[00:24] <TheMuso> evand: re dmraid, sounds good. What bug did you file, if you indeed had to file one?
[00:38] <evand> TheMuso: I have not yet, but will do so tonight and send you a link.
[00:41] <TheMuso> evand: Ok thanks./
[01:42] <CIA-1> ubiquity: jriddell * r2911 ubiquity/ (debian/changelog ubiquity/frontend/kde_ui.py): merge in lp:~andreas-wenning/ubiquity/releasenotesfix
[01:45] <CIA-1> ubiquity: jriddell * r2912 ubiquity/ (debian/changelog ubiquity/frontend/kde_ui.py): Use Oxygen style in KDE UI
[01:45] <Riddell> evand: are you planning a ubiquity upload before tomorrow (i.e. toot sweet)?
[01:52] <evand> Riddell: yes
[01:52] <evand> later this evening
[01:54] <Riddell> groovy, well there's two more changes for it
[01:54] <evand> thanks!
[06:58] <CIA-1> apt-setup: evand * r147 apt-setup.ubuntu/ (debian/changelog generators/40cdrom):
[06:58] <CIA-1> apt-setup: Check to make sure the CD filesystem is iso9660 before remounting it
[06:58] <CIA-1> apt-setup: as such. Do not write a CD identifier if we're not working with
[06:58] <CIA-1> apt-setup: iso9660 so pkgsel does not try to unmount such disks.
[07:00] <CIA-1> apt-setup: evand * r148 apt-setup/debian/changelog: releasing version 1:0.37ubuntu6
[07:01] <evand> cjwatson: ^ If you have a free moment today, could you just double check that is sane.  I was slightly confused by your previous suggestion as there is no [ -d /hd-media ] block (there is a ! one though).
[07:01] <evand> This worked fine in my testing.
[07:46] <ganes> is there any command to grasp the grub output to any file
[07:51] <ganes> is there any command to grasp the grub output to any file
[08:29] <ganes> is there any command to grasp the grub output to any file
[09:04] <CIA-1> ubiquity: evand * r2913 ubiquity/debian/changelog: Tabs are evil.
[09:09] <CIA-1> ubiquity: evand * r2914 ubiquity/ (d-i/manifest debian/changelog):
[09:09] <CIA-1> ubiquity: Automatic update of included source packages: apt-setup
[09:09] <CIA-1> ubiquity: 1:0.37ubuntu6, base-installer 1.86ubuntu7, kboot-installer
[09:09] <CIA-1> ubiquity: 0.0.1ubuntu7, user-setup 1.20ubuntu10.
[09:12] <superm1> evand, could you squeeze one more thing into that upload, or is it already on it's way up?
[09:16] <superm1> well debian/changelog still has UNRELEASED, i'll drop it in, if it's too late for it, just mark it for the "next" upload in the changelog
[09:16] <CIA-1> ubiquity: superm1 * r2915 ubiquity/ (debian/changelog scripts/mythbuntu/apply-drivers):
[09:16] <CIA-1> ubiquity: Simplify the xorg.conf that is written out in apply-drivers. (LP: #181564)
[09:16] <CIA-1> ubiquity:  The ideal situation is to enable these via Jockey, but Jockey doesn't
[09:16] <CIA-1> ubiquity:  have a good interface for turning on during installation and turning
[09:16] <CIA-1> ubiquity:  on and off tv-out.
[09:20] <evand> superm1: just in the nick of time :)
[09:20] <evand> I ctrl-c'ed the upload
[09:20] <superm1> evand, awesome.  i hadn't pushed it earlier since i didn't have real hardware to verify it on until i got home a few minutes ago.
[09:20] <superm1> thanks
[09:21] <evand> anytime
[09:22] <evand> ugh, I hate syslinux.
[09:23] <evand> It seems content not overwriting the boot sector.
[09:28] <ganes> is there any command to grasp the grub output to any file
[09:31] <cjwatson> evand: apt-setup> I just misspoke - your commit looks fine
[09:31] <cjwatson> ganes: please don't simply repeat the same question over and over again. What do you mean? "grasp ... to any file" doesn't really make sense, so try rephrasing.
[09:32] <evand> perhaps I'm missing something but dd if=/dev/sdb of=test bs=446 count=1; syslinux /dev/sdb; dd if=/dev/sdb of=test2 bs=446 count=1; cmp test1 test2; should state that the files differ, no?
