[04:50] <ScottK> If there's a buildd admin around that could have a look at https://launchpad.net/ubuntu/+source/brasero/0.9.1-0ubuntu3~intrepid1 - The binaries have been stuck at 'accepted' for almost half a day now, so I think the system has done what it is going to do.
[06:25] <constantine> hey I'm new to ubuntu, where can I find out what gnome is and why I keep seeing it mentioned
[07:13]  * DBO pokes calc 
[08:22] <cjwatson> ScottK: the publisher's crashing again - similar symptom to a while back
[08:23] <cjwatson> pitti: ^- I thought we fixed the broken pkgstriptranslations that was causing that?
[08:24] <cjwatson> pitti: (glib2.0_*_i386.translations.tar.gz being invalid due IIRC to multiple pkgstriptranslations running at once, if you remember
[08:24] <cjwatson> )
[12:32] <apw> where are bugs in the alternative installer reported?
[12:40] <ogra> apw, launchpad :P
[12:41] <ogra> https://bugs.launchpad.net/ubuntu/+source/debian-installer/+filebug ;)
[12:47] <apw> so the image contents is debian-installer too, cool
[14:41] <calc> how do i turn ctrl-alt-backspace back on?
[14:42] <calc> gnome just froze up on me and using sysrq-k let the video in a weird state that c-a-b doesn't cause
[14:42] <calc> so sysrq-k is definitely not a good solution if you need access to your console ever again (before rebooting)
[14:43] <calc> bryce: ping? ^ :)
[14:43] <tjaalton> man xorg.conf
[14:43] <calc> aiui there is a command line tool that does it automagically
[14:43]  * calc knows it is called DontZap
[14:43] <ogra> apt-get install dontzap
[14:44] <calc> ah the command is called that too, ok
[14:44] <ogra> sudo dontzap -d is the command you want iirc
[14:44] <calc> ok
[14:44] <calc> i was thinking use of c-a-b was a bit uncommon myself, until gnome just completely died on me, heh
[14:45]  * calc rebooting to get his console back working
[14:45] <ogra> tjaalton, even with the vblank .drirc fix my i965 is darn slow ... i just switched on Legacy3D which makes it nearly usable though
[14:46] <ogra> seems forcing XAA doesnt work anymore even though its mentioned in the intel manpage
[14:48] <calc> i just disabled compiz to work around the extremely slow intel driver :-\
[14:50]  * calc imagines dontzap will be added to automatix or whatever it is being called now
[14:53] <directhex> http://www2.apebox.org/wordpress/wp-content/gallery/00-single/yay_for_evolution.png
[14:53] <directhex> ¬_¬
[14:53] <calc> directhex: i'm guessing those numbers are wrong or you are really evil ;-)
[14:54] <directhex> calc, the numbers are a smidge wrong
[14:54] <calc> lol
[14:54] <directhex> calc, always happens to me though - e.g. http://www2.apebox.org/wordpress/wp-content/gallery/00-single/gw001.jpg
[14:55] <calc> heh
[14:56] <ogra> directhex, send your computer back to math school ?
[14:57] <directhex> ogra, even my first mobile phone did that - insisted i had space for 65534 text messages on my sim card
[14:58] <ogra> well, evo isnt really written for mobile phones :P
[14:58]  * ogra hides
[15:04] <tjaalton> ogra: having it in ~/.drirc doesn't work, as mentioned on the bug
[15:05] <ogra> tjaalton, bug 320813 ?
[15:06] <tjaalton> yes
[15:10] <ogra> hmm, i did read it differently then :)
[15:46] <Amaranth> calc: automatix is gone, unless you know someone else started working on it?
[15:50] <directhex> Amaranth, isn't that a good thing?
[15:50] <Amaranth> ;)
[15:51] <ogra> directhex, depends whom you ask :)
[15:51]  * ogra guesses arnieboy would disagree *g*
[16:23] <ScottK> Actually I think they've morphed into something called ultimatix or some such, I don't recall.
[16:27] <Nafallo> ScottK: sounds right to me
[16:28] <ScottK> calc: If you want dontzap and a gui option to configure it, it's already done in Kubuntu ....
[16:29] <ogra> ScottK, do you plan to keep it ? (it was removed from ubuntu again)
[16:29] <ScottK> ogra: As far as I know we do.  All the controversy was about Gnome U/I.   AFAIK the KDE U/I is untirely uncontroversial.
[16:30] <ogra> well, i thought for consistency between the desktops ...
[16:30] <ScottK> I don't think KDE should remove choice from the user because Gnome thinks that's a good idea.
[16:30] <ScottK> I think it's the same for Kubuntu/Ubuntu.
[16:31] <ogra> gnome doesnt think anyting here, thats my point :)
[16:31] <ScottK> Given the philosophical differences between the two DEs I think the current situation is entirely reasonable.
[16:31] <ogra> it was an ubuntu decision, not a gnome one
[16:31] <ScottK> I'm just saying I think the difference is consistent with the differences between the two.
[16:32] <ogra> yup
[16:33] <SupernalTriad> !ops | SupernalTriad
[16:33] <SupernalTriad> ubuntu is the worst distro i've ever used
[16:33] <SupernalTriad> where do i file a bug report
[16:34] <ScottK> SupernalTriad: You probably should just switch to another one then.
[16:35] <ScottK> SupernalTriad: https://bugs.launchpad.net/ubuntu/+filebug
[16:38] <calc> ogra: heh yea KDE wasn't founded on taking away choice from users ;-)
[16:38] <ogra> calc, ubuntu wasnt either :P
[16:38] <calc> well actually that didn't start happening until gnome 2.0, but for 8yr+ now anyway
[16:39] <calc> ogra: heh, aiui this was a upstream change that we just left as is :)
[16:39] <ogra> right
[16:40] <ScottK> It does seem to be in line with what Ubuntu is specifically intending these days though.
[16:40] <ogra> i agree that its very annoying to not have zapping during development
[16:40] <ScottK> I think the dontzap U/I is one example and I think the idea that taking away actions on notifications is somehow good is another.
