[00:01] <cjwatson> mathiaz: r488
[00:02] <mathiaz> cjwatson: great - thanks
[00:11] <Guest14892> what are the key differences btween karmic and jaunty ps iM A LINUX NOOB
[01:37] <Turl> hi
[01:38] <Turl> apturl package is broken
[01:38] <Turl>  error creating directory `./usr/lib/python2.6/dist-packages/AptUrl/gtk': No such file or directory
[05:42] <dholbach> good morning
[05:43] <dholbach> vorian: bug 386428 is assigned to you, when will you take care of it?
[06:09] <StevenK> ArneGoetje: I'd like to change the ibus.desktop to also have NoDisplay=true to hide it in the Applications menu. I've confirmed with both desktop and UNR that the preferences under System->Preferences is a seperate .desktop file. What do you think?
[06:58] <ArneGoetje> StevenK: IMHO the System->Preferences entry is sufficient, so we don't need any other .desktop file. IMHO it should be removed by upstream.
[06:59] <StevenK> ArneGoetje: That sounds fine too. Shall we patch it out first ourselves?
[07:01] <ArneGoetje> StevenK: let's get the package maintainer's opinion first... I'll comment on the bug
[07:02] <StevenK> ArneGoetje: Sounds good
[07:05] <porthose> are alpha5 iso images available?
[07:09] <pitti> Good morning
[07:14] <tkamppeter> pitti, good morning
[07:15] <tkamppeter> pitti, one additional thing about CUPS:
[07:15] <pitti> hi tkamppeter
[07:15] <pitti> good morning
[07:16] <tkamppeter> Installing it, adds a blacklist against the usblp module, but if a user updates and the module is loaded, the module keeps dangling around blocking the printers.
[07:16] <tkamppeter> pitti, Perhaps we should do "rmmod usblp" in postinst.
[07:17] <pitti> no
[07:17] <pitti> rmmod is evil, bad, and wrong
[07:17] <pitti> and it might even fail
[07:17] <pitti> but perhaps I was missing an update-initramfs call
[07:17] <pitti> to make the blacklist go into the initramfs, too; otherwise there's a chance that it gets loaded there
[07:18] <tkamppeter> pitti: Can modprobe more safely remove modules?
[07:18] <pitti> tkamppeter: no, modules can't guaranteed to be removable
[07:18] <pitti> something can still use the device, then it will just fail
[07:18] <pitti> but in the worst case it causes kernel panics and all that
[07:19] <pitti> tkamppeter: but after a jaunty->karmic upgrade you have to reboot anyway
[07:28] <tkamppeter> pitti: Should we perhaps set this special bit which is also set on kernel updates and shows the user a pop-up telling him that he needs to reboot?
[07:30] <pitti> tkamppeter: we can do that, yes; the postinst can check if upgrading from 1.3.x, and check if /usr/share/update-notifier/notify-reboot-required exists, and call that
[07:36] <pitti> :q
[07:36] <pitti> meh, compiz' focus handling sucks
[08:09] <ogra> can someone rescore empathy on armel ?
[08:18] <pitti> ogra: done
[08:19] <ogra> thanks
[08:20] <ogra> pitti, could you bump oo.o too ?
[08:21] <pitti> ogra: done
[08:21] <ogra> thanks again :)
[08:21] <pitti> except... lazr.restfulclient.errors.HTTPError: HTTP Error 503: Service Unavailable
[08:22] <pitti> ogra: web UI times out, too; sorry
[08:23] <ogra> gah
[08:23] <pitti> crank faster?
[08:24] <ogra> there is only so much lever i can apply
[08:25] <ogra> oo.o takes at least 36h
[08:34] <ogra> cjwatson, did you bump dove to -203 ABI too (your changelog entry doesnt say so)
[08:42] <cjwatson> ogra: no
[08:55] <tkamppeter_> pitti, your AppArmor fixes on CUPS do not work, the usb backend only works with "sudo aa-complain cupsd".
[08:56] <pitti> tkamppeter_: please reopen the bug and attach dmesg after a failed attempt
[08:56] <pitti> I'm off for some two hours for some errands
[09:11] <YokoZar> ArneGoetje: Can I ask for your opinion on https://bugs.launchpad.net/ubuntu/+source/wine1.2/+bug/412195 please?
[09:12] <ogra> cjwatson, dove bump comitted and pushed in platform seed and d-i ... in case you do another upload, please include it (doesnt look good for armel anyway since OO.o failed and is just being retried)
[09:23] <ogra> cjwatson, thanks :)
[09:40] <ArneGoetje> YokoZar: I will test the font myself and give you feedback
[09:40] <YokoZar> ArneGoetje: Thanks.  I'm hoping it's something I can just fix with an upstream patch
[10:25] <smb> pitti, Would you have time to have a quick look at bug 399877? Could that be a side effect of moving away from hal? To me it looks as there is a nautilus/banshee/whatever problem with the fact that the first partition of that device seems to be unusable but the second contains data.
[11:39] <t0bi1> hey guys, i disabled nautilus at startup in jaunty, but after the karmic-update its back again...i checked gnome-session-properties, but it aint there...where does it get started and how can i disable it?
[11:40] <seb128> t0bi1, /desktop/gnome/session/required_components_list
[11:40] <seb128> in gconf-editor
[11:40] <t0bi1> thx!
[11:40] <seb128> you're welcome
[11:41] <t0bi1> brb
[11:56] <iulian> pitti: Hi.  It seems that libyaml-perl is in dependency wait state due to libtest-cpan-meta-perl being in Universe.  I was thinking of writing a MIR for it but I noticed that libyaml-perl has 3 other dependencies in Universe.
