[12:09] <mdz> doko: for example?
[12:10] <doko> mdz: python-2.4.2, OOo-2.0.2, ...
[12:14] <mdz> doko: it makes sense to have them on the radar, but we won't make a decision until they're actually released and we know what is entailed
[12:24] <Kamion> daniels: xkb/console> no
[12:24] <Kamion> doko: live CDs are produced daily
[12:25] <Kamion> doko: if you're asking about Flight CD releases, you keep asking me this, but I'm sorry, no. In general I'll try to announce a few days in advance when I want to get one out.
[12:25] <Kamion> if I haven't said something, I generally don't have one planned.
[12:26] <doko> Kamion: ok, just still trying to find the OOo build failure on powerpc :-/
[12:27] <Kamion> I still have some flight 1->2 installer regressions to fix, not to mention a giant pile of ubuntu-express code to write
[12:27] <Kamion> as a rough guide, maybe near the end of next week? that's about the best I can do for you
[12:28] <doko> Kamion: just wanted to know, when I have to be ready not to block others ;)
[12:37] <daniels> Kamion: http://lists.alioth.debian.org/pipermail/pkg-kbd-devel/2005-September/000004.html http://lists.alioth.debian.org/pipermail/pkg-kbd-devel/2005-October/000013.html
[12:37] <daniels> Kamion: i'm thinking about specing it up, because the console keymaps are shit
[12:40] <jdub> "Supporting basicly dead chips is a waste of resources." <- on ubuntu-devel, *not* in that thread
[12:51] <ijuz> is there an easy way to deactivate the validation of packages in the debian-installer?
[12:51] <daniels> Kamion: (that would mean that the d-i keymap stuff could just pick an xkb map, and then everyone could stfu about differences in the installer and x map; the only problem would be an incorrect map)
[12:53] <Seveas> ijuz, why would you EVER want that?
[12:54] <ijuz> Seveas: i'm experimenting with using a different compression for the packages and therefore the md5 sums changed
[12:54] <Seveas> then update the md5sum lists :)
[12:55] <elmo> oh, that reminds me, I need to post about 7zip compression, it's kinda scarily good
[12:55] <ijuz> hehe...
[12:55] <ijuz> really good
[12:56] <daniels> elmo: actually it's far from the best
[12:56] <daniels> elmo: pales in comparison to dpkg v3's algorithm
[12:56] <elmo> daniels: har har
[12:56] <daniels> the downside is that you need flash and a jvm in the installer
[12:57] <Robot101> dpkg v3? :P
[12:57] <jdub> elmo: you'd pimp 7z for dapper+1?
[12:57] <daniels> doogie ftw
[12:57] <elmo> seriously, 7zip-ing the entire archive would save 2.3Gb _per-arch_ in breezy
[12:57] <ijuz> elmo: full 7zip isn't so easy
[12:57] <daniels> jdub: 7z for d+1 would require 7z support in d?
[12:57] <jdub> wow
[12:57] <elmo> err, actually not per-arch strictly speaking because  it includes arch: all, but still
[12:57] <jdub> daniels: urgh, yeah
[12:57] <elmo> and it's better than bzip2 in cpu and memory too, IIRC
[12:57] <elmo> I did stats and graphs and stuff
[12:58] <jdub> elmo: sshot!
[12:58] <daniels> elmo: are they BLING?
[12:58] <jdub> elmo: blog!
[12:58] <ijuz> 7zip (only lzma) breezy install CD is: 245601 extents written (479 MB)
[12:58] <daniels> elmo: i want the graph in chrome
[12:58] <elmo> ijuz: why not?
[12:58] <ijuz> lzma is easy, i'm just doign this
[12:59] <ijuz> you first have to compile all the 7zip stuff and get it working in debian-installer
[12:59] <jdub> elmo: actually, while i've got you, how doable are the firewall changes for humboldt->rookery (http)?
[12:59] <ijuz> elmo: did you try the "executable" encoding?
[01:00] <elmo> jdub: oh, that's find
[01:00] <elmo> ijuz: I just did '7z a' :P
[01:00] <elmo> jdub: fine
[01:01] <ijuz> elmo: that doesn't give you all the possible benefit, so you can as well use just lzma
[01:01] <jdub> elmo: doable soon without being in the dc?
[01:01] <elmo> jdub: yeah, I'll do it in a bit
[01:01] <jdub> thanks!
[01:02] <jdub> ELMO SAVES PLANET FROM UNCOMFORTABLE DOOM-LIKE SITUATION!
[01:02] <ijuz> one cool thing that brings extra benefit is that 7zip can "transcode" binaries platform specific so that they are better compressable
[01:02] <daniels> jdub: starring THE ROCK as James Troup
[01:03] <jdub> haw haw haw haw
[01:03] <jdub> is zakame a member?
[01:03] <jdub> oh, he has @ubuntu
[01:07] <slomo> ijuz: how does this "transcoding" work?
[01:08] <ijuz> slomo: it reorders the code somehow, i'll look it up
[01:09] <slomo> ijuz: but how is it possible that the compression ratio is better when doing some magic for some platforms? or do you mean (un)compression speed on that specific platform?
[01:09] <jdub> jsgotangco: ping
[01:10] <jsgotangco> jdub, pong, good morning
[01:10] <jdub> yo!
[01:10] <jdub> jsgotangco: you're happy with your current planet subscription?
[01:10] <jsgotangco> jdub, sure it seems to work unless you want me to change it
[01:11] <jsgotangco> should i send you my hackergotchi?
[01:11] <psusi> god I love simple fixes
[01:11] <jdub> jsgotangco: url?
[01:11] <psusi> spent several hours reading hundreds of lines of kernel code to finally solve the problem by adding two lines
[01:11] <whiprush> jdub: pls add mpt to planet!
