[00:37] <billybigrigger> how do i find out when vbox 3.0 will be hitting repos?
[00:38] <billybigrigger> is there anyway to track the status of it on launchpad or anything? or is it whenever an motu uploads it?
[00:39] <ogra> ScottK, directhex, mono uploaded, the testbuild is still runing, but got far enough beyond the failure point that i'm confident it will get through
[02:36] <billybigrigger> can someone help me out debugging a problem with my webcam and the new kernel?
[02:36] <billybigrigger> it worked fine yesterday, now after upgrading to 2.6.31 today it doesn't work
[02:36] <billybigrigger> Jul  2 19:31:39 cabo kernel: [11502.374422] gspca: usb_submit_urb [0] err -28
[02:36] <billybigrigger> Jul  2 19:31:39 cabo kernel: [11502.398542] ohci_hcd 0000:00:02.0: leak ed ffff88003781f2d0 (#81) state 2
[02:37] <billybigrigger> i get that from syslog when running cheese, the webcam light goes on...but just purple garbled screen
[02:38] <billybigrigger> i looked up gspca on launchpad and see that gspca-source
[02:38] <billybigrigger> Deleted in karmic-release  (Reason: (From Debian) ROM; mainlined in Linux >= 2.6.28)
[02:38] <billybigrigger> so do i file a bug against the new kernel? or where do i go from here?
[02:42] <TheMuso> billybigrigger: Filing a bug against the linux package for version 2.6.31 would likely be a start, yes, giving the dmesg stuff you quoted above.
[02:42] <billybigrigger> no dmesg, just syslog
[02:42] <billybigrigger> gonna start digging around in dmesg i guess
[02:42] <billybigrigger> should i just attach my whole dmes and syslog?
[02:44] <TheMuso> It would probably be better to attach the bits that seem relevant.
[04:48] <TheMuso> c
[05:13] <TheMuso> woohoo. LVM snapshots are broken with kernel 2.6.31.
[05:15] <TheMuso> ...or I wonder if the format for LVM stuff has changed, which causes the breakage...
[05:24] <liw> TheMuso, yo!
[05:25] <TheMuso> Hi liw.
[05:26] <liw> TheMuso, I failed to get my disks setup in a raid -- the bios raid configurator does not see the disks (they do work), whether I follow the manual's instructions or cycle through all combinations of all relevant settings -- any suggestions?
[05:27] <TheMuso> The BIOS doesn't see them. Are they connected to the right ports on the motherboard?
[05:27] <liw> they bios sees them
[05:27] <liw> I can install ubuntu on them and the installed ubuntu boots, for example
[05:27] <liw> the bios lists them in the post, etc
[05:29] <TheMuso> Ah but you can't configure them as RAID?
[05:29] <RAOF> Is there any particular reason to be using the bios raid configurator?
[05:29] <liw> TheMuso, yes, exactly
[05:29] <TheMuso> RAOF: He has been given the unfortunate task of maintaining dmraid.
[05:29] <liw> RAOF, yes
[05:29] <RAOF> TheMuso: You've managed to palm that off?  Sweet!
[05:30] <TheMuso> liw: I think the jmicron controller operates in a few different modes. Either IDE, AHCI, or rAID, depending on the BIOS/motherboard manufacturer, and the available ports on the board.
[05:30] <TheMuso> RAOF: I didn't manage to do anything. Since changing teams, it was reassigned to someone else.
[05:30] <liw> TheMuso, yeah, that's one of the settings I've cycled through
[05:30] <RAOF> So, I guess "why not just use md softraid" isn't an approprate response, then :)
[05:30] <TheMuso> I know the jmicron controller on my gigabyte board does that.
[05:30] <TheMuso> liw: You want it in RAID mode.
[05:31] <liw> TheMuso, I've tried it in each mode, to no effect, alas
[05:31] <TheMuso> RAOF: Try telling all those users out there who want to sit WIndows and Linux on a bad RAID setup. :p
[05:31] <TheMuso> liw: When in RAID mode, you have to press a key combination during boot to get into the RAID configuration utility.
[05:31] <TheMuso> Thats what I have to do with mine.
[05:31] <liw> TheMuso, yup, control-j. no effect :)
[05:32] <TheMuso> liw: What is the motherboard?
[05:32] <TheMuso> i.e manufacturer etc.
[05:32] <liw> Asus P5K
[05:32] <TheMuso> Ok.
[05:33] <TheMuso> One thing I found with mine is that for some reason, when the system boots, capslock is toggled on internally, although this is not shown, well not at least on my wireless keyboard. Prior to the key command neetding to be entered, try pressing capslock twice, then pressing control + J at the appropriate time.
[05:33] <TheMuso> Mine is a gigabyte board, which I think I've already said.
[05:33] <liw> that's... an interesting behavior
[05:34] <liw> I'll try that
[05:34] <TheMuso> Yes, it is.
[05:34] <TheMuso> Thats what I had to do with mine.
[05:34] <liw> unfortunately, doesn't help
[05:34] <TheMuso> grrr
[05:34] <liw> (I have a wired keyboard, though)
[05:34] <TheMuso> you could try control + shift + j
[05:35] <TheMuso> but other than that, I am out of ideas.
[05:35] <TheMuso> Does the board also have an intel controller?
[05:35] <liw> I thought that was it
[05:35] <TheMuso> If so, I would suggest using it instead, as it also supports RAID, and is somewhat better supported by dmraid./
[05:38] <liw> just went through everyting in the bios. the jmicron one seems to be the only one.
