/srv/irclogs.ubuntu.com/2008/03/18/#ubuntu-devel.txt

cjwatson(zapping the '^' check and copying the entire string one-by-one breaks too)00:00
keescookthat leaves the "^" though, right?00:00
cjwatsonright00:01
cjwatsonoh, hmm, I wonder if 'put' needs to be given a pointer type00:01
cjwatsonI bet it's type-sensitive ...00:01
keescookcjwatson: it's performing a forall against menuitemmap.text while changing the contents of menuitemmap.text...00:19
cjwatsonthe forall is against the string returned by translate, actually00:19
keescooker, or not? the forall is the translation -- yeah00:20
cjwatsonI'm trying to replace it by for though in case forall is the problem00:20
cjwatsonhmm, that sentence could have used some punctuation00:20
cjwatsonnope, using a for loop instead doesn't help00:20
keescookit was easier to parse than gfxboot code ;)00:20
cjwatsoncoffee00:21
slangasek"by for though in case" - who needs punctuation, just declare it to be a new dialect of python00:23
cjwatsonkeescook: as far as I can see, any attempt to write into menuitemmap.text using 'put' breaks it00:43
mathiazjdstrand: re bug  202706 - did you forward the patch to Debian ?00:44
ubotuLaunchpad bug 202706 in mysql-dfsg-5.0 "MySQL 5.0.51: ORDER BY not working with GROUP BY" [High,Fix released] https://launchpad.net/bugs/20270600:44
jdstrandmathiaz: not yet, plan to00:44
keescookcjwatson: odd, very odd00:45
keescookcjwatson: are there other cases of string-putting working?00:47
cjwatsonkeescook: loads00:48
cjwatsonit's used all over the place00:48
keescookcjwatson: what happens if you drop "translate" and leave everything else?00:49
ion_FilenameTypeSizeUsedPriority00:49
ion_/dev/compcache0                         partition304162238010000:49
keescookI don't really see why the source of the put should matter, though00:49
ion_Uh, sorry.00:49
cjwatsonkeescook: I think that worked, would have to check again00:49
emgentheya people00:54
keescookcjwatson: does forall include the terminating null?00:54
keescookoh, nm, it's nulled at the end.  hrmf00:55
cjwatsonshouldn't00:55
cjwatsongfxboot is at 4.0.1 upstream00:55
cjwatsonyarrr00:55
cjwatsonthough no changes in bincode.asm so I guess that isn't relevant00:58
mathiazIs there a way I can figure out which packages depends on mysql-server (excluding Recommends and Suggests) ?00:59
jdstrandmathiaz: apt-cache rdepends?01:00
jdstrand(don't know how it deals with Recommends and Suggests off-hand)01:01
keescookcjwatson: interesting update to /split def01:02
mathiazjdstrand: hum - it includes Recommends and Suggests01:03
cjwatsonkeescook: hmm, where?01:03
keescookin the upstream system.inc01:04
keescook+  % split does not work if str1 is in a special memory region (where01:05
keescook+  % 'cvp length' does not work). So we dup it first.01:05
keescookcjwatson: you want the full diff pasted somewhere?01:05
cjwatsonkeescook: no, I have it. I think that might be referring to string literals in the bytecode01:06
keescookyeah... ow my head01:06
cjwatsonkeescook: stripping out a bunch of keymaps also works around it, as Riddell suggested earlier on #ubuntu-release01:07
cjwatsonI think we're overflowing gfxboot's available memory, perhaps01:07
VolansHi, lool can I ask you a thing about the automount of a crypted partition of a USB HD?01:07
cjwatsonI might be able to fix it by stripping out some unused strings01:08
jdstrandmathiaz: apt-rdepends?01:08
jdstrandmathiaz: I think it goes too far though01:10
* Riddell hugs colin and kees and goes to bed01:10
keescookI don't understand why reducing the memory footprint would solve it -- both strcpy and put operate on the same already-malloc'd region.  hmpf01:11
cjwatsonyeah, I know01:12
mathiazjdstrand: awesome - that's what I was looking for01:12
cjwatsonthis is all very, very weird01:12
cjwatsonif I rip out dia_install.inc and everything related to it (since it was unused in Ubuntu anyway), it all works01:13
keescookcjwatson: the bootloader defines available memory? it's not yet obvious to me where that happens.01:13
cjwatsonerr. somewhere in syslinux and/or gfxboot, I'm not entirely certain where01:14
keescooklooks liek gfxboot reads it from a "sysconfig" area of the boot loader01:14
keescookI've haven't looked at the memory areas yet, but maybe the stack is crashing into malloc? strcpy is smaller than forall01:15
keescookI have no idea where code vs stack vs data vs malloc is stored yet :{01:16
cjwatsonit could be that, yeah01:16
cjwatson 58 files changed, 3 insertions(+), 3042 deletions(-)01:17
cjwatsonI have to say I'm happy with that kind of change on principle01:17
keescookhah -- what's that? 4.0 release?01:17
cjwatsonno, ripping all this stuff out01:17
keescookoh haha01:17
cjwatsonincluding translations01:18
keescookeek01:18
keescookcjwatson: so, what produces the syslinux config that launches gfxboot?01:18
cjwatsonhttp://people.ubuntu.com/~cjwatson/bzr/debian-cd/ubuntu/01:18
keescookcjwatson: my brain is melted enough for one day -- sounds like we have a viable workaround.  :P01:27
Hobbseemmmm....brains....01:27
cjwatsonkeescook: me too01:27
keescook(I don't see any hard-coded memory limits in config files -- perhaps isolinux has one internally, but I'm going to stop looking)01:27
keescookcjwatson: I feel kind of up to speed on gfxboot now, though.  ;)01:27
cjwatsonyay01:28
cjwatsona victim01:28
cjwatson:-)01:28
keescookhehe01:28
cjwatsonthe discussion was useful even if the solution was a lot cruder than we might have hoped01:28
cjwatsonwe might need those associative memories again in the future ...01:28
keescookyup, I figure I might actually be useful Next Time.  :)01:29
cjwatsonyou were :)01:30
keescookheh01:30
bryceslangasek: aforementioned fix posted to bugs 197645 and 199960 for validation of the fix before uploading.02:21
ubotuLaunchpad bug 197645 in gnome-settings-daemon "gnome-settings-daemon crashed with SIGSEGV in rw_screen_list_outputs()" [Medium,In progress] https://launchpad.net/bugs/19764502:21
ubotuLaunchpad bug 199960 in gnome-settings-daemon "error starting GNOME Settings Daemon" [High,Confirmed] https://launchpad.net/bugs/19996002:21
vlowtherbug #198808, what is up with that?  Who though it would be a good idea to pass all the video quirks unless specifically told not to?02:25
=== lamont` is now known as lamont
Amaranthhmm, seems i spend 10 seconds in initramfs on boot03:09
jdonglol you're really motivated to fix that, huh? :)03:09
RAOFAmaranth: Fujitsu was also seeing something like that (although 30 sec or something)03:09
Amaranthjdong: was watching a movie :)03:09
Amaranthjdong: and i want my < 20 second boot :D03:09
jdongAmaranth: lol03:09
jdongeveryone wants my 19s boot :)03:09
Amaranthjdong: not fair that you get 19s with 19MB/s and I get 32s with 47MB/s :P03:10
Amaranthi can see why you have me start udev on starting readahead though, i guess I have the IO to spare :P03:11
pwnguinis this something that shows up before bootchart begins loggin?03:12
Amaranthno03:12
Amaranthhttp://www.realistanew.com/random/hardy-20080317-3.png03:13
Amaranthjdong: also, you start a lot less stuff03:13
pwnguinoh wow03:13
Amaranthpwnguin: yeah, it's upstart scripts calling init scripts03:13
Amaranthbut in parallel and event driven03:13
Amaranthjdong did it03:13
pwnguinwhat?03:14
pwnguinmade it tank?03:14
jdongAmaranth: nice, your real boot is 22s if initramfs didn't suck.03:14
Amaranthoh, the 10 second stall isn't jdong03:14
Amaranthi'm pretty sure it's been doing that all along03:14
jdongpwnguin: the solid blue from 23 to 30 is all me :)03:14
pwnguinbut look at that green line03:14
pwnguinthat's all seek =(03:14
Amaranthjdong: yeah, i think if i trim down what i start and try that defragment thing i can get 12-15 seconds :)03:14
jdongAmaranth: try an easy cheap defrag replacement first03:15
Amaranthpwnguin: the whole boot is seek except for readahead03:15
pwnguineven readahead03:15
Amaranththat's why readahead is there :P03:15
Amarantheh? during readahead I peak at 47MB/s03:15
jdongAmaranth: tar up all the files in /etc/readahead/boot03:15
jdongAmaranth: then unpack it back into /03:15
jdongactually cpio probably works better with a list of files ;-)03:15
Amaranththat's...sneaky03:16
pwnguinheh03:16
jdongAmaranth: for me that halved readahead time.03:16
pwnguinif it works03:16
jdongAmaranth: cheaper than defragging03:16
Amaranthi see why it works03:16
jdonglol it is really sneaky and cheap :)03:16
Amaranthgimme a script or something? :)03:16
pwnguinxargs?03:16
* Amaranth lazy03:17
jdongAmaranth: lol one sec03:17
* jdong whips out the command he used03:17
jdongAmaranth: cat /etc/readahead/boot | xargs sudo tar cvpf /tmp/init.tar03:17
ion_Wasn't someone working on actually rearranging blocks on the partition based on usage patterns? That would entirely remove the need for readahead, i'd think.03:18
jdongAmaranth: sudo tar xvpf /tmp/init.tar -C /03:18
Amaranthtar cf foo `cat /etc/readahead/boot | xargs`03:18
Amaranthalright, i'll use yours03:18
jdonglol yours works too :)03:18
jdongI'd use p with tar though03:18
jdongyou never know what permissions some of those files might need :)03:18
pwnguinthen pipe that back out to tar ;)03:18
jdongpwnguin: LOL that sounds like... fun :D03:19
pwnguinjdong: so how expensive is ext3 defrag?03:19
pwnguinim not even aware of an application that does that03:19
jdongpwnguin: e2defrag03:19
jdongpsusi patched it to work with ext303:19
jdongit works... just is a lot more destructive than this03:19
jdongthis operation is more or less atomic03:19
pwnguindefine destructive03:20
Amaranthwow that went way too fast03:20
jdongpwnguin: lose power in the middle of the process03:20
pwnguinah03:20
jdongAmaranth: yeah it's instantaneous and yields a measurable performance for me03:20
jdongAmaranth: btw I'm really impressed by how deep your bootchart is03:20
jdongthat's amazing03:20
Amaranthok that was scary03:21
Amaranthdeep?03:21
jdongmust be at least 50 processes in parallel03:21
Amaranthah03:21
pwnguintall03:21
Amaranthyeah, my CPUs freak out for a bit :P03:21
Amaranthbrb, seeing if i destroyed anything03:21
jdong:)03:21
pwnguinnow for an after picture03:21
pwnguinone hopes that green line shows considerable improvement03:22
jdongindeed03:22
jdongmy green line is still rpetty abysmal03:22
jdongpwnguin: http://jdong.mit.edu/~jdong/macbook/bootchart-upstart.png03:22
jdongactuallly03:22
jdonghttp://jdong.mit.edu/~jdong/macbook/hardy-upstart-defragged.png03:22
jdongthat's after the "defrag"03:22
pwnguinnice03:23
jdongpwnguin: yeah, at this point unless readahead or modprobe (udevtrigger) drastically improves, I think I'm at my peak boot speed03:23
pwnguinlike someone else said, you could potentially group files together based on access patterns03:24
pwnguinie, config files at boot grouped togethre03:24
jdongpwnguin: I generated a dependency diagram of my upstart boot at http://jdong.mit.edu/~jdong/macbook/depchart/upstart-dep.png03:24
jdongpwnguin: it's nicely massively parallel03:24
Amaranthwhee03:24
Amaranthjdong: http://www.realistanew.com/random/hardy-20080317-4.png03:25
Amaranthpeak went down but throughput seems to have gone up overall03:25
Amaranthonly cut a second off though03:26
pwnguinreadahead's much shorter that time03:26
jdongshows some improvement03:26
pwnguintheres a wierd exe though03:26
Amaranthi think that's upstart03:26
jdonghow the hell is your readhead so short?03:26
* jdong cries03:26
pwnguinbecause it's faster?03:26
jdongAmaranth: move up udev03:26
* lamont wonders why it is that, when I have "shadow: compat db" in nsswitch.conf, the gnome screen lock needs read access to /var/lib/misc/shadow.db before I can unlock the screen.03:26
Amaranthjdong: i don't have the sad excuse of a harddrive that apple puts in their laptops03:27
pwnguinheh03:27
pwnguin12? ouch03:27
jdongAmaranth: /etc/event.d/udev, "start on starting readahead"03:27
Amaranthjdong: this HD was the fastest you could get for a laptop until a year after i bought it03:27
jdongAmaranth: try that for size. That should reduce the readahead bottleneck03:27
jdong(hopefully no races :D)03:27
pwnguini should probably rerun profile03:28
jdongAmaranth: same for lrm-manager03:28
Amaranthyeah but i still can't go fast until i fix initramfs03:28
* jdong looks at dep chart03:28
Amaranthwhat could make it slow?03:28
pwnguinreiserfs?03:29
Amaranth/etc/event.d/udev is already "start on starting readahead"03:29
Amaranthyou told me that one before03:29
Amaranthpwnguin: no, ext303:29
jdongAmaranth: oh, do the same for lrm-manager03:29
Amaranthalright, done03:29
Amaranthi wish this was a pdf so i could search through it03:30
jdongAmaranth: are you booting verbosely?03:30
Amaranthyeah03:30
Amaranthdon't do that?03:30
jdongAmaranth: you should be able to see where it hangs up in initramfs then03:30
jdongright?03:30
jdongI'm guessing probably the probing of some sort of hardware, either USB or disk03:30
jdonginitramfs runs a udevtrigger IIRC, which could be costly03:31
Amaranthit sits at "Starting up" then "Loading", then it seems to start for a little bit at setting up cfq03:31
jdongyou might win by removing all USB stuff from initramfs03:31
Amaranthbut it does sit for a second or two on "Starting up"03:31
jdongAmaranth: ok, I *DARE* you to....03:31
jdongAmaranth: http://jdong.mit.edu/~jdong/macbook/event.d.new.tar.gz03:32
Amaranthmy keyboard and touchpad are ps/2 so... :P03:32
Amaranthwhat is that?03:32
jdongAmaranth: diff that against your event.d03:32
jdongAmaranth: it's my updated event.d after shaving 2 secs off my boot03:32
jdongAmaranth: it sets the IO scheduler to AIO and disabled pdflush during boot03:32
jdongAmaranth: btw, you're responsible in rc.local to revert those two changes :D03:32
jdongsee event.d/pdflush03:33
jdongAmaranth: that should further reduce seek and write pressure during boot03:33