[01:12] <jdub> whiprush: have to get a request from him, but i'll ping him about it
[01:12] <ijuz> slomo: the author doesn't get specific, but says it's very fast, you have to tell the compiler the architecture for this, so no idea how it reacts to !binary files
[01:13] <jsgotangco> jdub, http://69.60.114.104/planet/images/heads/jsgotangco.png
[01:13] <Seveas> jsgotangco, finally :)
[01:14] <ijuz> btw. debootstrap doesn't write the logfile to /target/var/log/debootstrap.log *sigh*
[01:14] <jsgotangco> heh..i finally graduate from looking like a chess pawn!
[01:14] <jsgotangco> (with a jdub-like beard)
[01:35] <jdub> tseng: ping
[01:35] <tseng> jdub: pong
[01:35] <jdub> tseng: what's up with your blog?
[01:35] <tseng> its gone
[01:35] <jdub> ~brandon doesn't exit
[01:35] <jdub> exist
[01:36] <tseng> i was using my friends server
[01:36] <tseng> and the guy kept getting owned over and over
[01:36] <jdub> "smarter" IT solutions...
[01:36] <tseng> i moved some stuff
[01:36] <tseng> i havent been motivated to blog since
[01:36] <jdub> got a new location?
[01:36] <tseng> linode baby
[01:37] <jdub> url me
[01:37] <tseng> i setup typo, but theres no content
[01:37] <jdub> oh
[01:37] <tseng> ill write you an entry
[01:37] <tseng> and find a feed
[01:37] <jdub> ok, i'll update the url, and you'll reappear when planet is moved
[01:37] <tseng> moved?
[01:37] <robertj> I have the world's most getto hosting plan. No backup, $12.13/year with MySQL
[01:38] <jdub> tseng: changing hosts
[01:38] <jdub> for great justice, etc.
[01:38] <tseng> oh
[01:39] <tseng> i have a good idea for a post
[01:39] <whiprush> jdub: kamion's last entry doesn't appear to be on planet btw.
[01:39] <daniels> tseng: post about malibu and coke
[01:39] <daniels> the wonder drink
[01:40] <jdub> whiprush: he's never been on it :)
[01:40] <whiprush> oh
[01:40] <whiprush> well dang dude.
[01:40] <jdub> (he actually requested it a few days ago though, so i assume his last entry is interesting)
[01:40] <whiprush> yep
[01:41] <daniels> p.d.o, yo
[01:41] <jdub> whoa, i have't read pdo for ages
[01:41] <whiprush> since I'm being a planet nazi lately, I think it would be cool if you blogged about the new X crack daniels. :)
[01:41] <jsgotangco> add mjg59 for p.u.c spice!
[01:42] <daniels> whiprush: what new X crack?
[01:42] <daniels> it's just an extension of the old
[01:42] <daniels> anyone who's been associated with ubuntu since any time after june 2004 will know exactly what's going on ;)
[01:42] <daniels> i was harping on about it then, and even had packages
[01:42] <whiprush> daniels: we masses want to know about this XGL tarball and how it won't even come close to making dapper. You know, so we don't get our hopes up. :p
[01:43] <daniels> why would I lie to the masses?
[01:43] <whiprush> word.
[01:43] <daniels> HiddenWolf: dude, where I am at the moment, I have numerous excellent bags of single-origin beans, a grinder, and a nice espresso machine
[01:43] <whiprush> daniels: I think an X update would be cool.
[01:43] <daniels> this is grating quite badly against my no-caffeine policy
[01:44] <HiddenWolf> daniels, :)
[01:44] <whiprush> daniels: I mean, a blog entry about X, not an X update as in, software.
[01:44] <daniels> whiprush: i'll try to fire one off that doesn't piss p.fd.o and p.d.o off
[01:44] <daniels> heh
[01:44] <jdub> whiprush: "no you can't have a pony" only ever comes after "can i have a pony", not before.
[01:44] <daniels> i'm working on an x update
[01:44] <daniels> jdub: PONIES FOR ALL
[01:44] <HiddenWolf> jdub,can I have a pony?
[01:44] <psusi> daniels, aren't you the kernel maintainer?
[01:44] <tseng> psusi: erm.
[01:44] <HiddenWolf> psusi, no, benc is
[01:45] <psusi> ohh... he around?
[01:45] <HiddenWolf> we could put x in the kernel of course. ;)
[01:45] <whiprush> jdub: nah, I just want to know where the ponies lie in the field. No need to attempt at owning one ... just admire from afar or somesuch.
[01:45] <tseng> jdub: http://blog.brandonhale.us/
[01:45] <daniels> HiddenWolf: you laugh, but we're working on that
[01:45] <psusi> I've been working on https://wiki.ubuntu.com/PacketCD and found two bugs in the kernel and patched them... one in the pktcdvd driver and one in the udf filesystem
[01:45] <HiddenWolf> daniels, ehm, unk!
[01:45] <tseng> daniels: i threw the malibu and coke bit in for you
[01:46] <psusi> so I need to know what to do to get the patches into the next dapper kernel
[01:46] <psusi> they are very small and simple
[01:46] <daniels> HiddenWolf: well, the bits that need to be, anyway.  like stomping all over the PCI bus.
[01:46] <daniels> psusi: file bugs
[01:46] <psusi> hrm.... file a bug and attach the patch to it?
[01:46] <daniels> yeah
[01:46] <HiddenWolf> daniels, hm. Cool.
[01:46] <whiprush> there's a kernel mailing list also
[01:46] <BenC> psusi: just email them to me
[01:46] <whiprush> or do what ben says. :)
[01:47] <BenC> btw, when is the great bugzilla->malone conversion going to happen?