[05:39] <TheMuso> Hrm ok.
[05:39]  * TheMuso googles the mb to see what it has.
[05:41] <TheMuso> liw: According to what I've found, it uses an Intel P35/ICH9 chipset, which means there is an intel controller. Let me reboot, and see what its called in my BIOS.
[05:41] <TheMuso> brb
[05:50] <TheMuso> lionel: The section you want to go into if you have it is "integrated peripherals", and the option you want is SATA/RAID AHCI controller mode, or something similar.
[05:50] <TheMuso> sorry, liw ^^
[05:50] <TheMuso> lionel: sorry, wrong person, tab completion caught me out. :p
[05:54] <liw> TheMuso, I have "Onboard devices configuration" with "J-Micron eSATA/PATA controller" (changing it has no effect), and "SATA configuration" with "Configure SATA as" and the only option ever there is "IDE", no "RAID"
[05:54] <TheMuso> hrm ok, no idea then.
[05:54] <liw> TheMuso, I now realize that the J-Micron setting only says eSATA, so might apply only to external SATA drives anyway
[05:55] <TheMuso> hrm thats possible.
[05:55] <liw> and PATA drives, of which I have none anymore
[05:55] <TheMuso> I guess you could go into Linux, and try and find out what controller your drives are connected to, but I think that means digging through sysfs.
[05:56] <liw> i'll do that
[05:57] <TheMuso> In other news, a fresh install of the latest karmic daily from the alternate disk on my notebook leaves me with no sudo access. :)
[06:01] <TheMuso> 8/c
[06:06] <liw> lshw shows the disks as being connected to the 82801IB (ICH9) 2 PORT SATA IDE Controller
[06:07] <liw> and there's nothing connected to the jmicron one, hmm
[06:07] <TheMuso> Right, I thought as much.
[06:08] <TheMuso> SO there is probably an option in your BIOS to turn it into a RAID controller, but i can't help you as to what you are looking for.
[06:08] <liw> if there is, it isn't in the bios menus...
[06:09] <TheMuso> So the "configure SATA as" option didn't have other choices other than IDE? TO find out, you should be able to press enter to get a list of options.
[06:09] <liw> that only shows IDE as the only option
[06:11] <liw> aha, I found a SATA port on the motherboard for the jmicron contoller
[06:11] <liw> a single port
[06:12] <liw> which is rather less useful for raid than I would hope
[06:13] <liw> aha
[06:14] <liw> "Tho configure a RAID 0, RAID 1, or JBOD set, install an external Serial ATA hard disk drive and an internal Serial ATA hard disk drive connected to the onboard Serial ATA connector labeled SATA_E2."
[06:14] <liw> that's... an interesting design decision
[06:17] <TheMuso> liw: SOunds like you got one of the basic boards then.
[06:17] <liw> yeah, it's very basic
[06:17] <TheMuso> And it sounds like the BIOS is not letting you run the intel controller in either AHCI or RAID mode, or at least so far as you have found.
[06:17] <lifeless> liw: fakeraid time :)
[06:17] <lifeless> liw: welcoe to my world
[06:18] <TheMuso> liw: Yes, he has been given the task of maintaining dmraid, now that I've switched teams.
[06:18] <TheMuso> s/liw/lifeless/
[06:18] <lifeless> TheMuso: WOO
[06:18] <lifeless> liw: hehehehe
[06:21] <liw> the manual claims the "Configure SATA as" setting has options "IDE" and "ACHI", but I only ever see "IDE"
[06:21] <liw> no "RAID" documented, though
[06:21] <lifeless> liw: are you looking at the ISM setup ?
[06:21] <lifeless> liw: sorry, IMSM
[06:21] <liw> what is that?
[06:21] <lifeless> the firmware
[06:22] <lifeless> which is onboard, not loadable
[06:22] <lifeless> the intel raid implemenetation is 'Intel Matrix Storage Manager'
[06:22] <liw> and where would I find that?
[06:22] <lifeless> it has three modes for the sata ports that are run by intel. The jmicron ones are a different controller.
[06:22] <lifeless> IDE - runs a vanilla IDE command set
[06:23] <lifeless> ACHI - reports a different device ID, and runs with more scsi like capabilities
[06:23] <lifeless> and RAID, which is ACHI + the ability to read raid sets from disk during boot - but as soon as the OS starts running controllers directly that gets sidelined
[06:24] <lifeless> liw: typically in the BIOS setup
[06:24] <lifeless> liw: there may be multiple options labelled and looking the same.
[06:24] <liw> lifeless, well, gee, the discussion so far, since yesterday, has been concentrated on the fact that the BIOS does not have that
[06:24] <lifeless> liw: yes, I get that.
[06:25] <lifeless> liw: another possibility is that your Motherboard has had the option disabled by the vendor
[06:25] <lifeless> liw: is SATA_E2 the jmicron port?
[06:26] <liw> yes
[06:26] <lifeless> then my bet is that you don't have an IMSM module in that machine build
[06:26] <lifeless> and instead only have the jmicron fakeRAID facility
[06:26] <lifeless> which I haven't worked with :)
[06:26] <liw> that's the way it's looking, yes
[06:27] <liw> oh well, I'll write this up and email robbie
[06:28] <lifeless> what motherboard is it ?
[06:28] <lifeless> the jmicron is a good one to be able to test too
[06:28] <lifeless> its just probably not the most common one
[06:29] <liw> Asus P5K
[06:29] <liw> the jmicron controller is utterly useless for this, since it only has one SATA port (plus an eSATA port)
[06:30] <lifeless> well, it will work
[06:30] <lifeless> just have to have an external chassis too:P
[06:31] <lifeless> which p5K specifically? deluxe? premium?