[16:40] <ogra> but i also agree that its not worth bothering my mother with a UI option in a stable released system
[16:40] <calc> ScottK: there is a fine line between taking away choice and having sane defaults
[16:40]  * calc is worried gnome doesn't know the difference
[16:41] <calc> not in this particular instance but still
[16:41] <calc> for the X case we should just get those bugs fixed, of course that is unlikely to happen IRL so i think disabling the feature was a bad idea
[16:41] <ScottK> Dunno.  I'm still seriously confused about the idea that if clicking on a notification would do something that makes them worse.
[16:42] <calc> there are so many things that can kill/hang X is why i think its not going to happen, heh
[16:42] <ogra> you have all options you can want available in gconf ... since such options are really mostly advanced tasks using something like gconf editor and having a clean config UI otherwose is a very wise decision imho :)
[16:42] <ogra> and the success of ubuntu over the last years somewhat proves that point right i think
[16:42] <calc> ogra: sometimes... my pet gweather rant can't be fixed by using gconf
[16:43] <ogra> thats a bug :)
[16:43] <calc> ogra: upstream just completely removed the ability to select weather stations in it
[16:43] <calc> ogra: but upstream closes it and refuses to do anything about it
[16:43] <ScottK> ogra: No.  Upstream says it's a feature.
[16:43] <calc> ScottK: exactly
[16:43] <ogra> fork it !! :)
[16:43] <calc> ogra: i would if i didn't already have accuweather in firefox
[16:44] <calc> ogra: i might fork it anyway if i get annoyed enough, but it makes weather in gnome useless now
[16:44] <ogra> works for me
[16:44] <Mithrandir> it's been useless for ages in lots of places since the weather in the airport does not match the weather in the city.
[16:44] <calc> ogra: the issue is that it only has one station (that you don't know which one) per 'city'
[16:44] <ScottK> Ultimately I think Gnome is trying optimize itself for a certain class of user.  If you're in that class, then great.  If not, oh well.
[16:44] <ogra> right, its only a guesstimate anyway
[16:45] <Nafallo> Mithrandir: LCY seems to match where I live quite well... but I'm kind of an exception, living two stops away :-P
[16:45] <ScottK> Time will tell if that fraction of the user base is larger or smaller.
[16:45] <calc> places like houston happen to be very large say 16K sqkm
[16:45] <calc> so only having one station that is on the other side of the city isn't very useful
[16:46] <calc> Mithrandir: now its just useless for a much larger group of users
[16:46] <ogra> ScottK, i assume its the majority, most users i know just want to use their computer and dont even know how to change the wallpaper (and dont attempt to)
[16:46]  * calc really didn't intend to bring this rant back up again
[16:46] <Mithrandir> Nafallo: OSL is about 50km away from where I live.. and I live in the somewhat-northern parts of Oslo.
[16:46] <ogra> calc, just have more airports :P
[16:46] <calc> there is a station about 10km away from my house and there is another one still in 'Houston' that is about 80km from my house
[16:46] <Nafallo> Mithrandir: hence why the weather changes a lot more... ;-)
[16:47] <Mithrandir> calc: that's a good thing, since it might be fixed then.
[16:47] <calc> Mithrandir: nope he refused to do anything and just closed the bug wontfix
[16:47] <Mithrandir> calc: what's your US zip code?
[16:47] <calc> 77014
[16:47] <Nafallo> Mithrandir: 1.6 miles from LCY with car ;-)
[16:47] <ScottK> ogra: Perhaps.  Time will tell.  Sometimes I sit down to help my kids with stuff on their computer and my eyes bleed because of the way they've customized it.  They like the choice.
[16:48] <ogra> right, i thin its also a matter of age
[16:48] <ogra> *think
[16:48] <calc> upstream wants to do some sort of position checking (gps?) to determine which to show the user without asking them
[16:48] <Mithrandir> hmm, yr.no doesn't seem to do do US zip codes.
[16:48] <ogra> many people using computers today still didnt grow up with using them on a daily base
[16:48] <ScottK> ogra: There is that.  I actually don't customize my KDE desktop much at all.  I find the defaults quite sane and usable.  Gnome OTOH, just feels totally wrong to me.
[16:49] <ScottK> Choice is a good thing.
[16:49] <calc> of course that wouldn't work very well since users would want to see the weather at their summer home, etc
[16:49] <ogra> if the generations switch it might totally change the numbers
[16:49] <SupernalTriad> how do you know if you are catering to the least common denominator
[16:49] <Mithrandir> ScottK: _can_ be a good thing.  Choice for choice's sake isn't.
[16:49] <ScottK> Mithrandir: Certainly.
[16:50] <calc> the main thing i change on my systems using gconf is focus under mouse
[16:50] <calc> since gnome won't let you do that via the gui
[16:50] <ogra> works for me
[16:50] <ScottK> That's a GUI option in KDE4.
[16:50] <Mithrandir> I've fixed that by using a real wm. :-P
[16:51] <Mithrandir> (and it really annoys me when people say "gnome" when they mean "metacity")
[16:51] <SupernalTriad> what do you do when you make things so simple to use that people never learn anything and never know how to fix something when your 'simplified system' breaks down
[16:52] <Mithrandir> you make it so they don't break?
[16:52] <SupernalTriad> you get 1500 users in your channel asking retarded questions
[16:52] <calc> Mithrandir: i'm pretty sure it does the same with compiz as well
[16:52] <SupernalTriad> Mithrandir, impossible with apt
[16:52] <calc> Mithrandir: in any case the sys->pref->windows won't let you do it :)
[16:52] <Mithrandir> calc: and?  I'm not using compiz, nor metacity.
[16:52] <SupernalTriad> it tries to do too much so that it always breaks something eventually
[16:53] <Mithrandir> SupernalTriad: you've yet to actually say what the problem is.
[16:53] <ogra> calc, i can only re-state ... works for me
[16:53] <SupernalTriad> Mithrandir, the problem is in creating a subset of linux that is not vanilla at all...where you are told that using anything but apt is a pita or will break stuff
[16:53] <calc> ogra: yea i remember your system is weird, other people couldn't determine why yours works that way
[16:54] <calc> ogra: the gconf description even says it shouldn't work the way it does for you :)
[16:54] <ogra> SupernalTriad, my mother is quite happy with her ubuntu since two years ... and se surely doesnt know what apt is
[16:54] <ogra> nor does she have any probs doing her daily tasks
[16:54] <calc> Mithrandir: well those are the two 'gnome' wm's aiui anyway, so saying gnome isn't really a misnomer
[16:54] <ScottK> SupernalTriad: Any distro will tell you that installing outside the distro packaging system is a risky thing that will break things.