[01:47] <psusi> oh?  a kernel mailing list?  hrm... I should get on that
[01:47] <HiddenWolf> psusi, ubuntu-kernel list
[01:47] <BenC> I'm starting to get bug reports in malone, and bug maint for the kernel is hard enough without dealing with dual bug systems
[01:47] <daniels> BenC: november 2004
[01:47] <psusi> what's wrong with bugzilla?  I like it
[01:48] <BenC> I don't
[01:48] <BenC> I hate bugzilla
[01:48] <psusi> why?
[01:48] <HiddenWolf> malone is cute, if they ever get the interface right.
[01:48] <BenC> because it doesn't have an email gateway
[01:48] <seth_k|lappy> I demand a way to change the bug and add a comment in ONE STEP via malone, without using the e-mail interface :P
[01:48] <BenC> so far debbugs is still my favorite bug system, despite its short comings
[01:49] <daniels> seth_k|lappy: word to that
[01:49] <BenC> seth_k|lappy: file  a bug on malone :)
[01:49] <BenC> seriously, they love that stuff
[01:49] <BenC> but I think we already discussed that particular problem at UBZ
[01:50] <seth_k|lappy> alright. I'll grok the buglist to make sure it's already there
[01:50] <psusi> ok... so... patch to benc, or ubuntu-kernel?
[01:50] <BenC> bcollins@ubuntu.com
[01:51] <psusi> ok
[01:51] <BenC> and it's kernel-team@lists.ubuntu.com
[01:51] <BenC> for future reference
[01:51] <psusi> I'll have to sign up for that one...
[01:51] <BenC> psusi: know anyone that has tested DVD+R DL writing?
[01:51] <BenC> my new powerbook has a DL
[01:51] <psusi> BenC, nope...
[01:52] <HiddenWolf> BenC, I should have, on x86
[01:52] <psusi> my drive supports it but I've never bothered ot get any dvd media
[01:52] <BenC> guess I need to get some disks and try it out
[01:52] <BenC> now I can copy DVD's without losing the bonus material! :)
[01:57] <infinity> BenC: I burned some dual layer DVDs on my i386 laptop just fine.
[01:57] <infinity> BenC: Can't recall if it was +R or -R, though, since I have a +/- drive, and I don't pay much attention to what media I buy.
[01:57] <BenC> infinity: nice
[01:58] <BenC> I have a +/- driver too, but for DL, it only supports DVD+R
[02:06] <jsgotangco> jdub, pia online?
[02:10] <jdub> jsgotangco: not atm
[02:16] <psusi> BenC, ok... bombs away... should I also mail the diff to the lkml?  or just let you review it first?
[02:24] <floam> psusi: sending it to LKML, if anyone notices it, will get you some very critical feedback
[02:24] <floam> they don't seem to mind hurtin feeling :)
[02:24] <floam> hurting
[02:24] <psusi> aye...
[02:24] <psusi> well, at least it's only 6 lines of code to fix 4 bugs ;)
[02:25] <psusi> can't be TOO much wrong with it right? :)
[02:41] <floam> I had no idea what your patch was
[02:41] <floam> , psusi 
[02:41] <psusi> huh?
[02:42] <floam> I had no idea what your patch was, psui
[02:42] <psusi> ohh... right... fixes a few bugs with pktcdvd and udf
[02:42] <floam> if it's that small it should be fine
[02:43] <psusi> you don't remember how to get emacs to insert a hard tab rather than spaces to the default indent level do you?
[02:45] <floam> C-q tab
[02:45] <psusi> thanks
[02:45] <floam> C-q is the 'quote' key; it quotes whatever you type next.
[02:45] <psusi> ahhh, good to know
[03:51] <BenC> anyone know why nautilus wont view my webdav folder (davs://)?
[03:51] <BenC> just says try another viewer
[03:57] <infinity> BenC: Kernel bug, clearly.
[03:58] <BenC> so obvious, I can't believe I missed it :)
[03:58] <infinity> :)
[04:56] <whiprush> daniels: awake?
[05:07] <daniels> whiprush: ?
[05:10] <whiprush> daniels: hey how long has xorg looked at ~ for a xorg.conf?
[05:12] <whiprush> I totally just wasted 45 minutes thinking my X was busted, so I've been grepping the log for EE. Turns out one of my old backup xorg.conf's in ~ was being read first.
[05:12] <infinity> Is it '~', or '.' it's looking in?  I don't recall.
[05:12] <elmo> oh, yeah, most annoying feature ever
[05:12] <whiprush> I know that seems like expected behavior, but man, that kind of sucks.
[05:12] <infinity> Anyhow, it's done it for a long while.
[05:13] <whiprush> infinity: ~
[05:14] <whiprush> didn't know X did such a thing.
[05:19] <daniels> whiprush: yeah, that shits me off too
[05:20] <whiprush> cool, as long as I'm not the only one. :)
[05:21] <daniels> i'm tempted to just smash the established behaviour in favour of behaviour that makes bloody sense
[05:23] <infinity> It would probably make perfect sense if it was a dotfile instead (~/.xorg.conf)
[05:23] <whiprush> agreed 
[05:24] <whiprush> that would at least follow a convention of sorts.
[05:25] <infinity> And be less prone to blunders, and give less ~ clutter if you actually WANT a user-defined config.
[05:25] <infinity> And make your hair smell nice.
[05:25] <infinity> Etc.
[05:25] <whiprush> heh
[05:27] <dilinger> xorg gives my hair volume, bounce, and shine!
[05:28] <daniels> i don't want user-defined configs
[05:28] <daniels> not while it's suid
[05:29] <daniels> i want the search path to be $(sysconfdir)/X11/xorg.conf, that's it
[05:29] <daniels> ideally the thing wouldn't have to *exist*
[05:29] <daniels> but that comes later
[05:30] <whiprush> it's hard to think of a use case where a normal user account would need a specific xorg.conf
[05:33] <psusi> infinity, hey... what was your real name again?  I was looking for someone the other day and can't remember who now... but someone said their nick was infinity
[05:33] <infinity> psusi: That's what /whois is for. :)  (Adam Conrad)
[05:34] <psusi> ahh.... hi... now why was I looking for you? hrm...