jdongand HERE, we cross into the realm of voodoo boot magic03:33
Amaranthyou changed udevtrigger?03:33
jdongAmaranth: I did?03:33
jdongAmaranth: my memory kind of fails03:34
Amaranth-start on stopped lrm-manager03:34
Amaranth+start on started udev03:34
jdongAmaranth: oh, I wanted to trigger a bit earler as I didn't have any LRM devices03:34
Amaranthah03:34
Amaranthbut i do :P03:34
jdongAmaranth: yeah so you shouldn't do that :D03:34
Amaranthgimme your rc.local :P03:35
jdongAmaranth: my rc.local is slightly scary :)03:36
Amaranthis your change in load-modules the same thing?03:36
Amaranth+start on stopped udevtrigger03:36
ScottK2jdong: As long as you don't say worksforme and then insist everyone should do it this way.03:36
jdongAmaranth: right, I think load-modules should start on udev instead.03:37
jdongAmaranth: http://jdong.mit.edu/~jdong/macbook/rc.local03:38
jdongScottK2: lol, this is hack central, anyone who dares try this is at minimum insane :D03:38
jdongAmaranth: useful stuff are the last line and the 2nd line03:38
jdongAmaranth: the rest of hacks to save about 2W of power on my system03:38
jdongand note the modprobe usbhid for when I couldn't get udev started correctly :D03:39
* jdong puts on his hack shame hat03:39
Amaranthjdong: should it really be initctl emit shutdown in rc6?03:39
emgentheya03:39
jdongAmaranth: well, shutdown and reboot meant the same thing when I just wanted Pulse to save mixer levels.03:41
Amaranthhehe03:41
jdong:)03:41
Amaranthi don't ever reboot anyway03:41
jdongAmaranth: good, I'm not sure rebooting works all that gently03:42
jdongAmaranth: it's kind of the extreme end of the "faster shutdown" spec from 2 releases ago03:42
jdong:D03:42
Amaranthyeah, i see lots of devmapper errors on shutdown03:42
Amaranthbecause stuff goes away in the wrong order03:42
Amaranthbut screw it, it's shutdown :P03:42
jdong:)03:43
jdonggood enough, tear things down, sync, umount, good neough03:43
jdongAmaranth: since you're conveniently syndicated on planet, you should probably do a quick blogpost once you're done :D03:43
Amaranthuh oh03:44
Amaranthi knew there was a catch03:44
jdong:)03:44
Amaranthbrb, i hope03:45
pwnguinheh03:45
jdonghe didn't have it nearly as bad as me03:45
pwnguinadventurous to do that one one machine03:45
jdonglet me  tell you, screwing up udev mount order on a USB keyboard system is NOT fun03:45
pwnguinhaha03:46
pwnguinyou mean apple notebooks dont have a serial debugger port?03:46
jdonggasp :)03:46
jdongpwnguin: I'm lucky to have an optical drive and a mouse button.03:46
pwnguinbut just the one button, no more03:47
jdongpwnguin: watch multitouch replace that 6 months from now.03:47
pwnguin?03:48
jdongpwnguin: I doubt trackpads will have buttons at apple's rate03:48
jdongeveryhting's gonna be encapsulated in touch and drag type gestures03:48
pwnguinlike a multitouch screen, or the trackpad just gets even smarter03:48
jdongthe trackpad getting smarter03:48
travisjdong: ok, something went bad03:50
jdongeep03:50
=== travis is now known as Amaranth
AmaranthI got a desktop but I've got three sleeps running and got no bootchart03:50
jdonghmm03:51
jdongAmaranth: is rc.local hanging?03:51
jdongAmaranth: initctl status, look for anything in starting phase03:51
Amaranthjdong: initctl: missing job name03:52
jdongAmaranth: sorry, list.03:52
Amaranthgrep says nothing03:53
Amaranthoh, wait03:53
Amaranthi've got a bunch of stuff in 'waiting' but i think some of it is supposed to be there and some of it isn't03:54
Amaranthoh, no, that's all for on stop03:54
jdongAmaranth: what's bootchart's status?03:54
pwnguinman. i have this five second "resume" crap03:54
Amaranthjdong: waiting :/03:55
jdongAmaranth: follow up the chain: http://jdong.mit.edu/~jdong/macbook/depchart/upstart-dep.png03:56
jdongAmaranth: usplash->gdm->acpi_support->laptop_mode->rc_local->stop_bootchart03:56
Amaranthjdong: http://rafb.net/p/ZelU1G27.html03:56
Amaranthit's waiting for gdm? that doesn't make sense03:57
jdongAmaranth: no stopped is the normal state for that03:57
Amaranthi should probably go through here and remove all the things i don't have init scripts for03:57
jdongAmaranth: those are all "script / end script" relationships, which enter stopped/waiting once they finish03:57
Amaranthoh, i did have a typo in my rc.local03:57
Amaranthcfg instead of cfq03:58
Amaranth:P03:58
jdongAmaranth: well that'll do it :)03:58
jdongAmaranth: yay set -e.03:58
Amaranthso those jobs should be started to not need ok03:58
jdongAmaranth: agreed03:58
Amarantherr, i hope that made sense :P03:58
jdongAmaranth: try "start on starting shutdown" for size :)03:58
Amaranthstop-bootchart should not depend on rc.local working03:59
jdongstratus: right, it should just depend on rc.local stopping03:59
jdongerr, wrong person entirely03:59
Amaranthwait, waht?03:59
Amaranthstart on starting shutdown? for what?03:59
jdongAmaranth: for an oxymoronic statement03:59
jdongAmaranth: it was quite awkward writing rules that sounded like that :)03:59
jdongAmaranth: that's where upstart's UI falls03:59
jdong(and yeah blame it on PEBKAC)04:00
Amaranthalright brb again04:00
jdonggood luck04:00
pwnguinhmm04:02
Amaranthjdong: man you're killing me here: http://www.reaistanew.com/random/hardy-20080317-5.png04:05
pwnguinhttp://people.cis.ksu.edu/~jld5445/after-readahead-tar.png04:05
pwnguinAmaranth: ye olde 40404:05
Amaranthhttp://www.realistanew.com/random/hardy-20080317-5.png04:05
jdongAmaranth: hmm04:05
Amaranthpwnguin: you did the upstart thing too?04:06
pwnguinAmaranth: no04:06
Amaranthwow04:06
Amaranthoh you start almost nothing04:06
jdongAmaranth: so perhaps we should concentrate all the modprobing together instead, and start readahead after that's done04:06
pwnguini start X and gdm04:06
jdongAmaranth: it seems like hwenever we do modprobing everyhing else grinds to a halt04:07
Amaranthjdong: sounds good to me04:07
pwnguinAmaranth: what am i not starting/04:07
Amaranthalthough i should really take some time to figure out where my initramfs bug is04:07
jdongAmaranth: so I guess move udev*, lrm load-modules to "start on stopped mount-kernel-filesystems ok"04:07
Amaranthpwnguin: i dunno, i seem to start about twice as many things04:08
jdongAmaranth: and move readahead to the last thing in the module loading chain.04:08
pwnguinAmaranth: keep in mind that bootcharts elide short lived processes04:08
pwnguinmaybe my computer's fast enough that things don't show up in the charts ;)04:08
jdongpwnguin: Amaranth seems to have a larger bluetooth stack, ntfs mount, and firestarter/iptables04:08
pwnguinthose are all valid points ;)04:08
pwnguinntfs in particular would be painful04:08
Amaranthyeah, firestarter is for connection sharing, ntfs can go away04:09
Amaranthand bluetooth? wtf04:09
jdongpwnguin: your blotchy stuff towards the end of bootup looks like could be optimized.04:09
pwnguinjdong: im sure it can be04:09
jdongpwnguin: it would seem like resources are heavily underutilized the last quarter of your boot04:09
jdongpwnguin: so.. either join the dark side or creatively use ampersands :)04:10
pwnguinjdong: i agree, but im more curious about the first quarter where nothing's going on and a process called resume is sitting there04:10
jdongpwnguin: yeah that's annoying too04:10
* pwnguin should drop bluetooth04:12
Amaranthdon't suppose you guys have any idea where to poke to figure out initramfs troubles04:14
wasabi/usr/share/initramfs-tools04:14
wasabior -utils04:14
wasabii forget. ;)04:14
pwnguini recall someone saying there was an interactive boot version of initramfs stuff04:14
Amaranthi think thinking more "what would likely cause the stall" :P04:14
jdongpwnguin: break={init, top, premount, mount, postmount, bottom, magic, other voodoo}04:15
wasabiAhh. Yes. All that stuff.04:16
wasabiWhen I'm debugging the initramfs I braek premount usually.04:16
wasabiThen just step through it manually.04:16
wasabiWhich is a bit hard with such a crappy shell.04:16
pwnguinhmm. well, just touching rc to use CONCURRENCY=shell seems to have broken hal04:24
pwnguinwhich is a shame, since it went about 5 seconds faster04:24
jdongpwnguin: haha see ya still need those dependency things :)04:24
pwnguinit was quick enough to undo :P04:25
pwnguinhmm04:26
pwnguin"/scripts/local-premount/resume: /scripts/local/premount/resume: 57: log_begin_msg: not found"04:26
jdongpwnguin: I wonder if we can put some sort of mini-upstart in initramfs :)04:27
pwnguinjust use make ;)04:27
pwnguinim not even sure what that message means, or which file to look for line 57 on =(04:28
wasabijdong: it's been mentioned a few times that I remember04:29
wasabiI'd like to see upstart used for desktop session stuff.04:29
jdongpwnguin: /usr/share/initramfs-tools/scripts04:29
jdongwasabi: well right now I'm wasting over 5s in bootup because udevtrigger is called once in initramfs and once in mount04:30
jdongwasabi: I'd like to see some of that duplication removed, as initramfs is nearly 1/3 of my bootup right now04:30
wasabiudevtrigger is async though, and does not reload modules... which is the slow part. No?04:31
jdongwasabi: seems like things that close to touching the kernel aren't async04:32
jdongwasabi: and right, it doesn't reload modules but it does waste time that could be spent loading the rest of the modules :)04:32
jdongwasabi: maybe instead I need a really big initramfs with all my modules :)04:32
wasabiHehe.04:32
wasabiDoes it really waste time? I find that odd.04:33
wasabiI'd think it'd be nearly instant to finish it's work after it's already run once.04:33
Amaranthjdong: http://www.realistanew.com/random/hardy-20080317-9.png04:34
Amaranthwe seem to be doing the same amount of work in the same amount of time just in a different order :P04:34
jdongAmaranth: I guess that's a good sign of hitting a bottleneck :)04:34
Amaranthhehe, yeah04:35
jdongAmaranth: that bootchart would be SO sexy if the first 11s were stripped.04:35
Amaranthso to get faster i just need to trim down services and fix initramfs04:35
Amaranthi wonder if bootchart is the one breaking my initramfs :P04:36
jdongAmaranth: lol that would be funny04:36
jdongAmaranth: have you tried what happens if you zap out readahead altogether? Does your disk seek fast enough ? :)04:36
Amaranthbecause i don't remember whether or not it paused this long before before starting usplash04:36
Amaranthhmm04:36
jdongAmaranth: easiest way to test would be to move aside /etc/readahead/boot and replace it with a blank.04:36
Amaranthor comment out the exec04:37
jdongAmaranth: replace it with exec true :)04:37
Amaranthjdong: yes, my HD can seek fast enough :P04:44
jdongreally?04:44
Amaranthjdong: 36 seconds04:45
jdongAmaranth: slight loss though not too bad04:45
Amaranthjdong: same location, -10 instead of -904:45
Amaranthi dunno how else to optimize this04:46
Amaranthwhat i should do is start from a clean install04:46
Amaranthi've got a lot of cruft i'm too lazy to clear out04:46
Amarantheh, i never boot anyway04:49
* Amaranth leaves if for now04:49
StevenKNever, you say?04:49
Amaranthsince 2.6.24 suspend is _really_ reliable04:50
* StevenK removes the battery from Amaranth's laptop and then cuts the power04:50
Amaranthi only reboot when i get a kernel upgrade or something04:50
nxvlis there any way to tell a .desktop file to use gkdu or kdesu depending on if it is KDE or GNOME?05:09
RAOFYou could ship 2 desktop files, one with OnlyShowInGnome, the other OnlyShowInKDE?05:10
pwnguinwell i found the problem with resume05:20
pwnguin# Default delay is 5s05:20
pwnguinwhich leads into an interesting question: what's the point of uuids if they change between releases?05:22
nxvlRAOF: and where should i add OnlyShowInGnome? on the desktop file?05:27
RAOFnxvl: It's actually a desktop file field, IIRC.  OnlyShowIn=GNOME05:31
AmaranthI would to NotShowIn=KDE; for the gksudo one and OnlyShowIn=KDE; for the kdesu one05:37
Amaranthbecause you're more likely to have gksudo (XFCE)05:37
Amaranths/to/do/05:37
pwnguinhmm05:46
pwnguinif [ "x${resume}" = "x" ]; then05:46
pwnguini presume this means to check if resume has something to it05:47
nxvlRAOF: thanx05:48
dholbachgood morning05:49
ion_good night05:50
pwnguininteresting05:59
pwnguinGoogle announced the summer of code organizations06:00
RAOFpwnguin: Do expound, dear chap.06:00
pwnguini dont see ubuntu on the list06:00
pwnguinhttp://code.google.com/soc/2008/06:00
Amaranthinteresting06:09
pwnguinsee?06:09
Amaranthbut yay, Xorg got accepted this year06:10
pwnguinwell that's good06:10
dmbno ubuntu SOC?06:10
pwnguini asked a few weeks ago06:10
pwnguinand someone pointed to brainstorm and that was about the end of the conversation06:11
pwnguinwell, at least debian's on there06:11
dmbubuntu is normally in it06:12
dmbthats weird06:12
* pwnguin wonders if ubuntu even applied06:12
dmbtheres no good debian ideas06:12
pwnguindebgraph looks good06:13
pwnguini thought i saw something like that done already though06:13
Amaranthso much for telling people to propose compiz ideas to ubuntu for soc :P06:13
pwnguinfrom now on i think i'll just put questions on the council agendas06:14
superm1_asac, most certainly didn't come up nicely from suspend06:15
superm1_but i dont know that's regressible for me.  i had video related suspend problems previously (that appear to be better now).  I haven't suspended for months due to them06:15
=== superm1_ is now known as superm1
RAOFMan.  I want to do the liboiljit GSoC project :)06:17
Amaranthliboiljit?06:17
RAOFhttp://gstreamer.freedesktop.org/wiki/LiboilJIT06:18
RAOFTurn liboil into a small JIT'd runtime.06:18
RAOFCompiler writing is always fun.06:19
RAOF* for certain values of "fun".06:19
=== Arby_ is now known as Arby
Mirvbryce: hi. do you have that i18n-fixed screen resolution tool coming up? after beta I assume?07:00