[16:54] <SupernalTriad> ogra, least common denominator...when something goes bad, what do they do
[16:55] <Mithrandir> calc: sure it is.  That most distros default to compiz/metacity if you use gnome is just a historical artefact.  It used to be sawfish.
[16:55] <SupernalTriad> ScottK, not true
[16:55] <calc> SupernalTriad: install into /opt or /usr/local
[16:55] <SupernalTriad> ScottK, almost true, and i see that as a major problem
[16:55] <ogra> SupernalTriad, well, nothing went bad for her the last two years, but she has my phone nuber i think :)
[16:55] <Mithrandir> SupernalTriad: no, apt is not a PITA.   Maybe you want to say what the actual problem you are running into is?
[16:55] <calc> SupernalTriad: installing into /usr will break any *nix period
[16:56] <SupernalTriad> calc, i admit i do ./configure, make. make install for years and years....into /usr/ and havent broken it
[16:56] <ogra> calc, well, sloppy behaves identical to mouse for me ... but since it does what i wan i dont really care ;)
[16:56] <Mithrandir> SupernalTriad: uhm, you think that apt should be magic and see that you've overwritten parts of /usr with locally compiled stuff?  First, that's not apt at all stumbling at that, secondly, I think that's a completely unreasonable requirement.
[16:56] <cjwatson> if that works for you, that's fine. This is not relevant to #ubuntu-devel
[16:56] <SupernalTriad> just bloated it a bit...too lazy to keep track of whats installed and uninstallers / self made packages
[16:56] <directhex> calc, ?
[16:56] <SupernalTriad> i dont use any package manager
[16:56] <calc> SupernalTriad: you have been lucky then :)
[16:56] <SupernalTriad> because they always got in my way
[16:57] <SupernalTriad> i dont want to wait for some guy to tell me what version i can use
[16:57] <Mithrandir> SupernalTriad: then you're not using ubuntu, so I don't see why you are complaining about it.
[16:57] <SupernalTriad> i just grab the latest source
[16:57] <calc> directhex: huh?
[16:57] <directhex> calc, system, preferences, windows: "select windows when the mouse moves over them"
[16:57] <SupernalTriad> how do you guys deal with it
[16:57] <SupernalTriad> its so crippled
[16:57] <SupernalTriad> it stifles you
[16:57] <calc> directhex: sloppy, doesn't deselect window when you go to desktop
[16:57] <directhex> stupid troll is trolling
[16:58] <SupernalTriad> sure im trolling because you dont agree with me?
[16:58] <SupernalTriad> when i say apt/* stifles your learning
[16:58] <Mithrandir> SupernalTriad: that's irrelevant, you're completely off-topic for this channel.
[16:58] <ogra> calc, i raraely see my desktop ;) so i cant confirm or deny
[16:58] <calc> SupernalTriad: its equivalent to trying to install into C:\WINDOWS and hoping you don't blow up your machine, aiui microsoft actually keeps you from doing that now in some manner
[16:58] <cjwatson> Ubuntu is fundamentally a package-managed distribution and this is not going to change. You're welcome to use another distribution or from-scratch instructions or whatever if you prefer that
[16:58] <directhex> you're trolling because you're in #ubuntu-devel saying "man, isn't using a distro where the computer does the menial stuff for you so shitty"
[16:58] <SupernalTriad> i do
[16:58] <tjaalton> calc: focus-follows-mouse is configurable from the "windows" capplet
[16:59] <SupernalTriad> directhex, it tells you what softwaer you can use too
[16:59] <cjwatson> SupernalTriad: this is not relevant to #ubuntu-devel; please take the discussion elsewhere
[16:59] <SupernalTriad> so if its not in the repo, you are fucked
[16:59] <calc> tjaalton: but not focus-under-mouse, focus-follows-mouse is 'sloppy' and f-u-m is 'mouse'
[16:59] <directhex> waa, 20k packages isn't enough :'(
[16:59] <directhex> mummy, save me from the bad men who package things
[16:59] <SupernalTriad> directhex, 10k outdated packages
[16:59] <SupernalTriad> penis
[17:00] <directhex> go go gadget cjwatson
[17:00] <ogra> heh
[17:00] <tjaalton> calc: huh, maybe I just don't know the difference then :)
[17:00] <calc> tjaalton: apps/metacity/general/focus_mode is what i was referring to in gconf
[17:00] <calc> tjaalton: focus under mouse or strict (whatver you want to call it) unfocus a window when you leave it regardless of if you moved to the desktop or another window
[17:00] <directhex> calc, it's easier in gconf than in vista... gotta add some hex to a big long byte to get focus follows mouse
[17:00] <tjaalton> calc: ah, ok
[17:00] <calc> tjaalton: focus follows mouse 'sloppy' only changes focus when you go into another window
[17:01] <Mithrandir> directhex: or just use powertools or whatever it's called those days, I presume.
[17:01] <calc> directhex: yea or install xmouse (is that still available for vista?)
[17:01] <directhex> Mithrandir, no official free one for vista
[17:01] <tjaalton> yeah the key description clarified the difference
[17:01] <calc> directhex: iirc powertoys from microsoft let you set that
[17:01] <cjwatson> (probably a bit harsh to ban on first offence)
[17:01] <directhex> Mithrandir, only paid 3rd party equivalents
[17:01] <Mithrandir> cjwatson: /knockout in irssi is nice.
[17:01] <ogra> cjwatson, he was banned in #ubuntu before
[17:01] <calc> ah no tweakui for vista, ugh
[17:02] <directhex> ogra, almost as if he's a troll?
[17:02] <cjwatson> ogra: I rarely consider other channels in ban decisions here
[17:02] <directhex> calc, good innit
[17:02] <ogra> he just came here to go on ranting
[17:02] <cjwatson> Mithrandir: nice
[17:02]  * calc needs to summon pmladek from suse to fix OOo split build, heh
[17:02] <calc> i'm getting obscure error messages using it :\
[17:03] <directhex> i wonder if he'll reappear in #ubuntu-motu
[17:03] <IntuitiveNipple> Do we have anywhere, a specification or policy on what block-device/file-system combinations are supposed to be supported by Ubiquity? For example partition > lvm, partition >encrypted > block, partition > encrypted > lvm, partition > lvm > encrypted
[17:03] <calc> IntuitiveNipple: all of it? :)
[17:03] <directhex> nbd!
