[05:40] <cjwatson> evand: I've been plugging on with partman-auto-loop today, and made decent progress I think, though still nothing I can show off. I had to sidetrack for a while into fixing a complicated corner case in partman-auto's recipe expansion code.
[05:41] <evand> ok
[05:41] <evand> I've been trying to figure out this partman hanging bug
[05:41] <cjwatson> evand: in other news, mdz has been after me for a while to switch d-i to dhcp3, so I'm just doing that now, now that Debian has made that switch
[05:41] <cjwatson> oh, any joy?
[05:41] <evand> none
[05:41] <evand> there doesn't appear to be anything crazy in env
[05:42] <evand> and strace'ing parted_server and partman shows it stopping on partman/choose_partition, but I can't figure out why yet
[05:42] <evand> keeping in mind that this could be a slightly different bug than what others have been encountering
[05:43] <evand> as I can get to the advanced page, but as soon as I check the format box everything goes into a wait state
[05:52] <cjwatson> I've lost track, but is this reproducible with --debug atm?
[05:54] <evand> cjwatson: it appears to be.  At least here it does.
[05:59] <cjwatson> what question is it asking at that point?
[05:59] <cjwatson> oh, partman/choose_partition you said
[06:04] <evand> hrmm, gksu ubiquity != ubiquity.
[06:05] <evand> and gksu ubiquity --debug doesn't write to ubiquity/debug.
[06:09] <cjwatson> ubiquity should be approximately equivalent to gksudo --desktop /usr/share/applications/ubiquity-gtkui.desktop -- ubiquity
[06:10] <cjwatson> with gksu ubiquity --debug, the --debug probably gets eaten by gksu. That's what the -- is for above.
[06:10] <evand> ah, right.  Whoops
[06:21] <evand> ok, that changes things.  It still hangs, but on next press, not the format checkbox.  And it is not triggered with --debug.