquity gets too many bugs for it to be possible to address them all with anything like current staffing.  But that doesn't mean we don't look at interesting ones called to our attention ...
[21:48] <GrueMaster> May want to see if this exists in the current beta as well.
[21:48] <cjwatson> (I suspect bterm isn't being hooked up properly)
[21:49] <GrueMaster> cjwatson: THe odd part is this is fairly recent.  It worked fine after you fixed the font compiler issue.
[21:49] <GrueMaster> I suspect a separate package update may be causing this, not oem-config directly.
[21:49] <cjwatson> Possible.  oem-config does a fair bit of its own console setup though ...
[21:50] <cjwatson> But yes, it's possible it could be broken by (say) plymouth, console-setup, kbd, kernel
[21:51] <GrueMaster> I have a little idle time this week.  I'll try to root out the issue before I file a bug.  If I can, I'll also test the beta on one of our servers, just in case.
[21:52] <cjwatson> Thanks, would be appreciated.  Bed soon here, trying to figure out some buildd QA work first
[21:52] <GrueMaster> overworked/understaffed.  Sounds like my job here.  :P