[17:04] <directhex> iscsi!
[17:04] <ogra> nbd ftw !
[17:04] <directhex> ooh, encryption on lvm on local AND iscsi disks
[17:04] <directhex> that'd be awesome
[17:04] <IntuitiveNipple> I'm trying figure out some 'bugs' in relation to these combinations, especially when cryptsetup is involved,  right now
[17:05] <IntuitiveNipple> I've fixed cryptsetup to cope with partition > lvm > encrypted > block
[17:05] <cjwatson> IntuitiveNipple: none of LVM, RAID, or encryption are supported by Ubiquity right now, only by d-i
[17:05] <IntuitiveNipple> And Ubiquity can run with that *if* lvm2 and cryptsetup have been added
[17:05] <IntuitiveNipple> cjwatson: Okay, well I was using "ubiquity" as the umbrella
[17:06] <cjwatson> IntuitiveNipple: we would like to fix this in the future but currently are unlikely to entertain bug reports on attempts to hack it up, since it's known to be unsupported
[17:06] <cjwatson> IntuitiveNipple: don't do that :)
[17:06] <cjwatson> if you say ubiquity, we will all interpret you as meaning specifically the desktop installer
[17:06] <calc> ah yea i got confused between the alternate and desktop cds
[17:06] <cjwatson> IntuitiveNipple: all the rest should be supported but it's entirely possible that there are problems with certain combinations
[17:08] <cjwatson> IntuitiveNipple: there was a bug in gutsy/hardy/intrepid regarding /etc/crypttab generation that will certainly have broken some cases; that much is fixed in jaunty
[17:09] <cjwatson> but otherwise detailed bug reports with full explanations and logs will help us figure it out
[17:09] <IntuitiveNipple> cjwatson: Yeah, I've posted some patches
[17:12] <cjwatson> IntuitiveNipple: with regard to your complaints over the last week or two about devices disappearing in ubiquity (you kept leaving before I could answer ...), I expect that this is due to the known problem that update-dev calls udevadm trigger
[17:13] <IntuitiveNipple> cjwatson: no... not a complaint, a confusion, and I found the solution and posted it to the bug report :)
[17:13] <cjwatson> the "solution" that it should notify the user of the problem?
[17:13] <cjwatson> that hardly seems like a real solution
[17:14] <cjwatson> if parted is failing to understand the disk, that needs to be dealt with regardless
[17:14] <cjwatson> and no, partman and cfdisk have no libraries in common above libc6
[17:15] <cjwatson> IntuitiveNipple: I think what would help me is a dump of the raw contents of the partition table(s), that I could feed to parted locally to reproduce the problem
[17:15] <IntuitiveNipple> bug #324987
[17:16] <calc> ah the bot is an example of why +text needs to continue to exist
[17:16] <cjwatson> yeah, looking at it, absolutely no idea why you marked it invalid
[17:16] <calc> at least until it is converted to the api
[17:16] <IntuitiveNipple> cjwatson: I found it easy to recreate the issue: create the partition table with fdisk -u and start each partition on the next availabe sector number... in 'cyclinder' mode the end and next-start will be in the same cylinder... and that was the root cause
[17:17] <IntuitiveNipple> cjwatson: I considered it "user error" - the missing partition entries... but I've left the "silently fails" bug #324976
[17:18] <IntuitiveNipple> cjwatson: Unfortunately, because the partition offsets are reported by partman in absolute byte offsets (not sectors or cylinders) the reason wasn't showing up in the partman log
[17:18] <cjwatson> I've reopened 324987
[17:20] <cjwatson> IntuitiveNipple: for 324976, one time-saver for us would be if you could run ubiquity in debug mode ('ubiquity --debug' from the command line, or 'debug-ubiquity' added to kernel boot parameters on the live CD) and post logs including /var/log/installer/debug
[17:20] <IntuitiveNipple> ok... I thought it'd get thrown out as a "won't fix" because it is a known issue with buggy fdisk
[17:21] <IntuitiveNipple> cjwatson: I thought I did... maybe I just used them for my own debugging... I recall they didnt't help :)
[17:21] <cjwatson> libparted should be raising an exception here, and it ought to come up as a debconf protocol message
[17:21] <cjwatson> didn't help *you* ;-)
[17:21] <IntuitiveNipple> cjwatson: I'll create a VM tomorrow with that scenario and capture the logs
[17:21] <cjwatson> thanks
[17:22] <cjwatson> it's presumably a real-world partitioning scenario that might arise in a variety of cases, and I consider failing to handle such to be bugs
[17:22] <IntuitiveNipple> cjwatson: my head felt like I was wading through treacle trying to trace the execution flow!
[17:22] <cjwatson> that's OK, the debug logs are for us not you :)
[17:22] <cjwatson> (primarily, anyway ...)
[17:22] <cjwatson> yes, it is difficult to trace at times
[17:23] <cjwatson> certainly I see that parted_server has crashed in your log, which probably invokes undefined behaviour in ubiquity
[17:23] <IntuitiveNipple> cjwatson: That's great... your response is gratifying especially after a patch I posted to fix a bug in pbuilder was rejected by the debian developer as an intended feature, not a bug... that causes pdebuilder to fail when building older releases :)
[17:23] <cjwatson> parted_server: exception_handler: Bad option: ""
[17:24] <cjwatson> that's odd ...
[17:26] <cjwatson> oh, meh, the partman-commit component in ubiquity doesn't catch partman/exception_handler
[17:27] <IntuitiveNipple> Is that the cause of the silent failure?
[17:27] <cjwatson> I would guess so
[17:28] <IntuitiveNipple> cjwatson: you're star! I'm glad someone has their head around that spaghetti code :D
[17:28] <cjwatson> wouldn't mind verification after we upload my change, mind you
[17:30] <IntuitiveNipple> If the package update is in the archive tomorrow I'll fetch it before running the test in the VM
[17:30] <IntuitiveNipple> If there's a diff in the bug report I can appy that directly
[17:31] <cjwatson> IntuitiveNipple: I wasn't planning an upload just yet, but how about http://bazaar.launchpad.net/~ubuntu-installer/ubiquity/trunk/revision/2999 if you're willing to give that a go?