[11:57] <iulian> pitti: These dependencies were added because some tests failed.
[11:57] <iulian> pitti: I'm not sure what to do.  Is a MIR needed for all dependencies?
[12:05] <pitti> juanje: at least a main inclusion request bug; for simple perl libs we don't need a full-fledged MIR wiki page, but we still need to review the pacakges
[12:05] <pitti> smb: I'll have a look
[12:11] <tkamppeter> pitti, I have fixed the remaining AppArmor problem of bug 420015 now.
[12:12] <pitti> tkamppeter: oh, nice; saw your dmesg mail, but didn't read it yet
[12:12] <tkamppeter> pitti: Still remaining problem is that the infinite loop of the usb backend (bug 420797) is still there.
[12:12] <iulian> pitti: I suppose that message was for me, right?
[12:13] <pitti> iulian: oops, sorry
[12:13] <pitti> yes
[12:13] <pitti> meh, the apport retracers fail due to a wadllib exceptions
[12:13] <pitti>   File "/usr/lib/python2.6/dist-packages/wadllib/application.py", line 809, in bind
[12:13] <pitti>     value = validated_values[param.name]
[12:13] <pitti> KeyError: 'content_type'
[12:13] <pitti> did anyone happen to see this?
[12:20] <tkamppeter> pitti, I trying to build CUPS, but the test suite prevents it:
[12:20] <tkamppeter> FAIL: 1 warning messages, expected 0.
[12:20] <tkamppeter> W [02/Sep/2009:13:17:43.805950 +0200] [CGI] Unhandled message: interface=org.freedesktop.DBus.Introspectable, path=/, member=Introspect
[12:20] <pitti> right, I saw that as well yestercay
[12:20] <pitti> it still built fine on the buildds, but I guess something changed in between
[12:23] <pitti> james_w: would it be possible to suppress creating bzr brnaches for packages which have a Vcs-Bzr:?
[12:23] <pitti> james_w: now people keep getting confused and send me merge requests against e.g. lp:ubuntu/apport/karmic
[12:23] <james_w> technically, sure
[12:23] <pitti> but they are totally useless, since they are unmergeable
[12:23] <james_w> we don't want to do it wholesale
[12:23] <james_w> we will fix it for cases like that and just use your branch
[12:23] <pitti> but if a package already is maintained in bzr, anyone who uses the auto-import can just wreak more havoc?
[12:23] <james_w> I'm waiting on some fixes to make that possible though
[12:24] <james_w> yes, and I'm keen to fix it
[12:24] <pitti> ok, thanks
[12:24] <james_w> it was easier to just not special case anything to start and then fix up these cases later
[12:24] <james_w> unfortunately we're still waiting on external changes though
[12:24] <pitti> ok, I see
[12:45] <geser> could someone please give-back: ibus-anthy liblauncher liblouis. They failed to upload on ia64 and once sparc as the upload happened when the source got promoted to main. Thanks
[12:45] <iulian> pitti: Hmm, there are loads of packages that need to be promoted.  If we only promote yaml's deps, those dependencies will go in depwait state.  Is there an easy way to find all packages that need to be promoted? apt-cache + grep is not useful.
[12:47] <pitti> iulian: in fact there is: http://people.canonical.com/~ubuntu-archive/component-mismatches.txt
[12:47] <cjwatson_> geser: done
[12:51] <geser> iulian: you could try to simulate a install of each package in a pbuilder (or any other clean environment), filter out the needed packages and check for them the component
[12:55] <tkamppeter> pitti, the CUPS upstream fix for bug 420797 has no effect. Seems that Mike has done a wild guess without testing.
[13:01] <geser> asac: is it a mistake that both xulrunner-1.9 and xulrunner-1.9.1 try to build a xulrunner-dev package? this caused a failed to upload error for xulrunner-1.9 on powerpc
[13:02] <asac> geser: yes. xulrunner-1.9 is probably quite outdated?
[13:02] <asac> hmm
[13:02] <asac> could be that i forgot to remove it ;)
[13:02] <asac> let me check
[13:07] <iulian> geser: Good idea.  That ought to be much easier.
[13:09] <geser> iulian: try something like that: for i in `apt-get -s install libtest-yaml-meta-perl | grep ^Inst | cut -d" " -f2`; do apt-cache madison $i | grep universe; done
[13:25] <lifeless> uhm
[13:25] <lifeless> where has runlevel and telinit gone ?
[13:26] <lifeless> or perhaps a better question, can someone tell me what package they are in, on your running karmic ?
[13:27] <pitti> $ dpkg -S /sbin/runlevel
[13:27] <pitti> upstart: /sbin/runlevel
[13:27] <pitti> lifeless: ^
[13:27] <pitti> (same for telinit)
[13:27] <lifeless> ah
[13:27] <lifeless> my upstart didn'tupgrade
[13:27] <lifeless> because of upstart-logd
[13:28] <lifeless> so I'll be making a boot disk for my laptop tomorrow
[13:29] <lifeless> is it worth filing a bug? if there were some way to say 'gotta have upstart > X || <old thing'
[13:32] <asac> geser: fixed in xulrunner-1.9.head branch for .14
[13:37] <EagleScreen> how can I change the text editor used by dch?
[13:38] <ion> select-editor
[13:50] <lool> Sorry what's the magic for calling cpp so that it dumps all macros it knows about?