[06:31] <lifeless> http://event.asus.com/mb/p5k/spec.html
[06:31] <liw> buying more boxes not an option for me, though; I have way too much computer stuff already, and I'm getting ready to move...
[06:32] <liw> plain P5K afaik
[06:32] <lifeless> its missing the magical R
[06:32] <lifeless> you need the ICH9R chipset
[06:32] <lifeless> not ICH9
[06:33] <liw> oh
[06:34] <liw> thanks, that's useful information: now I at least know that I don't ahve the right chipset from Intel and don't need to look at that further
[06:34] <lifeless> I was aiming for helpful :)
[06:34] <liw> you often are :)
[06:35] <lifeless> and checking my facts - http://en.wikipedia.org/wiki/I/O_Controller_Hub
[06:35] <lifeless> I'm correct
[06:35] <lifeless> none of the p5k's listed ICH9D*
[06:36] <lifeless> oly ICH9/ICH9R
[06:36] <lifeless> so ifyou have a plain p5K, its ICH9 and can't do intel matrix storage manager raid
[06:40] <TheMuso> Which IMO is the better of the 2 controllers.
[06:40] <TheMuso> I have not had the best time with jmicron controllers.
[06:40] <TheMuso> But it does work well enough.
[06:41]  * TheMuso has a gigabyte board with the same 2 controllers, with intel and jmicron both able to do RAID. I use the jmicron controller for my DVD drive, and intel for disks. :)
[07:14] <dholbach> good morning
[07:14] <pitti> Good morning
[07:15] <dholbach> hi pitti
[07:18] <pitti> TheMuso: my mixer level always start with 0 and muted after a clean boot; is that just me?
[07:20] <Sarvatt> not just you
[07:20] <Sarvatt> same thing here
[07:24] <Sarvatt> http://pastebin.ca/1482735
[07:31] <Sarvatt> its just pcm and beep in alsamixer getting reset to 0 every boot, master stays where it is
[08:02] <TheMuso> pitti: is this 0.9.16?
[08:10] <pitti> TheMuso: yes, but it also happened with 0.9.15
[08:10] <pitti> TheMuso: http://pastebin.com/f41bc1413
[08:11] <pitti> TheMuso: I think its an alsa problem, not a pulse problem (ICBW, no idea how pulse fiddles with mixers nowadays)
[08:11] <pitti> -rw------- 1 martin martin 780252 2009-07-03 09:10 gvfs_1.3.1-0ubuntu3_amd64.deb
[08:12]  * pitti appreciates bzr-buildpackage to keep my s3kr1t code really secret, but that's a little over-ambitious :)
[08:12] <pitti> james_w: ^
[08:13] <pitti> james_w: (in ../build-tree/ they are okay, but the copies in ../ are 0600 now)
[08:13] <pitti> james_w: I'll file a bug later, just for amusement
[08:16] <TheMuso> pitti: you could try removing /var/lib/alsa/asound.state, however you ahve to unload all modules for sound before doing so, so that a new file is created with fresh settings. I did a fresh install on my notebook today and am not having that problem, whereas I was with the previous karmic install I had on there.
[08:16] <pitti> TheMuso: I'll check out today's live CD
[08:17] <pitti> TheMuso: thanks; if it's just local bad state, it's fine
[08:19] <StevenK> pitti: Which failed to build, since seb128 broke it with gdm
[08:19] <pitti> oh, will look into it
[08:20] <StevenK> gdm: Conflicts: fast-user-switch-applet but 2.24.0-3ubuntu1 is to be installed
[08:20] <StevenK> pitti: ^
[08:22] <pitti> ah, need to unseed
[08:23] <StevenK> pitti: Should I do the same for UNR?
[08:24] <StevenK> pitti: And would you mind sharing your thoughts on bug 395030, when you have a sec?
[08:27] <TheMuso> pitti: If the touchpad on my notebook is not working, where should I start debugging that? I
[08:27] <TheMuso> pitti: It will be Monday, but a pointer would be good.
[08:28] <pitti> StevenK: seeds updated; if you use gdm in unr, you need to do the same (merge?)
[08:28] <pitti> TheMuso: your sentence was cut off
[08:28] <pitti> TheMuso: I'm not familiar with the synaptics driver, I'm afraid; ubuntu-bug xserver-xorg-input-synaptics is all I can give you :/
[08:29] <Sarvatt> 2.6.31 got really touchy about touchpads huh? i need to add i8042.reset=1 ever since around 2.6.30-git13 on this aspire one
[08:29] <pitti> TheMuso: do you see it in sudo lsinput at least?
[08:29] <pitti> or isn't it even detected?
[08:29] <chrisccoulson> StevenK - i wrote a patch to gnome-session to expose restart and shutdown methods over DBus
[08:29] <chrisccoulson> you could use that :)
[08:30] <pitti> TheMuso: if it is in lsinput, please check lshal whether it has input.x11_driver = 'evdev'
[08:30] <chrisccoulson> StevenK - gnome-power-manager has nothing to do with shutdown or restarting now in Karmic
[08:31] <pitti> StevenK: 395030> looks like it needs to call ConsoleKit instead?
[08:31] <hyperair> chrisccoulson: then what does?