[17:31] <cjwatson> http://bazaar.launchpad.net/%7Eubuntu-installer/ubiquity/trunk/diff/2999 might be easier to deal with
[17:31] <IntuitiveNipple> I'm guessing the change is in a python file or shell script ?
[17:32] <cjwatson> yes, /usr/lib/ubiquity/ubiquity/components/partman_commit.py
[17:32] <IntuitiveNipple> okay... that's easy to apply then
[17:32] <IntuitiveNipple> the server is down atm so I can't look at the launchpad diff
[17:33] <cjwatson> IntuitiveNipple: I've attached the diff to the bug
[17:34] <cjwatson> not that this will necessarily make the failure *informative* ...
[17:34] <cjwatson> you'll just get "Error informing the kernel about modifications to partition /dev/sda5 -- Device or resource busy.  This means Linux won't know about any changes you made to /dev/sda5 until you reboot -- so you shouldn't mount it or use it in any way before rebooting.
[17:34] <cjwatson> " with "Ignore" and "Cancel" or some such
[17:35] <IntuitiveNipple> that's better than the silent failure :)
[17:35] <cjwatson> I'm still not seeing any sign of any other libparted exception, which is extremely weird
[17:35] <cjwatson> and leads me back to my earlier udevadm trigger hypothesis
[17:36] <IntuitiveNipple> quick question on a slightly different topic - in the initrd, when custom crypto scripts are run, which shell hosts them?
[17:36] <IntuitiveNipple> I was testing some modifications to my crypto script that works in Hardy and Intrepid, and it throws some errors relating to not finding busybox... weird.
[17:37] <cjwatson> IntuitiveNipple: if you fancy doing another test after the partman_commit one, try deleting the "udevadm trigger" line from /bin/update-dev
[17:37] <cjwatson> IntuitiveNipple: shell> busybox sh (busybox/shell/ash.c)
[17:38] <IntuitiveNipple> okay, will try that... with VMs its easy and quick.
[17:38] <cjwatson> in the initramfs you must confine yourself to utilities available in busybox
[17:38] <IntuitiveNipple> cjwatson: Yes, it is. failing for use of 'basename'
[17:39] <IntuitiveNipple> I'll pin it down anyhow. My head is turning to mush right now I'm juggling too many inter-related issues
[17:39] <cjwatson> IntuitiveNipple: "$(basename "$x")"  ->  "{x##*/}"
[17:40] <cjwatson> where "x" must be a shell variable name
[17:41] <IntuitiveNipple> yeah... I need to reproduce it... it was, I think, caused by being passed a weird path from /sys/ in its hunt for USB devices
[17:41] <IntuitiveNipple> I've replaced it anyhow... I discovered that vol_id is hidden in /lib/udev/ :)
[17:42] <IntuitiveNipple> All the script wants to do is identify likely USB memory devices
[17:43] <cjwatson> that said, basename *is* available in jaunty's busybox
[17:44] <IntuitiveNipple> Yes, at the prompt I used it. never mind, the vol_id discovery negated the need for it
[17:48] <IntuitiveNipple> How to kill a system: type . instead of , in /etc/crypttab entries that get copied to initrd!
[18:19] <Riddell> calc: you pinged the other day?
[18:24] <calc> Riddell: yea, it looks like mandriva still depends on kdelibs even though they have --disable-kde set, so i emailed to find out the status, haven't heard back from them yet
[18:26] <Riddell> calc: ok, thanks, it's entirely possible the Mandriva openoffice guy hasn't done what Helio said he had done :)
[18:26] <calc> ok
[18:26] <calc> Riddell: they appear to be using ooo-build so if it actually works it shouldn't be too hard to get it for ubuntu
[18:28] <slangasek> superm1: no mythbuntu alpha-4 test results reported yet?
[18:30] <ogra> slangasek, how long do i have to get armel versatile netinstall in before its to late =
[18:30] <ogra> ?
[18:30] <calc> anyone happen to have a linksys WRT610N?
[18:30] <slangasek> ogra: too late for what?
[18:30] <ogra> slangasek, to get a delayed alpha4 :)
[18:31] <slangasek> ogra: hmm.  By this point there will be significant package skew relative to the other alpha-4 images - I would suggest focusing on alpha-5 at this point?  Or do you need an alpha published so that you can have others do more install testing?
[18:31] <ogra> it worked fine with the previous netboot d-i, but i didnt test the actual final one
[18:32] <ogra> slangasek, well, its netboot anyway, there is no guarantee for package consistency anyway for that i assume
[18:32] <slangasek> ogra: oh, netboot rather than netinst; then we don't actually copy those anywhere that we mark them 'alpha', anyway.
[18:32] <ogra> ah, k
[18:33] <slangasek> (I guess I should've known you meant netboot, since Ubuntu doesn't have 'netinst' images like Debian :)
[18:33] <ogra> err, no i meant netinst
[18:33] <ogra> but there is no difference, the d-i blob is the same
[18:33] <ogra> at least for qemu :)
[18:34] <ogra> qemu actually boots the kernel and initramfs directly (no tftp/dhcp needed here) but the d-i starting then is indeed pretty much identical to netinst
[18:35] <ogra> and since arm cant boot isos thats the only way for me to test
[18:35] <slangasek> er, please don't use the term "netinst" - that refers to a specific class of Debian CD images
[18:35] <slangasek> I'm pretty sure that's not what you have here
[18:35] <slangasek> you have "local netboot" or something :)
[18:36] <ogra> hmm, i always thought netinst is just netboot/install d-i components rolled into an iso
[18:37] <ogra> but then it was woody when i used one last :)
[18:43]  * ogra grumbles about nslu2 ... 
[18:58] <Adri2000> does anyone know what's the plan for samba in jaunty? has the server team already decided something? (slangasek?)
[19:10] <scientes> i cant figure out what module i need to build to active this one little change i made
[19:11] <scientes> i edited drivers/hid/hid-input.c to change the behavior of apple keyboard
[19:11] <scientes> but in my modules folder there is no hid-input.ko only hid.ko and then the usb folder
[19:17] <Keybuk> did you check the Makefile?
[19:17] <Adri2000> hi Keybuk!
[19:18] <Keybuk> hi
[19:19] <Adri2000> did you speak with kees about my latest changes to the comments patch?
