[12:28] <xivulon> cjwatson ping
[12:35] <xivulon> cjwatson the patched version of autopartition-loop goes a bit further but it stops when doing mount -o move /isodevice /target
[12:35] <xivulon> error is: mount: /isodevice is not a block device
[12:38] <xivulon> evand mentioned it might be a problem with the version of mount used
[12:41] <evand> well, that it might be the same bug that cjwatson is seeing
[12:44] <xivulon> any good reason not to use mount --bind?
[12:45] <xivulon> I ma using it right now and it is moving forward... Let's see if I hose anything
[12:46] <xivulon> zeroing the image file (not the spediest thing I have seen)
[01:09] <evand> the current version of wubi does not seed ubiquity/summary, but xivulon is adding this to build 296 shortly
[01:42] <cjwatson> evand: still around?
[01:42] <cjwatson> I'm sorting out xivulon's issues
[01:43] <cjwatson> evand: I found a couple of m-a bugs
[01:43] <cjwatson> evand: aside from ubiquity/summary not being preseeded, the Install button shows up disabled
[01:43] <cjwatson> evand: any idea why?
[01:44] <cjwatson> for the beta, I'm going to change the zeroing to just create a sparse file
[01:44] <cjwatson> it'll probably fragment like a bastard and it won't preallocate space properly, but it's SO SLOW at the moment it's unbearable
[01:44] <evand> cjwatson: yes, brb though
[01:47] <cjwatson> also found another partman-auto-loop bug and an os-prober bug
[01:48] <cjwatson> but I'm getting there
[01:54] <evand> cjwatson: sorry about that, I would be very interested in the m-a bugs.  I'm not sure about the install button, but I recall seeing it before.  I just haven't noticed it in a long while as I've been preseeding that question.  I seem to recall going back a page and going forward fixes it, but that wont work here
[01:54] <evand> as we disable the back button
[01:54] <evand> I'll look into it tonight though
[01:54] <cjwatson> I'm in an awkward state for sending patches at the moment
[01:54] <evand> heh, I can imagine
[01:54] <cjwatson> but basically the windowsxp special case in mount_os gets it a bit wrong
[01:54] <cjwatson> need to kill the /etc/mtab stuff because that isn't accurate (or at least move it below /proc/mounts)
[01:54] <evand> it's detecting vista as xp, isn't it?
[01:55] <evand> ok
[01:55] <cjwatson> and stick a head -n1 in those pipelines
[01:55] <cjwatson> and remove the backslashes at the ends of those lines 'cos I'm not quite sure why they're there :)
[01:55] <evand> line continuations?
[01:55] <cjwatson> but the next lines are separate commands
[01:55] <cjwatson> so it doesn't make sense to continue the lines there
[01:56] <evand> oh, yeah, curious
[01:56] <cjwatson> it's the sequence that goes grep /etc/mtab, if that failed grep /proc/mounts, etc. to find the mountpoint
[01:56] <cjwatson> after that though we still have a problem
[01:56] <cjwatson> wubi preseeds migration-assistant/partitions to "Windows XP Professional (/dev/sda3)"
[01:56] <cjwatson> but it actually needs to be "Windows Vista/Longhorn (loader) (/dev/sda3)"
[01:56] <evand> that shouldn't matter
[01:57] <evand> It doesn't actually do anything with the first part
[01:57] <cjwatson> well, the m-a page pops up ...
[01:57] <evand> that might just be a bug in the ubiquity m-a code
[01:57] <cjwatson> /dev/sda3 is in choices
[01:57] <evand> what question is it asking?
[01:57] <cjwatson> I'm not sure, I'll have to rerun with --debug
[01:58] <cjwatson> I'm getting used to the pkill parted_server; rm -rf /var/lib/partman /var/run/parted_server.pid; losetup -d /dev/loop2; losetup -d /dev/loop3; losetup -d /dev/loop4; rm /host/ubuntu/disks/*.disk; umount /host dance
[01:58] <evand> hahaha
[01:58] <cjwatson> you're not really meant to do this ...
[01:59] <evand> it's a good thing you know what you're doing, or I'd consider that brand new windows partition a goner
[01:59] <evand> s/or/otherwise
[02:00] <cjwatson> I *think* I know what I'm doing
[02:00] <cjwatson> there's a subtle difference
[02:00] <cjwatson> oh fuck, no no no
[02:00] <cjwatson> framebuffer died
[02:01] <cjwatson> remind me how I suspend from a shell, assuming the shell is listening?
