[10:33] <CIA-3> partman-partitioning: cjwatson * r702 ubuntu/debian/changelog: releasing version 64ubuntu6
[10:39] <cjwatson> superm1: should bug 290400's dell task be closed too? I see the oem-priority task is closed
[11:12] <evand> mpt: no worries, appreciated as always :)
[11:13] <mpt> evand, on the bright side I cleaned up the time-zone-clicking bug reports -- I think there's only three actual bugs there
[11:13] <mpt> albeit that all three of them are pretty dire
[11:14] <evand> indeed, I'm definitely looking into them
[11:15] <CIA-3> debian-installer: cjwatson * r1064 ubuntu/ (5 files in 2 dirs): Move mainline architectures to 2.6.28-11 kernels.
[11:16] <evand> at some point I'd like to figure out netconsole enough to write a guide for debugging hard locks when running ubiquity.  Gave it a shot yesterday, but my networking skills are a bit lacking.
[11:16] <CIA-3> debian-installer: cjwatson * r1065 ubuntu/ (5 files in 2 dirs): Move ports architectures to 2.6.28-5 kernels.
[11:16] <evand> rather than having to go with the often difficult serial console approach
[11:17] <davmor2> Guys you know when you do 70% of a drive for lvm.  It is 70% of whatever linux says the drive is correct?
[11:18] <davmor2> as in actual size and not human readable size
[11:36] <cjwatson> davmor2: it's 70% of the available size within the volume group
[11:37] <cjwatson> davmor2: it's computed based on its accurate internal idea of the size and not based on the rounded value displayed in the user interface, if that's what you mean, yes
[11:38] <davmor2> cjwatson: Okay cool thanks that explains the strange size then :)
[11:52] <CIA-3> debian-installer: cjwatson * r1066 ubuntu/ (3 files in 3 dirs): Re-enable netboot/386 as a ports architecture, since it builds again.
[12:53] <CIA-3> ubiquity: evand * r3109 ubiquity/ (5 files in 3 dirs):
[12:53] <CIA-3> ubiquity: When the user selects "go back" in response to the weak password dialog,
[12:53] <CIA-3> ubiquity: select the text in the first password entry (LP: #344402).
[13:24] <superm1> cjwatson, when the seed changes are deployed internally, yeah I'll close it. holding off for a post a6 DVD to get made to deploy them
[13:26] <cjwatson> righto
[13:34] <superm1> thanks for the improvements on the other speed bug.  it's a big improvement, and there will be thanks flowing in from several of the burn teams in other regions later this year as it makes it through the different test phases in deployment
[13:39] <cjwatson> I'm *very* glad that's out of the way :-) the profiling tools I could find weren't working so I had to use the old-fashioned approach
[13:41]  * evand really wishes there was a stronger link between bzr change and launchpad bug.  --fixes is nice, but I really wish it was handled on commit.
[13:42] <cjwatson> it is, isn't it?
[13:42] <cjwatson> well, push, anyway
[13:44] <cjwatson> fader: this has my full attention, it's just taking me a little while to trace
[13:44] <fader> cjwatson: Cool, I am here for whatever you need on it
[13:45] <evand> ah, a bit of a misstatement.  I don't really need it (I can grep bzr log easily), but it would still be cool if launchpad bugs noted the bzr revision that fixed a bug.
[13:45] <cjwatson> oh yes
[13:49] <cjwatson> fader: would you be able to run a test for me, but not fully automatic, instead stopping at the hostname prompt to edit a file?
[13:49] <fader> cjwatson: I should be able to... it'll take me a few minutes to edit the preseed and reboot
[13:51] <cjwatson> fader: I'd like 'set -x' inserted on the second line of /lib/partman/lib/lvm-remove.sh
[13:51] <cjwatson> I need to trace the logic of device_remove_lvm and see exactly what it's doing
[13:52] <cjwatson> because currently the answer appears to be "nothing" and I want to know why :-)
[13:53] <cjwatson> I speculate that it might be a cciss-specific problem but it's hard to say for sure right now
[13:54] <fader> cjwatson: I'll let you know when I'm at that point
[13:54] <cjwatson> thanks
[14:33] <CIA-3> oem-config: cjwatson * r635 trunk/debian/ (changelog control):
[14:33] <CIA-3> oem-config: Tighten frontend dependencies on oem-config to prevent problems when one
[14:33] <CIA-3> oem-config: architecture is out of sync.