bryceMirv, indeed I have the package posted and am just waiting on seb128 to upload it:  http://people.ubuntu.com/~bryce/Testing/XrandrGui/gnome-control-center_2.21.92-0ubuntu3_source.changes07:02
* Fujitsu is impressed to have found a page that crashes both webkit- and xulrunner-based browsers.07:04
=== hunger_t is now known as hunger
Mirvbryce: hmm, it seems seb128 has already uploaded version 2.22.0-0ubuntu1 without those patches on Mar 11th07:11
bryceMirv, oh yeah; I updated it today:  http://people.ubuntu.com/~bryce/Uploads/gnome-control-center_2.22.0-0ubuntu2_source.changes07:14
pittiGood morning07:17
TheMusoMorning pitti.07:18
* bryce waves to pitti07:19
ion_Good night07:19
Mirvbryce: ah, ok. great!07:20
warp10Good morning07:38
superm1bryce, gah, displayconfig-gtk's showing up in the "Other" menu all alone now.  I thought the goal of that last upload was to disable it in the menu system :)07:47
bryceyeah it was.  I removed all the categories for it07:47
brycedoes the desktop file need some other modification to get rid of it?07:48
superm1i'm not sure..07:49
superm1maybe just dont install the desktop file?07:49
superm1or at least install it somewhere that it doesnt get populated for now until you want it to show up07:49
* TheMuso checks the desktop file variable needed to hide an item.07:50
TheMusoHang on a sec.07:50
TheMusoNoDisplay=true should do the trick.07:51
TheMusoThats used in the gnome-orca package, as well as others to hide the desktop icon.07:51
bryceTheMuso: aha, thanks07:51
TheMusobryce: Welcome.07:51
superm1much more appropriate solution :)07:52
sorencjwatson: I just took a peek in http://people.ubuntu.com/~cjwatson/bzr/cdimage/mainline/etc/crontab to figure out when I could expect the new JeOS image to be built, but it's not listed. I could have sworn that it used to be?08:08
mdkeslangasek: was there a problem with the ubuntu-docs upload? I don't see it in Launchpad08:12
kagouhey seb128, i will redo login time test with the daily-live iso, to see impact of prefetch etc.08:14
seb128hello kagou, thanks08:14
=== \sh_away is now known as \sh
kagouthere are no daily(live) isos today ?08:38
asacsuperm1: maybe try ;)08:44
slangasekmdke: no problem, I just had an X crash and lost track of the fact that I hadn't pushed it through the approved queue yet08:49
pittierk09:11
pittiseb128: 'sudo users-admin', does that work for you? I can't do anything in it09:11
pittiseb128: I just noticed that on the current live system all PK stuff is broken due to that (also gnome-mount -vbd /dev/sda1, etc.)09:11
seb128pitti: admin tools use policykit and need a bus session to connect, etc, don't run those using sudo?09:12
pittiseb128: longer context: on the live CD we change PolicyKit.conf to grant all access to 'ubuntu'09:13
pittito avoid asking for an empty password09:13
seb128looks good09:13
seb128still you should not need sudo?09:13
pittiI konw09:14
pittiseb128: it just more or less emulates the situation09:14
seb128hum09:14
seb128I'm not sure I understand09:14
seb128you should be able to run users-admin09:14
pittiah, nevermind09:14
seb128and click on unlock09:14
pittiI get a parse error09:14
seb128and it should work09:14
pittiseb128: right, but ideally root wouldn't need to unlock09:14
pittiit should just work for root09:14
seb128the sudo case it different, you run with an user with has no dbus session available09:15
pittiand indeed root doesn't need to unlock09:15
pittihm, why does it need a session dbus?09:15
seb128not sure, does policykit use it for authentification?09:15
pittinevermind, I found the bug in casper09:15
seb128ok09:16
pittithere's an extra blank line at the start of PolicyKit.conf09:16
seb128still the question on why it needs a bus stands09:16
pittithus XML parsing error09:16
pittiseb128: right, I wonder about that, too09:16
pittislangasek: I need to upload a new casper to unbreak PolicyKit in the live system, sorry09:16
pitticjwatson: that'll include your committed-but-not-uploaded fix for bug 202699, I guess that's ok?09:17
ubotuLaunchpad bug 202699 in casper "[Hardy] casper-snapshot wrongly labels squashfs as cpio.gz" [Undecided,Fix committed] https://launchpad.net/bugs/20269909:17
seb128pitti: the session bus is used for policykit apparently09:17
seb128pitti:09:17
seb128message = dbus_message_new_method_call ("org.gnome.PolicyKit",09:17
seb128"/org/gnome/PolicyKit/Manager",09:17
seb128"org.gnome.PolicyKit.Manager",09:17
seb128"ShowDialog");09:17
cjwatsonpitti: should be, your call of course09:17
seb128dbus_connection_send_with_reply (priv->session_bus, message, &pending_call, -1);09:17
seb128pitti: ^09:17
cjwatsoncasper-snapshot is only ever run by hand so I can't see how it'd break anything09:18
pittiright, was just going to say09:18
pittiseb128: interesting09:18
slangasekpitti: ok, eta on that upload?09:18
pittislangasek: give me 5 minutes, so that I can test it09:18
slangasekok09:19
RAOFasac: Why is nspluginwrapper (Universe) maintained in a bzr repository only core-devs can push to?09:19
asacRAOF: i guess thats a bug :)09:20
RAOFasac: I've finally picked that merge-from-debian up.  Shall I push bzr to somewhere ubuntu-dev can write to? :)09:20
pittislangasek: uploaded09:20
asacRAOF: yes, please push to ubuntu-dev and change control accordingly09:21
slangasekpitti: cheers09:21
asacRAOF: can you try to resurrect the .debian branch as well?09:21
asace.g. so we can merge using bzr means?09:21
Fujitsuasac: Or you could just delete the -core from the owner.09:21
asaci can?09:21
asaclet me see09:21
RAOFasac: Not right now, but that'd be nice.09:21
* pitti does some test installs09:22
Fujitsuasac: `Change registrant' if you hadn't found it already.09:23
pittislangasek: I take it it gets a little late for new langpacks?09:25
asacFujitsu: is that field really used for permission control?09:25
asacFujitsu: ok i changed the author to just ubuntu-dev, but i don't think that this helps09:26
asaci think the easiest way is just to push that to -dev and let me know so i can abandon the other branch09:27
stgrabermoin09:27
asacFujitsu: oh. you were i right i guess09:28
asacno all is fine09:28
asacthanks09:28
Fujitsuasac: Yes, that controls it.09:28
slangasekpitti: yes :/09:28
Fujitsuthat's good09:28
Fujitsuarrrgh, x on the other hand is not good.09:28
asacdoing the same for the upstream and debian branch09:28
pittislangasek: ok, then I better upload fresh ones after the beta; the current CDs don't have a lot of langpacks anyway, right?09:29
asacbut someone should push at least a bzr revision with the adapted Vcs header09:29
slangasekpitti: yeah, currently we're down to around 5 languages due to lack of space09:30
pitti:(09:30
* cjwatson blinks at the last comment in bug 19904809:30
ubotuLaunchpad bug 199048 in partman-partitioning "swapoff before swapon" [Medium,In progress] https://launchpad.net/bugs/19904809:30
cjwatsontook me a moment to get the reference :-)09:30
* pwnguin guesses the karate kid09:31
* pitti wonders whether he'll ever get used to those new progress bar09:31
cjwatsonyes09:31
pittislangasek: casper is in unapproved, and my ssh is currently broken again; can I ask you to accept it?09:33
slangasekyes, looking09:33
pittislangasek: thanks09:34
james_wDoes the installer ever ask anything about the root account? (I don't know whether they used the alternate or not).09:36
cjwatsonjames_w: the alternate/server installer will ask if and only if you also use expert mode09:37
james_wcjwatson: thanks, I'll ask if they used the alternate then.09:37
cjwatson(otherwise known as "I am an installer developer and want lots of noise" mode)09:37
slangasekpitti, cjwatson: this casper change to rename the images according to the image type isn't going to break builds, is it?09:39
cjwatsoncasper-snapshot is only run by hand09:39
slangasekok09:39
cjwatsonnothing in the CD build process uses it09:40
stgraberslangasek: Is 20080318 good for early testing or is there another set of ISO to come in the next hours ? (gfxboot fix ?)09:42
slangasekstgraber: there'll be another set for the gfxboot fix09:46
slangasekthe Ubuntu ones should be ok for early testing despite being superseded later; the Kubuntu and Xubuntu ones are known not to boot09:46
slangasek(well, presumed not to boot, given the gfxboot breakage)09:47
pwnguini've been redirected to here with a questions, so I'll pose it: what on earth is the initramfs-tools resume script doing waiting five seconds for?09:47
stgraberoh, and daily-live is oversized on i38609:48
slangasekstgraber: ack; that was one of the things I was checking for with this run, ohwell...09:49
slangasekpwnguin: according to the source, it only sleeps that long if that's how long it takes your resume device to become available09:52
pwnguinslangasek: the trouble im having is that i didn't hibernate09:56
pwnguinand it's still going off09:56
sorencjwatson: Your tasksel fix for JeOS worked perfectly. Thanks very much!09:57
slangasekpwnguin: well, yes, that seems like a plausible bug to me; if it's waiting for a disk that will never exist, for instance09:59
pwnguinslangasek: hmm09:59
slangasekpwnguin: the script I'm looking at *implies* that this only happens if you've booted with a 'resume=' option10:00
pwnguinslangasek: thats the thing i cant figure out -- where is resume getting set from?10:00
pwnguinkernel          /boot/vmlinuz-2.6.24-12-generic root=UUID=4cd91330-31b1-4a9b-aa9f-0f8b0c729420 ro quiet splash10:01
slangasekpwnguin: afraid I don't know the answer to that10:03
TheMusopwnguin: It gets set from initramfs configs in /etc/mkinitramfs10:03
TheMusoerr /etc/initramfs-tools10:03
TheMusopwnguin: /etc/initramfs-tools/conf.d/resume to be exact.10:04
sorenpwnguin: /etc/initramfs-tools/conf.d/resume10:04
cjwatsonsoren: woo. I left the JeOS project bug task open, so I guess you can close that bit now10:04
pwnguini swear10:04
pwnguinUUIDS are worthless10:04
sorenpwnguin: Eh?10:04
pwnguinlooks like that's a uuid to a device that doesn't exist10:05
pwnguinnever did10:05
pwnguinor at least, doesnt anymore10:05
pwnguinthe other day, i discovered that swap was using the wrong uuid, so i had no swap10:06
pwnguinerr, fstab was using the wrong uuid for swap. dont know why10:06
cjwatsonwhich is because something remade the swap partition; unfortunately we don't know what10:07
pwnguinat any rate, thanks for the help here10:08
pwnguincjwatson: ive always wondered where uuids come from -- are they in the partition table?10:10
cjwatsonpwnguin: they're in the filesystem itself10:13
pittimdke: hm, https://translations.edge.launchpad.net/ubuntu-doc/ is obviously not where I should go to fix a translation bug in ubuntu-docs?10:38
pittimdke: ah, found it10:47
=== cjwatson_ is now known as cjwatson
pittiah, my test installs look reasonably well10:55
mvoogra_: have you seen bug #201230 yourself at some point? I was not able to reproduce it11:04
ubotuLaunchpad bug 201230 in kdeedu "kverbos "breakes" the gui-installer" [Undecided,Incomplete] https://launchpad.net/bugs/20123011:04
=== \sh is now known as \sh_away
=== \sh_away is now known as \sh
Riddellmvo: could we add ksplash-engine-moodin to dist upgrade tools list of packages to remove?12:07
=== Sebast1an is now known as Sebastian
=== chuck_ is now known as zul
=== Hobbsee is now known as Spiky
=== Spiky is now known as Hobbsee
mvoRiddell: sure, added12:27
Riddellthanks12:29
Riddellmvo: oh, its probably too late but app-install-data is preferring KDE 4 .desktop files over KDE 3 ones12:29
Riddellmvo: ideally it would list both, but I guess that would be notable changes12:30
mvoRiddell: meh, sorry for that. do they have the same name?12:30
Riddellmvo: yes, in most cases12:30
mvoRiddell: ok, I have a look later12:32
Riddellmvo: if its complex to fix, don't worry about it.  but if there's an easy way to have both kde 3 and 4 files in there that would be nice12:32
mvook12:34
Riddellstgraber: kubuntu 20080318.1 are up and needing testing12:38
stgraberRiddell: ok, thanks12:40
kagouseb128, WoW 17.4sec for gnome login (as fast as Mandriva)12:40
seb128kagou: excellent ;-)12:40
seb128kagou: how much did we have before the changes?12:40
kagouseb128, yes :) login time as increased too but just for 2 sec12:40
kagoubefore : 39 + 2912:41
kagounow : 41 + 1712:41
kagouknowing that compiz is not activated, and no more deskbar+tracker12:41
ion_What was changed?12:42
kagouseb128, i should retest last mandriva to compare with hardy12:43
seb128ion_: update-manager and jockey delay their work, gdm does preloading and we stopped using deskbar and tracker12:47
seb12829 seconds to 17 is not too bad ;-)12:48
ion_Nice12:49
=== FreeNode is now known as herb
=== cjwatson_ is now known as cjwatson
kagoumvo, seems that apturl don't work with firefox beta4 ?!12:55
kagousomeone can confirm this bug ?!12:55
seb128pitti, slangasek: would it be ok to uploade a yelp with updated translation now for beta?12:55
seb128kagou: no need to use extra ponctuation ;-) How do you I test apturl?12:56
pittiseb128: an upload is certainly ok, but I'm not sure whether we'll get it on the CD; IIRC Riddell will/did trigger the last/next CD rebuild12:56
kagouopen apt://inkscape12:56
bardyrdoes not work12:56
kagouunder firefox it would prompt to install inkscape12:57
bardyrkagou, it prompts to select a handler for apt://12:57
kagouok i open a bug. thanks12:57
bardyrbut it works if you select to open it with /usr/bin/apturl12:58
kagoubardyr, please could you confirm Bug #20353813:01
ubotuLaunchpad bug 203538 in apturl "Don't work with Firefox3 beta4" [Medium,New] https://launchpad.net/bugs/20353813:01
bardyryea 2sec13:01
bardyrkagou, confirmed13:05
sorenogra_cmpc: Do you have a moment to fill in a few blanks for me?13:12
sorenogra_cmpc: I'm working on iscsi support in initramfs.13:12