[08:31] <hyperair> oh whoops
[08:31] <chrisccoulson> pitti / StevenK - the issue with ConsoleKit is that the session is terminated without being torn down properly
[08:32] <pitti> ah, so it does want to do that
[08:32] <pitti> org.gnome.SessionManager.CanShutdown()
[08:32] <pitti> org.gnome.SessionManager.RequestReboot()
[08:32] <pitti> org.gnome.SessionManager.RequestShutdown()
[08:32] <pitti> if you use gnome-session
[08:32] <chrisccoulson> those are the ones:)
[08:33] <chrisccoulson> but the Request*() methods are Ubuntu patches though
[08:33] <pitti> uh? how does that work upstream then?
[08:33] <chrisccoulson> without those, I don't know of any other proper way now for an application to shut down the machine and terminate the session correctly
[08:34] <chrisccoulson> pitti - i don't know about upstream. i forwarded my gnome-session patch upstream, but I've had no comments so far
[08:34] <pitti> then I don't know either, I'm afraid :/
[08:35] <pitti> gnome-power-cmd reboot perhaps?
[08:35] <chrisccoulson> hmmm, i wonder what that does
[08:35] <pitti> (don't want to try right now, have editors open, etc.)
[08:37] <chrisccoulson> pitti - is the plan to migrate everything to polkit-1 this cycle?
[08:37] <pitti> chrisccoulson: as much as possible, anyway
[08:37] <pitti> http://fedoraproject.org/wiki/Features/PolicyKitOne looks pretty good already
[08:38] <chrisccoulson> thanks. yeah, i was just taking a look at that
[08:38] <chrisccoulson> it seems polkit-1 really breaks shutdown/restart
[08:46] <tseliot> soren: I've just noticed that your requested that you branch was merged into lp:broadcom-sta instead of lp:~broadcom-sta-hackers/broadcom-sta/ubuntu . Can you merge it into lp:~broadcom-sta-hackers/broadcom-sta/ubuntu , please?
[08:51] <pitti> hm, does bluetooth obexftp browsing work for anyone else in karmic? I can pair, but then it just says "service not supported by remote device"
[08:56] <pitti> ah, the G1 doesn't like obexftp, works fine with my old one
[09:15] <vise> hi yall
[09:17] <dunham> Hi all! I'm packaging a software that has a COPYING (GPL2) and copyright information, but some source files miss license headers. some of them only report the author name, others haven't copyright/license info at all. Is such a package acceptable in the archives? Should I ask upstream to fix it?
[09:18] <liw> dunham, a) there's probably thousands of such packages in the archive already b) it's always good to politely ask upstream to fix that, since ambiguity in licensing does nobody good any good
[09:20] <dunham> liw: a) ok. So, in case upstream is not that responsive, can I suppose archive-admins would accept it? b) indeed! hopefully he will address my points and make everything easier
[09:20] <liw> dunham, I would expect the archive admins to accept it, yes
[09:21] <dunham> liw: ok, good. Thank you!
[09:23] <vise> Im compiling a set of files (around 30) and then linking them together using my own makefile. In order to speed up, I use -j2. But while it reaches the link stage, there are unfinished compilations which result in the linker not finding 1 or more object files. Is there some way I could tell make to do the linking after all the concurrent compilation jobs are over?
[09:24] <liw> vise, do you have the target of the link depend on all the object files?
[09:24] <liw> if so, make should do the right thing
[09:26] <vise> Okay! I have no dependencies for the link step.. Adding all the object files as dependencies there should do the job... Il try... tyvm
[09:27] <vise> works... ty @liw
[09:28] <liw> you're welcome
[09:30] <dholbach> kirkland: why do I have Private/THIS DIRECTORY HAS BEEN UNMOUNTED TO PROTECT YOUR DATA -- ...?
[09:33] <cjwatson> TheMuso: don't suppose you have an installation log from that system that was left without sudo access?
[09:33] <cjwatson> dholbach: run ecryptfs-mount-private?
[09:34] <cjwatson> I think that file has some useful contents, even
[09:34] <dholbach> cjwatson: I was just surprised - I never had that file before
[09:34] <cjwatson> I don't know why ecryptfs unmounted it, just explaining how to get things back :)
[09:34] <dholbach> right :)
[09:34] <dholbach> thanks
[09:37] <pitti> it's even supposed to be clickable in nautilus, isn't it?
[09:39] <dholbach> hum
[09:39] <dholbach> Jul  3 10:38:31 bert kernel: [   33.496877] Unable to allocate crypto cipher with name [aes]; rc = [-2]
[09:39] <dholbach> Jul  3 10:38:31 bert kernel: [   33.496882] Error attempting to initialize key TFM cipher with name = [aes]; rc = [-2]
[09:39] <dholbach> Jul  3 10:38:31 bert kernel: [   33.496885] Error attempting to initialize cipher with name = [aes] and key size = [16]; rc = [-2]
[09:39] <dholbach> Jul  3 10:38:31 bert kernel: [   33.496888] Error parsing options; rc = [-22]
[09:39] <dholbach> I have no idea what's going on here
[09:49] <dholbach> pitti, cjwatson, kirkland: nevermind, it was entirely my fault and is too embarassing to tell :)
[09:55] <pitti> dholbach: it was your dog, right?
