[02:40] <CIA-4> ubiquity: cjwatson * r1833 ubiquity/ (4 files in 3 dirs):
[02:40] <CIA-4> ubiquity: * New partitioner: Handle partman/exception_handler and
[02:40] <CIA-4> ubiquity:  partman/exception_handler_note (LP: #82848).
[03:38] <CIA-4> ubiquity: cjwatson * r1834 ubiquity/ubiquity/components/partman.py: tell debconffilter to catch partman/exception_handler and partman/exception_handler_note
[04:03] <cjwatson> I seem to have forgotten about that up to now
[04:21] <evand> heh
[04:28] <CIA-4> ubiquity: cjwatson * r1835 ubiquity/ (debian/changelog ubiquity/frontend/ubiquity.glade):
[04:28] <CIA-4> ubiquity: * New partitioner: Note that unadorned sizes are assumed to be in
[04:28] <CIA-4> ubiquity:  megabytes (LP: #82848).
[04:30] <cjwatson>  components/partman.py   |   77 ++++++++++++++++++++++++++++++++++++++++++++++--
[04:30] <cjwatson>  frontend/gtkui.py       |   36 +++++++++++++++++++---
[04:30] <cjwatson>  frontend/ubiquity.glade |   58 +++++++++++++++++++-----------------
[04:31] <cjwatson>  3 files changed, 139 insertions(+), 32 deletions(-)
[04:31] <cjwatson> easier than I thought, but I wonder how many bugs there are in that ...
[04:35] <evand> I'd wager none in that last file
[04:53] <cjwatson> I hope not, otherwise I'm slipping
[04:55] <cjwatson> what's the usual rate, one bug per 20 lines of code?
[05:09] <evand> haha
[05:16] <cjwatson> evand: how's m-a mergeability looking? feature freeze is Thursday ...
[05:24] <evand> cjwatson: I still have to take a look at that old partitioner issue, but I intend to get a new release of m-a out the door this weekend.  So it should be mergeable after that, provided you're willing to let it in while I try to fix the old partitioner in advanced mode + m-a issue.
[05:27] <cjwatson> I'm happy to merge without the old partitioner
[05:27] <cjwatson> I can just disable m-a for the old partitioner. no problem
[05:27] <evand> great!
[05:27] <evand> The Ubiquity code is complete (except for that one issue), it just depends on the unreleased version of m-a, to which I want to add one or two more features before releasing.
[05:29] <evand> I'll make sure I have m-a stable by the end of this weekend so a version can be uploaded and then anything I add after that (provided it's far enough before the freeze that it can be uploaded) is uploaded in a new version.
[05:30] <evand> cjwatson: I'm assuming I have to fill out a main inclusion request when the time comes.
[05:45] <cjwatson> yeah
[08:05] <cjwatson> argh, fucked up the Partman state machine in some non-intuitive way AGAIN
[08:17] <cjwatson> ahh, backing up from partman-partitioning/new_size jumps back to choose_partition rather than stepping back to active_partition, for some reason
[08:18] <evand> not a problem, it livens this place up a bit
[08:28] <cjwatson> my thought too
[08:29] <cjwatson> one benefit of all this, it's certainly making partman's control flow in d-i more regular
[08:29] <cjwatson> particularly as regards backup
[08:29] <cjwatson> because the ubiquity component works by stepping into all of the partitions and lots of the options for each partition, and backing up rather than actually doing anything
[08:30] <cjwatson> if backup behaves oddly, you really notice
[08:33] <evand> I can imagine
[08:33] <evand> That reminds me, at some point m-a needs to support that capability, not that it's needed in Ubiquity.
[08:33] <evand> But for d-i
[08:40] <cjwatson> backup's fairly easy once you get the hang of it, as long as your code is a state machine already
[08:42] <evand> that's the problem, we actually briefly discussed this at UDS
[08:42] <evand> basically, it has an inner loop or two
[08:43] <evand> and that's hard to roll into a state machine in bash without some somewhat ugly code, afaik
[08:43] <evand> but I'm sure I'll figure it out, there are just more pressing issues
[08:44] <cjwatson> oh yes, we did
[08:45] <cjwatson> you can do it, it's just fiddly :)
[08:45] <cjwatson> actually, I'm not sure why it's really harder with nested loops than without
[08:46] <cjwatson> the only wart with nested loops is that you need to propagate backup when you back up off the start of an inner loop, and therefore keep track of how to back up in each loop independently
[08:46] <evand> I meant ash, but I'm sure you knew that
[08:46] <evand> indeed
[08:47] <evand> yeah, that doesn't sound too bad, I think I was just having a hard time conceptualizing it when I first tackled it
[08:49] <cjwatson> I tend to think of it as "busybox sh" to avoid confusion with the shells available in the normal system
[09:14] <CIA-4> ubiquity: cjwatson * r1836 ubiquity/ (4 files in 3 dirs):
[09:14] <CIA-4> ubiquity: * New partitioner: Implement resizing. Ask for the size using a spinbutton
[09:14] <CIA-4> ubiquity:  (in megabytes) rather than a text entry field.