ogra_cmpcsure, if i can :)13:12
sorenogra_cmpc: I want to try to configure iscsi if I have network setup.13:13
ogra_cmpccheck for a default route ?13:13
sorenogra_cmpc: Is there a simple way to determine whether I have that?13:13
sorenHm..13:13
sorenYes, I suppose that's a reasonable heuristic13:13
sorenDo you have an ipconfig command line handy to do that?13:14
ogra_cmpcNETWORKUP=$(route -n |grep -c ^0.0.0.0)13:14
ogra_cmpcif its 1 youre online13:14
* soren hugs ogra_cmpc 13:14
sorenAwesome! Just what I needed. Thanks very much.13:15
ogra_cmpc:)13:15
kagouthanks bardyr13:16
=== herb is now known as herb__
=== herb__ is now known as herb
RiddellUbuntu Desktop beta candidate CDs up for testing( 20080318.1)13:31
cgreganHello all, I wanted to introduce myself here on the community pages. I just joined Canonical as Ubuntu Mobile QA Engineer. If there is anything I can do to improve the relationship between community devel and test, please let me know. Thanks13:33
* _MMA_ waves hi to cgregan.13:34
pedro_hey cgregan welcome13:34
stgraberhi cgregan13:35
cgreganThanks....glad to be a aboard13:35
Hobbseehey cgregan13:43
sorenogra_cmpc: How does the nic driver land in the initramfs? Do you manually add it to /etc/initramfs-tools/modules?13:44
ogra_cmpcdepends what kind of initramfs you use13:45
ogra_cmpcif your config uses MODULES=most it should be in13:45
ogra_cmpcsame for MODULES=netboot13:46
sorenOh, and that's the default, I suppose.13:46
sorenmost, that is. Not netboot :)13:46
jpatrickwelcome cgregan :)13:48
mvoasac: could you tell me about bug #203538 ? does firefox need a different location than /usr/share/firefox/defaults/pref in beta4?13:48
ubotuLaunchpad bug 203538 in apturl "Don't work with Firefox3 beta4" [Medium,Confirmed] https://launchpad.net/bugs/20353813:48
asacmvo: yes. we have now /usr/lib/firefox-3.0b4/defaults/syspref -> /etc/firefox-3.0/pref13:52
asacis it ok to put it in /etc/..13:52
asac?13:52
mvoasac: aha, ok. I will change the package then13:52
asacmvo: please test before upload. but it should work13:53
kagouRiddell, ok let's go testing :)13:55
asacmvo: you can keep the old place and depend on '| firefox-2' if you want13:56
Riddellogra_cmpc: you're wanting CDs for beta presumably?14:09
ogra_cmpcsure14:10
pittimvo: do you have some time in the next days for verifying bug 198129?14:18
ubotuLaunchpad bug 198129 in tzdata "Chile delay in 3 weeks the daylight time transition" [Undecided,Fix committed] https://launchpad.net/bugs/19812914:18
pittimvo: (right, deja vu, we had to update tzdata again for that)14:18
mvopitti: aha, sure. how urgent is it? if super urgent, I squeeze it in today14:19
pittimvo: no, it's not; by mid next week would be good, though14:20
pittimvo: the bug will occur on March 2914:20
pitti(tzdata is probably the only package which allows us to predict future bugs that precisely :-P)14:20
mvoheh :) we should have more of those14:21
pwnguinhmm. its too bad CONCURRENCY=shell doesn't work, because it brings my boot down below 20 seconds =(14:21
pittiTheMuso, cjwatson: as "happy" broadcom wifi users, do you have an idea about bug 197819?14:22
ubotuLaunchpad bug 197819 in b43-fwcutter "[Hardy]b43-fwcutter install script fails to fetch firmware in first run" [Undecided,Confirmed] https://launchpad.net/bugs/19781914:22
Riddellsoren: ready for ubuntu server beta candidates to be built?14:25
Keybuk*blink*14:32
Keybukmy gpg public key just vanished off my box14:32
pittiKeybuk: secret key still intact?14:33
Keybukyes14:33
jdongmeh you never need your own public key anyway ;-)14:33
sorenRiddell: Yes, I think now is a good time.14:42
superm1asac, well yeah no video issues this time around, but still the wifi doesn't come up nicely on 0.6.614:42
sorenRiddell: Er.. Hang on.14:42
sorenRiddell: When is the beta due out?14:42
Riddellsoren: thursday14:42
sorenRiddell: Ah, I somehow convinced myself it'd be today. In that case, I have more stuff I'd like to sneak in, so not today.14:43
=== dholbach_ is now known as dholbach
sorenRiddell: Is that cool? I'm not familiar with the processes involved.14:43
Riddellsoren: hmm, we're in quite deep freeze already14:43
zdzichuBGwill hardy get mobile broadband support like in fedora? maybe via updates/backports?14:43
Riddellsoren: it should be important changes only by now14:44
sorenRiddell: Oh, it's important alright.14:44
sorenRiddell: I'll run it by slangasek, when he shows up. Cool?14:44
asacsuperm1: so even not with the patch?14:44
asacsuperm1: is there an improvement at all?14:44
Riddellsoren: fine with me but he won't be awake for another 4 or so hours14:45
superm1asac, the improvements with the patch get me connected14:45
superm1but not after suspend14:45
asacok14:45
asacplease attach that information to your bug and post a new syslog14:45
seb128asac: do we need to have this network manager editor in the menus?14:45
asacsuperm1: ^^14:45
superm1asac, alright14:45
sorenRiddell: I'm not the one who's in a hurry :)14:45
asacseb128: well ... except that there i bugs i don't see why we wouldn't want that14:46
asacseb128: ?14:46
asaci see that its in multiple places. we can probably fix that14:46
seb128asac: it doesn't really seem to be useful, isn't the context menu entry on the applet enough?14:47
pittiryanakca: since you fixed some jockey-gtk bugs (thanks!), are you interested in some more cosmetical things?14:49
pittiryanakca: (in particular bug 197777)14:50
ubotuLaunchpad bug 197777 in jockey "Need space in error message" [Undecided,Confirmed] https://launchpad.net/bugs/19777714:50
asacseb128: its bug 27066314:50
asacoops14:50
asacseb128: bug 20112714:50
asacfeel free to milestone if you want it removed14:51
seb128asac: thanks14:51
seb128asac: well, rather discuss it, I got comments about the preferences menu not fitting on a 1024x768 screen in the default installation14:51
seb128asac: so I'm trying to see what we can move away from there14:52
asacseb128: i have such a screen on x61 ... it fits here14:52
seb128asac: ok, somebody told me that it was touching the bottom panel, I didn't verify14:53
seb128asac: maybe it depends of the dpi you are using ;-)14:54
asachmm ... i have both panels on auto-hide14:54
asacdidn#t change dpi14:54
_MMA_seb128: Maybe remove "Main Menu" (Alacarte) as right-click on the menu itself launches it also?14:54
asacseb128: without autohide its stiill a bit above the bottom panel14:55
seb128ok14:55
asacseb128: this is a clean hardy install i did about 3 weeks ago14:55
seb128_MMA_: I think people argued it was not discovrable enough14:55
seb128asac: anyway I don't want to argue about a few pixels, we have lot of items there and I was just looking if we could simply it a bit14:56
seb128maybe we should use applications, system tools again14:56
asacsure. if you milestone it on behalf of the desktop team, i will go on and remove it14:56
seb128looks like something is installing the hardware testing application there already14:56
seb128asac: ok, thanks14:56
asacbut for final plesae :)14:56
asacbeta is too close. if i upload anything i will include this fix of course14:57
seb128right, not for beta14:57
seb1288.04 is good14:57
seb128milestoned14:57
_MMA_seb128: Ok. One could argue that windows users are doing the same thing now to edit the Start menu. They seem to find it fine. *If* they are our target audience. I don't mind either way. :P14:58
seb128;-)14:58
mathiazRiddell: what do you think about bug 197476 ?15:00
ubotuLaunchpad bug 197476 in mysql-dfsg-5.0 "akonadi  does not work with the apparmor rules introduced for /usr/sbin/mysqld on hardy." [Low,Incomplete] https://launchpad.net/bugs/19747615:00
mathiazRiddell: "akonadi starts a user-specific mysqld to store it's PIM data.15:01
Riddellmathiaz: I think it should be fixed :)15:01
Riddellon the apparmour side15:01
mathiazRiddell: It's unusual that a program would start its own mysqld process15:02
mathiazRiddell: why can't akonadi use the system databases ?15:02
\shrelmgr: please approve claws-mail and claws-mail-extra-plugins to hit the buildds/archives15:02
wasabiI don't find that really unusual. Completely removes any consideration of security.15:02
wasabiTo an extend I'd prefer more apps worked like that.15:03
wasabiextent15:03
sorenCould someone please give back virt-manager?15:03
ogra_cmpcwasabi, but then its really cleverer to use a serverless db like sqlite15:04
mathiazwasabi: I agree with ogra_cmpc15:04
mathiazwasabi: sqlite has been designed for that15:04
Riddellmathiaz: let me try and ask someone15:05
wasabiSo?15:05
wasabiNot everybody wants to use sqlite. ;)15:06
mathiazRiddell: I think the concerns raised by the reporter are valid15:06
mathiazRiddell: I question the use of mysql as a storage backend to fix their concerns.15:06
Hobbseesoren: given back15:06
sorenHobbsee: Thanks very much.15:06
Hobbseeyw15:06
Riddellmathiaz: sqlite doesn't work well enough15:09
mathiazRiddell: could you be more precise about what doesn't work well enough ?15:11
mathiazRiddell: or point me to a channel where these things are discussed ?15:11
Riddellmathiaz: not really, I'm not an akaonadi developer15:11
Riddellmathiaz: #kontact15:11
wasabiI think arguing about what technology a given app uses is not helpful.15:11
wasabiThey chose to use MySQL. They might have had more experience with it.15:12
wasabiOr a hundred different factors.15:12
mathiazwasabi: well - I'd like to know why they're using it.15:12
ogra_cmpcwasabi, if it doesnt fit our security model its at least friendly t0o suggest something that works right15:12
wasabiI agree... I just wonder what's different from using MySQL per-user vs SqlLite per-user.15:13
ogra_cmpcor to find out why they cant use something that fits in15:13
SupaFlyhey how can i participate in ubuntu's development?15:13
ogra_cmpcwasabi, one starts a server daemon, the other doesnt ?15:13
wasabiBut he said it was per-user, no?15:13
wasabiStarts a per-user instance of MySQL.15:13
wasabiMaybe I misread.15:14
ogra_cmpcyes thats what he said15:14
wasabiOh. Does it listen on external ports or something?15:14
Riddellogra_cmpc: you have 20080318.2 CDs to test15:15
ogra_cmpcno idea ... but if the policy is to lock down servers to a certain extent for security reasons and such a corner case comes around that doesnt work through that, do you drop the whole policy or do you ask the dev if he cant use something thats supported ?15:15
ogra_cmpcRiddell, gracias15:16
pittiryanakca: nevermind, I fixed it myself15:17
pittiyay mastering pyqt ;)15:17
emgentheya pitti :)15:18
=== azeem_ is now known as azeem
pittiRiddell: for bug 193985: is there only myWindowObject.exec_() or a static method like gtk.main() in PyQt?15:18
ubotuLaunchpad bug 193985 in jockey "jockey-kde crashed with AttributeError in ui_main_loop()" [Undecided,Confirmed] https://launchpad.net/bugs/19398515:18
cjwatsonKeybuk: is MINKVER="2.6.17" in /usr/share/initramfs-tools/hooks/udev still accurate?15:19
Keybukcjwatson: I think so15:19
pittiRiddell: basically I need to process events for the tray icon until it gets clicked; would KApplication.kApplication().exec_loop() work for that?15:19
Keybukuntested, obviously15:19
Keybukbut there's no reason it shouldn't work15:19
Keybukwe include both sets of rules for the old usb devices, etc.15:20
Riddellpitti: not for a qt4 app, you need   QApplication.exec_()  I think15:20
pittiRiddell: oh, great; I'll try that15:21
cjwatsonKeybuk: ok, just following up on bug 19028115:24
ubotuLaunchpad bug 190281 in linux "Segfault in initrd with 2.6.24-7 on intel x86_64" [Medium,In progress] https://launchpad.net/bugs/19028115:24
Keybukerr?15:24
seb128pitti: do you have any idea of why the policykit-gnome translations are not in the current language packs?15:24
cjwatsonoh, it's the other way round15:25
cjwatsonhardy kernel on feisty15:25
seb128pitti: danilo said that the template has been imported in rosetta some weeks ago and the translations should be available15:25
pittiseb128: oh, I see; I was just going to check for a .pot build15:26
seb128pitti: I did check that, and talked with him, the template has been imported correctly15:26
pittiargh argh ssh15:26
pittiseb128: do you see them in rookery:/srv/language-packs.../hardy/sources-base/ somewhere?15:27
seb128pitti: yes15:27
seb128pitti: for some locales but not the ones I'm interested in ;-)15:28
pittiseb128: upstream package only has Danish translations, nothing else15:29
seb128pitti: right, but it has been translated in french on rosetta for example15:30
Mirvseb128: and they're translated in Rosetta you're looking at? at least Finnish has made its way to language packs.15:30
seb128pitti: though that seems to be recent, so maybe in next update round15:30
pittiseb128: right; the previous export came too late and was broken, so I'll do another update on Friday, after beta15:30
seb128pitti: ok, seems to not be a bug15:30
seb128pitti: ok, thanks15:30
* pitti hugs seb12815:30
* seb128 hugs pitti15:31
Riddellevand: ready for gobuntu CDs for beta candidates?15:39
evandRiddell: hrm, allow me to give them a quick look over.  Pulling downnow.15:47
\shcjwatson: bug #107326 do we have any possibility to fix this in time for hardy?15:51
ubotuLaunchpad bug 107326 in parted "non working gpt labels" [High,Confirmed] https://launchpad.net/bugs/10732615:51
cjwatson\sh: are you in a position to test the result?15:52
cjwatsonit should be doable, and it's definitely on at least one "fix by hardy" list15:52
\shcjwatson: hopefully next week yes...I'm getting back a machine with 7TB...15:52
\shcjwatson: amd64 that is15:52
cjwatsonI'll see if we can have a look after beta15:53
\shcjwatson: if it's only a matter of replacing parted with a new version on the alternate cd..I can try over the weekend to tweak the alternate cd we have...15:54