[02:01] <evand> haha, you're asking the wrong person
[02:01] <evand> btw, if you'd like to skip m-a for now, you can always preseed partitions to nothing, though I imagine you know and have done this
[02:02] <evand> I'm going to grab some programming dinner in the form of taco bell, and I shall return to release a new version of m-a with the fixes you suggest.  Feel free to keep me up to date.
[02:03] <cjwatson> damnit, had to give up and power off
[02:03] <cjwatson> there go all the in-place customisations
[02:03] <cjwatson> evand: I'll make a branch of m-a with the fixes I tested in-place
[02:04] <evand> ...months, if not years, of vmware has taught me to scp at every possible free moment
[02:04] <evand> though you probably don't have a network device
[02:04] <evand> cjwatson: thanks
[02:04] <cjwatson> oh, I mounted my normal root filesystem for a bit and stuck the changes in there
[02:04] <cjwatson> and I remember the rest
[02:04] <cjwatson> it's just that I have to apply them all to the live filesystem again by hand
[02:04] <evand> so then it's just the painstaking process of reapplying them
[02:04] <evand> indeed
[02:05] <cjwatson> d-i hacking is pretty much the same story
[02:06] <evand> gah, I can imagine
[02:15] <cjwatson> evand: bzr+ssh://bazaar.launchpad.net/~kamion/migration-assistant/wubi-fixes - please review
[03:09] <evand> reviewing now
[03:16] <evand> looks good, thank you
[03:16] <CIA-18> migration-assistant: evand * r61 migration-assistant/ (debian/changelog ma-script-utils):
[03:16] <CIA-18> migration-assistant: * In the Windows case in mount_os:
[03:16] <CIA-18> migration-assistant:  - Look in /proc/mounts, not /proc/partitions.
[03:16] <CIA-18> migration-assistant:  - Look in /proc/mounts before /etc/mtab.
[03:16] <CIA-18> migration-assistant:  - Make sure to select only the first mountpoint for a given device (in
[03:16] <CIA-18> migration-assistant:  case of bind mounts).
[03:18] <cjwatson> evand: hmm, /var/log/installer/debug says it set migration-assistant/sda3/users but doesn't seem to have any INPUT
[03:18] <evand> and the interface is showing up?
[03:18] <cjwatson> yeah
[03:18] <cjwatson> oh, hang on
[03:19] <cjwatson> it did INPUT high ubiquity/run-ma-again
[03:19] <cjwatson> and that was seen false
[03:19] <evand> curious
[03:19] <evand> oh, hrm
[03:20] <cjwatson> that seems to always be asked assuming ma-ask succeeded
[03:20] <cjwatson> so I think it needs to not trigger a page load
[03:21] <evand> ok
[03:21] <evand> the question then becomes where do I trigger the page load then.
[03:22] <cjwatson> surely that isn't what triggers it normally
[03:22] <cjwatson> normally it would just be migration-assistant/partitions
[03:22] <evand> it needs to continue after migration-assistant/partitions
[03:23] <evand> oh
[03:23] <evand> I'm going to read my own code now instead of assuming I know what I'm talking about
[03:23] <cjwatson> oh, hm, that really is where you trigger it
[03:23] <evand> heh
[03:23] <cjwatson> nothing else does FilteredCommand.run
[03:24] <evand> well, unless there isn't anything to import from, but yeah
[03:25] <cjwatson> would if 'UBIQUITY_AUTOMATIC' in os.environ: return FilteredCommand.run(blah); else: return self.succeeded
[03:25] <cjwatson> work?
[03:25] <evand> yeah, but that line is ending up everywhere in the source code
[03:25] <cjwatson> er except the other way round
[03:25] <cjwatson> I know :-/
[03:25] <evand> it's becoming quite the special case
[03:26] <evand> hrmm, I suppose that can work until a better solution is discovered
[03:26] <cjwatson> I'm not fond of it, if we can think of something better that'd be great
[03:29] <cjwatson> hmm
[03:29] <evand> yeah, continue with that in there for now and I'll see if I can find a way to get around it.  No sense in that tying you up any more than it has.
[03:31] <cjwatson> ok, I'm thinking maybe instead you should display the interface at some point after self.firstrun becomes False
[03:31] <cjwatson> half-formed thought though
[03:32] <evand> another half formed thought: you can only display the interface cleanly through FilteredCommand.run and that's a blocking operation (calls enter_ui_loop)
[03:33] <evand> I really need to delete comments as they become irrelevent.