[13:52] <lool> Ah -dM
[13:52] <lool> Found it in the man page, like last time </slap>
[14:15] <soreau> I am writing a script that installs some trivial components (extra compiz plugins in this case) and needs certain dependencies. I would like to have the dependencies in the script checking if they're installed before running apt-get install blah blah. Is there an easy way to check if a package is already installed via bash script?
[14:17] <soreau> I was thinking about grepping dpkg -l but that seems a bit cumbersome
[14:29] <lool> cjwatson: Would love if you could apply the minor fix I just pushed to lp:~ubuntu-core-dev/console-setup/ubuntu in Debian if it's not there already (s/dpkg --print-installation-architecture/dpkg --print-architecture/)
[14:31] <cjwatson> lool: already done in console-setup 1.37
[14:31] <lool> k that's what I guess thanks
[14:31] <lool> guessed
[14:39] <Chipzz> soreau: dpkg --get-selections
[14:41] <soreau> Chipzz: What's the difference between 'install' and 'deinstall' there?
[14:43] <soreau> That is a very nice looking list though ;)
[14:44] <ScottK> soreau: A more elegant way to solve the problem would be to put your script into a Debian package and have that package depend on the packages that need to be installed.
[14:45] <soreau> ScottK: If that were the case, I would just make a deb package compiz-plugins-experimental. The reason for the script is that it should work on other distros too provided the dependencies are installed for that system
[14:46] <ScottK> Then a dpkg solution won't work either.
[14:46] <soreau> I'm just tailoring it to fit ubuntu since it's the most popular and more noobs use it
[14:52] <cjwatson> soreau: it wouldn't be all that much less efficient to just 'apt-get install' them unconditionally, and probably easier
[14:52] <cjwatson> anyway, dpkg --get-selections is wrong; that lists desired states, not current states
[14:52] <soreau> cjwatson: That's what it does currently
[14:53] <soreau> and that's how I'm thinking of keeping it, just tinkering with the script and will write a how-to on the forums
[14:53] <cjwatson> you probably want something more like the third column of:  dpkg-query -W -f '${Status}\n' $packages
[14:53] <soreau> cjwatson: ok
[14:53] <soreau> Thanks for all your help guys, gotta go to work
[14:53] <cjwatson> you could use 'apt-get --no-upgrade install', even
[15:42] <tkamppeter> pitti, I have found a fix for bug 420797 now. Can you upload CUPS, bug is probably severe enough to let the fix go into Alpha 5.
[15:47] <pitti> tkamppeter: yay you
[15:49] <pitti> tkamppeter: it might not make alpha-5, since the candidate images are already built, but of course people can dist-upgrade after installation, so we should still get it uploaded
[15:49] <cjwatson> we're going to need a respin
[15:50] <cjwatson> really needed to sort out a grub issue
[15:50] <pitti> ok, uploading right away then
[15:50] <cjwatson> Keybuk: did you say you'd run into fsck failing due to timestamp skew again?
[15:54] <pitti> tkamppeter: did you commit the fix? bzr pull gives nothing new
[15:55] <pitti> tkamppeter: forgot to push? in the bug you said you committed
[15:59] <tkamppeter> pitti, now it is pushed.
[15:59] <pitti> tkamppeter: hm, now I get a "bzr: ERROR: mismatched lock context and write group. None, <bzrlib.transactions.WriteTransaction object at 0xa118dac>"
[15:59] <pitti> tkamppeter: ah, works now
[16:06] <Keybuk> cjwatson: a couple of people have reported it
[16:06] <Keybuk> but so far the reasons have been explainable by other problems
[16:06] <cjwatson> cr3: meet Keybuk
[16:07] <cjwatson> I've noticed it myself, though only once and it didn't happen next time; I wonder if it's something to do with unreliable NTP sync
[16:07] <Keybuk> the two I've seen were
[16:07] <Keybuk> - installing over top of Windows, UTC=yes but hardware clock was in localtime
[16:07] <Keybuk> Live CD doesn't "save back" the time
[16:07] <Keybuk> so the first boot, you have the wrong time (double timezone)
[16:07] <Keybuk> then NTP fixes that
[16:08] <Keybuk> then you shutdown and save the hwclock
[16:08] <Keybuk> and then the second boot, your mount time is in the future
[16:08] <Keybuk> (and after that, it's fine)
[16:08] <pitti> tkamppeter: uploaded
[16:08] <davmor2> Keybuk: I'm having issues on every first reboot
[16:08] <Keybuk> the second is Scott Ritchie, who dual boots between jaunty and karmic, and one is UTC=no and the other is UTC=yes which is clearly not going to work ;)
[16:08] <Keybuk> davmor2: first reboot after what?
[16:09] <cjwatson> FWIW I've seen it on a server install in kvm with a blank disk
[16:10] <cjwatson> which is about the simplest scenario imaginable
[16:10] <cjwatson> unfortunately, as I say, it didn't happen the next time when I was trying to look at it
[16:10] <Keybuk> yeah that's the problem
[16:11] <davmor2> Keybuk: install
[16:11] <Keybuk> a couple of people have hit it
[16:11] <Keybuk> then "fixed" it and carried on
[16:11] <Keybuk> to debug, I need them to stop at the fsck prompt and reach for help
[16:11] <Keybuk> since that's the only time we can see the disparity between the hardware and system clocks ;)
[16:11] <cjwatson> fsck didn't prompt in the case I saw - it just exited 1
[16:11] <Keybuk> right, it'll say on the screen what the difference between the timezones is though
[16:11] <davmor2> Keybuk: tell me what you need and give me 20 minutes
[16:12] <gnarl> Not sure this was somehow involved in my other issue but it seems apturl_0.4.0ubuntu3 ist not installable on my system
[16:12] <Keybuk> davmor2: do an install, and the reboot, and it should fail to boot
[16:12] <Keybuk> with fsck outputting information
[16:12] <Keybuk> need the info fsck outputs
[16:12] <Keybuk> and "hwclock --debug --show", "grep UTC /etc/default/rcS" and "date"
[16:12] <davmor2> Superblock last mount time (Wed Sep  2 16:18:53 2009, now = Wed Sep  2 15:31:31 2009) is in the future this info or all of it?