cjwatson\sh: no, it's not just a matter of that15:56
cjwatsonActual Coding is involved.15:56
\shcjwatson: you mean the sync problem on macs?15:58
cjwatsonyes15:58
cjwatsoncuriously, http://developer.apple.com/technotes/tn2006/tn2166.html claims that Apple uses a protective MBR15:59
cjwatsonwhereas (we think) it's actually a legacy MBR for legacy BIOS compatibility15:59
\shwhen I reported the bug regarding broken GPT lables, these days the problem was parted, not writing correctly the partition tables...because I never booted from partitions >2TB...but all partitions >2TB were not usable, because they had a msdos label and not GPT label16:01
\shdapper worked correctly16:01
\emgentheya16:01
cjwatson\sh: thank you, I understand the problem16:01
\shcjwatson: well, I'm confused now ... I'm trying to figure out what grub has to do with it (reading some comments)16:02
cjwatson\sh: not much these days16:02
cjwatsonwell, it impedes better solutions16:03
cjwatsonI don't really want to explain it now, I'm trying to figure out other things16:03
cjwatsonlike whether there's actually a standard for Intel Mac >2TB disks16:03
\shcjwatson: and I'll try to stop being confused with further readings16:06
=== Kopfgeldjaeger is now known as Off
=== Off is now known as Off|Tuhpik
\shcjwatson: and if you need someone to test on amd64 hw, next week I'm able to do it16:07
maswanyou need some testing with >2TB partitions? anything I can help out with?16:08
maswanI can find a server or two with a decent ammount of storage16:08
\shmaswan: always the very same people, I think ;)16:09
maswanWe usually don't partition those filesystems though16:09
cjwatson2TB is the (relatively) easy bit; the problem is that what you have to do to cope with >2TB and what you have to do with Intel Macs conflict, and I'm trying to work out what will happen if somebody puts a >2TB disk in an Intel Mac and tries to dual-boot OS X and Ubuntu on it16:09
maswanAh, no macs around here16:09
cjwatsonwhich will very likely be the case in the lifetime of Hardy16:09
\shmaswan: well, I had to, and it looks like I have to do this in the very near future again :(16:10
pittiRiddell: ok, I think I got the jockey-kde notifications fixed; the kde test suite is still a mess, but at least the app itself works now16:11
\shmaswan: and again with *censored* areca raid16:11
cjwatson(note that I am not saying that the above should impede the >2TB case in general from being fixed in hardy, but it seems like something we need to consider)16:12
maswan\sh: oh, they are still bad? hm. hope ours won't be too horrible, we'll get a few delivered soon.16:12
\shmaswan: well, the ones we had during combots time ;) which were crashing because of broken firmware...and old kernel drivers...the native ones in newer linux kernels are much better, but firmware was again very bad these days...let's see what the new firmware gives me16:14
maswan\sh: ok, hope we'll have less issues. anyway, delivery seems to be delayed, so we might run hardy on them from the start. :)16:14
\shmaswan: when you get hands on your controllers, just send me the model number :) so I can think of replacing those old ones :)16:16
maswan\sh: sure, if they work. now off.16:16
=== hefe_bia_ is now known as hefe_bia
cjwatsonI think it may be that >2TB disks on Intel Macs simply have to be an error :-/16:21
* ogra_cmpc wonders why his rsync is lagging so heavily16:23
ogra_cmpcdoes anyone else have slow rsyncs or is it just me ?16:24
mvosiretart: if you don't mind, I would like to upload http://paste.ubuntu.com/5811/16:25
\shcjwatson: >2TB disks in intel mac laptops, for sure...;)16:29
mvobryce: when you uploaded displayconfig-gtk, did you put your changes in bzr?16:32
Riddellogra_cmpc: other people are grumping too16:39
Riddellevand: shall I build gobuntu?16:40
evandRiddell: it seems the seeds need to be updated16:40
evandthe install is failing on the missing hwdb-client-gnome issue from a few days back16:40
ogra_cmpcmy first one was flying ... the second one sits at 4k with 28h to go since 30min already16:40
evandworking on that now16:40
Riddellevand: ok, let me know if you need uploads approved16:41
evandRiddell: will do, thanks16:41
pittibryce, tjaalton: hm, can you please tell me if http://launchpadlibrarian.net/12724433/xorg.conf is actually valid? I particularly mean the "  screen 0 "aticonfig-Screen[0]" 0 016:43
pitti" line in ServerLayout16:43
ogra_cmpcdid anyone try alternate yet ? i'm getting a ton of perl errors here17:02
ogra_cmpchmm, but seems that only spams the log17:02
evandRiddell: gobuntu-meta 1.11 whenever you have a free moment.17:07
mvopitti: I updated the information in bug #174128  - if you are happy with the debdiff I will upload it17:17
ubotuLaunchpad bug 174128 in dhcp3 "asks debconf question on dapper->hardy upgrade" [Undecided,Incomplete] https://launchpad.net/bugs/17412817:17
jdstrandhi lamont17:17
jdstrandlamont: just sent git patch for bug #20352817:18
Riddellevand: accepted17:18
ubotuLaunchpad bug 203528 in bind9 "apparmor profile should be in complain mode on certain upgrades" [Undecided,In progress] https://launchpad.net/bugs/20352817:18
jdstrandlamont: this follows the (updated) ApparmorProfileMigration17:19
evandRiddell: thanks17:19
lamontjdstrand: ok.  for beta, or post-beta?17:20
jdstrandlamont: basically, it just makes sure pre-apparmor bind9 upgrades are in complain mode, as well as forcing complain mode the apparmor-profiles is unchanged17:20
jdstrandlamont: that was why I wanted to talk to you17:20
jdstrandlamont: I *doubt* there will be many dapper-hardy upgrades for beta, but there will be gutsy-hardy17:20
lamonttrue17:21
jdstrandlamont: therefore, it probably should go in before17:21
lamontok.  I'll upload and request the sync tonight-ish17:21
jdstrandlamont: cool, thanks!17:21
* jdstrand now goes to update mysql and openldap in the same way17:22
pittimvo: looking17:22
lamontjdstrand: if you have a patch for the host-doesn't-print-nameserver list bug too, that'd be wonderful. :-)17:22
lamontjdstrand: bug 19153017:22
ubotuLaunchpad bug 191530 in bind9 ""host" cannot see sites in .org" [Medium,Triaged] https://launchpad.net/bugs/19153017:22
pittimvo: ah, so it was never really a conffile prompt?17:22
jdstrandthe one we talked about on saturday? sorry, no17:23
pittimvo: does the default DTRT?17:23
* lamont yawns at bug 203476 - does anything actually _use_ that inet_ntoa>?17:23
ubotuLaunchpad bug 203476 in bind9 "[libbind9] [CVE-2008-0122] off-by-one error in the inet_network function" [Undecided,New] https://launchpad.net/bugs/20347617:23
jdstrandlamont: rh said they don't, in their bug report17:24
mvopitti: frankly, I don't know if the initial thing was "needs-start" or "files-moved-around", I could be mixing the initial report up17:24
mvopitti: the default should be fine, its just a debconf note17:24
lamontjdstrand: yeah, and debian has the bug sitting there stale for some time as well.17:24
jdstrandI saw that, went 'oh my', and then went, ohh-- backburner17:24
pittimvo: ah, great17:25
lamontyeah17:25
lamontjdstrand: every now and then, as I read the bug page, I go "HUH!?!?!" and then go "yeah. right."17:26
jdstrandhaha17:26
mvopitti: ok, if all is cool with the diff I will upload (at this stage of the freeze I like to always double check :)17:26
pittimvo: sure17:26
lamontjdstrand: maybe I'll backport that patch from 9.4.3 (or the report) just to shut that up17:29
jdstrandoh the fedora one? I haven't looked at it17:30
lamontthe CVE-2008-0122 thing17:31
ubotuOff-by-one error in the inet_network function in libbind in ISC BIND 9.4.2 and earlier, as used in libc in FreeBSD 6.2 through 7.0-PRERELEASE, allows context-dependent attackers to cause a denial of service (crash) and possibly execute arbitrary code via crafted input that triggers memory corruption. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2008-0122)17:31
lamontah u\botu got smart about CVE too17:32
jdstrandyeah, it certianly seems harmless enough17:32
jdstrandcertainly17:32
* lamont isnt' on the good mailing list anymore17:32
jdstrandhttps://bugzilla.redhat.com/attachment.cgi?id=29203017:32
jdstrandnot private17:33
jdstrandhttps://bugzilla.redhat.com/show_bug.cgi?id=429149 is the bug17:33
ubotuRed Hat bug 429149 in vulnerability "CVE-2008-0122 libbind off-by-one buffer overflow" [Low,Verified: ]17:33
infinitylamont: Aww, quitting HP got you knocked off vendor-sec?17:33
lamontinfinity: well, it was through hp, so yeah17:35
ryanakcapitti: sorry, I'm at school :)17:36
=== emgent_ is now known as emgent
=== thegodfather is now known as fabbione
pittiScottK: I don't quite understand why we keep the remaining delta for cyrus-sasl2-heimdal18:43
pittiScottK: libdb-dev is by and large equal to libdb4.6-dev?18:43
slangaseksoren: what are you running by me? :)18:47
millilamont: you get my patch against 9.4.2?18:48
pittiryanakca: NP, I fixed them all :)18:58
emgentasac: ping18:59
emgentthere is a critical problem in nm-applet on hardy18:59
emgentseems dont work with new Intel PRO/Wireless 3945ABG19:00
emgentwith iwconfig working fine in hardy, but GUI dont work19:01
asacemgent: hidden network?19:01
emgentasac: no19:01
asaceveryone i know that has 3945 could connect19:02
ogra_cmpcasac, can we get rid of the NM console messages on shutdown for final ?19:02
ogra_cmpcthey look scary19:02
emgentit's an open A.P.19:02
asacogra_cmpc: i never shutdown my system. how do they look like?19:02
asacemgent: hmm19:02
asacemgent: can you try WPA?19:03
ogra_cmpcasac, console messages, debug stuff19:03
emgenttried first, same problem19:03
emgenti think that it's a problem with device name19:03
ogra_cmpcit only blinks up shortly before usplash kicks in with the shutdown screen19:03
emgentnew ipw driver generated "wlan0_rename" device19:03
ogra_cmpcbut on first glance it looks like an error msg19:03
asacemgent: i have several reports that it works with WPA19:04
=== kylem_ is now known as kylem
asacemgent: how far does nm get?19:04
emgentasac: false i used WPA TKIP19:04
asacdo you see one green light at least?19:04
emgentnope led too dont work.19:05
asacplease try to connect now and post your complete syslog somewhere19:05
emgentbut ipw driver working fine19:05
emgentasac: it's impossible, i can only connect with iwconfig, but not in GUI19:06
asacemgent: most likely it doesn't work fine; otherwise nm would work :)19:06
emgenti cant see the possibility to connect with it :P19:06
asacah19:06
asacpost /etc/network/interfaces then19:06
emgenthttp://nopaste/p/aJWKMiQGH19:07
KeybukI had that the other day19:07
Keybukit was because something had installed the -386 kernel for me19:07
emgentseems see eth0 and lo19:07
Keybukand since -386 < -generic, that's what grub picked19:07
Keybukemgent: what does uname -a say?19:07
slangasekbryce: can you help me understand the gnome-control-center upload?  199549 is about settings not taking effect on an i965, doesn't that support xrandr 1.2?  and bug #198951 isn't assigned to gnome-control-center at all; are these fixes that are important for beta?19:07
ubotuLaunchpad bug 198951 in gnome-desktop "gnome-display-properties crashed with SIGSEGV in screen_info_new() - i855, fglrx, radeonhd" [High,Fix released] https://launchpad.net/bugs/19895119:07
emgent2.6.24-12-generic19:08
emgenton i68619:08
Keybukok, not the same problem then19:08
emgentonly GUI dont work19:08
Keybukwhat does iwconfig say?19:09
emgentdriver working fine to me19:09
Keybuk(when run without arguments)19:09
bryceslangasek: they could wait until after beta, I think the gnome-settings-daemon fixes may cover us sufficiently.  the gnome-control-center changes just make the new xrandr gui more useful19:09
slangasekbryce: ok, I'll sit on it, thanks19:09
Keybukemgent: driver is clearly not working fine, or the GUI would work19:09
emgentKeybuk: http://nopaste/p/aI3ckLUNw19:09
Keybukok19:10
emgenti'm connected to wifi no, but only with iwconfig.19:10
Keybukemgent: sudo mv /etc/udev/rules.d/70-persistent-net.rules /root19:10
Keybukthen reboot19:10
emgentk19:11
emgentjust a moment.19:11
lamontKeybuk: /root... is that your version of a persistent /tmp ? :-)19:11
Keybuklamont: exactly19:11
emgentheheh :)19:11
asac_emgent: sorry reconnect. did you answer what Keybuk asked you?19:13
emgentyes19:13
Keybukhe did, just waiting for him to reboot19:13
asac_so why does -386 break NM?19:13
emgentnow working fine..19:13
emgentudev rules..19:14
Keybukasac_: it doesn't have the iwl driver in it if you don't install the modules package :)19:14
Keybukasac_: I only got the kernel by some upgrade bug19:14
asac_strange19:14
asac_but it appears to have ipw394519:14
emgentKeybuk: ok thanks19:14
asac_or is that independent?19:14
Keybukasac_: isn't that in linux-ubuntu-modules ?19:14
asac_yes i think so19:15
Keybukright19:15
Keybukthat's what I didn't get installed19:15
asac_but not 100% source19:15
Keybukjust linux-image19:15
asac_yeah, but emgent appeared to have at least some driver :)19:15
asac_thats why i wondered19:15
KeybukI suspect he just had some transient problem19:16
Keybukthe reboot cured it19:16
Keybukmaybe caused or not caused by that udev rename bug19:16
asac_right. lets keep it that way :)19:16
dneary Hi19:22
dneary I've got a .dsc, .diff.gz and _orig.tgz from a Hardy package that I'd like to build on Gusty19:22
dneary Anyone mind running me through generating a .deb from the package files?19:22
dnearyI'm not exactly starting from scratch, and I've been having lots of trouble going in circles in the packaging guide19:22
jdstrandbryce: hi!19:23
brycejdstrand: hey19:23
Keybukdneary: dpkg-source -x DSCFILE ; cd DIRECTORY ; debuild19:23
jdstrandbryce: are you aware of any weird bugs with /tmp/.ICE-unix ?19:23
jdstrandaccess to it19:23
Keybuk(with build-essential, dpkg-dev, devscripts and fakeroot installed)19:23
