[09:21] <tjaalton> ev, xnox: I was debugging an issue with ubiquity last friday, it doesn't recognize uefi-installation of win8 here, so can't resize the disks to make room for saucy
[09:21] <tjaalton> so was wondering you might have a clue of what's missing?
[11:25] <tjaalton> cjwatson: 'WINOSDATA=true os-prober' doesn't return anything here, so I guess os-prober still needs some fixing. dunno if this is used strictly for migration-assistant only, and not partman on ubiquity..
[11:30] <tjaalton> right, it's the check for /sys/firmware/efi in 20microsoft, which makes it skip the rest, so m-a doesn't work on uefi systems at all
[11:36] <tjaalton> adding a condition for '&& [ ! -n "$WINOSDATA" ] seems to work
[11:37] <cjwatson> It's certainly not about m-a, since that's been removed
[11:37] <cjwatson> (Mostly; there may be a few pieces left)
[11:37] <tjaalton> oh
[11:38] <cjwatson> The only reference I see to WINOSDATA in ubiquity is AFAICS vestigial
[11:38] <tjaalton> so maybe the WINOSDATA-stuff in os-prober can be dropped now
[11:38] <cjwatson> But I don't see WINOSDATA in migration-assistant either ...
[11:38] <tjaalton> https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/772470
[11:38] <ubot2> Ubuntu bug 772470 in ubiquity (Ubuntu Precise) "os-prober doesn't detect Windows partition but the recovery partition instead" [High,Fix released]
[11:39] <tjaalton> that's where it was added
[11:39] <cjwatson> That wasn't (just) for migration-assistant
[11:39] <tjaalton> right
[11:39] <cjwatson> ubiquity needs to know where the Windows partition is so that it can display it appropriately in the simplified partitioning dialogs
[11:41] <tjaalton> looks to me it should just blindly go through ntfs partitions?
[11:41] <cjwatson> Yeah
[11:41] <cjwatson> Fixing.  Is there a bug# for this?
[11:41] <tjaalton> no, I could create one
[11:42] <cjwatson> No, don't worry
[11:42] <cjwatson> Just for the changelog if there already was one
[11:42] <tjaalton> right
[11:43] <cjwatson> Uploaded
[11:44] <cjwatson> Now, I wonder where the client code setting WINOSDATA went
[11:44] <tjaalton> this was the removal of the diff?
[11:44] <cjwatson> ?
[11:45] <tjaalton> what got uploaded
[11:45] <tjaalton> I lost track :)
[11:45] <cjwatson> http://paste.ubuntu.com/5683444/
[11:46] <tjaalton> oh that
[11:46] <tjaalton> could've merged 1.16 too ;)
[11:46] <tjaalton> er, 1.60
[11:46] <cjwatson> meh, will do later
[11:47] <cjwatson> This way I get touched-it-last back and don't have to ask anyone ;-)
[11:48] <tjaalton> heh, true
[11:48] <cjwatson> Hm, it does seem that WINOSDATA was only ever added for migration-assistant, though
[11:48] <cjwatson> Maybe ubiquity is supposed to do it some other way
[11:56] <cjwatson> stgraber: ^- Honestly I am hopelessly confused at this point.  Do you remember how any of this works?
[15:12] <stgraber> cjwatson: so looking at the uploads I did around the same time as os-prober back in precise, I think that env variable was only ever used by migration-assistant
[15:14] <stgraber> there apparently was also a change to ubiquity to have it use a different cache of os-prober results when called by m-a then when called from the partitioning step
=== kentb is now known as kentb-out