[14:34] <davmor2> cjwatson: whats up with lvm now?
[14:35] <cjwatson> davmor2: I don't know yet. Bug 341928
[14:36] <cjwatson> definitely different from earlier problems
[14:39] <davmor2> cjwatson: I'll have a look at it too I've not seen any issues since so this might be a server thing.  I've got the iso so I'll give it a go after and see if I can confirm it
[14:46] <cjwatson> davmor2: fader said he'd get me more detail, so I'm happy to work with that once it arrives
[14:48] <davmor2> cjwatson: np's I'm just interested too :) encrypted lvm worked fine this morning on kubuntu and also after when I wiped it :)
[14:48] <cjwatson> it could well be only reproducible on cciss or something
[14:48] <cjwatson> I don't want to speculate too much with insufficient data though
[14:50] <CIA-3> ubiquity: cjwatson * r3110 ubiquity/debian/ (changelog control):
[14:50] <CIA-3> ubiquity: Tighten frontend dependencies on ubiquity to prevent problems when one
[14:50] <CIA-3> ubiquity: architecture is out of sync.
[14:58] <cjwatson> oh, this could also be due to autopartitioning rather than manual partitioning
[14:58] <cjwatson> that's actually more plausible
[14:59] <cjwatson> fader: any progress?
[14:59] <davmor2> cjwatson: on a plus side it would confirm that it is on specific hw and not general which would make me happy
[15:00] <davmor2> cjwatson: does the preseed not just give out the info that the partition needs to continue with?
[15:01] <cjwatson> doesn't seem to reproduce here though
[15:01] <cjwatson> davmor2: it's a different code path, unfortunately
[15:01] <cjwatson> preseeding the partitioner is very unusual by d-i standards
[15:01] <cjwatson> the partitioner's normal interactive mode is not preseedable because it asks the same questions repeatedly as you go back and forward, and you need to provide different answers each time
[15:02] <cjwatson> whereas preseeding operates by setting the answer to each question up-front and then setting the whole thing going
[15:02] <cjwatson> so far my initial attempt to reproduce this with autopartitioning has failed, though
[15:02] <fader> cjwatson: Sorry, no... the machine has decided to be difficult about pxe booting
[15:02] <fader> I'm troubleshooting it now
[15:23] <fader> cjwatson: Do you know of a way to cause a preseed to stop and ask a question?  My preseed-fu is week and commenting out the hostname didn't do it.
[15:24] <fader> s/week/weak/
[15:25] <cjwatson> fader: just commenting out the hostname prompt is the right answer. However, it's only asked at priority high, so you will also need to remove whatever you have that's setting the debconf priority to critical (probably priority=critical on the kernel command line)
[15:25]  * cjwatson -> out for 30 minutes
[15:25] <fader> cjwatson: thanks, I'll give that a shot
[15:27] <davmor2> evand: London is closer now it's in devon/cornwall now :)
[15:31] <evand> heh
[15:31] <evand> odd though, can you please take a screenshot
[15:31] <evand> it should be in the channel
[15:31] <evand> and if it's not (in gtk), then there might be an additional bug
[15:33] <davmor2> evand: I can next time round gone past it now :(
[15:34] <evand> no worries
[15:34] <evand> but sure
[15:57] <fader> cjwatson: I'm in a shell before partitioning but I don't see /lib/partman
[16:01] <cjwatson> fader: *blink*
[16:01] <cjwatson> fader: what question is on the main installer screen
[16:01] <cjwatson> ?
[16:01] <fader> Well it was prompting for the hostname before I did a "go back" to open a shell
[16:04] <cjwatson> oh, this is netboot, isn't it
[16:04] <cjwatson> doh
[16:04] <cjwatson> yeah, partman isn't installed yet, that's a bit awkward
[16:04] <fader> Hmm
[16:05] <cjwatson> give me a moment
[16:07] <CIA-3> usb-creator: evand * r90 usb-creator/ (debian/changelog usbcreator/gtk_frontend.py): Disable WrapLabel until its bugs are worked out.