brycenope19:23
jdstrandhrmm19:23
=== asac_ is now known as asac
zulmvo: is there a howto to do a LTS->LTS on the wiki somewhere?19:26
jdstrandbryce: I can't open terminals, vim hangs, and probably other things19:27
jdstrandbryce: strace xterm gives:19:27
jdstrandconnect(4, {sa_family=AF_FILE, path="/tmp/.ICE-unix/9709"}, 21) = 019:27
jdstrandfcntl(4, F_SETFD, FD_CLOEXEC)           = 019:27
jdstrandwrite(4, "\0\1\0\0\0\0\0\0", 8)         = 819:27
jdstrandread(4,19:27
jdstrandthen hang forever19:27
jdstrandbryce: I really have no idea what could be causing that-- ever seen anything like it?19:27
brycejdstrand: no I haven't seen bugs with errors in /tmp/.ICE-unix.19:28
* jdstrand scratches his head19:28
brycejdstrand: I would guess to look first at what changes have happened to your system recently...  we've not done many significant X uploads lately, so my first guess would be something non-X related, but who knows.19:29
* jdstrand nods19:29
brycejdstrand: also I'd doublecheck resources are okay, clear tmp, restart, etc. etc.19:29
jdstranddid all that19:29
mvozul: yes: https://help.ubuntu.com/community/HardyUpgrades#head-e7f287c730b93116f89de7ea7e05efbe95fa6dd119:30
brycepossibly the .ICE-unix is just a symptom of an underlying issue19:30
dnearyKeybuk: Thanks19:32
zulmvo: thanks!19:32
dnearyWhat package does debuild come in, I don't think it's installed?19:32
mvodevscripts19:32
jdstrandbryce: ok thanks19:32
mvocommand-not-found should know that too19:32
dnearyDo I also need fakeroot?19:34
dnearythanks mvo19:34
jdstrandbryce: was X built since the latest libc6 update?19:34
mvocheers19:34
dnearySo, yes, I need fakeroot :)19:34
dnearyor I have to build as root19:35
dnearyWoohoo! autoconf output!19:35
brycejdstrand: last upload was Mar 13th it looks like19:35
slangasekjdstrand: fuser -vm /tmp/.ICE-unix/9709?19:35
mvoRiddell: could you please check http://paste.ubuntu.com/5819/ ? this is dapper->hardy19:36
jdstrandslangasek: http://paste.ubuntu-nl.org/60069/19:37
Riddellmvo: mm, right19:37
dnearyKeybuk, mvo: Thanks a lot, that's just what I needed.19:37
Riddellmvo: any other koffice issues?19:37
dnearykeysign didn't work because I'm missing the private key, it says19:37
slangasekjdstrand: er, oops - I meant fuser -v /tmp/.ICE-unix/970919:37
dnearyAnyone know where I can get that?19:37
mvoRiddell: not that I'm aware of19:38
dneary(OK, I was joking there...)19:38
jdstrand/tmp/.ICE-unix/9709: jamie      9709 F.... x-session-manag19:38
jdstrandslangasek: ^^19:38
jdstrandslangasek: it exists, perms are ok19:38
jdstrandI checked the obvious stuff...19:38
slangasekjdstrand: yes; the question is really, what's happening at the other end of the socket - x-session-manager may be broken on your system19:38
slangasekI don't know if it's sane to try stracing that process, but that would be my next step19:38
mvoRiddell: I can have a look later too if you are too busy19:39
slangasek(don't know if it's sane -> don't know if your session will survive ;)19:39
jdstrandslangasek: I'll wait til my non-screen'd build finishes19:39
jdstrand:)19:39
Riddellmvo: I can do it19:40
jdstrandslangasek: $ strace -p 970919:43
jdstrandProcess 9709 attached - interrupt to quit19:43
jdstrandrestart_syscall(<... resuming interrupted call ...>19:43
jdstrandtried launching stuff, but nothing19:43
jdstrandoh, let me try -f19:44
jdstrandaha!19:44
jdstrand[pid  9790] read(24,19:44
jdstrandhangs there19:44
jdstrandI don't have a pid 9790...19:45
slangaseka thread, perhaps19:45
slangasekwhat is x-session-manager on this system? gnome-session?19:45
slangasek(x-session-manager is an alternative)19:46
jdstrandslangasek: yes19:46
jdstrandgnome-session19:46
slangasekok19:46
jdstrandI do have /usr/bin/seahorse-agent --execute x-session-manager19:46
slangasekwell, all I can say is that the two sides of the socket don't agree about whose turn it is to write :/19:47
Keybukmathiaz: really don't add a status to udev19:48
Keybukit'd encourage people to think that "start" and "stop" are useful commands19:48
Keybukwhen they're not19:48
Keybukif you run either on a real system, you'll break it19:48
jdstrandslangasek: somewhat disturbing, as I wasn't actively fiddling with anything.  just 'working'19:48
mathiazKeybuk: sure - for udev it doesn't make sense19:48
jdstrandslangasek: this happened once before-- around the libc6 madness (but I never installed the bum version on the system)19:49
* jdstrand scratches his head more19:49
Keybukmathiaz: it also struck me that your list is pretty much the list of packages that are likely to be the first to go to upstart19:49
Keybukso I'm not sure how much effort it's worth putting into a complex status() command19:49
mathiazKeybuk: well - my target is really hardy19:50
slangasekKeybuk: well, I don't think the proposed command was terribly complex either?19:50
Keybukright, but you're adding a potentially complex thing to debug and support19:50
Keybukespecially if you're doing more than just pidof()19:50
mathiazKeybuk: as you mentionned, Intreprid would use upstart which is fine by me19:50
mathiazKeybuk: Hardy is going to be around for some time and adding status action to init script is valuable IMO19:51
mathiazKeybuk: I wouldn't defend this idea if hardy wasn't an LTS19:51
Keybukironically it's precisely because it's going to be around for a long time that it makes me nervous ;)19:51
=== Traxer is now known as Traxer|on
=== Traxer|on is now known as Traxer
mathiazKeybuk: I see your point.19:52
mathiazKeybuk: IMO it's worth the tradeoff - the code to add a status action is small and isolated.19:53
cjwatsonis it good to have a status option for just a few init scripts in an LTS, thereby confusing people because it isn't there across the board?19:53
mathiazKeybuk: I don't think it's worth adding a complex status action that does more than check if the pid is running.19:53
cjwatsonI don't think you should do a half-baked job of it for an LTS, and I don't think we have time for you to do a full job19:54
kirklandcjwatson: it's already in a couple...  postgres, mysql19:54
Keybukmathiaz: what pid?19:54
mathiazcjwatson: well - I think it's better than nothing. Moreover it's already available in some of them.19:54
cjwatsonkirkland: right, but a tiny minority19:54
Keybukwhere do you get the pid ?19:54
mathiazKeybuk: it depends on the init script.19:55
slangasekmathiaz: btw, your count is off, zul already took the liberty of adding this option to /etc/init.d/samba in his latest upload..19:55
mathiazKeybuk: some init scripts define a PID file.19:55
Keybukmathiaz: exactly, then we start to get into more complicated territory19:55
Keybukand you'll have race conditions with when pid files appear/disappear19:55
cjwatsonI don't have a problem with it being done gradually, but I don't think it's worth breaking feature freeze for19:55
mathiazKeybuk: some don't.19:55
mathiazslangasek: I know - but I think you mentionned that it reports only smbd.19:55
mathiazslangasek: nmbd should also be checked.19:56
slangasekmathiaz: hrm, if you mean my bug reopen, that was about the if-up.d script that was added that doesn't do anything useful19:56
mathiazcjwatson: well - some people would argue that it's a bug.19:57
cjwatsonit's not19:57
mathiazcjwatson: now I do think it's a new feature19:57
* lamont mutters about cups/printers.conf having &*)*^*)_ timestamps in it from when cups was last started...19:57
lamontthat's not something that belongs in /etc19:57
mathiazcjwatson: but looking at the required changes and the testing to be done, it's worth the tradeoff.19:58
kirklandslangasek: i had already posted patches for samba, apache, and cron to Launchpad before we started thinking about how to solve this more generally, and posted the lsb- note to ubuntu-devel19:58
mathiazcjwatson: as I said, it's really something that stands out when compared to other distros such as redhat or suse19:58
KeybukI guess I also wonder ...19:58
Keybuk*why* would cron ever not be running?19:59
cjwatsonit won't stand out if we do a partial job19:59
lamontwhat binary is it that implements the gnome session screen lock?19:59
Keybuklamont: gnome-screensaver19:59
lamontdanke19:59
lamontKeybuk: I just restarted cron on a dapper machine where it was not running... nfc why it decided to walk away from the table, though.19:59
mathiazcjwatson: well - some scripts already have a status action. so it's already a partial job.19:59
cjwatsonwhy is it worth pushing through feature freeze if it's not going to be complete?20:00
kirklandKeybuk: cron, udev -- exactly should always be running.  it would be nice to determine through a common mechanism if they're not, such that a monitor or ebox or a human or some such could take evasive action ;-)20:00
Keybukkirkland: I'd prefer to fix things so that they're never not running20:00
lamontnagios?20:00
lamontKeybuk: that'd be wonderful20:00
Keybukpeople actually building things on these status arguments makes me even more against them than before :)20:00
MithrandirKeybuk: software has bugs, systems run out of memory and processes get killed, those kinds of things.20:01
mathiazcjwatson: when we posted the patch to add status to samba init script to the Debian bts, we got a reply right away from users to include it.20:01
KeybukMithrandir: "respawn"20:01
mathiazcjwatson: so I take the approach that users are really requesting it.20:01
MithrandirKeybuk: people hack on those daemons too, in which case respawn often isn't what you want. :-P20:01
mathiazcjwatson: and I don't think to do everything in one go - it won't be possible.20:01
KeybukMithrandir: "stop" :-)20:01
mathiazcjwatson: in that particular case, we can really implement it step by step20:02
kirklandKeybuk: is it preferable to have people building things on `ps -ef | grep foo | awk bar` ?20:02
mathiazcjwatson: package by package - it's not a big rollout. We just fixe annoyance in everyday life or sysadmin20:02
Keybukkirkland: I'd prefer people waited for the proper fix20:02
cjwatsonmathiaz: and potentially introduce little niggling bugs20:03
Keybukmathiaz: that's fine during the usual development stages20:03
Keybukbut the whole point of freezes is that after them we concentrate on fixing bugs20:03
Keybukrather than the continued deployment of features20:03
lamontKeybuk++20:04
siretartmvo: ugh darn. good catch! thanks!20:04
=== Off|Tuhpik is now known as Kopfgeldjaeger
* lamont grumbles at jdstrand for his commit messages20:06
jdstrandlamont: ?20:06
lamontfix for LP: #203528 (force complain for apparmor on certain upgrades Signed-off-by: Jamie Strandboge <jamie@canonical.com>20:06
jdstrandhey, I didn't touch changelog this time!20:07
lamontheh20:07
mvosiretart: thanks, no problem20:08
lamontso the change is that if the package is upgraded, then it's just complain-mode, and enforce-mode on clean installs?20:08
jdstrandlamont: not exactly20:08
jdstrandlamont: it is complain on pre-feisty package to enforcing package upgrade20:09
jdstrandlamont: it is also complain on pre-enforcing but apparmor aware OS20:09
jdstrandlamont: and the apparmor profile on the system prior to upgrade is in complain mode20:10
jdstrandthis is documented in ApparmorProfileMigration20:10
lamontapparmor: force complain-mode for apparmor on certain upgrades20:11
lamontSet complain mode on upgrades from 1:9.3.4-2ubuntu2 and earlier,20:11
lamontrather than enforcing mode.  See also20:11
lamonthttps://wiki.ubuntu.com/ApparmorProfileMigration20:11
lamontAddresses-Ubuntu-Bug: 20352820:11
lamontSigned-off-by: Jamie Strandboge <jamie@canonical.com>20:11
lamontso like that?20:11
jdstrandlamont: well, that covers the first case, yes20:11
lamontsee, if you wrote real commit logs, then I wouldn't have to rewrite  them and get it wrong... :-)20:12
jdstrandlamont: but not when upgrading from feisty or gutsy and apparmor-profiles is installed, or usr.sbin.named exists20:12
lamontthe first line gets added as the changelog entry, with the relevant Addresses-{Ubuntu,Debian}-Bug: entries.20:12
lamontthe Signed-off-by comes from git commit -s, if you're actually using git to do the formatting... :-)20:13
lamonts/formatting/committing/20:13
lamontand if there was a bzr UI that actually read _and_wrote_ git backends, then you could just use bzr20:13
jdstrandI used git20:13
lamontok.  just throw the -s in and it'll add the signed-off-by line20:14
jdstrandthat's what I did20:14
lamontthen something smashed it all together.20:14
lamontbecause it was one line after the git-am20:14
mvoRiddell: could you please check http://paste.ubuntu.com/5821/  too ? this is dapper->hardy20:14
jdstrandbut, I didn't want to be too wordy on the commit, as I quoted the bug, that talks about it and has the link.  *shrug*20:14
jdstrandI'll admit I forgot you used the commit message to get the changelog20:15
lamontthe changelog creation script uses the first line, and the addresses-* lines.  the rest of the log is for posterity and can be as verbose as you want20:15
jdstrandok20:15
jdstrandsomewhat terse first line, the rest as descriptive as need be20:15
cjwatsonheno: any word on bug 199129?20:16
ubotuLaunchpad bug 199129 in ubiquity "Auto-resize install fails to mount drive" [Undecided,New] https://launchpad.net/bugs/19912920:16
lamontwhich also means that the first line wants a 'apparmor:' or such prefix to say what part of bind9 it's mucking with20:16
cjwatsonheno: especially https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/199129/comments/8, if possible20:16
ubotuLaunchpad bug 199129 in ubiquity "Auto-resize install fails to mount drive" [Undecided,New]20:16
Riddellmvo: ok, fixing20:16
affluxscreen brightness does not change at all, even echoing to /sys/class/backlight/acpi_video0/brightness does not work. Is this a bug in the kernel?20:18