[16:13] <mathiaz> cjwatson: hi - I'm testing the new daily -server iso.
[16:13] <mathiaz> cjwatson: it seems that the node-preseed file is not generated correclty
[16:13] <cjwatson> mathiaz: ok, details?
[16:14] <mathiaz> cjwatson: /etc/eucalyptus/node-preseed.cfg doesn't exist at all once the system is installed
[16:14] <davmor2> Keybuk: right re-installing now the line above was from my last install
[16:14] <mathiaz> cjwatson: /etc/eucalyptus/node-preseed.conf to be correct
[16:15] <cjwatson> mathiaz: can I see logs?
[16:15] <mathiaz> cjwatson: sure - one moment
[16:16] <sistpoty|work> Keybuk: imho fsck shouldn't panic in this situation, would that be a solution?
[16:17] <mathiaz> cjwatson: http://people.canonical.com/~mathiaz/euca-cc-install.log.tgz
[16:18] <cjwatson> mathiaz: what does euca-cc.preseed contain?
[16:19] <davmor2> keybuk: do you want all the info dropping in bug 422869?
[16:19] <mathiaz> cjwatson: http://people.canonical.com/~mathiaz/euca-cc.preseed
[16:24] <cjwatson> mathiaz: you aren't installing eucalyptus-udeb
[16:24] <cjwatson> -ianna/choose_modulesstring modules
[16:24] <cjwatson> oops
[16:24] <cjwatson> d-i anna/choose_modules string modules
[16:24] <cjwatson> that should be:
[16:24] <cjwatson> d-i anna/choose_modules string eucalyptus-udeb
[16:24] <NCommander> Stupid question, but if I do apt-get dist-upgrade in the live environment, then install, do those updates translate across?
[16:24] <mathiaz> cjwatson: oh right - paste failure
[16:25] <mathiaz> cjwatson: I'll try again
[16:25] <cjwatson> NCommander: no
[16:25] <NCommander> That's what I thought :-/
[16:25] <cjwatson> (intentionally)
[16:25] <mathiaz> cjwatson: with the corrected line in the preseed
[16:25] <maco> NCommander: use net install. then youre always up to date ;)
[16:25] <NCommander> maco, I need to use live images in this case
[16:25] <maco> oh boo
[16:26] <maco> chroots?
[16:26] <Keybuk> sistpoty|work: disagree
[16:27] <Keybuk> davmor2: interesting that your software and hardware clocks are not out by a timezone multiple
[16:27] <davmor2> Keybuk:  but it is from utc to timezone
 Superblock last mount time (Wed Sep  2 16:18:53 2009, now = Wed Sep  2 15:31:31 2009) is in the future this info or all of it?
[16:28] <Keybuk> ?
[16:28] <Keybuk> oh, sorry
[16:28] <cjwatson> that isn't software and hardware clocks, is it? it's time when filesystem was last touched (software) to current time (hardware)
[16:28] <cjwatson> or possibly vice versa software/hardware, I forget
[16:28] <Keybuk> right you took 13 minutes to reboot
[16:28] <Keybuk> used to seeing those numbers closer together
[16:29] <davmor2> Keybuk: I was running more than one install :)
[16:30] <davmor2> Keybuk: do you want us to use that bug or would you like us to open a new one for this?
[16:32] <Keybuk> open a new one
[16:32] <Keybuk> scott ritchie's bug is very clear and obvious
[16:32] <Keybuk> and is his own config error
[16:33] <davmor2> np's
[16:34] <sbeattie> Keybuk: I'm seeing it guests on karmic's virtualbox, where it's taking the host's setting (in PDT -0700) and exporting it to guest as utc (-0000) which means on reboot it thinks the current hardware time is nearly 7 hours earlier than the superblock last modified time.
[16:35] <sbeattie> This results in being dropped to a shell to manually fsck.
[16:37] <Keybuk> sbeattie: that should be ok though surely
[16:37] <Keybuk> assuming the install has UTC=yes
[16:41] <kirkland> bryce: i think you added Ubuntu to the spell-check database in Karmic, one of the paper cuts, right?
[16:41] <kirkland> bryce: i was hoping to add virtualization
[16:41] <sbeattie> UTC is set to yes. The dropping to a shell for a manual fsck happens on every cold boot.
[16:41] <kirkland> bryce: can you give a pointer?
[16:42] <Keybuk> sbeattie: what is the exact fsck message when it happened?
[16:45] <davmor2> Keybuk:  bug 423247
[16:46] <davmor2> keybuk: is that everything you need or is there anythind else?