[09:33] <persia> evand, Not if someone ran syslinux against the device previously.
[09:33] <cjwatson> evand: don't we still need the partman-target change (150872)?
[09:33] <cjwatson> oh, I suppose technically that won't affect ubiquity as such, because it'll be in finish-install or whatver
[09:34] <cjwatson> whatever
[09:35] <ganes> cjwatson,the command  "find /boot/boot/stage2" in grub prompt gives output that i need in a file
[09:35] <evand> 150872> it's done in ubiquity now
[09:36] <davmor2> ganes: just add > foo.bar to the end to grab text from a stream where foo.bar is the name of the text file
[09:37] <cjwatson> davmor2: that works at the grub prompt, does it? I didn't think so.
[09:37] <davmor2> cjwatson: sorry didn't realise that bit :)
[09:38] <ganes> cjwatson, you are correct , it will not work at grub prompt
[09:38] <cjwatson> ganes: try "echo 'find /boot/boot/stage2' | sudo grub --batch >/some/file/name'
[09:38] <cjwatson> err, with sensible quoting. you get the idea.
[09:38] <evand> persia: ah, thanks!  noted.  I'll have to play with that some more in a few hours. (Trying to work through the cause of the no boot device bugs on usb-creator, noticed it doesn't like to overwrite grub for one)
[09:39] <cjwatson> evand: 150872> that isn't good enough for d-i installs surely.
[09:39] <cjwatson> as I wrote in that bug:
[09:39] <cjwatson> Instead, in the USB install case I think we need to remove /cdrom from /etc/fstab at the end of the installation, in a finish-install script (for d-i) and in custom scripts/install.py code (for ubiquity).
[09:39] <evand> indeed, I can try to quickly cook something up.
[09:39] <persia> evand, I was having lots of trouble with syslinux earlier in the cycle.  My experience was that doing it multiple times helped.  Since this is counter to my experience with most such tools, I mention it.
[09:39] <ganes> cjwatson, at console mode or grub
[09:40] <evand> persia: multiple times?  like running syslinux over and over again in a loop?
[09:40] <persia> It seems to work better when applied to something that hasn't been previously bootable.
[09:40] <cjwatson> ganes: does it look like a command you'd type at the grub prompt?
[09:40] <cjwatson> ganes: (hint: no)
[09:40] <ganes> cjwatson, yeah
[09:40] <persia> evand, Well, I did it manually, but yes, sometimes the second or third time would work.  From my limited reading of the code, I have no idea why this would be true.
[09:41] <ganes> cjwatson, the command which run in grub that output i want in a file
[09:41] <cjwatson> ganes: I've already given you the answer. You run it at an ordinary shell.
[09:41] <evand> so I should be able to dd the first 446 to zero and remove ldlinux.sys and it will be happy to write, no?
[09:41] <evand> persia: scary
[09:43] <ganes> cjwatson, no output
[09:44] <ganes> cjwatson, its giving empty file
[09:44] <persia> evand, At least if you zero it first, you can check if it wrote.
[09:45] <cjwatson> ganes: you'll have to experiment locally
[09:45] <cjwatson> ganes: for further help, please ask #ubuntu
[09:45] <evand> persia: true
[09:45] <ganes> cjwatson, yeah, i can pass the value from console to grub but reverse is not possible by me
[11:00]  * evand goes to quickly test http://evalicious.com/tmp/partman-target-remove-bad-cd.patch
[11:15] <evand> hrm, perhaps that should be 02remove_broken_cdrom
[11:35] <evand> hrm, no vol_id
[12:05] <evand> woohoo, it works
[12:09] <CIA-1> partman-target: evand * r733 partman-target/ (4 files in 3 dirs):
[12:09] <CIA-1> partman-target: Remove the cdrom mounpoint if the filesystem is not iso9660 or udf
[12:09] <CIA-1> partman-target: (LP: #150872).
[12:09] <cjwatson> evand: hmm, I'm not sure I like that
[12:10] <cjwatson> since it fiddles about with the whitespace on the line
[12:10] <cjwatson> purely cosmetic I know
[12:10] <evand> ah, hrm
[12:10] <cjwatson> can I suggest instead doing 'while read line; do ...' and then extracting fields from $line with sed?