* lamont slaps 191530 around, before going back to work for another 20 minutes20:35
mvoasac: could you please have a look at http://paste.ubuntu.com/5823/ ? (dapper->hardy)20:41
heno_cjwatson: looking20:43
asacmvo: looking20:43
=== heno_ is now known as heno
* heno tests it again20:44
asacmvo: does dapper have a network-manager-gnome package or is it "all-in-one"?20:45
mvoasac: dapper has a network-manager-gnome20:46
asachmm20:46
asacmvo: ok that file appears to be shipped by nm inself now20:48
asacstrange20:48
amitkafflux: yes it is known, fix will go in after the beta AFAICT20:48
affluxamitk: oh, okay, thanks20:49
asacmvo: would a versioned conflict ensure that the package getes removed _before_ unpacking is attempted?20:51
mvoasac: I haven't looks at it, but all you need is probably a "Replaces: n-m-gnome (<< version-where-the-file-moved-from-nm-to-nm-gnome)20:53
mvoasac: I can have a closer look if you want me20:53
asacmvo: yes its NN replaces or conflicts nm-gnome < XXX20:54
asacmvo: just not sure if conflicts would be enough20:54
mvoasac: ok, replaces should be fine20:54
mvoasac: a conflicts is the needed, a replaces really means "replaces files " and that is what happens20:54
mathiazslangasek: I don't see any ubuntu-server iso in iso.qa.ubuntu.com. What's your testing plan for the server isos ?20:54
mvoasac: a conflicts is too strong if its just files moving around20:54
slangasekmathiaz: to click this 'add build(s)' button right now so they appear20:55
asacmvo: well, but we don't want version of NM-gnome considerably lower than NM :)20:55
mathiazslangasek: ah ok. I was just to impatient then...20:55
asacmvo: thats why i thought that its more accurate to use conflicts.20:55
asacmvo: but i will go with whatever you prefer20:56
slangasekmathiaz: no, your comment was what reminded me that I hadn't clicked the button yet, thanks :)20:56
mathiazslangasek: is this the first iso candidates ?20:56
mvoasac: heh :) I have no idea aobut the nm package relations - let me have a look first before I make more unqualified comments20:56
mathiazslangasek: I remember heno sending an email about two phases testing or something like that. I wonder how would that apply to the ubunu-server iso20:57
slangasekmathiaz: we're in phase two at this point20:57
slangaseki.e., full validation of all images20:57
mathiazslangasek: hum... was there something planned for phase one for ubuntu-server iso ?20:58
asacmvo: indee20:58
asacmvo: independent of that its questionable if that file belongs to NM at ll20:58
slangasekmathiaz: I don't know; if so I guess it was coordinated just among #ubuntu-testing20:59
mvoasac: yeah, that may actually be the right solution to move it to the -gnome package - the nm package does not have a glade or gtk dependency AFAICS21:00
asacmvo: i think the upstream source split led to that. no idea if they forgot to move it21:03
asacmvo: added to my todo21:03
mvoasac: looks like bug #123772 caused the move21:06
ubotuLaunchpad bug 123772 in network-manager-openvpn "network-manager-applet no longer produces/provides usr/bin/nm-vpn-properties" [Low,Fix released] https://launchpad.net/bugs/12377221:06
mvoasac: that is what the old changelog says21:06
mvoasac: I think http://paste.ubuntu.com/5826/ should be enough21:07
=== Traxer is now known as Traxer|on
mvoasac: if you are happy with it, I will commit and upload21:08
asacmvo: let me do that21:10
asacmvo: i need to add bugs to the other changelog entry thats already on the branch21:11
amitktop line from 'top':  5959 amit      20   0 1756m 1.4g 3012 S    0 70.9  12:52.29 compiz.real21:12
mvoasac: ok, thanks (even better :) - my changelog is a bit short, you may want to add somehting like "fix file overwrite problem on dapper->hardy upgrades" or something like this21:12
* mvo hugs asac21:12
mvoamitk: is that with nvidia?21:12
amitkumm... is it normal for compiz to show 70.9% memory?21:12
amitkmvo: no... I have a poor old Intel graphic adapter21:13
amitkmvo: Intel Corporation 82G33/G31 Express Integrated Graphics Controller, FWIW21:14
mvo1.4g is quite impressive21:18
mvoAmaranth: ---^ have you seen numbers like this already?21:19
Keybukdoesn't compiz directly map video memory on newer drivers?21:19
Keybukamitk: can you nopaste /proc/5959/maps somewhere?21:19
Amaranthno idea21:19
amitkKeybuk: http://pastebin.ubuntu.com/5828/21:21
mvoamitk: how long is that system running (i.e. do you always suspend/resume it?)21:21
* mvo wonders if exa may have anything to do with that21:22
amitkKeybuk: result of uptime ->  23:22:01 up  9:22,  6 users,  load average: 0.09, 0.17, 0.3421:22
Keybukamitk: it's all heap21:23
amitkKeybuk: no, this machine doesn't suspend/resume, only reboots when there are upgrades to the kernel21:23
Keybukso looks like a genuine leak21:23
amitkKeybuk: want me to file a bug?21:23
Keybukamitk: --> mvo ;)21:24
mvoamitk: please21:26
huntekewhere can I find more information about upstart?  will it be included in hardy?21:26
mvoamitk: anything special in the plugin settings (i.e. have you played with ccsm)?21:26
amitkmvo: I might have, though I could restore it to defaults. Any way to dump current CCSM settings?21:28
Keybukhunteke: what would you like to know?21:28
mvoamitk: yeah, it has save current settings in the left-hand side under preferences21:29
mvoamitk: the more information you can gather, the better :)21:29
huntekeKeybuk: what's holding it back, how a noob can help/get involved with only about 3 hrs/wk to commit to it, why only a few devs care about it . . . I guess some of the esoteric things of "more about it"21:30
sorenslangasek: I wanted to get your opinion on whether to build the beta now or to wait until I had finished my iscsi changes. It's moot now anyway, though, as I'm not done, and we should get the iso built sooner rather than later.21:30
SwissPhoenixHi folks, I'm using hardy and getting a PANIC on my RAID Controller, files a bug ten day ago, but nobody noticed - just wondering if anybody cares to fix it to the release or if I have to do something more....21:31
Keybukhunteke: to answer your last question first, I guess the primary reason only a few people care about it is that it's beneath the level most people tend to look21:31
Keybukwe have some kind of service management today, it just sucks21:31
Keybukand only a few very strange people actually care that it sucks, and want to do something about it to fix it21:31
Keybukand much of the energy of those kinds of people is already distracted by things at the same level like D-Bus, HAL, udev, etc.21:32
huntekeKeybuk: I know. :-(  although I'm looking at it from an end-user perspective of "I want linux to beat windows"21:32
hunteke:-)21:32
SwissPhoenixhunteke: Never use the force to do harm ;21:32
=== Traxer|on is now known as Traxer
Keybukhunteke: Upstart, at best, gives us feature parity with Windows21:33
KeybukWindows (NT branch, at least) has always had a service management framework21:33
huntekeKeybuk: well, then, call it more a "I want a faster boot because sleep is so sketchy right now" less harm oriented for SwissPhoenix ;-)21:33
amitkmvo: https://bugs.edge.launchpad.net/ubuntu/+bug/20371521:33
ubotuLaunchpad bug 203715 in ubuntu "compiz eats lots of memory" [Undecided,New]21:33
Keybukhunteke: Upstart won't give you a faster boot21:34
Keybukit will just give you a boot where the order that things happen is no longer important21:34
asacslangasek: network-manager 0.6.6-0ubuntu2 knocks on your door21:34
Keybuk(which means we can take out a lot of sleep statements and busy loops)21:34
huntekeno? I thought it was event based so a lot of the serialization could be parallelized21:34
huntekebut I suppose also sleep statements removed21:34
Keybukit's event based so we can do things when we can do them21:34
Keybukrather than doing things in the hope that we can do them21:34
huntekehehe, point21:35
Keybukthe first question is that it's not being held back21:35
Keybukhttp://bazaar.launchpad.net/~keybuk/upstart/trunk/changes21:35
Keybukthere's been a vast amount of development in bzr21:36
Keybukthis hasn't reached Ubuntu because Ubuntu is doing an LTS release21:36
Keybukand some of the development is a bit exciting21:36
Keybuk(I've been hanging around cjwatson too long and am getting conservative in my old age)21:36
SwissPhoenixIf anybody cares: Its bug#199934, so far I'm using an gutsy kernel which works...21:36
huntekeoh gosh, Keybuk, you're main dev on upstart (just whoised you).  Before I go any farther, let me say thank you for your work.21:37
SeveasKeybuk, how's they grey hair doing? :)21:41
KeybukSeveas: heh, if I was taking that much after Colin it'd be going red, not grey21:41
Seveaslol, true that21:42
Seveasit'd also be a tad longer21:42
SeveasKeybuk, do you plan to break intrepid the minute hardy has been released, or will we have to wait for upstart goodness a bit longer?21:45
Keybukdepends21:45
KeybukD-Bus and I have yet to become friends21:45
Seveasthat's tough love21:46
SeveasI'm fighting the pythin bindings now, there seems to be something in there that's less than optimal when sending a message21:46
Seveasand I'm sending a lot of messages :)21:47
jdongwhat are the hard dependencies of starting gdm?21:47
jdongof course x11-common and checkroot'ed21:47
jdongbut is dbus actually needed before login?21:47
Keybukabsolutely21:48
Keybuksince the gdm greeter will need dbus21:48
Keybukand hal, and policykit, and consolekit, etc.21:48
jdongoh ok21:48
jdongso I'm getting too greedy :)21:48
jdongKeybuk: is there a way to use initctl to block until some event has happened? (i.e. start X first then have early gdm init block until dbus is ready)21:49
Keybukgdm starts X21:49
Keybukthat's kinda the point21:49
KeybukGNOME *DISPLAY MANAGER* :)21:49
jdongKeybuk: well the VT switching into X on my macbook takes like 3-4 bloody seconds21:50
jdongKeybuk: I'm wondering if that could be taken care of in a more parallel manner :)21:50
Keybukjdong: that's fixed by other bits of work in progress21:50
jdongthat's good to hear21:50
wasabiI hear mode switching will move into the kernel, so modes won't have to be switched when changing VTs and such, unless resolution actually changes.21:51
Keybukexactly21:52
Keybukgetty-gtk ;)21:52
wasabiusplash should leave it in the right res for GDM21:52
wasabigetty-gtk sounds awesome.21:52
wasabiHow far are we from all that?21:52
KeybukI've seen people actually using it now21:53
wasabiSeriously? Wow.21:53
Keybukwhich is a good step up from seeing people demo it21:53
Keybukso end of the year maybe before it's in shippable drivers21:53
Keybuk(where it = kernel mode setting/TTM/Gallium3D/etc.)21:53
sorenI've seen interfaces get renamed to eth?_renamed a few times before... What's that all about?21:55
wasabiThat's exciting stuff. Can't wait to see it all come together.21:55
Keybuksoren: interface renaming21:55
sorenKeybuk: No kidding? :)21:55
Keybuksoren: it's about renaming interfaces from one name to another :)21:55
Keybukso, the kernel is a bit thick21:56
sorenKeybuk: Oh. This is very enlightening.21:56
soren:)21:56
Keybukit has a counter21:56
sorenRight.21:56
Keybukevery time it sees a new interface, it increments that counter, and names it that21:56
sorenI realise why we rename them.21:56
Keybukso the first interface the kernel sees is eth021:56
Keybukthe second one is eth121:56
Keybukand so on21:56
sorenSure, sure.21:56
sorenI got that.21:56
sorenI also get the persistent naming and all that.21:56
Keybukunfortunately the kernel doesn't see them in any kind of consistent or useful order21:56
wasabiSo usplash starts, puts monitor in right mode, initramfs finishes quickly, upstart comes up, something can send job status text from upstart to usplash if we care, then X appears. No clicking or anything.21:56
Keybukso we have to rename them21:56
Keybukthe problem is when the kernel sees A=eth0 and B=eth121:57
Keybukbut we want B=eth0 and A=eth121:57
Keybukthe only way to resolve that deadlock is to rename them to something else21:57
sorenKeybuk: Ah.21:57
Keybukso udev names A=eth0_rename B=eth1_rename21:57
Keybukthen renames A=eth1 and B=eth021:57
sorenMakes sense.21:57
Keybuksometimes this comes a bit unglued when it tries to rename A and B to the same thing21:57
Keybukso you end up with A=eth0 and B=eth1_rename21:57
Keybukthis tends to happen with mac80211 devices which have both a wmaster0 and wlan0 interface21:58
sorenRight. This is in the installer with two identical NIC's (different MAC's though).21:58
Keybukand udev thinks they should both be wlan0 or eth121:58
Keybukit shouldn't ever try and rename two things with different MACs to the same name21:58
sorenThese are realtek 8139 nic's.21:58
Keybukaddress matching is on MAC21:58
ion_Obvious solution: create a program that changes the bytes directly in /dev/kmem.21:58
sorenYeah, I know.21:58
sorenHmm..21:58
Keybukwasabi: upstart starts in initramfs, hands over to upstart in real system21:58
wasabiWe're going to get it to that?21:59
sorenMaybe the peristent-names-generator is racy somehow.21:59
Keybuksoren: shouldn't be21:59
wasabiI saw your latest upstart release thing fly by. I was excited to see it coming along.21:59
sorenNo, persistent-net.rules looks fine.21:59
sorenHm...21:59
* soren scratches his head21:59
Keybukit happens *inside* the installer?21:59
sorenYeah.21:59
Keybukthat's not doing anything crazy like killing udev, right?21:59
sorenAh, sorry. I've booted into rescue mode on the server cd, but I belive I saw the same in the real installer.22:00
Keybukdo you have /var/log/udev you can nopaste for me?22:00
soren70-persistent-net.rules *is* odd, now that I think about it.22:00
sorenThe rules look like:22:00
Keybukcan you nopaste that too? :)22:00
sorenSUBSYSTEM=="net", ACTION=="add"; ATTR{type}=="1", NAME="eth0"22:00