[09:56] <dholbach> no, Murphy's far too clever :)
[09:56] <dholbach> ... to have an overeager clean up command running that deletes /lib/modules
[09:56] <dholbach> lalala
[09:56] <dholbach> anyway, it's fixed :)
[09:57] <pitti> :)
[09:57] <pitti> kees always does that for security
[09:57] <dholbach> and good to see that Ubuntu runs stable enough without it
[09:57]  * dholbach hugs y'all
[09:57] <pitti> thanks to the kernel team for building in the most important stuff :)
[09:57] <dholbach> that's fantastic, really
[09:57] <pitti> i915                  208744  3
[09:57] <pitti> drm                   192800  3 i915
[09:57] <pitti> those are really the only two you can't really do without
[09:58] <pitti> well, there's still wifi/sound, but if you are on eth, you indeed might not even notice
[10:03] <Keybuk> pitti: having all that in the initramfs is temporary
[10:03] <pitti> Keybuk: ok, so the fbcon issue is a bug then?
[10:03] <Keybuk> what's the issue?
[10:07] <pitti> Keybuk: try booting without an initramfs
[10:07] <Keybuk> right
[10:07] <pitti> as soon as udevadm trigger kicks in, and thus KMS gets active, screen goes black
[10:07] <Keybuk> though I suspect that's the most minor thing that goes wrong if you boot without an initramfs ;)
[10:07] <pitti> Jesse suspects that it's due to fbcon not being loaded
[10:07] <pitti> Keybuk: why? works fine
[10:07] <Keybuk> by stripping back the initramfs to only mounting the root filesystem and nothing else, we'll solve most of those in one fell swoop
[10:08] <Keybuk> pitti: there's a whole bunch of modprobes in the initramfs that aren't duplicated outside
[10:08] <Keybuk> fbcon is just one of them
[10:09] <pitti> Keybuk: so it seems we need to duplicate that fbcon modprobe after the udevadm trigger, to make VTs work if KMS doesnt' get flipped on in the initramfs?
[10:10] <Keybuk> no, we need to take the kms drivers back out of the initramfs
[10:11] <pitti> together with fbcon then?
[10:11] <Keybuk> right
[10:11] <pitti> I know that this worked a few weeks back, but something in the last few weeks has broken it
[10:22] <cjwatson> Keybuk: (optionally; they'll need to stay in in some cases)
[10:25]  * ogra wonders what mvo broke on strace to make it ftbfs on arm :P
[10:27] <ogra> hmmm linux/arm/syscallent.h contains funny code thats aparently supposed to make builds fail ...
[10:27] <ogra> #if SYS_socket_subcall != 400
[10:27] <ogra>  #error fix me
[10:27] <ogra> #endif
[10:27] <ogra> thats just mean
[10:27] <pitti> well, it doesn't say #error MUHAHAYOUSUCK
[10:28] <pitti> it looks intentional?
[10:28] <ogra> yes, apparently
[10:29] <ogra> but that means it will never build
[10:29]  * ogra tries to find out why it was added
[10:29] <Keybuk> cjwatson: the "need a password for the root filesystem" case only, surely?
[10:34] <ogra> aha !
[10:34] <cjwatson> Keybuk: yes
[10:35] <Keybuk> cjwatson: which has the nice property of not booting without an initramfs anyway, thus pitti's bug isn't relevant to that particular case <g>
[10:36] <cjwatson> quite :)
[10:36] <pitti> Keybuk: well, my bug is not really about "boot without initramfs", it's about when KMS gets flipped on
[10:36] <pitti> apparently "kms after fbcon" is the problem
[10:37] <Keybuk> yeah, not 100% sure how to handle fbcon yet
[10:37] <Keybuk> haven't done the necessary testing
[10:37] <Keybuk> I guess it's needed, otherwise you won't have a console
[10:37] <pitti> but booting without initramfs is a nice and easy way to test it temporarily
[10:37] <Keybuk> so we'd need to load it at some point
[10:37] <Keybuk> and then, as you say, does it need to be loaded before or after the kms module?
[10:37] <pitti> Keybuk: ah, it has no modaliases indeed
[10:37] <pitti> after
[10:38] <Keybuk> easiest way is probably going to be to just add "depend i915 fbcon" to module aliases somewhere
[10:38] <pitti> would it work to add it to /etc/modules?
[10:38] <pitti> is that run after udevadm trigger?
[10:38] <Keybuk> /etc/modules is before right now
[10:38] <pitti> oh, ok
[10:38] <Keybuk> and I tend to dislike adding hardcoded things there
[10:38] <pitti> *nod*
[10:39] <TheMuso> cjwatson: I do, I'll file a bug and throw them in there.
[10:42] <geser> could someone please unsubscribe u-m-s from bug #386428. subscribed the wrong team
[10:42] <pitti> geser: done
[10:44] <TheMuso> cjwatson: what package is responsible for setting user groups? I'll file it against that with a tarball of the install logs.
[10:45] <cjwatson> TheMuso: user-setup
[10:45] <TheMuso> ok thanks
[10:45] <cjwatson> TheMuso: if you could just attach syslog, please, not as a tarball
[10:45] <cjwatson> /var/log/installer/syslog - I'm unlikely to need anything else
[10:45] <TheMuso> oh ok
[10:46] <cjwatson> (tarball => extra faff)
[10:48] <TheMuso> yep
[10:55] <Laney> I just full-upgraded and now have a GDM greeter displaying which I cannot interact with in any way :(
[10:55] <Laney> didn't even restart...
[10:55] <TheMuso> cjwatson: bug 395082 filed
[10:56] <Laney> had to SSH in and kill it
[10:57] <pitti> meh "branch linked" spam
[11:57] <Keybuk> I've said this before, but one thing I love about Upstart is that in single user mode, I can run "start tty2" ;-)
[11:58] <ogra> pfft, excessive use of terminals
[11:59] <ogra> :)
[13:59] <Xhema> can someone please fill me in on one thing?  is there any requirement for people to publish changes to open office? is the PO files? is there any requirement for them to publish the sources?