[16:10] <CIA-3> installation-guide: cjwatson * r454 ubuntu/ (debian/changelog en/appendix/preseed.xml): Document partman/default_filesystem.
[16:18] <cjwatson> fader: I think your best approach is to go back to the main menu, select "change debconf priority" and set it to medium; walk through the menu step by step (just pressing enter) until you're about to enter "Partition disks"; get a shell again and do that edit; then select "change debconf priority" again to set it back to critical, at which point it will continue
[16:18] <fader> cjwatson: will do
[16:18] <cjwatson> easier to do than to describe :)
[16:18] <fader> :)
[16:20] <fader> cjwatson: argh, the preseed values kicked in and I wasn't prompted again... I'm going to remove the partitioning info from the preseed and start again
[16:21] <fader> (Sorry, it'll take a few minutes as it has to be reinstalled as a PPA machine before I can tickle the bug)
[16:21] <cjwatson> fader: no
[16:21] <fader> Okay
[16:21] <cjwatson> fader: if you do that it will take a completely different code path which may not reproduce this bug
[16:21] <cjwatson> fader: can you describe to me the sequence of prompts you saw?
[16:22] <fader> Sure... I went into "change debconf priority" and set it to medium.  Then I resumed with the network detection
[16:22] <fader> It picked up a DHCP address and prompted me for a hostname and a domain name
[16:23] <CIA-3> usb-creator: evand * r91 usb-creator/usbcreator/ (gtk_frontend.py wrap_label.py): Clean that last commit up a bit.
[16:23] <fader> After that it went on and automatically set up the mirror, detected disks, and attempted to partition
[16:24] <cjwatson> fader: hmm, odd. Try it with priority=medium on the kernel command line from the get-go
[16:24] <cjwatson> and then just set it to critical immediately before entering partitioning
[16:25] <fader> cjwatson: Won't the values specified in the preseed fill in most of that again?
[16:25] <cjwatson> fader: the main menu shouldn't be preseeded, so it should still prompt you for each menu step
[16:25] <fader> I'm nervous that if having it on priority=medium this time didn't stop before partitioning that it won't again
[16:26] <fader> Okay, I'll give it a shot :)
[16:26] <fader> Out of curiosity, what is the default priority?  (I removed the priority altogether when I ran it last)
[16:26] <cjwatson> fader: I suspect what happened is it got confused about why it set the priority to medium, and thought it was because of an error
[16:26]  * fader nods.
[16:26] <cjwatson> and therefore reckoned that it could bump the priority back up again after a successful step
[16:26] <cjwatson> main-menu is a bit magic
[16:26] <cjwatson> fader: default: high
[16:27] <fader> cjwatson: Good to know, thanks
[16:27] <fader> I'll let you know when it's restarting... probably 30 minutes or so :/
[16:38] <CIA-3> usb-creator: evand * r92 usb-creator/po/ (28 files): Update translations.
[16:44] <davmor2> evand: http://www.davmor2.co.uk/tzmap.png
[16:45] <evand> ok, that matches up with what I have
[16:45] <davmor2> It's closer than denark :)
[16:46] <CIA-3> usb-creator: evand * r93 usb-creator/debian/changelog: releasing version 0.1.14
[16:50] <davmor2> evand: how is the full screen map going?
[16:52] <evand> full screen map?
[16:52] <davmor2> evand: ubiquity only install
[16:53] <evand> oh, that should be inaccurate to the same level as the map in windowed mode, if that's what you mean
[16:53] <evand> so to answer from that perspective, I'm working on it, and hope to have a solution tonight (as beta freeze is tomorrow)
[16:55] <CIA-3> usb-creator: evand * r94 usb-creator/ (debian/changelog desktop/usb-creator.desktop.in): Set the proper gettext domain. Thanks Timo Jyrinki (LP: #331061).
