[02:11] <CIA-3> ubiquity: cjwatson * r3058 ubiquity/ (d-i/manifest debian/changelog):
[02:11] <CIA-3> ubiquity: Automatic update of included source packages: partman-base 128ubuntu8,
[02:11] <CIA-3> ubiquity: user-setup 1.23ubuntu13.
[02:19] <CIA-3> ubiquity: cjwatson * r3059 ubiquity/debian/changelog: remove superseded changelog entry
[02:29] <CIA-3> ubiquity: cjwatson * r3060 ubiquity/debian/changelog: releasing version 1.11.13
[07:53] <tjaalton> has the debconf option to remove an existing lvm setup been renamed again?
[07:53] <tjaalton> can't install jaunty atm because it fails to remove the old vg
[09:13] <cjwatson> tjaalton: not AFAIK ...
[09:14] <tjaalton> cjwatson: ok, I wonder why it fails then
[10:14] <tjaalton> cjwatson: partman failed on a server that doesn't have any disks on /dev/cciss: http://users.tkk.fi/~tjaalton/foo/partman-cciss
[10:14] <tjaalton> but removing the module made it continue
[10:29] <cjwatson> two problems here: one is that /dev/cciss/c0d0 appears to exist (why, if you have no disks?), one is that parted_server has some bogus exception handling
[10:29] <cjwatson> the latter is within my ability to fix, but I'm not sure about the former
[10:51] <tjaalton> ok, the first one might be a driver bug
[11:00] <cjwatson> tjaalton: you should have got a select-type question with "Ignore" and "Cancel" options - did you? if so, which option did you select?
[11:09] <tjaalton> cjwatson: I did, they both resulted in the same
[11:09] <tjaalton> hmm, or maybe "cancel" threw me in the menu, where I could spawn the shell (and modprobe -r cciss)
[11:10] <cjwatson> very interesting, clearly exception handling is busted
[11:10] <cjwatson> but you didn't select some kind of blank option or anything?
[11:11] <tjaalton> I could try again to be sure
[11:12] <cjwatson> if you could try with DEBCONF_DEBUG=developer, that wouldn't hurt
[11:12] <cjwatson> just to make sure I know exactly what I'm seeing
[11:12] <tjaalton> yeah, will do in a minute
[11:13] <cjwatson> oh, and is this jaunty?
[11:14] <tjaalton> yes
[11:15] <cjwatson> ok
[11:19] <cjwatson> I was wondering if r120 in lp:~ubuntu-core-dev/partman-base/ubuntu might be at fault, but I think it will just change the error message in the log (from 'No data in infifo' to 'Bad option: ""')
[11:20] <cjwatson> hmm, although error_handler claims that it writes the string "unhandled"
[11:21] <cjwatson> oh, I wonder ...
[11:21] <cjwatson> -        if (1 != iscanf(" %a[^\n]", &str))
[11:21] <cjwatson> and elsewhere
[11:21] <cjwatson> +                iscanf("%a[^\n]", str);
[11:21] <cjwatson> note the missing &
[11:21] <cjwatson> oh, no, never mind me, str is a different type in the latter case
[11:46] <davmor2> evand: when's the new wubi/umenu going in?
[11:48] <evand> davmor2: I'm in the process of sending a mail to xivulon about it.
[11:49] <davmor2> evand: Ah okay cool :)
[11:49]  * evand kicks Google Apps Gmail - quite down at the moment.
[11:51] <cjwatson> http://news.bbc.co.uk/1/hi/technology/7907583.stm
[11:52] <evand> oh, well then
[11:52] <evand> glad to know I'm not the only one then
[11:53]  * davmor2 is happy he now has a dedicated server :0
[11:54] <davmor2> :) even
[12:11] <cjwatson> tjaalton: right, caught it in the act with strace ...
[12:11] <tjaalton> cjwatson: cool
[12:14] <tjaalton> I'm still fighting with iLO, so didn't get the log yet
[12:15] <cjwatson> oh, looks like the problem is that parted_server doesn't block on infifo - it just sees that there's nothing in the fifo and gives up straight away without waiting
[12:15] <cjwatson> goodness, there must be a zillion bugs about this. I wonder where they all are
[12:16] <tjaalton> hehe .)
[12:16] <tjaalton> :)
[12:16] <cjwatson> bug 317435 is certainly the same thing though
[12:16] <tjaalton> indeed it is
[12:20] <cjwatson> the weird thing is that strace doesn't even show it *trying* to read from the relevant fd
[12:21] <cjwatson> 10230 write(3, "parted_server: OUT: \n\n\n"..., 23) = 23
[12:21] <cjwatson> 10230 write(3, "parted_server: exception_handler: Bad option: \"\"\n"..., 49) = 49
[12:22] <cjwatson> and yet it unconditionally goes through fgetc(infifo)
[12:24] <soren> fgetc wouldn't result in a syscall if the FILE was already closed. Could that be it?