[05:34] <whiprush> psusi: he's got a cool hat 
[05:35] <psusi> infinity, ahh, right... I have a bug that has been pending since breezy was released for dmraid support... it's currently assigned to you for some reason, I guess because you maintain mkinitramfs
[05:35] <psusi> infinity, could you glance at the attached scripts to see if they look ok, and reassign the bug to the dmraid package sometime?
[05:36] <psusi> bug #15897
[05:37] <infinity> Probably little need for me to glance at all.  The bug definitely belongs to dmraid.
[05:37] <psusi> I think that the idea is that those packages are supposed to go into the dmraid package rather than mkinitramfs... once that's done, next task is to get the package moved to main and added to the seed I think
[05:37] <psusi> s/packages/scripts
[05:38] <infinity> dmraid-in-main (and in the installer, oh my) are entirely different issues, but initramfs integration would certainly be nice.
[05:38] <infinity> I believe fabbione was maintaining (or fiddling with) dmraid in Ubuntu.
[05:38] <psusi> aye...
[05:38] <infinity> If you're looking for a scapegoat to harass and reassign your bug to. :)
[05:38] <psusi> yes, please do ;)
[05:39] <psusi> I understand he has such hardware, but doesn't have it working for some reason... works fine for me though and 3 or 4 others who have spoken to me
[05:39] <infinity> I just bought a dmraidable controller recently, but I dunno if I'll have time to do anything fun with it.
[05:39] <psusi> it sure is nice to boot directly from a raid0 between two 10,000 rpm drives ;)
[05:39] <psusi> and be able to dual boot with winxp no less ( not that I ever use it any more )
[05:40] <psusi> but I've been hoping for dapper to be able to install normally instead of having to debootstrap from the livecd
[05:41] <psusi> though... I guess if dapper is going to have a combined install/livecd, an apt-get install dmraid is easy enough
[08:23] <pitti> Good morning
[08:28] <desrt> hello.
[08:28] <desrt> how are you?
[08:31] <Keybuk> hey desrt
[08:31] <desrt> hey.  haven't heard from you in a while.
[08:32] <desrt> i actually had a question you could have helped with the other day... someone was here asking about how hotplug works
[08:32] <pitti> hi desrt 
[08:32] <desrt> i was able to explain it all except for one point (which i told him my guess)
[08:32] <desrt> does udev call modprobe with some tag like usb-id-1234-5678 and modprobe does the lookup?
[08:32] <Keybuk> yes
[08:32] <Keybuk> pretty much exactly like that
[08:33] <desrt> that's really nice.
[08:33] <Keybuk> the tag comes from the kernel and is known as a "MODALIAS"
[08:33] <desrt> sweet
[08:33] <desrt> udev and the kernel both get to be oblivious
[08:33] <desrt> i love that :)
[08:33] <Keybuk> and the lookup table comes from the modules themselves (they have alias config options, use modinfo on one) and is stored in modules.alias by depmod
[08:34] <desrt> oh wow.  big file.
[08:34] <Keybuk> descent scott% modinfo tg3
[08:34] <Keybuk>   :
[08:34] <Keybuk> alias:          pci:v000014E4d00001644sv*sd*bc*sc*i*
[08:34] <Keybuk> alias:          pci:v000014E4d00001645sv*sd*bc*sc*i*
[08:34] <Keybuk>   :
[08:34] <Keybuk> etc.
[08:34] <desrt> i assumed that it parsed all of the modules/modules.*map files each time
[08:34] <Keybuk> nah, that's how we used to do it in breezy
[08:34] <desrt> oh well
[08:34] <desrt> this is really neat :)
[08:35] <Keybuk> in breezy, the kernel provided a bunch of information about the device in various environment variables; we then had a shell script that collected those up, called grepmap to find the modules using the modules.*map files, and then loaded the modules
[08:35] <Keybuk> the dapper way is MUCH easier
[08:35] <desrt> indeed.
[08:35] <Keybuk> ENV{MODALIAS}=="?*", RUN+="/sbin/modprobe -Q $env{MODALIAS}"
[08:35] <desrt> seems like there might be some benefit to pre-parsing/caching all of the modprobe.d files and the alias map into a single quick binary format...
[08:36] <Keybuk> possibly, though I expect most of the time is spent on string matching, rather than file parsing
[08:37] <desrt> right.  i mean a hash or a tree of some kind
[08:37] <Keybuk> that'd mean you'd need to know the format of the modalias, no?
[08:37] <Keybuk> the nice thing currently is that it's just a random string
[08:38] <desrt> i mean take all of the modprobe files and make a hash table (or such) of all of the aliases
[08:38] <Keybuk> you can stick "alias desrt-cd-* mod_foo" in there, and modprobe desrt-cd-wibble, etc.
[08:38] <desrt> oh.
[08:38] <desrt> wildcards in alises?
[08:38] <desrt> no thx :)
[08:38] <Keybuk> right
[08:38] <Keybuk> that's how they work
[08:38] <desrt> oh man.  ok.  much harder problem.
[08:38] <Mithrandir> nah, not really.  You can do glob trees.
[08:38] <desrt> true.
[08:39] <Mithrandir> it's just a state machine.