[16:46] <davmor2> anything even
[16:46] <Keybuk> don't know, please leave it at that root prompt for a moment
[16:47] <Keybuk> err
[16:47] <Keybuk> you haven't pasted all the hwclock output
[16:47] <Keybuk> more specifically, you haven't pasted THE MOST IMPORTANT BIT! :D
[16:47] <davmor2> keybuk: no probs
[16:47] <davmor2> I'll add the rest
[16:48] <Keybuk> interesting that your "last mount time" is an hour in the future
[16:48] <Keybuk> the real, genuine future
[16:48] <Keybuk> your hardware clock looks like it has the right value (15:40ish UTC)
[16:48] <Keybuk> UTC=yes matches that
[16:49] <Keybuk> your timezone file looks like it has the right value (16:40ish BST)
[16:49] <Keybuk> and your resulting system clock is right (16:40ish UTC)
[16:49] <Keybuk> what's _wrong_ is that your filesystem was unmounted at 17:30 BST
[16:49] <Keybuk> which is an hour from now
[16:49] <Keybuk> unless Wolverhampton is being dragged into a space time vortex?
[16:50] <Keybuk> ok
[16:51] <Keybuk> let's try something else
[16:51] <Keybuk> can you do the install again
[16:51] <Keybuk> (wiping it)
[16:51] <sbeattie> Keybuk: here's the fsck message I see in a vbox guest: http://www.nxnw.org/~steve/tmp/fsck-on-boot.png
[16:51] <Keybuk> but don't reboot, open a terminal instead
[16:51] <Keybuk> sbeattie: again, output of "hwclock --debug --show", "grep UTC /etc/default/rcS" and "date" plz
[16:52] <sbeattie> Keybuk: yep, one sec
[16:52] <davmor2> Keybuk: added
[16:52] <Keybuk> davmor2: thanks - I think this is a problem somewhere in the installer
[16:52] <Keybuk> davmor2: this is your *first* reboot, right?
[16:53] <Keybuk> you've not rebooted the real system yet?
[16:53] <cjwatson> I'm happy to help if it is, but will need advice
[16:53] <Keybuk> cjwatson: well, the filesystem was unmounted in an hour's time
[16:53] <Keybuk> and the hardware and system clock are otherwise correct
[16:53] <Keybuk> so that suggests to me (assuming this is the first boot of the installed system) that the system clock was wrong during the installer session
[16:53] <Keybuk> (live I guess)
[16:54] <cjwatson> hmm
[16:54] <cjwatson> log-output -t clock-setup chroot /target hwclock --systohc --debug &
[16:54] <davmor2> Keybuk:  On this box this was first reboot from the installed system
[16:54] <cjwatson> in clock-setup's finish-install script
[16:54] <cjwatson> I wonder if that's at all related
[16:54] <Keybuk> errr
[16:54] <davmor2> Keybuk:  normally it is the first reboot after installation though
[16:54] <Keybuk> shouldn't think so
[16:55] <sbeattie> Keybuk: http://www.nxnw.org/~steve/tmp/hwclock-info.png
[16:55] <Keybuk> davmor2: ok, so we need to go through an install
[16:55] <Keybuk> and keep checking the hardware and system clocks as we go through each step
[16:55] <davmor2> yes
[16:55] <Keybuk> see if there's some point it goes wrong
[16:55]  * cjwatson places a small bet on it being in finish-install
[16:55] <davmor2> Keybuk: would this be easier from an alternate cd in expert mode?
[16:55] <Keybuk> sbeattie: fresh install as well?
[16:56] <cjwatson> though if it's not there, it'll be in clock-setup when it calls rdate
[16:56] <Keybuk> davmor2: and alternate install is an entirely different installer ;-)
[16:56] <cjwatson> davmor2: have you already reproduced it with an alternate CD?
[16:56] <cjwatson> the alternate and desktop installers do have the bits of code I consider most likely to be causing this problem in common
[16:56] <cjwatson> but that's an estimate ...
[16:57] <davmor2> just checking with kirkland about his
[16:57] <cr3> I can kick off an alternate install if anyone likes
[16:57] <sbeattie> Keybuk: roughly; like I said it happens every cold boot (ubuntu-server install from 24 hours ago).
[16:57] <Keybuk> cjwatson: oddly, the hardware clock looks *right* though
[16:58] <Keybuk> sbeattie: weird, because again, your hardware and system clock are right
[16:58] <cjwatson> presumably the unmount time will be set from the system clock
[16:58] <Keybuk> welkl
[16:58] <cjwatson> so if the system clock is wrong, and the hardware clock isn't set (properly) from it on shutdown
[16:58] <Keybuk> it's very odd that your hardware clock thinks it's 8am ;)
[16:58] <Keybuk> cjwatson: that's true
[16:59] <Keybuk> oh
[16:59] <Keybuk> no
[16:59] <Keybuk> sbeattie: it's obvious what your problem is
[16:59] <Keybuk> sbeattie: your hardware clock says 8am, which is PDT not UTC
[16:59] <Keybuk> but your system has UTC=yes
[16:59] <Keybuk> so the system clock is set 7 hours wound back from that
[16:59] <Keybuk> then you boot
[16:59] <Keybuk> NTP runs
[16:59] <Keybuk> and resets your system clock forwards
[17:00] <Keybuk> and then you unmount on shutdown
[17:00] <Keybuk> with a "future" time
[17:00] <Keybuk> your "hardware clock" would normally be saved (at least mucking up the clock)
[17:00] <Keybuk> but since it's a virtual machine... that's discarded
[17:00] <Keybuk> so you hit it every boot
[17:00] <Keybuk> sbeattie: change UTC=yes to UTC=no
[17:01] <davmor2> cr3: are any of your issues on alternate?
[17:01] <sbeattie> Keybuk: right; but note this only happens on karmic, do to (apparently?) more strict checking by fsck, jaunty and earlier guests don't hit it on boot.