[12:24] <soren> At least I don't think it would.
[12:29] <cjwatson> no, I've figured it out with the help of #chiark
[12:30] <cjwatson> a previous fscanf pushed back the trailing newline from a previous comment, so fgetc just gave me that, which confused iscan_line into seeing a blank line
[12:31] <soren> Ah.
[12:31] <cjwatson> which leads me to wonder how the hell I rejig parted_server so that I can unambiguously tell when I get a blank line, which was half the point of r120
[12:32] <cjwatson> maybe I just have to arrange for that not to matter
[12:37] <davmor2> evand: gmail should be back up now :)
[12:37] <evand> davmor2: it's up and down for me
[12:38] <evand> but thanks
[12:38] <davmor2> :)
[12:53] <cjwatson> something like http://paste.ubuntu.com/122365/ might do the trick
[12:59] <CIA-3> partman-base: cjwatson * r133 ubuntu/ (debian/changelog parted_server.c):
[12:59] <CIA-3> partman-base: Deal with leading newlines while reading exception options and partition
[12:59] <CIA-3> partman-base: flags passed to SET_FLAGS; regression introduced in partman-base
[12:59] <CIA-3> partman-base: 128ubuntu4 (LP: #317435).
[13:01] <CIA-3> partman-base: cjwatson * r134 ubuntu/debian/changelog: releasing version 128ubuntu9
[13:01] <cjwatson> tjaalton: ^- that should do it, thanks for the nudge and the logs
[13:01] <tjaalton> cjwatson: great, thanks for a quick fix. I'll try it out once it's available
[13:11] <CIA-3> partman-target: cjwatson * r747 ubuntu/ (debian/changelog finish.d/fstab_hd_entries): Use labels rather than UUIDs if they exist (LP: #320871).
[13:12] <CIA-3> partman-target: cjwatson * r748 ubuntu/debian/changelog: releasing version 58ubuntu4
[13:25] <kirkland> cjwatson: could you have a quick look at http://launchpadlibrarian.net/23052247/331085.debdiff
[13:25] <kirkland> cjwatson: fix for https://bugs.edge.launchpad.net/ubuntu/+source/ecryptfs-utils/+bug/331085
[13:27] <kirkland> cjwatson: please leave any comments in that bug;  i'm working on site at a partner's office today
[13:30] <cjwatson> done
[13:42] <evand> cjwatson: Given the removal of /etc/adjtime, would it be ok if I add "|| true" in the clock-setup finish-install script to guard against the file not existing?
[13:43] <cjwatson> evand: yes please, or perhaps better explicitly check for its presence to avoid error message noise
[13:43] <evand> will do
[13:49] <kirkland> cjwatson: thanks.
[13:51] <CIA-3> clock-setup: evand * r203 clock-setup/ (debian/changelog finish-install.d/10clock-setup):
[13:51] <CIA-3> clock-setup: Check for the existence of /etc/adjtime before running sed over it,
[13:51] <CIA-3> clock-setup: given util-linux 2.14.2-1ubuntu4.
[14:05] <CIA-3> clock-setup: evand * r204 clock-setup/debian/changelog: releasing version 0.97ubuntu2
[14:33] <davmor2> cjwatson: did the lvm issue from yesterday get resolved do you know?  I know you said the numerical display did.
[14:38] <cjwatson> davmor2: Keybuk believes he's fixed it, although we haven't yet uploaded a new d-i so it may be a bit wonky during installation still since it'll use the old udev. It should be happier after the first reboot now
[14:39] <davmor2> I'll give it a try then be nice to get it in before a5
[15:50] <davmor2> Meh no lvm is the same :(
[16:07] <davmor2> cjwatson: when you get chance I've replicated the lvm removal and the install folder is here http://www.davmor2.co.uk/install
[16:08] <cjwatson> davmor2: so would an example reproduction recipe be to do two guided LVM installations in a row?
[16:08] <cjwatson> I think it will be easiest to fix this if I can reproduce it
[16:09] <davmor2> yes or just a whole guided - whole drive after an lvm
[16:09] <cjwatson> ok
[16:09] <cjwatson> the logs don't seem all that informative, I'll see what I can do by way of reproduction
[16:10] <davmor2> cjwatson: I did the lvm at 20% as before in yesterdays tests.  That didn't work again and I want to test apps so I did a whole drive after
[17:05] <charlie-tca> live cd is failing towards the end of an install; clock setup. Is this known?
[17:06] <cjwatson> yes, fixed in bzr
[17:06] <charlie-tca> Thanks
[17:06] <cjwatson> evand: are you going to upload this before end of day?