[12:10] <cjwatson> then you can just echo "$line" in the passthrough case and you definitely won't change those lines at all
[12:11] <evand> ok
[12:11] <evand> will do
[12:13] <evand> cjwatson: is 02remove_broken_cdrom OK, or should I pick a different number?
[12:14] <cjwatson> evand: I'd have thought it would need to go after 50fstab_removable_media_entries
[12:14] <cjwatson> oh, hang on, finish-install.d not finish.d
[12:14] <cjwatson> ignore that last
[12:15] <cjwatson> evand: I think your change is going to interact messily with finish-install.d/90cdrom in the finish-install source package
[12:15] <cjwatson> hmm, actually no that's rubbish too, the checks there are good
[12:15] <cjwatson> in that case I think 02 is fine
[12:16] <evand> ok, thanks
[12:50] <evand> cjwatson: is http://evalicious.com/tmp/partman-target-remove-bad-cd2.patch more what you had in mind?
[12:50] <evand> I'm going to test that now
[12:55] <cjwatson> evand: yeah
[13:45] <persia> evand, On a machine with no network, I just got a somewhat interesting error: "HwDetect failed with code 1".  What's the best information to preserve to file this as a bug?
[13:47] <cjwatson> /var/log/syslog /var/log/partman
[13:47] <cjwatson> (default for installer bugs)
[13:48] <cjwatson> have a look in /var/log/syslog and see if it has anything obvious; if not, it might be worth reproducing with 'ubiquity -d' (or debug-ubiquity on the kernel command line) and capturing /var/log/installer/debug too. If you do that then don't use a valuable password.
[13:54] <persia> /var/log/syslog has a traceback.  Although I didn't run with -d /var/log/installer/debug has a GTK warning, and a report of X crashing (possibly unrelated).
[13:55] <persia> I'll see if I can reproduce.
[13:59] <cjwatson> traceback is probably just "hw-detect failed"
[13:59] <cjwatson> the underlying problem is that a shell script exited non-zero which is not going to be elucidated by a python traceback
[14:02] <persia> I think the interesting thing is console-kit-daemon reporting "GLib-GObject-CRITICAL: g_object=unref: assertion 'G_IS_OBJECT (object)' failed", which I suspect is related to the debugging james_w is currently chasing.  If I can reproduce with -d, perhaps this can be confirmed.
[14:03] <cjwatson> I suppose that *might* be related to hw-detect failing but it seems pretty tenuous!
[14:03] <cjwatson> you could edit /bin/hw-detect towards the start of installation and put 'set -x' near the top.
[14:04] <persia> Ah.  I was guessing that it couldn't get the rights it needed, and that's why it crashed, but yes, tenuous.  Reproduction is the key important part, I suspect.
[14:05] <cjwatson> hw-detect is a d-i component and so wouldn't be intentionally interacting with consolekit
[14:05] <cjwatson> I suppose it could be something it calls that does so
[14:51] <persia> cjwatson, I couldn't reproduce.  I'm guessing the log files aren't very helpful?
[14:51] <cjwatson> not if you couldn't reproduce it ...
[15:04] <charlie-tca> Hey, what's the user name and password to log in to today's live cd?
[15:07] <cjwatson> err. you shouldn't ever have to log in to a live CD
[15:07] <cjwatson> did you do something unusual?
[15:08] <charlie-tca> Nothing unusual, testing for a couple of weeks now, and today it did not log in, Xubuntu LiveCD
[15:09] <cjwatson> the question is not "what's the username and password" but "why is it so horribly broken" :-)
[15:09] <charlie-tca> I see. I'll try rebooting it, then file the bug report?
[15:09] <cjwatson> did Xubuntu stop using gdm or something?
[15:09] <cjwatson> charlie-tca: try 'ubuntu' as the username, with a blank password
[15:09] <charlie-tca> Not that I know of. It's the same login screen.
[15:10] <charlie-tca> did that, wrong username or password
[15:10] <cjwatson> then your CD is completely buggered
[15:10] <charlie-tca> Yea, that's what I was afraid of. Thanks
[15:11] <cjwatson> my suspicion would actually be along the lines of broken CD drive or dodgy burn. Find out if somebody else can reproduce it?
[15:11] <charlie-tca> will do. I'll have to burn it again, too.
[16:03] <superm1> xivulon, I finally got to testing 5.13 wubi with mythbuntu 8.10.  I did it without being connected to the web though, and the wubi exe crashed at the end before i got to reboot.