[14:10] <pbn> Xhema: hi, better ask on #openoffice.org
[14:10] <Xhema> thanks
[14:13] <ogra> asac, my cursor keys and pgup/pgdn dont work on LP pages in shiretoko, is that LPs fault or shiretoko ?
[14:14] <ogra> hmm, the do work but only after some time
[14:23] <asac> ogra: i dont think its a 3.5+ dependent
[14:24] <asac> seems to be lower down the stack
[14:24] <ogra> hmm, k
[14:24] <asac> i will use 3.0 for a while today to see if it happens there too
[14:28] <ogra> but ou notice the same ?
[14:28] <ogra> *you
[14:32] <james_w> does anyone have a bug number for the "unable to mount external media in karmic bug"?
[14:33] <james_w> I thought I remembered someone saying it was a known bug the other day
[14:33] <james_w> people keep filing it against dbus because you get a DBus timeout error
[14:37] <pitti> james_w: "the" bug? works fine here
[14:37] <pitti> they might not have upgraded and don't have policykit-1-gnome yet?
[14:37] <Laney> I was seeing it with my iPod the other day, but I cleverly managed to lose it... :(
[14:37] <james_w> pitti: thanks, I'll ask
[14:37] <Laney> phone mounts fine though
[14:37] <ogra> works here as well
[14:37] <james_w> e.g. bug 393306
[14:37] <ogra> "to well" actually :)
[14:38] <Laney> is it .31? I heard that there may be problems there
[15:33] <amitk> liw: should binutils-static be in computer-janitor?
[15:36] <PPDP12> Hi, how do I get started developing with Ubuntu
[15:36] <PPDP12> Maybe like tutorials and stuff?
[15:37] <PPDP12> I have Experience with Microsoft Programming
[15:37] <ion> The topic would be a good start
[15:37] <PPDP12> Topic?
[15:38] <ogra>  /topic
[15:38] <PPDP12> O ok
[16:07] <ogra> uuuh, where does my fusa applet go with this upgrade ?
[16:09] <ion> gdm provides it now.
[16:10] <ogra> ah
[16:10] <ogra> i just saw it being deleted
[16:10] <ogra> as long as shutdown and reboot are still functional all is fine :)
[16:14] <ogra> err
[16:16] <ogra> now that was irritating ... ending up on a login screen in the middle of an upgrade
[16:16] <pitti> amitk: FYI, latest gnome-panel brings back shutdown options in system menu
[16:17]  * ogra trusts his system when it tells to need to reboot after upgrade and just does that 
[16:20] <ogra> eeek
[16:20] <ogra> so the gdm fusa forces me to look at mz face all the time
[16:20] <ogra> !
[16:21] <ogra> my
[16:21] <ogra> where is mz kbd mapping gone _
[16:21] <ogra> grr
[16:22] <pitti> ogra: https://bugs.launchpad.net/bugs/395103
[16:22] <pitti> most probably
[16:22] <ogra> ha
[16:22] <pitti> you can switch it back in gnome prefs
[16:22] <ogra> zep
[16:22] <ogra> err
[16:22] <ogra> yep
[16:22] <pitti> but really, US-layout is LOVE
[16:23] <ion> pitti: Amen, brother
[16:23] <Tm_T> is not ):
[16:23] <ogra> depends what zou are used to
[16:23] <pitti> well, if a German guy had invented LaTeX and vim, it wouldn't be
[16:23] <ion> pitti: I only use the fi layout for typing Finnish language.
[16:23] <pitti> (and C, and shell, etc.)
[16:24] <ogra> zzyyy
[16:24] <ogra> hmm
[16:25]  * pitti watches ogra practicing
[16:25] <ogra> doesnt switch on the fly
[16:25] <pitti> and enjoy shift/alt-gr less /, \, =, `, etc. :)
[16:30]  * ogra tries a re-login
[16:30] <geser> ogra: don't forget to reset your keyboard layout after the re-login
[16:31] <ogra> hmm, why does nothing happen if i click quit in fusa
[16:31] <ogra> oh, nice it automatically pops up behind all open windows ... how convenient
[16:32] <ogra> äöüyyzz
[16:32] <ogra> ha
[16:33] <Tm_T> å ø æ too
[16:34] <ion> ☺
[16:34] <pitti> ♥
[16:34] <pitti> ^ ubuntu-love key
[17:25] <slangasek> mathiaz: what things other than maintainer scripts do you possibly need that horrible shell library for?
[17:25] <slangasek> mathiaz: also, if it was provided separately it would also still need to be embedded in the maintainer scripts
[17:27] <mathiaz> slangasek: right - I was looking for a standard way to create a new db backend
[17:27] <mathiaz> slangasek: for example I was writting a ldap-based project and was looking into testing it
[17:28] <mathiaz> slangasek: in order to do so I needed to setup up my own ldap directory and started to rewrite all the code from the maintainer script in my testing suite
[17:28] <slangasek> hmm, is it that you're specifically looking for debconf prompting as part of it?
[17:28] <slangasek> otherwise I wouldn't think those scripts would make a very good base
[17:28] <mathiaz> slangasek: I don't think so.
[17:29] <mathiaz> slangasek: right - I was just looking for a generic way setup a directory
[17:29] <mathiaz> slangasek: one of the idea I had was to get something similar to the postgresql-common package.
[17:30] <mathiaz> slangasek: where you have generic commands to create a new postgresql cluster, etc...