[17:07] <evand> cjwatson: indeed, was just waiting for clock-setup to build
[17:07] <evand> I'll start the process now
[17:08] <davmor2> please say yes testing starts tomorrow :)
[17:09] <cjwatson> FWIW ubiquity uploads don't need to block on d-i elements building, only on source being available in the archive
[17:10] <evand> ah
[17:10] <evand> noted
[17:15] <CIA-3> ubiquity: evand * r3062 ubiquity/debian/po/ (79 files): debconf-updatepo
[17:15] <CIA-3> ubiquity: evand * r3061 ubiquity/ (d-i/manifest debian/changelog):
[17:15] <CIA-3> ubiquity: Automatic update of included source packages: clock-setup
[17:15] <CIA-3> ubiquity: 0.97ubuntu2, partman-base 128ubuntu9, partman-target 58ubuntu4.
[17:29] <CIA-3> ubiquity: evand * r3063 ubiquity/debian/changelog: releasing version 1.11.14
[18:16] <CIA-3> apt-setup: cjwatson * r161 ubuntu/ (3 files in 2 dirs):
[18:16] <CIA-3> apt-setup: Allow preseeding apt-setup/partner to enable the partner repository;
[18:16] <CIA-3> apt-setup: requested by Brian Thomason.
[18:17] <CIA-3> apt-setup: cjwatson * r162 ubuntu/debian/po/ (60 files): debconf-updatepo
[18:20] <CIA-3> apt-setup: cjwatson * r163 ubuntu/debian/changelog: releasing version 1:0.37ubuntu10
[18:59] <CIA-3> debian-installer: cjwatson * r1051 ubuntu/debian/changelog:
[18:59] <CIA-3> debian-installer: No-change rebuild to pick up new components (particularly udev and
[18:59] <CIA-3> debian-installer: rescue).
[19:00] <cjwatson> davmor2: ok, reproduced a hang on attempting to remove existing LVM
[19:00] <cjwatson> davmor2: I'll look into it some more tomorrow
[19:01] <cjwatson> vgremove seems to be blocked on stdin
[19:01] <davmor2> cjwatson: do you want me to throw a rough bug together?
[19:01] <cjwatson> probably wouldn't hurt
[19:01] <davmor2> np's
[19:01] <cjwatson> throw it on debian-installer
[19:02] <davmor2> okay
[19:04] <CIA-3> debian-installer: cjwatson * r1052 ubuntu/debian/changelog: releasing version 20081029ubuntu20
[19:04] <davmor2> already one in place
[19:04] <davmor2> bug 319150
[19:05] <cjwatson> ok
[19:06] <davmor2> cjwatson: I've confirmed and updated with todays info
[19:07] <cjwatson> err
[19:07] <cjwatson> I wonder if that is in fact the same bug
[19:07] <cjwatson> it is *similar*
[19:07] <cjwatson> you're using automatic LVM removal, Chris is doing it manually
[19:08] <davmor2> no chris did it manually because automatic didn't work
[19:09] <davmor2> I was unable to simply remove the LVM immediately
[19:09] <cjwatson> sure, but the automatic and manual symptoms are not necessarily the same cause ...
[19:09] <cjwatson> anyway, whatever, I have to go now :)
[19:09] <davmor2> sort it out tomorrow then bye
[19:10] <davmor2> hey evand did you hear back from xivulon?
[21:21] <xivulon> evand hi, do you have further questions on wubi progress?
[22:22] <CIA-3> wubi: Agostino Russo * r67 trunk/ (763 files in 72 dirs):
[22:22] <CIA-3> wubi: * Compile grub4dos from source as opposed to using precompiled
[22:22] <CIA-3> wubi:  binaries
[22:22] <CIA-3> wubi: * Added source files for grub4dos and grubutil
[22:31] <superm1> xivulon, is python wubi going on live disks as of yet?  when should the mythbuntu folks start trying it?
[22:32] <xivulon> should be not sure what version though
[22:33] <xivulon> r67 should be ok (haven't fully tested yet)
[22:33] <superm1> xivulon, okay well i'll tell folks to give it a brief look and see how things are looking and what needs changing at a5 then
[22:35] <xivulon> superm1 thanks, in case the version on the live cd is not up to date, it should be possible to compile wubi from source and use the exe with the CD or ISO (in the same folder)
[22:35] <xivulon> bzr branch lp:wubi && cd wubi && make
[22:35] <superm1> xivulon, okay we'll see how things are looking
[22:36] <superm1> i'm thinking the preseed will need some adjusting likely since there have been a lot of changes in the mythbuntu ubiquity frontend
[22:36] <superm1> mostly tearing out unnecessary questions that just cause confusion
[22:36] <xivulon> need to edit the mythbuntu preseed template slightly one sec
[22:39] <CIA-3> wubi: Agostino Russo * r68 trunk/ (data/preseed.cdboot data/preseed.mythbuntu debian/changelog): Updated mythbuntu and cdmenu preseed templates
[22:39] <xivulon> superm1 you might want to use r68+
[22:40] <superm1> xivulon, okay will try to take a look