Keybukthere's a ";" there?22:01
sorenMistyped.22:01
Keybukok22:01
sorenThe interesting bit is the lack of a mac.22:01
Keybukyeah that's VERY interesting :)22:01
Keybukthat means there was no address node in sysfs for that card22:01
Keybukis there one now? :)22:01
Riddellogra_cmpc: edubuntu dvds for beta?22:02
sorenKeybuk: Yep.22:02
Keybukweird22:02
Keybukcan you nopaste both for me?22:02
sorenQuite.22:02
sorenKeybuk: Do you have kvm capable hardware?22:02
Keybuksoren: how do I find out?22:03
sorenKeybuk: grep -E '(vmx|svm)' /proc/cpuinfo22:03
Keybukflags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow pni lahf_lm cmp_legacy22:03
sorenGreat.22:03
sorenThen you do. install kvm, and fire it up like so:22:03
seb128soren: I guess that virt-manager limiting cursors move to the wrong area until you move it on the right is a known issue?22:03
Keybuksoren: I do?22:03
Keybuksoren: neither of those things is in there22:03
mvoamitk: thanks!22:04
sorenKeybuk: Oh, I thought that was the output of the grep.22:04
sorenKeybuk: (the output would be the empty string if you didn't have them)22:04
Keybuk:)22:04
soren..so I didn't really look all that closely.22:04
Keybukno, I was mentally grepping and pasting in case I had anything else useful in there that you recognised22:04
Keybukin fact22:05
Keybuknone of my machines have either of those22:05
sorenShame.22:05
Keybukthis is an Athlon 64 X222:05
Keybukthe laptops are Intel Core Duo U2500 and Core 2 Duo T555022:05
Keybukthey don't have it either22:05
mvoogra__: do you handle moodle? I get a upgrade problem from stock dapper->hardy with it22:05
sorenKeybuk: Yeah, there are a few models of C2D's that don't. Those might be the ones :)22:06
sorenKeybuk: You want the udev log and the persistent-net.rules?22:07
sorenKeybuk: Was that it?22:07
Keybukyup, please22:07
LaserJockmvo: what kind of problem?22:07
sorenwtf... There's no /var/log/udev.22:08
crimsunLaserJock: keep in mind the delta between dapper and dapper-{security,updates}22:09
sorenKeybuk: I'm finding myself increasingly unable to construct coherent sentences. I need sleep :) I'll look at it tomorrow morning.22:09
crimsunLaserJock: namely, debian/{postinst,templates}22:09
mvoLaserJock: not sure yet, it seems to got stuck at a ucf prompt at dapper->hardy, I need to investigate further22:10
LaserJockcrimsun: sure. I was going to say that a lot has changed in moodle since dapper22:10
LaserJockmvo: ucf?22:10
crimsunmvo: dapper's stock moodle has that nasty debconf bit (i.e., fails to install)22:11
mvocrimsun: interessting, I have it in my dapper image, that indicates that it somehow got installed without failure22:12
mvoLaserJock: package urc: Description: Update Configuration File: preserve user changes to config files.22:12
mvoLaserJock: meh, make that "package: ucf"22:12
crimsunmvo: was moodle installed before or after apache?22:12
crimsunI vaguely remember discussing this with Martin a couple years ago22:13
ScottK2pitti: On cyrus-sasl2-heimdal, yes, today libdb-dev and libdb4.6-dev is the same, but there's no guarantee in the future.  Since cyrus-sasl2-heimdal needs the same version as cyrus-sasl2, we needed an ubuntu1 version in any case and so we might as well keep the better build-dep.  That's my view anyway.22:13
mvocrimsun: that is a very good question :)22:14
pittiScottK2: ah, right, that version identity; I see22:14
pittiScottK2: thanks for confirming22:14
pittiScottK2: btw, I just fixed that 'apport crashes on guidance' bug (I hope), in unapproved now22:14
mvocrimsun: hm, it seems like moodle is a fresh install, not a upgrade (brought in by edubuntu something)22:15
Riddellevand: gobuntu oversized :(22:16
evandARGH!22:16
ScottK2pitti: Great (I hope).  At least I'll know the truth then ..22:19
sorenKeybuk: No /var/log/udev in the proper installer either (I was in the installer's rescue mode before)22:20
Keybukthat's odd22:20
Keybukmaybe they don't make one22:20
mvocrimsun: I can not install moodle in a pbuilder chroot here, it hangs and the process list shows me there is something ucf releated going on (but no prompt or anything)22:29
mvo(not sure that I should talk to you about it :)22:29
mvoor is ogra a better person to ask here?22:30
mvohm, pressing "\n" helped22:30
* mvo scratched his head22:30
crimsunmvo: the last time I touched moodle was for dapper-updates22:32
crimsunso no, I'm probably not the best WRT hardy22:32
mvook, thanks crimsun22:33
LaserJockmvo: the newer packages were done by moquist22:33
mvoLaserJock: does he do irc?22:33
LaserJockin #edubuntu sometimes22:33
mvolast update in gutsy ... hmmm22:33
sorenKeybuk: Oh.. I think I may have an idea.22:34
LaserJockmvo: yeah, that's when we did a major repackaging and got it into Main22:34
mvoLaserJock: thanks a lot, I will look into it and see what I can do22:34
LaserJockmvo: we did a lot of changes to debconf stuff, it's possible something broke when we did so22:34
Keybuksoren: what is your idea?22:35
sorenKeybuk: I just noticed the persistent-net-generator.rules weeds out nic's with locally assigned mac addresses.22:36
sorenENV{MATCHADDR}=="?[2367abef]:*", ENV{MATCHADDR}=""22:37
moquistmvo: LaserJock clued me in. What's up?22:38
sorenKeybuk: Hm.. The macs start at 52:54:00:12:34:56 where the last octet is increased for each nic.22:38
sorenKeybuk: They should be alright.22:38
mvohello moquist! I was having trouble with moodle and was wondering if you might help, I can not install it in a pbuilder chroot, it hangs and only continues if I press "enter" (but there is no prompt)22:39
mvomoquist: it hanged in my auto dapper->hardy upgrade test, this is why I came to it22:39
mvomoquist: I run with DEBIAN_FRONTEND=noninteractive, so it might be something that is not easily triggered22:39
sorenKeybuk: Er... No. I'm on crack.22:39
sorenKeybuk: That *does* match that pattern, so the mac is cleared.22:40
sorenAnyhow.. Bed time.22:40
sorenLong over due.22:40
mvomoquist: I suspect its something to do with debconf and/or ucf - something like a leftover file descriptor or something equally crazy22:40
mvo(in the postinst)22:41
moquistmvo: You're working with a version of the package that I've never touched; I didn't work on it until gutsy, really.22:42
mvomoquist: this happens in a hardy pbuilder chroot too22:44
mvomoquist: but give me a bit, I'm looking into it - I *think* the call to dh_stop might be the problem22:45
moquistmvo: ah! And thanks for all the diagnosis.22:45
mvomoquist: so if I move handle_config before db_stop all seems to be fine. I have little clue about this package, so if you could double check that I'm not on crack, that would be much appreciated22:46
moquistmvo: I'm checking into it now.22:47
mvomoquist: great, thanks a lot!22:47
nxvli'm having a problem with Bug #20371022:49
ubotuLaunchpad bug 203710 in mysql-dfsg-5.0 "mysql-server-5.0 does not prompt for conffile update on upgrades" [High,Confirmed] https://launchpad.net/bugs/20371022:49
nxvli can't find out why is that it isn't recognizing the configfile as changed22:50
jdstrandnxvl: conffile22:50
Keybukis it a conffile or a config file?22:50
nxvlconffile22:50
Keybuksure?22:50
jdstrandKeybuk: it's a confile as seen in /var/lib/dpkg/status22:50
jdstrandhrm conffile22:50
Keybukjdstrand: what are the exact two packages involved here?22:51
Keybuk(URLs to .deb files, please)22:51
jdstrandKeybuk: well, the 'upgrade to' package was one I was working on, but I didn't change anything22:52
jdstrandKeybuk: with the conffiles22:52
nxvlKeybuk: http://paste.ubuntu.com/5837/22:52
jdstrandKeybuk: nxvl has some IIRC22:52
Keybukjdstrand: do you have the one you were working on and the one you had before?22:52
nxvli do it upgrading from gutsy to hardy22:53
jdstrandKeybuk: nxvl confirmed it with gutsy version to hardy version22:53
jdstrandKeybuk: if you give me a minute, I can fire up a gutsy vm and upgrade22:53
nxvli did it from: https://edge.launchpad.net/ubuntu/gutsy/+source/mysql-dfsg-5.0/5.0.45-1ubuntu322:54
nxvlto https://edge.launchpad.net/ubuntu/hardy/+source/mysql-dfsg-5.0/5.0.51a-3ubuntu222:54
nxvlusing pbuilder environment22:54
Keybuknxvl: i386?22:54
nxvlKeybuk: yep22:54
jdstrandKeybuk: mine was amd6422:54
Keybukwhich package?22:54
nxvlmysql-server-5.022:55
jdstrandnxvl: did you also see it with mysql-common (with my.cnf)?22:55
nxvljdstrand: with my.conf it asked22:55
Keybukd'ling...22:55
Riddellogra_cmpc: your dvds are nastily oversized22:57
mvolamont: this http://paste.ubuntu.com/5841/ is probably a bug for you, right?22:59
nxvljdstrand: it seems that it isn't recognizing it as a conffile -> http://paste.ubuntu.com/5843/23:03
ScottK2doko: If you're around I'd like to discuss python-xml transition for a moment?23:05
jdstrandKeybuk: just confirmed bug bites on a gutsy to hardy upgrade (both clean installs) on amd6423:15
jdstrandKeybuk: during install happened to get:23:15
jdstrand$ md5sum ./debian-*23:15
jdstrand7ac6cab7c0c2efaaaca2691b4ec3d6ec  ./debian-start23:15
jdstrand4272e4d740c8ae651ac35bbf4d2ed6dc  ./debian-start.dpkg-new23:15
jdstrandand of course /var/lib/dpkg/status had:23:16
jdstrand /etc/mysql/debian-start 4272e4d740c8ae651ac35bbf4d2ed6dc23:16
nxvljdstrand: how did you get that?23:17
nxvlgot*23:17
jdstrandnxvl: just lucky timing :)23:17
Keybukso you installed 5.0.4523:18
Keybukmodified debian-start23:18
Keybukthen installed 5.0.51a ?23:18
jdstrandnxvl: ran md5sum after unpack but before postinst23:18
Keybukand it didn't prompt?23:18
nxvlKeybuk: yep23:18
jdstrandKeybuk: yes23:18
jdstrandit's kooky23:18
nxvljdstrand: oh!23:18
jdstrandnxvl: -v23:19
nxvljdstrand: yes, i already did that23:20
jdstrandKeybuk: I did not know of a way to prevent dpkg from prompting on a conffile23:21
Keybukjdstrand: ?23:21
jdstrandKeybuk: I am simply puzzled by the behavior23:21
jdstrandKeybuk: I have not seen it before23:22
Keybuksomeone remind me how I disable debconf? :p23:22
mvoexport DEBIAN_FRONTEND=noninteractive23:22
Keybukthanks23:22
KeybukD000020: process_archive conffile `/etc/mysql/debian-start' no package, no hash23:23
Keybukok, so that's sensible on first install23:23
nxvlKeybuk: where did you get that output?23:25
Keybuknxvl: dpkg23:25
jdstrandnxvl: --debug23:25
nxvlthnx23:25
nxvl:D23:25
Keybuk-D77777 actually23:26
KeybukD000020: deferred_configure `/etc/mysql/debian-start' (= `/etc/mysql/debian-start') useredited=-1 distedited=-1 what=20423:26
Keybukok23:26
Keybuksensible on first configure23:26
KeybukI have edited that file23:27
KeybukConffiles:23:27
Keybuk /etc/init.d/mysql 4f0c573e38f141149bd19e4a929305b923:27
Keybukand edited to23:27
Keybuk347d7585c6ba7936e4d11ed0b1df8b0e  /etc/mysql/debian-start23:27
Keybukerr23:28
Keybuk /etc/mysql/debian-start 4272e4d740c8ae651ac35bbf4d2ed6dc23:28
* Keybuk pastes the right line from status23:28
dokoScottK: discuss what? just remove it ;)23:29
Keybukok, so now let's unpack 51a23:31
KeybukD000020: process_archive conffile `/etc/mysql/debian-start' package=mysql-server-5.0 same hash=4272e4d740c8ae651ac35bbf4d2ed6dc23:31
Keybukand install23:32
chadmillerHi.  I haven't made a new .deb in /years/.  What's the recommended tool these days to boostrap making a new package?23:33
chadmillerI mean, initialization with example debian/* files and such.23:34
ScottK2chadmiller: Packaging questions are best on #ubuntu-motu23:34
KeybukD000020: deferred_configure `/etc/mysql/debian-start' (= `/etc/mysql/debian-start') useredited=1 distedited=0 what=223:34
Keybuknxvl, jdstrand: I am seeing no bug here23:35
Keybukcould you be a little more specific about what you're expecting to happen23:35
Keybukand more importantly, *why* you're expecting that to happen23:35
chadmillerScottK2: Thanks.23:35
jdstrandKeybuk: why didn't it prompt on upgrade?  the file changed, the package installed a new version23:35
jdstrandscratch23:35
nxvlKeybuk: if i edited debian-start, dpkg should ask me if i want to keep the edited file or update it23:35
Keybukfile hasn't changed23:35
jdstrandist didn't install a new version, but should have prompted to install the new version23:36
nxvlKeybuk: and it's not asking, just ignoring it23:36
Keybukno it shouldn't23:36
nxvlso23:36
jdstrandKeybuk: it isn't the same as the old one, so it is non-default23:36
jdstrandgranted, it didn't overwrite it23:37
Keybukwhy should it prompt?23:37
nxvlwhat you ar saying is that as the packaged debian-start from old and new version are the same, dkpg shouldn't ask?23:37
Keybukright23:37
nxvls/ar/are23:37
jdstrandand I agree23:37
jdstrandit is that the file on the system changed and is like neither the old installed version, or the new23:38
nxvlso, if i change debian-start on the new package and try to install it, then it should promt, didn't it?23:38
jdstrands/new/new installed version/23:38
nxvljdstrand: yes, but as the 2 packaged conffiles are the same, there is no problem with the changes i have done to the file (if i have understood correct)23:39
nxvlKeybuk: didn't it?23:40
Keybukright23:40
jdstrandoh-- because the md5 of 5.0.45 and 5.0.51a were the same, we can keep the changes and be reasonably sure things are ok23:43
nxvljdstrand: yep23:43
jdstrandand that just happened to also be the case for the file I was working with23:44
jdstrandman, I feel like a dumbass23:44
jdstrandKeybuk: thanks for your time on this23:44
nxvli'm rechecking packing a changed version of debian-start, just to be sure23:44
nxvlwell, at least i learn a lot about packages wasting my time on this bug :D23:45
* jdstrand goes and makes a note of this subtelty23:45
* nxvl HUGS Keybuk and jdstrand 23:45
* jdstrand and appreciates dpkg even more23:45
jdstrandnxvl: I invalidated the bug23:47
jdstrandnxvl: thanks for your time23:48
jdstrandtoo23:48
nxvljdstrand: heh, i was doing it right now23:48
nxvl:P23:48

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!