[17:01] <cr3> davmor2: I haven't noticed, I'm preparing an installation now
[17:01] <Keybuk> sbeattie: not so much
[17:01] <Keybuk> fsck upstream had slipped in a hack for Ubuntu to not fail on clock errors
[17:02] <Keybuk> I undid that hack
[17:02] <sbeattie> Keybuk: oh, hah.
[17:03]  * sbeattie wanders off to file a bug against virtualbox, for not exporting the actual UTC to the guest.
[17:04] <dholbach> Ubuntu Developer Week in #ubuntu-classroom - NOW! :-)
[17:04] <kirkland> davmor2: what was you question for me?  is this happening on server install?  yes, absolutely.
[17:04] <davmor2> cool so it should happen on alternate too :)
[17:04] <jjardon> pitti, I've added some comments about gnome-power-manager in https://wiki.ubuntu.com/Halsectomy; maybe they are of your interest
[17:05] <davmor2> Keybuk: I'll run an expert install and save it all to a file it should all > to a txt file shouldn't it
[17:12] <mathiaz> jdstrand: hey - does bug 423246 make sense?
[17:13] <LaserJock> slangasek: if I do some work on the Edubuntu seeds is that going to mess up Alpha 5 release management at all?
[17:13] <jdstrand> mathiaz: it does in general to me, we do for several other services
[17:13] <mathiaz> jdstrand: great - thanks
[17:14] <jdstrand> mathiaz: I peeked at the debdiff, but haven't really gone over it. conceptually, I like it though :)
[17:20] <cjwatson> LaserJock: if in any doubt, wait; it's certainly possible
[17:21] <LaserJock> cjwatson: yeah, I'd rather not hold up progress trying to get the Edubuntu DVD seeds worked out. I can wait a couple days
[17:22] <cjwatson> commit to a branch ...
[17:23] <ogra> yippie, my notifications are back where they belong :)
[17:29] <ebroder> Anybody from ubuntu-sru around who could look at bug #330766?
[17:34] <pitti> jjardon: thanks for your help with that
[17:34] <jjardon> pitti, :) . I've added some upstream bugs too
[17:36] <davmor2> right running from live cd it's too much of a headache from alt :(
[18:12] <bryce> kirkland, yep, snag the aspell-en and hunspell packages, and append your word(s) to the extrawords.txt files in each
[18:13] <bryce> kirkland, whoops, should be hunspell-en-us
[18:20] <hyperair> dpkg: error processing /var/cache/apt/archives/apturl_0.4.0ubuntu3_all.deb (--unpack): error creating directory `./usr/lib/python2.6/dist-packages/AptUrl/gtk': No such file or directory
[18:20] <hyperair> is anyone getting that?
[18:21] <james_w> hyperair: check the bug reports
[18:21] <hyperair> ah
[18:22] <hyperair> hmm three of the same bug eh
[18:22] <ogra> fis was uploaded 1h ago
[18:22] <ogra> *fix
[18:23] <hyperair> ah
[18:38] <davmor2> Keybuk: Right I've added clock.txt to the bug it list when I took the info.  I went for key points I could think of through the install and then post install until the fsck issue.
[18:39] <Keybuk> what was the bug# ?
[18:40] <davmor2> Keybuk: bug 423247
[18:41] <davmor2> Keybuk: hope it makes sense
[18:42] <davmor2> Keybuk: On a plus note you know it's reproduce-able :)
[18:45] <Keybuk> hmm
[18:45] <Keybuk> Wed 02 Sep 2009 06:18:42 PM UTC  -0.935498 seconds
[18:45] <Keybuk> UTC=yes
[18:45] <Keybuk> Wed Sep  2 17:21:36 UTC 2009
[18:46] <Keybuk> --
[18:46] <Keybuk> that's very odd
[18:47] <Keybuk> davmor2: did the install take 10 minutes or 1h10m ?
[18:47] <davmor2> probably 10 minutes give or take
[18:47] <Keybuk> ok
[18:47] <Keybuk> so the hardware clock is being set during the install
[18:48] <Keybuk> and being set an hour forward
[18:48] <Trewas> btw I hit that fsck ~bug today after upgrading a karmic installation in virtualbox (a couple of days worth of updates)... but I didn't take notice of any details
[18:48] <Keybuk> Trewas: thanks for your help ;-)
[18:49] <davmor2> Keybuk: so this happens not always on first reboot but certainly on second reboot.  That is a reboot from the installed system.  rather than the one that spits out the cd if you get me :)
[18:54] <Keybuk> davmor2: yeah
[18:54] <Keybuk> is definitely an installer bug
[18:54] <Keybuk> it's storing local time in the hwclock, not UTC
[18:55] <davmor2> Keybuk: Glad my log was useful then
[18:56] <Keybuk> davmor2: OOI grep GMT /etc/default/rcS
[18:57] <davmor2> Keybuk: Nothing shows up
[18:57] <Keybuk> ok, just checking
[18:58] <torkel> Keybuk: is that something that will affect upgraded systems too? Because I have seen it on my laptop which is upgraded from jaunty to karmic?
[18:59] <Keybuk> no, shouldn't think so
[18:59] <torkel> changing to UTC=yes seems to have fixed it though
[19:02] <lool> Keybuk: Not sure what you said to Cody Russell on xsplash; I think you suggested using system wide config files for the list of apps to wait for, but that doesn't work to wait for different components when multiple environments are installed and user-selected on the gdm screen
[19:03] <lool> Keybuk: Did you take that into account and decided it was not worth supporting, or did I miss how it would be supportable?