[19:26] <scientes> sweet it worked
[19:26] <scientes> awesome
[19:26] <scientes> good i had a ps/2 keyboard
[19:27] <maxb> ooi, what's the latest on the crashing publisher?
[19:27] <scientes> static int hid_apple_fnmode = 2; should be the default
[19:33] <IntuitiveNipple> Is there a way to 'tell' the live-CD not to eject the CD when restarting?
[19:46] <Keybuk> Adri2000: I didn't know there was any further changes?
[19:46] <Keybuk> you should speak to kees directly about them
[19:47] <ogra> [42950525.820000] NOHZ: local_softirq_pending 100
[19:47] <ogra> Finding valid ramdisk creators.
[19:47] <ogra> Using mkinitramfs-kpkg to build the ramdisk.
[19:47]  * ogra scratches head ...
[19:49] <Adri2000> kees: see ^. looks like Scott isn't aware of the changes and your review
[19:49] <ogra> shouldnt update-initramfs be used by default if its installed ?
[19:56] <ogra> hmm, apparently not in debian armel kernels
[20:20] <calc> what is 'R' state for a process?
[20:20] <calc> a webpage wedged firefox into that
[20:20] <calc> and it seems to be non-killable
[20:20] <slangasek>        R    Running or runnable (on run queue)
[20:20] <slangasek> ps(1)
[20:20] <calc> so it should be killable then right?
[20:21] <slangasek> normally, yes
[20:21] <calc> is there a i really mean die argument to pass kill?
[20:21] <ogra> -9
[20:21] <calc> ogra: hmm yea that doesn't work
[20:21] <slangasek> check dmesg
[20:22] <slangasek> because -9 is "don't ask the process for permission, just obliterate it at the kernel level"
[20:22] <ogra> what process is it btw ?
[20:23] <ogra> geez, generating ssh keys on a nslug takes 20min
[20:24] <calc> shortly after it locked up hard
[20:24] <calc> sak didn't even kill X
[20:24] <calc> i had to sysrq sub to reboot
[20:24] <calc> i'll see if i have anything in logs
[20:25] <calc> interesting
[20:25] <calc> Feb  8 09:11:04 laptop-c2d kernel: [ 1511.744906] npviewer.bin[5890]: segfault at 6c2f6b6e ip 000000006c2f6b6e sp 00000000ff9a709c error 14
[20:25] <calc> then a dump
[20:25] <calc> Feb  8 14:19:15 laptop-c2d kernel: [20002.740004] BUG: soft lockup - CPU#0 stuck for 61s! [firefox:12682]
[20:25] <slangasek> that's all?
[20:25] <calc> full call trace as well
[20:25] <slangasek> oh, 'soft lockup'
[20:26] <calc> wtf is that?
[20:26] <slangasek> that's a non-diagnostic error message (it can have any number of causes), but it pretty much explains why you couldn't kill the process
[20:26] <calc> ok
[20:26] <slangasek> any chance that firefox itself died with a SEGV, and apport tried to catch it
[20:26] <slangasek> ?
[20:27] <calc> i didn't see apport doing anything visible
[20:27] <calc> i tried kill -9 a few times on firefox
[20:27] <calc> then tried to run another program and it locked too
[20:28] <calc> couldn't switch to a vt, and sysrq k failed, so i just did a sysrq sub to get back
[20:30] <Blacksun> Nabend an alle
[20:33] <Blacksun> Ich hätte mal ne frage zu ner phyton Datei, und zwar wollte ich abc torrent installieren, doch irgend wie bekomme ich das nicht hin hab schn mit abc.py probiert geht aber nicht
[20:33] <Blacksun> hat jamand neine lösung wie sowas zu installieren ist ?
[20:33] <Tm_T> Blacksun: olen samaa mieltä
[20:33]  * ogra grins ...
[20:34] <ogra> Blacksun, normalerweise spricht man hier nur englisch ;)
[20:34] <Tm_T> Blacksun: det är rolight
[20:34] <Tm_T> ogra: in how many languages you like me to comment on this ?
[20:34]  * ogra guesses Blacksun wanted to go to #ubuntu-de instead ;)
[20:34] <Blacksun> hups da hab ich mich im raum vertan
[20:35] <ogra> Tm_T, try one that has no vowels ;)
[20:36] <Tm_T> bah, too late
[20:38] <ogra> so, why cant i login with normal users to a freshly bootstrapped system ... i get motd, no error message and get dropped back to the login prompt immediately
[20:39] <ogra> and apparently it doesnt seem to matter if i'm on a serial console or try a ssh session
[20:39] <cjwatson> ogra: "netinst" specifically refers to Debian CD images that include d-i and the Debian base system - as slangasek says, we have no such images in Ubuntu
[20:39] <ogra> no indication in auth.log
[20:39] <ogra> i can create brandnew users as well ... same behavior
[20:40] <ogra> cjwatson, ah, right, i forgot about the base system ... i was thinking i remembered its just all the udebs in one iso
[20:41] <cjwatson> that's called "businesscard" in Debian
[20:41] <ogra> ah, right ... :)
[20:42] <ogra> removing the rootpw from /etc/passwd gets me in as root on the serial console though
[20:43] <ogra> homedir permissions seem ok ... filesystem isnt full, disks are mounted rw ... i really start to run ut of ideas
[20:43] <ogra> *out
[21:00] <Adri2000> slangasek: do you know what's the status of samba in jaunty?
[21:14] <Renfield> Where can I get the md5sum for the mini.iso images?
[21:17] <cjwatson> I don't think we publish it anywhere that comes to mind
[21:17] <cjwatson> you could use rsync to make sure you have a good coppy
[21:17] <cjwatson> copy
[21:39] <ogra> root@nslu2:/home/ogra# sudo -i -u ogra
[21:39] <ogra> Killed
[21:39] <ogra> mumble
[21:40] <ion_> Ouch
[21:40] <ogra> yeah, i dont get what it is
[21:40] <ion_> strace?
[21:41] <ajmitch> process limits?
[21:41] <Renfield> cjwatson: How would I do that?
[21:41] <Mithrandir> Renfield: rsync -c
[21:41] <Renfield> To what host/directory?
[21:42] <ogra> ion_, ...