[03:33] <evand> I'm all for it
[03:33] <evand> but I imagine that's a decision for Mark to rubber stamp
[03:33] <cjwatson> dive in and redesign bastard bits like that
[03:34] <cjwatson> I believe Matt has authority to approve sprints
[03:34] <cjwatson> I'll run it by him at some point
[03:34] <evand> FilteredCommand.run?  Heh, technically it just was, though I'll accept possibly poorly.
[03:34] <cjwatson> it wasn't well-designed to start with, which can't have helped
[03:36] <evand> the sprint idea might help with the massive ubiquity bug fix that I believe you have planned for 8.04, no sense putting band-aids on open wounds.
[03:36] <evand> er that metaphor was terrible, but I'm sure you get the idea :)
[03:38] <evand> hrmm
[03:45] <cjwatson> ok, my box crashed again, but it was doing better that time
[03:46] <cjwatson> I'm going to try to get everything I have in before the LP downtime and then go to bed
[03:46] <evand> ok, goodnight
[03:46] <evand> I may have a solution for the m-a problem
[03:46] <cjwatson> are you ready to have m-a uploaded?
[03:46] <evand> yeah, might as well
[03:46] <cjwatson> oh?
[03:47] <cjwatson> I was going to commit that workaround, but if you have something better, that's good
[03:47] <evand> no, by all means, commit that, this can always be added after the fact
[03:47] <evand> after all, it may fail miserably
[03:51] <evand> is there some trick to bzr+ssh I'm missing
[03:51] <evand> AssertionError: end of file reading from server.
[03:51] <cjwatson> #ubuntu-devel
[03:51] <cjwatson> it's down
[03:51] <cjwatson> (argh!)
[03:52] <evand> argh indeed!
[04:00] <cjwatson> ok, 15 minutes until it comes back up, let's make sure we have all our ducks in a row
[04:01] <evand> I ran pbuilder while we were waiting just to play it safe.  But as far as m-a goes it's just that two-liner.  I can't imagine that breaking anything.
[04:01] <cjwatson> partman-auto-loop: sparse files, remove 'mountpoint' file so that ubiquity doesn't get confused and claim /dev/sda is busy
[04:01] <cjwatson> migration-assistant: upload the stuff already in bzr
[04:02] <cjwatson> ubiquity: m-a workaround by checking UBIQUITY_AUTOMATIC
[04:02] <cjwatson> am I missing anything?
[04:02] <evand> not sure as my ubiquity branch is potentially out of date
[04:02] <evand> do you want to upload m-a straight, or should I put it on rookery
[04:03] <cjwatson> put it on rookery please
[04:03] <evand> you fixed the ubiquity-dm file to call with --automatic, right?
[04:04] <cjwatson> yes
[04:04] <cjwatson>   * Use the --automatic option when running ubiquity from the init script
[04:04] <cjwatson>     (thanks, Evan Dandrea).
[04:04] <cjwatson>   * Work around migration-assistant always being shown in automatic
[04:04] <cjwatson> that's the stuff in the changelog right now
[04:04] <cjwatson>     installs.
[04:05] <evand> http://people.ubuntu.com/~evand/upload/migration-assistant_0.5.1.dsc
[04:17] <evand> my god...I take over ubuntu-devel-discuss and I find messages from 2006 in the queue.
[04:17] <cjwatson> common with unmoderated lists, I think
[04:17] <cjwatson> the queue tends to be nearly all spam
[04:17] <cjwatson> so checking it isn't rewarding
[04:18] <evand> these are just people who were not subscribed
[04:18] <evand> I think
[04:18] <evand> the messages are not spam, at any rate
[04:18] <CIA-18> ubiquity: cjwatson * r2252 ubiquity/ (debian/changelog ubiquity/components/migrationassistant.py):
[04:18] <CIA-18> ubiquity: * Work around migration-assistant always being shown in automatic
[04:18] <CIA-18> ubiquity:  installs.
[04:23] <cjwatson> evand: m-a uploaded
[04:24] <evand> cjwatson: thanks!
[04:26] <CIA-18> ubiquity: cjwatson * r2253 ubiquity/ (d-i/manifest debian/changelog):
[04:26] <CIA-18> ubiquity: * Automatic update of included source packages: migration-assistant 0.5.1,
[04:26] <CIA-18> ubiquity:  partman-auto-loop 0ubuntu9.
[04:28] <evand> cjwatson: you let it through the freeze block, right?
[04:28] <evand> just making sure
[04:28] <cjwatson> I will
[04:29] <CIA-18> ubiquity: cjwatson * r2254 ubiquity/debian/changelog: releasing version 1.5.17
[04:29] <evand> ok
[06:28] <CIA-18> migration-assistant: evand * r62 migration-assistant/debian/changelog: releasing version 0.5.1