[19:04] <lool> Keybuk: #418716
[20:20]  * ccheney is signing the contract on his new home today, buying a new house is scary, heh
[20:34] <soren> cjwatson: I'm working on making the eucalyptus configuration not a conffile. I'm not sure how to handle this on upgrades. Can you offer some advice on that?
[20:49] <soren> mathiaz: I've merged your two Eucalyptus branches. Can you mark them as merged, please?
[20:53] <davmor2> bryce: are you about?
[20:54] <bryce> davmor2, need something?
[20:55] <davmor2> I'm just tracking down the bug number
[21:00] <davmor2> bryce: bug 421225 I'm about to try an install from the latest iso is there anything that you need info wise?
[21:03] <bryce> davmor2, apport-collect <bug-num> and steps to reproduce,
[21:04] <bryce> davmor2, you're seeing this on -ati?
[21:04] <davmor2> yes
[21:04] <davmor2> but I haven't tried out todays images yet only yesterdays
[21:05] <bryce> yeah there's a new -ati uploaded yesterday which would be worth doublechecking
[21:05] <davmor2> bryce: right np's I let you know :)
[21:14] <davmor2> bryce: same thing as yesterday.  I knocked splash off the kernel line so at least I can now access the terminal
[21:18] <mathiaz> soren: they've already been marked as merged.
[21:21] <slangasek> smoser: do we have candidate EC2 AMIs for alpha5 yet that I can link from the tracker?
[21:22] <soren> mathiaz: Sort of :)
[21:22] <soren> mathiaz: They're marked as Merged, but their status is still development (presumably because they were merged into a branch that is not the trunk).
[21:22] <mathiaz> soren: ok - fixing this as well then
[21:23] <mathiaz> soren: the merge proposal was marked as merged
[21:24] <smoser> slangasek, just now, yes.
[21:24] <slangasek> smoser: ok, cool - what are the AMIs?
[21:25] <smoser> us-east-1.i386.karmic.ami:ami-3520c05c
[21:25] <smoser> us-east-1.x86_64.karmic.ami: ami-99df3ff0
[21:25] <smoser> slangasek,  but those aren't public at the moment
[21:25] <slangasek> smoser: should they be made public for testing?
[21:26] <smoser> yes, except
[21:26] <smoser> the name
[21:26] <smoser> ami-99df3ff0   ubuntu-alphas-us/karmic-x86_64-alpha5.manifest.xml
[21:26] <smoser> thats how i bundled/registered it
[21:27] <smoser> what do you think? should i re-register under a different name?
[21:27] <smoser> or is it ok that it says 'alpha5'
[21:28] <smoser> slangasek, ^^
[21:28] <slangasek> smoser: that doesn't seem critical IMHO
[21:28] <smoser> ok. then i'll flip them to public
[21:31] <slangasek> smoser: sounds good
[21:38] <davmor2> bryce: bug 421225 apport-collect wasn't all that  good :(  I've added my lspic-vvnn Xorg.0.log and syslog anything else I can do for you?
[21:40] <bryce> davmor2, oh xsplash must not have apport hooks
[21:40] <smoser> slangasek, those are public
[21:40] <mathiaz> cjwatson: hey - while going through the generated preseed for the node installation, all of the questions asked during the cluster controller installation are included but commented out in the generated preseed file
[21:41] <mathiaz> cjwatson: it looks like a logic problem in the question function from eucalyptus-udeb.finish-install
[21:41] <mathiaz> cjwatson: the answers are correct though
[21:42] <davmor2> bryce: the end of the syslog file is interesting
[21:42] <bryce> davmor2, add your gdm log files too
[21:43] <bryce> davmor2, RT kernel?
[21:45] <davmor2> bryce: gnome-session timing out
[21:45] <bryce> no, what's the RT business there?
[21:45] <davmor2> bryce: No idea default ubuntu iso
[21:46] <davmor2> 20090902.1 or .2
[21:46] <bryce> davmor2, can you explain how you're determining that you're getting the same bug as #421225 ?
[21:46] <bryce> (X freeze problems can arise from a variety of different root causes, but all look the same at first glance)
[21:47] <slangasek> smoser: cheers
[21:47] <slangasek> smoser: are European AMIs also in progress?
[21:48] <davmor2> bryce: It looked the same bug when it was reported.  I can report a separate one if you want
[21:48] <smoser> slangasek, i'm gonna sniff these really quick
[21:48] <smoser> then i'll migrate
[21:48] <bryce> davmor2, yeah thanks, let's do that
[21:48] <bryce> better to have a dupe than to have two bugs and get only one fix
[21:49] <davmor2> bryce: no probs I'm having issues copying the gdm logs
[21:50] <davmor2> bryce: Okay I got it :)
[21:54] <cjwatson> soren: it gets complicated - the "Config file handling" section of debconf-devel(7) might help
[21:54] <cjwatson> mathiaz: hmm, I thought I'd fixed that - logs again?
[21:54] <mathiaz> cjwatson: sure - just one moment
[21:55] <cjwatson> Keybuk: what do you think about http://paste.ubuntu.com/264016/ ?
[21:55] <davmor2> bryce: any preference on package I file against?
[21:55] <bryce> davmor2, just file against xorg
[21:56] <davmor2> np's
[21:56] <bryce> I've got scripts that will move the bug to the appropriate place :-)
[21:58] <taavikko> #422989 should be dublicate of #419126 ?