[17:11] <evand> cjwatson:  hrm, cdebconf doesn't support the escape capability does it?  So if I want to subst a list into a debconf question I need to separate the items by commas, and not newlines, right?
[17:12] <cjwatson> evand: right
[17:13] <cjwatson> though I had to check :)
[17:15] <evand> heh
[17:15] <evand> ok, that makes things slightly easier
[17:20] <fader> cjwatson: I'm up to "detect disks" as the next step in the installer
[17:21] <fader> You still just want set -x' inserted on the second line of /lib/partman/lib/lvm-remove.sh?
[17:21] <fader> 'set -x'
[17:25] <cjwatson> right
[17:26] <fader> Argh, not even vi is there :P
[17:32] <fader> cjwatson: Okay, you want partman and syslog again?
[17:42] <cjwatson> yes please
[17:42] <cjwatson> you have nano
[17:43] <fader> cjwatson: Ah, I love nano... I should have looked for it.  head and tail and echo worked though :)
[17:44] <evand> ...love...nano ;)
[17:44] <CIA-3> oem-config: evand * r636 oem-config/ (debian/changelog lib/timezone_map.py):
[17:44] <CIA-3> oem-config: Update the timezone_map to match changes in ubiquity (more accurate
[17:44] <CIA-3> oem-config: city placement).
[17:45] <fader> nano brings back fond memories of pine :)
[17:45] <evand> heh
[17:46] <cjwatson> don't get me wrong, I prefer vi ... but I concede that users unfamiliar with it are probably better off with an editor that doesn't have beep-mode and corrupt-mode
[17:47] <fader> cjwatson: Yeah, vi gives me the howling fantods even though I can use it in a pinch.  I just figured that if vi weren't there there was *no* editor, since in my mind it's the stone knife and bear skin editor
[17:47] <fader> cjwatson: Logs attached to the bug
[17:51] <cjwatson> fader: thanks. One hopefully last question: what does 'pvs --noheadings --nosuffix -o pv_name' say?
[17:51] <cjwatson> it's failing because it thinks there is no PV on /dev/cciss/c0d0
[17:52] <fader> ~ # pvs --noheadings --nosuffix -o pv_name
[17:52] <fader>   /dev/block/104:6
[17:54] <cjwatson> oh FFS LVM
[17:54] <cjwatson> right
[17:59] <evand> cjwatson: why does the code in components/partman.py preseed false when the user selects the continue button?  That seems wrong to me, and indeed it's messing up partman/unmount_active, but I suspect it has a good reason for being there.
[17:59] <evand> by code I mean, the code to handle generic boolean questions in run
[18:00]  * evand breaks out bzr blame
[18:00] <cjwatson> evand: because some questions are "do you want to continue?" and some are "do you want to return to the partitioning menu?"
[18:00] <cjwatson> evand: hence the answer_reversed stuff around line 1080
[18:00] <evand> ah
[18:00] <evand> so then I should just add it to that list
[18:01] <cjwatson> right
[18:01] <evand> noted; thanks
[18:04] <CIA-3> partman-base: evand * r141 ubuntu/ (debian/changelog init.d/parted): Fix the unmount_active code to work in d-i.
[18:05] <CIA-3> ubiquity: evand * r3111 ubiquity/ (debian/changelog ubiquity/components/partman.py):
[18:05] <CIA-3> ubiquity: Add partman/unmount_active to the list of questions that mean "do you want
[18:05] <CIA-3> ubiquity: to continue?" in the partman component.
[18:06] <CIA-3> partman-base: evand * r142 ubuntu/debian/po/ (64 files): debconf-updatepo
[18:11] <CIA-3> partman-base: evand * r143 ubuntu/debian/changelog: releasing version 129ubuntu3
[18:12] <cjwatson> fader: thanks, I think I have plenty to work with now
[18:13] <cjwatson> fader: although I'll want verification once I've fixed it, of course
[18:13] <fader> cjwatson: No problem.  Plesae let me know if there's anything else I can get
[18:13] <fader> Do you want me to leave this machine in this state?
[18:17] <cjwatson> fader: would that be a pain?