[16:03] <superm1> all the content was prepared though, and i was able to install
[16:31] <davmor2> who worked on system cleaner?
[16:32] <evand> davmor2: liw
[16:32] <davmor2> evand: ta :)
[16:32] <nebuchadnezzar> hello
[16:34] <nebuchadnezzar> is there a way to specify the "first disk whatever it is" in partman-auto/disk ?
[16:34] <nebuchadnezzar> I try the devfs syntax but seems not working
[16:34] <cjwatson> yes, certainly; just 'd-i partman-auto/method string regular' and *omit* partman-auto/disk
[16:35] <cjwatson> actually, that will use the first disk provided that there is only one
[16:35] <cjwatson> if you want "first disk", then the only really meaningful label for that is /dev/sda ...
[16:36] <nebuchadnezzar> ok, because I saw some RAID driver defining some /dev/rd, IIRC
[16:37] <nebuchadnezzar> I have a LVM recipe wich works well, but having 2 disks creates problems, acording to the documentation I must declare partman-auto/disk
[16:37] <cjwatson> ah, if you have more interesting disks then you need to work harder, yes
[16:37] <cjwatson> hardy or intrepid?
[16:37] <nebuchadnezzar> hardy
[16:38] <CIA-1> partman-target: evand * r734 partman-target/finish-install.d/02remove_broken_cdrom: Cleaner version of 02remove_broken_cdrom.
[16:38] <cjwatson> damn. it's annoyingly fiddly in hardy.
[16:38] <nebuchadnezzar> :-)
[16:39] <nebuchadnezzar> ok, I don't find fiddly in google translate ;-)
[16:40] <cjwatson> d-i preseed/early_command string (echo '#! /bin/sh'; echo 'set -e'; echo '. /usr/share/debconf/confmodule'; echo 'if [ -f /var/lib/partman/initial_auto ]; then exit 0; fi'; echo 'db_set partman-auto/disk "$(list-devices disk | head -n1)"'; echo 'exit 0') > /lib/partman/display.d/01firstdisk; chmod +x /lib/partman/display.d/01firstdisk
[16:40] <cjwatson> untested
[16:40] <nebuchadnezzar> ouch
[16:40] <CIA-1> partman-target: evand * r735 partman-target/debian/changelog: releasing version 55ubuntu5
[16:41] <nebuchadnezzar> thanks, I'll test that
[16:41] <nebuchadnezzar> tomorrow
[16:42] <cjwatson> the other approach of course is just to deliver slightly different preseed files to different machines
[16:43] <nebuchadnezzar> ok
[18:00] <xivulon> superm1 do you have logs for me?
[18:01] <superm1> xivulon, I wasn't sure wubi's exe left logs?
[18:02] <superm1> at least about the crash.  i got an offer to submit the details to MS, but i didnt' think that would be useful..
[18:03] <xivulon> the log is in your user temp folder, type %temp% in windows explorer
[18:03] <superm1> xivulon, okay i'll have to retry later today when i have that windows VM by me again then
[18:03] <xivulon> running with --skipmd5checks should do as a workaround
[18:04] <superm1> you think the crash was from it trying to do md5 checks without knowing the md5 then?
[18:04] <xivulon> no it's likely an old and annoying bug very difficult to track, simply changing code path makes it go away in most cases
[18:05] <superm1> oh yuck
[18:06] <xivulon> I didn't bother with that since 8.10 was supposed to use a complete rewrite, but didn't finish it in time, so had to resume the old code
[18:07] <xivulon> it's --skipmd5check singular
[23:08] <CIA-1> grub-installer: cjwatson * r753 ubuntu/ (debian/changelog grub-installer):
[23:08] <CIA-1> grub-installer: Set a sensible default boot device when /cdrom is not iso9660, as this
[23:08] <CIA-1> grub-installer: is probably a USB install and (hd0) does not make sense when installing
[23:08] <CIA-1> grub-installer: from a removable disk (LP: #282037).
[23:11] <CIA-1> grub-installer: cjwatson * r754 ubuntu/debian/changelog: releasing version 1.32ubuntu10
[23:20] <CIA-1> debian-installer: cjwatson * r980 ubuntu/ (35 files in 2 dirs): Update help text translations from Launchpad.
[23:36] <CIA-1> debian-installer: cjwatson * r981 ubuntu/debian/changelog: releasing version 20080522ubuntu22