[21:59] <mathiaz> cjwatson: http://people.canonical.com/~etienne/euca-cc-install.tar.gz
[22:00] <cjwatson> gar, no visible errors
[22:00] <cjwatson> mathiaz: can I get node-preseed.conf too?
[22:01] <davmor2> bryce: https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/423415
[22:04] <bryce> davmor2, thanks
[22:05] <davmor2> bryce: anything else before I move the cd onto another machine so I can continue to test.  Ps had no issues testing on that box last week just thought of  that
[22:06] <mathiaz> cjwatson: http://people.canonical.com/~etienne/node-preseed.conf
[22:10] <bryce> davmor2, dmesg please?
[22:12] <davmor2> bryce: np's
[22:12] <bryce> davmor2, also your gdm0.log appears to be an Xorg.0.log - can you check that you've attached /var/log/gdm/:0.log and maybe :0.log.1
[22:15] <cjwatson> mathiaz: hmm
[22:17] <davmor2> bryce: that is the right one.  It took me a while to figure out how to copy it.  In the end I did cat :0.log > /mnt/drive/...
[22:18] <bryce> davmor2, hmmmm that is sounding wrong
[22:19] <davmor2> bryce: you are on about the logs that are in /var/log/gdm right?
[22:19] <cjwatson> mathiaz: don't wait for me, but I'm trying to work this out - I'll need to attempt to reproduce it locally
[22:19] <mathiaz> cjwatson: right - not critical for alpha5
[22:19] <bryce> davmor2, that's right
[22:19] <mathiaz> cjwatson: should I file a bug in LP?
[22:19] <cjwatson> mathiaz: yes please
[22:19] <mathiaz> cjwatson: ok
[22:22] <davmor2> bryce: figured out why I couldn't copy it cp didn't like the : at the begining
[22:23] <bryce> ah yeah gotta escape that
[22:23] <bryce> it's a shame gnome includes that in the filename
[22:23] <bryce> guess they're just reusing $DISPLAY for it tho
[22:24] <davmor2> bryce: right added 0.log which is a direct cp of :0.log and dmesg
[22:25] <mathiaz> cjwatson: bug 423424
[22:26] <bryce> davmor2, okay thanks.  I don't think we have a good error message yet, but am not sure where to look next.  I'll send the bug upstream and maybe they can give guidance
[22:27] <davmor2> bryce: no worries just ping me if you need me for anything
[22:43] <slangasek> smoser: European AMIs?
[22:45] <smoser> slangasek, give me 2 minutes
[22:45] <slangasek> okie
[22:49] <ion> ½.
[22:50] <ion> Whoops
[22:50] <ion> That would be me killing ssh, thinking 0) the us keyboard layout is active and 1) the connection is dead.
[22:53] <ogra> what is it though ? my font is so small
[22:53] <ogra> does it mean "half a dot" ?
[22:54] <mathiaz> cjwatson: what's the state of the node acknowledgement process on the cluster controller?
[22:55] <mathiaz> cjwatson: is there a command that needs to be run to register/accept the nodes on the cluster controller?
[22:55] <ion> Yeah, the `~ key is §½ in the (awful) fi layout.
[22:55] <cjwatson> I haven't written that bit yet
[22:55] <mathiaz> cjwatson: ok - cool.
[22:55] <cjwatson> if somebody would like to do it, it shouldn't be hard
[22:55] <soren> ion: Danish as well.
[22:55] <cjwatson> euca_conf --discover-nodes was the agreement, I think
[22:55] <cjwatson> it'd basically be a wrapper around avahi-browse
[22:56] <cjwatson> mathiaz: I just had a brainwave about the preseeding thing - you're preseeding the cluster installation, right?
[22:56] <mathiaz> cjwatson: hm - not in this use case
[22:56] <cjwatson> oh, ok
[22:56] <cjwatson> scratch that brainwave then
[22:56] <mathiaz> cjwatson: ie the last installer logs I send you
[22:57] <mathiaz> cjwatson: it's true that I have other installations that I preseed - however this time around it was an install from the iso on plain hardware
[22:57] <cjwatson> I've traced it back as far as the seen flag simply not being set in questions.dat
[22:57] <cjwatson> at least not for all but a very few questions
[22:59] <mathiaz> cjwatson: one question that was asked was the timezone
[22:59] <mathiaz> cjwatson: the node-preseed generated file has the correct answer (Us/Central)
[23:00] <cjwatson> right, the value is correct but it doesn't have the flags set properly
[23:00]  * mathiaz nods
[23:00] <cjwatson> flags are set for the questions that were preseeded, but not for those asked interactively
[23:00] <cjwatson> which implies a cdebconf bug
[23:01] <davmor2> bryce: it only effect ubuntu and maybe xubuntu kubuntu is working fine on that ati box
[23:03] <smoser> slangasek, i'll get those to you, but its going to be a while
[23:05] <slangasek> smoser: ok.  how's testing on the US ones, in the meantime?
[23:05] <davmor2> bryce: it appears that there was an issue with the livefs so is the older ati driver
[23:06] <smoser> well, so far fine other than bug 419306
[23:07] <smoser> which we think we have to just document as known issue for this release.
[23:09] <cjwatson> oh, wait, cdebconf deliberately doesn't save the seen flag in d-i. argh. I'd forgotten about that.
[23:51] <mathiaz> ttx_: http://paste.ubuntu.com/264075/
[23:55] <slangasek> mathiaz: ubuntu-server build is up
[23:58] <kirkland> slangasek: we're already rsync'ing
[23:58] <kirkland> slangasek: ;-)
[23:59] <slangasek> kirkland: good, good