[08:39] <Keybuk> the module declares something like usb:v0499p100Ed*dc*dsc*dp*ic*isc*ip*
[08:39] <desrt> and it looks like for most of these aliases here the beginning part is almost unique
[08:39] <Keybuk> and the device has a much longer string
[08:39] <Keybuk> desrt: often the first couple of fields are vendor/device ids ... but not always
[08:39] <desrt> so you could do alookup of the part before the first * real quick then do a quick-check for if it is a real match
[08:40] <Mithrandir> Keybuk: does it do best-match or first-match?
[08:40] <Keybuk> Mithrandir: all matches
[08:40] <Mithrandir> ah, ok
[08:40] <Keybuk> if a device matches multiple module aliases, all of the modules are loaded
[08:40] <desrt> actually
[08:40] <infinity> For better or worse.
[08:40] <Keybuk> so you can get usbmouse, usbhid, usbcore, etc. for a device
[08:40] <Keybuk> and evbug, which we blacklist :p
[08:40] <desrt> if we were going for maximum speedup we could add modprobe/insmod functionality to udevd :)
[08:41] <Mithrandir> we don't blacklist usbmouse?
[08:41] <Keybuk> desrt: that would be the way to do it, have udevd parse and cache modprobe.d and modules.alias, and watch with inotify for changes
[08:41] <Keybuk> pretty much like we do for rules.d
[08:41] <Keybuk> Mithrandir: shush, I was examplifiying
[08:41] <Mithrandir> heh
[08:41] <desrt> Keybuk; i filed a bug about loading modules against alsa recently.  don't suppose that's anything you're involved with?
[08:42] <Keybuk> what was your bug?
[08:42] <desrt> the order in which the sound modules are loaded needs to be changed
[08:42] <infinity> Ooo.
[08:42] <infinity> Keybuk: This should make you happy:
[08:42] <infinity> The patch creates sysfs device links for athX interfaces and changes
[08:42] <infinity> the ARP type of the wifiX interface to ARPHRD_IEEE80211. This is
[08:42] <infinity> needed by tools like NetworkManager or SUSE's ifup for proper
[08:42] <infinity> operation, and hopefully fixes tickets #70 and #205.
[08:42] <desrt> infinity; ship new fglrx!!!!
[08:42] <infinity> desrt: s/new/fixed/ you mean?
[08:42] <desrt> infinity; yes.  fixed would be new :)
[08:42] <infinity> desrt: Working on LRM right now (hence the above quote)
[08:43] <Keybuk> pitti: btw, I have a /etc/udev/rules.d/libsane.rules ... why?
[08:43] <desrt> Keybuk; http://bugzilla.ubuntu.com/show_bug.cgi?id=21801
[08:43] <infinity> Keybuk: Of course, that's in the madwifi-ng branch, which we've not switched to because it scares me (we're still shipping madwifi-old, cause, well, it works, and I don't have atheros hardware to test the new stuff on)
[08:43] <Keybuk> infinity: I don't get what that does?
[08:44] <pitti> Keybuk: oh, ouch - it should be 45-libsane.rules, right?
[08:44] <Keybuk> sounds like they just fixed madwifi-ng to actually show up the card as a wifi card
[08:44] <desrt> infinity; i have an atheros -and- a powerpc if you need testing on anything
[08:44] <Keybuk> pitti: yes, and it shouldn't call /etc/hotplug.d/usb/libsane.hotplug, cause that doesn't exist? :p
[08:45] <pitti> Keybuk: I'll ask Mithrandir about it; he did the change and he is able to test it (no scanner here)
[08:45] <Keybuk> in fact, it still ships things in /etc/hotplug which it should not do
[08:45] <desrt> oddly i think the atheros is in my coat pocket...
[08:45] <Mithrandir> Keybuk: probably because I suck, then.
[08:45] <infinity> Keybuk: You were complaining that your atheros and NM didn't get along.  I'm assuming from that log message that (with madwifi-ng), they should.  <shrug>
[08:45] <pitti> Mithrandir: oh, heh - I think the RUN rule can safely be removed
[08:45] <Keybuk> infinity: nope, that won't fix it
[08:45] <Keybuk> allegedly madwifi-ng fixes it out of the box anyway
[08:46] <infinity> Ahh.
[08:46] <pitti> Mithrandir: it seems to set mode and group just fine
[08:46] <Keybuk> the problem is that the madwifi driver doesn't do "background scanning"
[08:46] <Keybuk> if you instruct the card to scan for new networks, it has to leap off the one its on
[08:46] <Keybuk> not good when n-m does a scan every 13 seconds :p
[08:46] <infinity> Keybuk: Well, if I can get a good response on a call for testing, I'll happily switch to -ng.  I'm just not keen on shipping code I can't test and finding out that it's broken after we press CDs. :)
[08:46] <Keybuk> infinity: give me a package and I'll test away
[08:47] <Keybuk> I'm Atheros-encumbered
[08:47] <Burgundavia> infinity, several of the laptop testing people of atheros, including myself
[08:47] <Mez> grr - what was the command to find out which package a file belongs to 
[08:47] <desrt> infinity; here's your response :p
[08:47] <Keybuk> Mez: dpkg -S
[08:47] <StevenK> dpkg -S
[08:47] <infinity> Alright, cool.  I'll roll up -ng in the next day or two, then.
[08:47] <StevenK> (Or dlocate)
[08:47] <Keybuk> do they have different module names, or did they helpfully use the same ones?
[08:48] <Mez> hmm
[08:48] <Mez> apparently not found
[08:48] <Keybuk> what's the file?
[08:48] <infinity> Keybuk: Looks similar enough.
[08:48] <Burgundavia> infinity, post a call on the laptop-testing mailing list
[08:48] <Mithrandir> pitti: I can't really test it here, since I need a patched sane-backends, as I need the genesys backend.  I was going to fix that at some point in dapper, but ENOTIME so far.