[17:30]  * slangasek nods
[17:30] <mathiaz> slangasek: I also need something similar to setup the directory infrastructure we talked about at UDS.
[17:31] <mathiaz> slangasek: in that use case I need to load an additional schema to support MIT krb5
[17:31] <ogra> slangasek, arent you supposed to pretend 3 is 4 today too ? (go holidaying!)
[17:31] <mathiaz> slangasek: and load a default DIT
[17:32] <slangasek> you might be able to derive from scripts-common for this, but I definitely wouldn't install it unmodified given that everything in it is specifically targeted for maintainer scripts
[17:32] <mathiaz> slangasek: it seems that these are functions that are already done in the maintainer scripts and was looking into a way to provide the same functionality outside the package
[17:32] <mathiaz> slangasek: sure.
[17:33] <mathiaz> slangasek: so I looked at it yesterday and tried to implement the same thing with puppet
[17:33] <mathiaz> slangasek: I'm almost done now.
[17:34] <mathiaz> slangasek: I was able to implement most of the functionality in puppet recipes.
[17:34] <mathiaz> slangasek: thanks to the no_configuration debconf question which doesn't setup the package at all.
[17:35] <mathiaz> slangasek: it's an interesting concept though.
[17:35] <mathiaz> slangasek: and quite different from using shell/perl/python scripts
[17:36] <mathiaz> slangasek: and I don't really know how this could be integrated since it would add a dependency on puppet
[17:36] <slangasek> -
[17:36] <slangasek> yeah, not keen on that idea :)
[17:37] <mathiaz> slangasek: right. However providing most of the functionality from the script-common in a package would be good idea?
[17:38] <slangasek> seems a reasonable thing to do if that's what you need
[17:43] <ion> Puppet is quite painful when you need functionality it doesn’t already offer. If the recipe syntax were an actual scripting language with a Puppet DSL, everything would be so much nicer. For instance, AFAIU, you can’t do something as simple as ‘if <query augeas for something> then <tell augeas to do something> else <tell augeas to do something else>’ without implementing your own ‘if’ structure separately within the Puppet Augeas module.
[17:47] <mathiaz_> slangasek: hm - I got dropped from my wireless network
[17:48] <mathiaz_> slangasek: did you add something after: < slangasek> seems a reasonable thing to do if that's what you need
[17:51] <mathiaz_> ion: right - however that's the point of puppet: not having a scripting langage
[17:51] <mathiaz_> ion: the great strenght is to have a DSL which is a descriptive language, and not a imperative language to express your configuration
[17:51] <mathiaz_> ion: it requires a change of mindset though.
[17:52] <mathiaz_> ion: and I also have to admit that I don't have enough experience with puppet - you may raise a valid point though.
[17:52] <ion> I simply haven’t found a way (changing the mindset or not) to make the augeas module do changes based on the result of an augeas query without hacking the module itself.
[17:53] <mathiaz_> ion: right. Why not use templates instead of augeas then?
[17:54] <mathiaz_> ion: I think augeas is a great piece of technology but I'm not sure about it fits exactly in the big picture.
[17:54] <mathiaz_> ion: I think augeas is a great piece of technology but I'm not sure *where* it fits exactly in the big picture.
[17:54] <mathiaz_> ion: what is your use case for augeas within puppet?
[17:54] <ion> When i wanted an automatically updated list of IRC server IP addresses as a whitelist for identd, instead of just doing DNS queries in the recipe, i had to write a Puppet module to do the queries. And even that can only be used in some very specific places in a recipe, not as a generic construct.
[17:56] <ion> mathiaz: I’d like to query whether there’s a “* hard nproc <anything>” rule in /etc/security/limits.conf and either add “* hard nproc 400” or replace the existing rule with that.
[17:56] <ion> mathiaz: And the equivalent for * soft nproc
[17:57] <mathiaz_> ion: right - wouldn't writting a fact to cover this be a solution?
[17:57] <mathiaz_> ion: and then use a template to generate the correct line?
[17:58] <mathiaz_> ion: the idea here is to *not* base the condition on the content of the limits.conf but rather on some environment facts
[17:58] <mathiaz_> ion: (such as this is a developer workstation, http production server, etc...)
[18:01] <ion> mathiaz: Whether limits.conf is touched by the thing would depend on a puppet manifest, but what i meant was “update ‘* hard nproc’ to 400 if a value already exists; otherwise add ‘* hard nrpoc 400’”.
[18:01] <ion> mathiaz: I should investigate http://wiki.opscode.com/display/chef/Home, perhaps it’s a worthy replacement for me.
[18:02] <mathiaz_> ion: right. so what you wanna achieve in this use case is to have a line that says '* hard nrpoc 400' in limits.conf on every host managed by puppet?
[18:03] <mathiaz_> ion: chef looks also promising - however it's a much younger project.
[18:04] <mathiaz_> ion: the main difference with puppet is that it uses ruby to express the configuration rather than a custom DSL.
[18:05] <ion> mathiaz: I’d make it a module (‘class proc-limit { if augeas-something then augeas-something else augeas-something }’) and do ‘include proc-limit’ within certain nodes.
[18:05] <mathiaz_> ion: some people think it's a good thing, others prefer to have a declarative langage to express configuration.
[18:05] <ion> If i could use any Ruby constructs within the recipe, i could do the equivalent easily.
[18:05] <mathiaz_> ion: why don't you use a template instead of augeas? if every system needs to the same line it seems to be the best option.