[18:18] <fader> cjwatson: I think it would be fine for a day or two, if it's longer it might be an issue
[18:18] <cjwatson> ok, then let's do that if that's ok
[18:18]  * cjwatson -> out
[18:18] <fader> (Though I can always get it back to this state in <1 hour or so)
[18:18] <fader> Sure, no problem
[21:07] <Riddell> evand: is ubiquity due to be uploaded before beta?
[21:18] <evand> Riddell: ja, why, what's up?
[21:21] <Riddell> evand: just I committed a fix and it's beta freeze sometime soon :)
[21:22] <evand> ah, indeed.  I intend to one one more ubiquity upload before tomorrow.  Was just holding out in case I found some time this evening to get a few more things done, but that's looking unlikely.
[21:34] <CIA-3> usb-creator: evand * r95 usb-creator/debian/changelog: releasing version 0.1.15
[21:39] <CIA-3> wubi: Agostino Russo * r100 trunk/ (debian/changelog src/wubi/backends/win32/eject.py):
[21:39] <CIA-3> wubi: Unicode win api call was using non-unicode argument preventing CD
[21:39] <CIA-3> wubi: from ejecting (LP: 339907)
[21:41] <CIA-3> ubiquity: evand * r3113 ubiquity/ (d-i/manifest debian/changelog):
[21:41] <CIA-3> ubiquity: Automatic update of included source packages: partman-base
[21:41] <CIA-3> ubiquity: 129ubuntu3, partman-partitioning 64ubuntu6, user-setup 1.23ubuntu15.
[21:41] <xivulon> evand, hi
[21:42] <evand> hi xivulon
[21:42] <xivulon> it would be good to have a second build which has alpha6 in isolist.ini
[21:42] <xivulon> alpha6 has a torrent! so we can fully test bittorrent that way
[21:43] <evand> not beta?  We're on the cusp of beta freeze.
[21:43] <xivulon> normally isolist points to either the final release or the daily build
[21:43] <xivulon> well this is only for testing purposes so we can test bt before the beta
[21:44] <xivulon> should not be in a commit
[21:44] <xivulon> if it works then we can point isolist.ini to the beta so that everyone tests by
[21:44] <xivulon> bt
[21:44] <xivulon> we will have to change that back to final release after the beta of course
[21:45] <evand> ok
[21:45] <xivulon> by the way have uploaded r100 which fixes a couple of bugs
[21:46] <xivulon> so is w7 fully supported now?
[21:46] <xivulon> thanks again for the fixes
[21:47] <evand> anytime
[21:48] <evand> so much easier to hack on when it's written in python
[21:48] <evand> indeed, I had r99 up on people.ubuntu.com, I'll upload r100 so that it gets in before beta freeze
[21:57] <CIA-3> ubiquity: evand * r3114 ubiquity/debian/changelog: releasing version 1.11.19
[22:01]  * evand wonders aloud if the inability for debchange to add a new entry below an existing author in debian/changelog (rather than create another copy of the author's name) in the so-called "braced maintainer" mode is a known bug.
[22:02] <cjwatson> it irritates me when other people do that :-) I just use vi
[22:02] <cjwatson> oh
[22:03] <cjwatson>        DEBCHANGE_MULTIMAINT_MERGE
[22:03] <cjwatson>               If  set  to yes, when adding changes in multiple-maintainer mode
[22:03] <cjwatson>               debchange will check whether previous  changes  by  the  current
[22:03] <cjwatson>               maintainer  exist  and add the new changes to the existing block
[22:03] <cjwatson>               rather than creating a new block.  Default is no.
[22:03]  * cjwatson sets that one
[22:04] <evand> ah, nice!
[22:04] <evand> very nice
[22:07] <evand> xivulon: ^ You might want to set DEBCHANGE_MULTIMAINT_MERGE=yes in ~/.devscripts.  It'll make sure you don't have duplicates of each author in debian/changelog as they often are in the wubi changelogs.
[22:12] <xivulon> ah will do was wondering about that
[22:15] <evand> wubi r100 is now marked as stable, will be on tomorrow's cds
[22:15] <evand> gotta run
[22:17] <xivulon> thx