[08:48] <Keybuk> infinity: bah; if they used different, we could've shipped both and just blacklisted one :p
[08:48] <Mez> Keybuk, /usr/bin/qmake
[08:48] <infinity> Keybuk: I can do that anyway.  <shrug>
[08:49] <StevenK> Mez: Is it a symlink into /etc/alternatives ?
[08:49] <Keybuk> http://packages.ubuntu.com/cgi-bin/search_contents.pl?word=qmake&searchmode=searchfiles&case=insensitive&version=breezy&arch=i386
[08:49] <infinity> Keybuk: I already do something similarly evil with nvidia/nvidia_legacy (which are both originally "nvidia")
[08:49] <Mez> StevenK, nvm- I found out I had to use dpkg -S qmake not dpkg -S /usr/bin/qmake
[08:50] <infinity> Keybuk: Given the changes between the two, though, I'd rather just switch wholesale to -ng if it works for most people.
[08:50] <Mez> oh wait - yes it is :D
[08:50] <Burgundavia> infinity, -ng doesn't work for USB atheros devices
[08:50] <infinity> Keybuk: ath_pci has never worked for EVERYONE anyway, so maintaining a smilar status quo is fine by me.
[08:50] <Keybuk> infinity: yeah, gimme a package and I'll see if it works for me
[08:50] <infinity> Burgundavia: The one we ship now does?
[08:51] <Burgundavia> infinity, not certain, don't own one, but I imagine that is implied
[08:51] <Keybuk> no, it doesn't
[08:51] <Keybuk> grep usb:.*ath /lib/modules/*/modules.alias
[08:51] <infinity> Burgundavia: I thought it never did, and USB was on the -ng radar.

[08:51] <infinity> Burgundavia: They mention is as a deficiency, not a regression.
[08:52] <Burgundavia> infinity, ah, ok
[08:52] <sivang> morning all
[08:53] <infinity> Keybuk: After tomorrow's LRM release, I'll put up parallel LRM builds that are identical, but use -ng instead of -old, and ask for ath testing.
[08:54] <sivang> woof, OO upgrade
[08:54] <desrt> Seveas; ping
[08:54] <sivang> yay, evolution back installable
[08:55] <Keybuk> infinity: cool
[08:56] <infinity> Ooo, and we get sparc64 support with -ng
[08:56] <Keybuk> man, I need some of those magic tablets that kick-start your day :-/
[08:56] <infinity> Which I'm sure no one will test or use.
[08:57] <Keybuk> mmm, tea
[09:00] <Mez> tea :D
[09:00] <Mez> and cherry coke
[09:00] <Mez> and cigarettes and mints! and hotdogs
[09:02] <Keybuk> heh
[09:02] <Keybuk> I had coffee, orange juice, bran flakes and a banana
[09:03] <sivang> the banana is the good part, it provide lots of energy to jump start the day
[09:05] <lifeless> Keybuk: on a diet ?
[09:06] <Amaranth> good stuff
[09:06] <Keybuk> lifeless: no, just my usual breakfast
[09:07] <Amaranth> and it would be good for me, if i didn't add 2-3 tbsp of sugar :P
[09:07] <Keybuk> I find anything larger sends me to sleep, and anything less leaves me asleep :p
[09:09] <Mez> is there  away to find out whether theres a file in a package if I dont have the package installed on my system
[09:09] <CarlFK> mez - I use http://packages.ubuntu.com
[09:09] <lifeless> Keybuk: ah ;0
[09:09] <Keybuk> Mez: use the web interface
[09:09] <Mez> It has a search files?
[09:09] <Keybuk> lifeless: and I am vaguely trying to get fit again this yeear
[09:13] <Amaranth> Mez: apt-file?
[09:45] <Mez> is it bad of me to be using initng ?
[09:48] <Keybuk> yes, you fool
[09:49] <Mez> why ? It makes my boot up time nice ;d
[09:49] <Mez> and is actually working quite well
[09:50] <Keybuk> because you're running scripts at the same time that are designed to follow each other
[09:51] <Keybuk> mounting the root filesystem before it's been fsckd, etc.
[10:19] <pitti> mjg59: sorry, I couldn't review pam_foreground yesterday any more, my head exploded (I still suffer from a cold)
[10:19] <pitti> mjg59: I'm reviewing it now
[10:20] <pitti> mjg59: shouldn't the package activate the module in pam somehow (conffile/postinst)? if not, where do you intend to do that?
[10:30] <Keybuk> pitti: meh, could you start an audit of libsepol for me
[10:30] <Keybuk> when I pointed out it wasn't needed, RH's response was to provide the missing call into it
[10:32] <pitti> Keybuk: could you please create an initial report for it? QA and bug research, etc.
[10:32] <pitti> Keybuk: I'll do the source review and security history research
[10:32] <pitti> Keybuk: but it should be in the MainInclusionQueue so that I don't forget
[10:33] <Keybuk> yeah, it's just going to block further udev uploads
[10:33] <Keybuk> including the one I need to do today
[10:33] <pitti> Keybuk: btw, could you make sense of that ppc hang?
[10:33] <Keybuk> yes, today's upload would fix that, if it wasn't going to block on libsepol :p
[10:36] <Keybuk> do you want historic CVEs?
[10:36] <pitti> Keybuk: I can do that myself, but if you want do look for that, sure :)
[10:36] <Keybuk> (not that there are any, but I thought I'd check)
[10:36] <pitti> Keybuk: I'm just interested about how vulnerable a new package was in the past, to see how much of a burden it will be
[10:40] <Keybuk> https://wiki.ubuntu.com/MainInclusionReportLibSepol
[10:40] <pitti> Keybuk: are you fine with the package's quality in general?
[10:40] <seb128> pitti: did you and Keybuk discussed floppy?
[10:40] <pitti> Keybuk: i. e. does it do what it's supposed to, does it have many upstream bugs, an active upstream, etc.?