[18:06] <ion> How about if another module also wants to touch limits.conf? The point of puppet modules is supposedly that they are reusable. :-P
[18:07] <ion> My puppet module for changing limits would supposedly be usable by anyone else with an OS that has limits.conf. Their OS packaging could even have preset rules that my puppet module shouldn’t touch.
[18:07] <mathiaz_> ion: right. It depends what the other module wanna do. The other module could call proc-limit definition with the relevant options.
[18:11] <mathiaz_> slangasek: so to summarize, having a slapd-common package seems to be a good idea?
[18:13] <mathiaz_> slangasek: the use case I'm trying to solve is the following: I need to be able to setup a slapd with a back-ldap + pcache overlay + nssov overlay with just one command
[18:14] <mathiaz_> slangasek: and another use case is: I need to be able to steup a slapd with a back-hdb + load schemas + load default DIT
[18:14] <mathiaz_> slangasek: all done from another package than slapd
[18:21]  * pitti uploads gdm with unbroken keyboard layout setting
[18:21]  * geser hugs pitti
[18:24] <pitti> cjwatson: ok, so two of the desktop RC bugs are fixed now \o/
[18:24] <pitti> cjwatson: thanks for fixing autologin in the installer
[18:24] <mathiaz_> pitti: hi - I'm looking at your postgresql-common package
[18:25] <mathiaz_> pitti: to get some inspiration for something similar to slapd
[18:25] <pitti> mathiaz: for multiple different instances?
[18:25] <mathiaz> pitti: hm - not necessarly
[18:25] <pitti> mathiaz: don't look too hard, it's Perl :)
[18:25] <mathiaz> pitti: it's true that sometimes we need to dump the bdb database
[18:26] <pitti> it was from the time when we didn't have Python in a defualt install
[18:26] <mathiaz> pitti: when we upgrade from one version of slapd to another one
[18:26] <pitti> (which is still current in Debian)
[18:26] <mathiaz> pitti: hm - that was one of my questions actually - if you had to restart now you'd code in python?
[18:27] <pitti> mathiaz: I guess yes; OTOH it keeps my Perl skills above "dead"
[18:27] <pitti> and you get pretty far with perl-base
[18:27] <mathiaz> pitti: right. I should talk to the debian maintainer team (ie slangasek)
[18:28] <mathiaz> pitti: most of the functionality is actually already written in shell script as it's embedded in the maintainer scripts
[18:29] <mathiaz> pitti: I'd like to extract these and provide them in a package (slapd-common) that could then be used by others things to setup different configuration of slapd.
[18:29] <mathiaz> pitti: that's why I'm looking at your postgresql-common package.
[18:36] <pitti> have a good weekend everyone!
[19:02] <mbana> why does my screen flash when i'm in fullscreen mode?
[19:05] <geofft> ogra, just out of curiosity, what caused your upgrade to make you end up at a login screen?
[19:08] <geser> probably gdm
[19:09] <sebner> it's definately gdm, here too ^^
[19:09] <sebner> huhu geser :D
[19:09] <geser> Hi sebner
[19:09]  * geser has a gdm without the keyboard layout bug :)
[19:10] <sebner> geser: /me in some hours too (updates) *hehe*
[19:26] <lfaraone> dtchen: hey, could you help with https://lists.ubuntu.com/archives/ubuntu-sugarteam/2009-July/001098.html by any chance?
[20:49] <mneptok> lifeless: woohaa NZ! - http://www.youtube.co.uk/watch?v=7-Mq9HAE62Y
[20:55] <ion> Great, gdm killed the user session on upgrade.
[20:58] <c_korn> why did gdm remove fast-user-switch-applet?
[21:01] <ion> It provides it.
[21:03] <c_korn> eh, I don't see any applet
[22:03] <geser> c_korn: the fusa changes didn't yet got updated for the new gdm (as far as the changelog tells)
[22:03] <ogra> yeah
[22:03] <ogra> and the gdm fusa thats replacing it is just weird
[22:04]  * ogra would really prefer if it wouldnt use his picture ... its strange to look at yourself on the panel all the time
[22:57] <Sarvatt> gdm restarting the gnome session might be a nasty surprise for someone upgrading from jaunty :D
[23:03] <Ng> Sarvatt: mine didn't restart, it switched me to a login screen, but my session was still running
[23:04] <Ng> but it's still not ideal ;)
[23:04] <Ng> imho we're always going to have problems trying to upgrade significant core compoenents of a running session
[23:05] <Sarvatt> oh nice, mine completely restarted which i imagine would stink because it'd close the installer. i'm in the middle of upgrading an ibook from hardy to karmic so i'll see how that goes in a few hours :D
[23:06] <Ng> release upgrades seem like highly suitable fodder for the "install updates on shutdown" idea :)
[23:08] <Sarvatt> wish i could have just installed karmic directly, but hardy is the newest release i could get to install on this thing. of course i found a way to do it on intrepid mid hardy install though
[23:38] <james_w> cjwatson: thanks for your post on subprocess_setup, fta2 just reported a bug caused by that and I never would have found the solution otherwise
[23:46] <YokoZar> So what happened to gnometris in karmic?
[23:49] <ion> On that note, why is there a “FreeCell Solitaire”, which seems to be just “AisleRiot Solitaire” in Freecell mode and without the mode selector? :-)
[23:50] <YokoZar> ion: because Aislerot is very unhelpful at helping you find other forms of solitaire to play
[23:50] <YokoZar> ion: select game doesn't give a preview or explanation, for instance
[23:54] <lifeless> mneptok: classic