[21:42] <ogra> rt_sigaction(SIGTSTP, {SIG_DFL}, NULL, 8) = 0
[21:42] <ogra> execve("/bin/bash", ["-bash"], [/* 15 vars */]) = -1 EACCES (Permission denied)
[21:42] <ogra> very intresting
[21:42] <ion_> Huh. :-)
[21:43] <ogra> root@nslu2:/home/ogra# ls -l /bin/bash
[21:43] <ogra> -rwxr-xr-x 1 root root 695904 2009-02-08 20:02 /bin/bash
[21:43] <ogra> heh
[21:43] <ogra> all fine int seems
[21:43] <ogra> *it
[21:44] <Mithrandir> Renfield: rsync -cavP archive.ubuntu.com::ubuntu/dists/jaunty/main/installer-amd64/current/images/netboot/mini.iso mini.iso
[21:45]  * ogra starts to get the feeling he hunts phantoms
[21:48] <ogra> hmm, identical result with dash
[21:49] <Renfield> Oh, great, that just overwrote my mini.iso, not verify that it is correct.
[21:50] <Renfield> I'm not trying to install jaunty. I'm trying to install 8.10.
[21:50] <eut> lol :/
[21:51] <Renfield> I guess I'm just going to have to give up on Ubuntu. It just will not install.
[21:52] <Mithrandir> Renfield: well, you need to adjust the path for your distro and arch.  I thought I wrote that, but apparently I didn't.
[21:53] <Renfield> No big deal. I can find the one I am using. I am just so frustrated at this point with the whole thing.
[21:54] <eut> yeah i know what you mean
[22:18] <cjwatson> Renfield: a corrupted mini.iso is far from the only possibility where installation failures are concerned. What are the symptoms?
[22:18] <cjwatson> eut: you too, if you have problems
[22:24] <Renfield> cjwatson: I get errors during the "install base system".
[22:24] <Renfield> It fails to download various packages.
[22:24] <alex-weej> Amaranth: you still do compiz work?
[22:29] <cjwatson> Renfield: that's very unlikely to be a problem with the mini.iso then; if it were, you wouldn't have got that far
[22:29] <cjwatson> Renfield: perhaps you could put the full syslog on paste.ubuntu.com or similar (or file a bug report); you can extract the log using "save debug logs" from the installer main menu
[22:30] <Renfield> Ok, I'll do that after it errors out.
[22:30] <alex-weej> can someone explain to me how to get a fix sponsored?
[22:30] <alex-weej> i've opened a bug
[22:30] <alex-weej> and i know the code change
[22:30] <alex-weej> i just have no idea how to make it so that a main developer can just go "yep, that's ok", click a button and have it in main
[22:30] <alex-weej> apparently it's possible
[22:31] <cjwatson> alex-weej: subscribe ubuntu-main-sponsors or ubuntu-universe-sponsors depending on the package
[22:31] <cjwatson> it will be taken care of from there
[22:31] <alex-weej> cjwatson: but what do i need to upload?
[22:31] <cjwatson> patch
[22:31] <alex-weej> what kind of patch?
[22:32] <cjwatson> unified diff
[22:32] <alex-weej> ah, someone's pointed me to https://wiki.ubuntu.com/SponsorshipProcess
[22:32] <alex-weej> cjwatson: is this integrated with PPA somehow?
[22:33] <cjwatson> no
[22:33] <alex-weej> aww
[22:33] <cjwatson> it might be nice if it were, but really all any competent developer needs is a unified diff
[22:33] <cjwatson> you will see some stuff about debdiffs and what-have-you but that is gravy
[22:34] <cjwatson> if the package is in bzr then you can create a branch from it and submit a merge request
[22:34] <directhex> mmm gravy
[22:34]  * directhex hands cjwatson a mug of bovril
[22:34] <cjwatson> then the sponsor-side process is 'bzr merge lp:~user/package/branchname'
[22:34] <cjwatson> (oh and resolve conflicts, bzr commit etc.)
[22:35] <alex-weej> basically i want people to be able to use and test my package easily
[22:35] <alex-weej> so i need to have it in my PPA
[22:35] <cjwatson> you are entirely welcome to put it in a PPA as well
[22:35] <cjwatson> it is not essential to the sponsoring process
[22:36] <alex-weej> ok, just saying it is kind of useful in case you had any leverage over the design of it all :P
[22:36] <alex-weej> it would be cool if the debdiff i upload could just automatically create packages
[22:36] <alex-weej> for testing
[22:36] <cjwatson> integrating PPAs better for review purposes is actually already on our long-term roadmap
[22:36] <lfaraone> Hi, I had an idea: would a file handler for RPMs that automagically alienates them into a deb and hands off the job to gdebi (with a gui "converting to Deb" page) be useful for jaunty+1?
[22:37] <alex-weej> lfaraone: the fact that any RPMs work at all is just luck. not all linuxes are equal.
[22:37] <cjwatson> we'd like it to be something like: (1) commit to personal bzr branch (2) hit button to build package from that in PPA (3) submit merge request which includes a reference to said PPA
[22:37] <lfaraone> alex-weej: Of course.
[22:37] <lfaraone> alex-weej: But LSB packages at least should work, right?
[22:37] <directhex> lfaraone, i'd call that actively harmful, personally
[22:38] <lfaraone> directhex: Why's that?
[22:38] <alex-weej> because they're never designed for ubuntu
[22:38] <alex-weej> if they were, they'd be Deb format
[22:38] <directhex> lfaraone, because people would use it by & large for completely incompatible packages which would serve only to break things
[22:38] <directhex> lfaraone, if it goes into /usr, it should be for ubuntu. perhaps for debian at a push.
[22:40] <lfaraone> directhex: The user will install the RPM whether there's a GUI or not, they'll just call up their "linux geek friend" to run alien from the command line.
[22:41] <lfaraone> directhex: We _should_ be at least supporting LSB packages.
[22:41] <calc> are we back to the same crap as earlier today?
[22:41] <calc> are LSB packages allowed to write somewhere other than /opt ?
[22:41] <calc> if so that should be a bug
[22:41] <directhex> lfaraone, if you can identify LSB ones, then fine. but every report i get from people trying to run suse mono on ubuntu, i'll hand them your mail address
[22:42] <calc> 3rd party stuff according to FHS (unless it changed) is only allowed to write into /opt
[22:42] <lfaraone> directhex: Please do. I use filters. :)
[22:42] <Renfield> cjwatson: http://paste.ubuntu.com/115796/
[22:44]  * cjwatson scratches his head at that log
[22:44] <cjwatson> Feb  8 21:47:51 debootstrap: Setting up libldap-2.4-2 (2.4.11-0ubuntu6) ...