[10:40] <seb128> pitti: any reason to have /dev/fd0 640 instead of 660? It makes gfloppy unhappy, it wants writing to format
[10:40] <pitti> seb128: no?
[10:41] <seb128> Keybuk said you requested that change, so to ping you about it :)
[10:41] <pitti> seb128: hm, we keep plugdev devices r/o for the user, but I don't particularly mind
[10:41] <Keybuk> pitti: no idea, haven't looked at it; it's part of the great selinux beast
[10:41] <pitti> seb128: 0640 would be consistent to USB sticks and the like, and I quickly discussed that ages ago with mdz
[10:41] <seb128> floppy group members should be allowed to format a floppy no?
[10:42] <Treenaks> pitti: also when we want to format USB sticks?
[10:42] <Keybuk> write access to a device allows formatting
[10:42] <Treenaks> pitti: people keeps asking that too
[10:42] <pitti> seb128: if we want to allow that, then we should allow plugdev members to format their usb sticks, too, IMHO
[10:42] <seb128> fine with me
[10:42] <pitti> with me, too
[10:42] <pitti> it shuold just be consistent
[10:42] <seb128> k, so please do it 660 consistant :)
[10:42] <pitti> the problem with that is that it would allow users to circumvent permissions on e.g. ext2 drives
[10:43] <seb128> how?
[10:43] <Kamion> only if they're in the group that owns the disk device
[10:43] <pitti> seb128: well, of course you can already read the whole device, so it's not a strong point
[10:44] <jk_work> mjg59: Is "seamless" wireless wpa support on the agenda?
[10:44] <jk_work> comments to https://bugzilla.ubuntu.com/show_bug.cgi?id=16551
[10:44] <pitti> it's mainly how we want to consider usb/fw devices - entirely untrusted, or we want to enforce r/o file permissions on file systems that support them
[10:44] <jk_work> point out that wpasupplicant is in universe
[10:45] <crimsun> jk_work: I've already fixed that in Dapper.
[10:45] <Keybuk> pitti: meh, don't worry about it actually; false alarm, was a different bug in selinux, it doesn't need libsepol
[10:45] <crimsun> jk_work: it now uses ifupdown
[10:45] <jk_work> nice
[10:45] <pitti> Keybuk: that means you don't need the package at all, or it isn't from selinux?
[10:45] <Keybuk> don't need it at all
[10:45] <crimsun> I need to consider the case where there are multiple wireless interfaces, but the single case is handled already
[10:45] <Keybuk> well, not yet anyway
[10:45] <pitti> Keybuk: cool :)
[10:48] <jk_work> crimsun: will wpa support be included in ubuntu proper, or will it remain in universe?
[10:48] <crimsun> jk_work: that I don't know
[10:51] <Keybuk> pitti: *shrug* if you're in plugdev, you theoretically have physical access to both disk and machine
[10:52] <Keybuk> is formatting a disk a root operation?
[10:52] <Treenaks> Keybuk: not for floppies
[10:52] <Keybuk> Treenaks: why not?
[10:52] <Keybuk> why are floppies treated differently to USB keys
[10:52] <Keybuk> if I'm not root, should I be able to format a floppy?
[10:53] <pitti> Keybuk: I don't understand - why does plugdev membership imply physical access?
[10:53] <Keybuk> pitti: because it's the access you give people so they can plug things in? :p
[10:53] <Keybuk> if they don't need to plug it in, they don't need plugdev
[10:53] <pitti> Keybuk: if you use an USB drive on e. g. a kiosk system, then the user can't write onto the raw disk
[10:53] <pitti> Keybuk: ok, true
[10:53] <Keybuk> I wouldn't give the user plugdev to solve that, they don't need to write to the raw disk
[10:53] <Kamion> pitti: has anyone asked you to look at notify-daemon?
[10:54] <pitti> Keybuk: such things are corner cases anyway, and as I wrote above, I'd be fine with 066
[10:54] <Keybuk> 660 you mean? :p
[10:54] <Kamion> pitti: I'm not certain whether it needs an inclusion report or not; it's a next-generation version of notification-daemon, but I'm not sure whether it's a continuation of the cdebase
[10:54] <pitti> Keybuk: I'm in a mirror mood today
[10:54] <Kamion> codebase
[10:54] <Keybuk> pitti: ok, 660 it is
[10:54] <pitti> Kamion: it's just a renaming, I'm fine with it
[10:55] <Kamion> ok, promoted
[10:55] <pitti> thanks
[10:55] <Kamion> that should make ubuntu-desktop happier, and thus the live CD etc.
[11:02] <mvo> Kamion: thanks a lot! I wanted to edit the seeds yesterday but didn't managed to
[11:02] <mvo> (timewise)
[11:06] <sivang> guten morgan dholbach :)
[11:06] <dholbach> hello everybody
[11:06] <dholbach> Guten Morgen, Sivan! :)
[11:07] <pitti> mjg59: I reviewed pam_foreground.c; I found a few small issues, I wrote them down on the report page
[11:12] <mvo> hello dholbach 
[11:12] <dholbach> hellas mvo
[11:12] <Keybuk> yay, found a good name for today's udev \o/
[11:13] <mvo> Keybuk: is it "guten morgen dholbach" :P ?
[11:14] <Keybuk> mvo: that wouldn't fit the naming scheme
[11:14] <pitti> Keybuk: btw, do you know which package is responsible for the lack of automatic ifup/ifdown invocation?
[11:14] <Keybuk> pitti: the "keybuk" package
[11:14] <Keybuk> I haven't written any new code for that yet, might actually tackle it this afternoon
[11:14] <Keybuk> there's plenty of dup bugs about it already
[11:14] <pitti> 'Keybuk does not type 'ifdown wlan0' for me any more if I pull out my USB wireless'
[11:16] <pitti> Keybuk: actually, I always need to go over to my gf (who sits at my laptop) whenever she needs the wireless :)
[11:18] <hunger> Is it possible not to compress all those X manpages that contain only ".so man3x/something"? They actually grow when compressed.