[22:44] <cjwatson> Feb  8 22:37:42 debootstrap:  libcurl3-gnutls depends on libldap-2.4-2 (>= 2.4.7); however:
[22:44] <cjwatson> Feb  8 22:37:42 debootstrap:   Package libldap-2.4-2 is not installed.
[22:45] <Renfield> I don't know what is going on.
[22:46] <cjwatson> oh, you restarted base system installation, that explains the confusion there
[22:46] <Renfield> I restarted a couple of times.
[22:47] <Renfield> But I only did that each time after wiping the hard drive. So, I expected that it would be smart enough to start cleanly.
[22:47] <Renfield> By wiping, I mean I went back to partion, and let it reformat.
[22:47] <lfaraone> directhex: But honestly, how many suse mono bugs do you get like that?
[22:47] <cjwatson> didn't actually wipe it from the looks of things, but anyway, that's really just a side issue
[22:48] <directhex> lfaraone, on irc? a few people asking about alien for that specific purpose. they're told where to shove it
[22:50] <cjwatson> Renfield: oh, actually it's not a side issue at all, you appear to have been given an explicit warning by the installer the second and subsequent times you ran base system installation to the effect that you were trying to install to an unclean partition and that this was liable to break
[22:50] <lfaraone> directhex: doesn't apport know not to reportbugs on alienated packages?
[22:51] <cjwatson> Renfield: I'm not sure why the *first* run failed, though
[22:51] <cjwatson> Feb  8 21:26:14 debootstrap: Setting up base-files (4.0.4ubuntu2) ...
[22:51] <cjwatson> Feb  8 21:26:14 debootstrap: chown:
[22:51] <cjwatson> Feb  8 21:26:14 debootstrap: invalid user: `root:root'
[22:51] <Renfield> cjwatson: Yes, it did say that. I then canceled the action, then went to the partition section and reran that part.
[22:51] <cjwatson> Renfield: apparently unsuccessfully
[22:52] <Renfield> I can just reboot and do this all over again.
[22:52] <cjwatson> Renfield: might be more reliable to reboot and go through it again ... yes, what you said
[22:52] <Renfield> This is the second reboot already.
[22:52] <Renfield> Ok.
[22:52] <cjwatson> none of this affected the first run that gave the chown error above, though
[22:56] <Renfield> Ok, this time I am not going to add any additional installer components to the list and see if that changes anything.
[22:56] <Renfield> Previously, I was adding SSH client/server.
[22:57] <cjwatson> Renfield: I note that the installer seems to have had some trouble retrieving its own components over the network
[22:57] <cjwatson> Jan  2 05:04:18 anna[6157]: DEBUG: retrieving partman-ext3 52ubuntu3
[22:57] <cjwatson> Jan  2 05:07:27 anna[6157]: wget: cannot connect to remote host (91.189.88.31): Connection timed out
[22:57] <cjwatson> that sort of thing
[22:57] <cjwatson> also:
[22:57] <cjwatson> Jan  2 04:40:03 net-retriever: gpgv: key 437D05B5 was created 85069852 seconds in the future (time warp or clock problem)
[22:57] <cjwatson> maybe some kind of clock problem affecting things?
[22:58] <Renfield> Hmmm.
[22:59] <Renfield> I wonder.
[22:59] <Renfield> I configured it to use my own ntp server. Maybe it didn't work, but neglected to tell me that it didn't work.
[22:59] <Renfield> This time when I get to setting the clock, I will use the defaults.
[23:06] <cjwatson> none of the explanations above are really satisfying to me, but they're the best I can do on a Sunday evening :-/
[23:06] <Renfield> Oh, when is the best time to get you? :)
[23:06] <cjwatson> European working hours
[23:10] <Renfield> Ah, still bombs.
[23:10] <Renfield> Warning: Couldn't download package debianutils
[23:31]  * calc thinks he found the current issue with his split build, need to learn to read rpm spec files better ;-)
[23:32]  * calc begins to realize how hacky the split build still is :-\
[23:36] <directhex> calc, split build? you don't enjoy 10 hours taken to test 1-line patches?
[23:39] <calc> directhex: heh
[23:40] <calc> directhex: well testing patches part won't be helped at least with the early rev of split build
[23:40] <calc> at least i don't think it will be that much
[23:40] <calc> we really need Sun to do the split properly for everything
[23:40] <calc> i'm here to get them to at least split languages properly
[23:41] <calc> here being in Hamburg
[23:45] <slangasek> Adri2000: sorry, "Status" how?
[23:48] <LaserJock> bah, I keep mixing up git and bzr :/
[23:48] <Adri2000> slangasek: is it going to be updated to a more recent 3.2 version or even 3.3?
[23:49] <jsmidt> ping james_w
[23:49] <slangasek> Adri2000: you're probably better off asking the server team; but unless I've overlooked something, 3.3 isn't really a series that we should be shipping in a stable release
[23:52] <Adri2000> slangasek: ok. I'm asking because I'd like to backport the patch for samba bug #5906 to hardy, and it's applied in version 3.2.6 whereas jaunty only has 3.2.5. so the jaunty version should be updated first
[23:53] <Adri2000> ubottu: you failed
[23:53] <slangasek> Adri2000: the jaunty version should be updated first, or should be updated in lieu of patching?
[23:53] <slangasek> oh, "hardy", ok
[23:54] <slangasek> well, I don't see any reason for us to not push a newer version of samba 3.2 to Debian unstable, really
[23:54] <Adri2000> because afaik the sru policy requires the bug to be fixed in the development release before fixing it in a stable release
[23:54] <slangasek> yes, that's correct
[23:55] <Adri2000> if we're going to wait for the package to be updated in debian unstable, that would mean waiting for lenny to be release wouldn't it?
[23:56] <directhex> in theory next week for that
[23:56] <directhex> (ha!)
[23:57] <Adri2000> but that would be only a few days before our feature freeze
[23:58] <directhex> i think i'm going to be filing a few FFE. stupid debian NEW is kicking my ass
[23:59] <slangasek> Adri2000: no, there's no reason it needs to be delayed until the Debian release