[11:18] <hunger> By the may: They point to non-existing files. The stuff moved to man3/something.
[11:20] <Kamion> hunger: it doesn't matter that they grow because they'll still be less than one filesystem block
[11:20] <Kamion> but you could file a wishlist bug against debhelper on bugs.debian.org asking for dh_compress to skip them, I guess; joeyh might do that
[11:20] <hunger> Kamion: I think it does in reiserfs... but that is a good point:-)
[11:20] <Keybuk> where the buggery-bollocks does selinux matchpathcon_init_prefix come from?!
[11:21] <hunger> Kamion: The wrong path is rather annoying though... I get mails complaining about that each night.
[11:21] <Kamion> sure
[11:22] <Keybuk> hmm
[11:22] <Keybuk> is there any particular reason that we're behind Debian on libselinux versions?
[11:23] <Keybuk> we have 1.26-1, Debian has 1.28-1
[11:23] <Keybuk> 1.28-2 in fact
[11:23] <Kamion> libselinux |     1.28-1 |        dapper | source
[11:23] <mdke> jdub, another suggestion for Ubuntu planet: give the links a different css, atm they are hidden
[11:24] <Kamion> it will have failed to build due to libsepol not being in main
[11:24] <pitti> so we do need it after all
[11:24] <maswan> Heh. About half the boottime for my flight-2 test seems to be waiting for modprobe.
[11:24] <Keybuk> yup, seems we do need it after all
[11:24] <Kamion> dunno why we haven't synced libselinux 1.28-2 yet though; it was uploaded on 1 Jan
[11:25] <Keybuk> dunno either, changelogs.debian.net doesn't see it either
[11:26] <ajmitch> it's not still in NEW, is it?
[11:26] <ajmitch> due to python bindings being added
[11:27] <Kamion> ajmitch: no
[11:27] <Keybuk> looks like it escaped NEW yesterday
[11:27] <Kamion> ajmitch: oh, it could have been in Debian NEW for a while, ok
[11:35] <hunger> I'd report way more bugs if I wasn't so confused by malone all the time!
[11:35] <hunger> "Report a bug" gets   you to several different pages depending on where you were when clicking it:-(
[11:38] <hunger> How can I find out the source package for some package installed/
[11:38] <crimsun> apt-cache showsrc, or apt-cache madison, etc.
[11:39] <pitti> slomo_: ping
[11:39] <hunger> crimsun: thanks!
[11:40] <pitti> seb128: do you know how/where I can reach lennart on IRC?
[11:41] <seb128> nop
[11:41] <janimo> crimsun, you were saying something about changing xfce4-terminal name in the desktop file?
[11:41] <janimo> right now it is xfce4-terminal not Terminal
[11:42] <crimsun> janimo: Exec=xfce4-terminal?
[11:42] <Seveas> desrt, pong
[11:42] <janimo> crimsun, lemme see
[11:42] <janimo> yes not it is xfce4-terminal
[11:42] <janimo> now it is
[11:43] <crimsun> janimo: ok, that's good. Thanks.
[11:43] <janimo> same in debian upstream
[11:43] <janimo> great
[11:45] <slomo_> pitti: pong
[11:46] <pitti> slomo_: nevermind, sorry
[11:46] <slomo_> pitti: you can reach lennart in #avahi (here on freenode) as mezcalero
[11:46] <pitti> ah, thanks
[11:50] <pappan> hi BenC
[11:54] <Kamion> BenC: 2.6.15-11 NEWed
[11:55] <pitti> slomo_: ok, I found a small potential issue in avahi, I contacted lennart
[11:55] <pitti> slomo_: but otherwise the package seems fine now
[11:56] <slomo_> pitti: cool, thanks :)
[11:57] <koke> there is a problem with releases.u.c
[11:57] <pitti> slomo_, seb128: avahi approved, please germinate it
[11:57] <koke> http://releases.ubuntu.com/kubuntu/5.10/kubuntu-5.10-install-powerpc.iso gives me "Banned extension: .iso  "
[11:58] <slomo_> pitti: germinate?
[11:58] <pitti> slomo_: seed, or make it a dependency of a main package
[11:59] <koke> ups, sorry it's my company firewall
[11:59] <seb128> pitti: thanks :)
[11:59] <seb128> I'll rebuild gnomevfs with it
[12:00] <slomo_> pitti: ah ok
[12:00] <seb128> but some binaries probably probably need to be seeded
[12:00] <seb128> like the server part
[12:00] <pitti> right, avahi-daemon, etc.
[12:00] <seb128> apps will only Depends on the libs no?
[12:00] <pitti> pelase don't seed libs
[12:00] <seb128> don't worry
[12:01] <seb128> slomo_: could you make a list of what could use avahi and what need to be promoted out of the libs which will be triggered by Depends automatically? :)
[12:01] <pitti> seb128: do you know librsvg a bit? can we build it against firefox-dev?
[12:02] <seb128> a bit. No pb to build against firefox afaik, I've a version update planned, I change that in the same time
[12:02] <seb128> s/I change/I'll change
[12:02] <pitti> cool, thanks
[12:02] <seb128> np
[12:03] <seb128> BTW is anybody working on eclipse?
[12:03] <seb128>   libswt3.1-gtk-java: Depends: mozilla-browser (>= 2:1.7.0) but it is not going to be installed
[12:03] <seb128> some users asked if anybody is working on that :)
[12:03] <slomo_> seb128: sure, i'll make a list... and afaik only avahi-daemon needs to be promoted... libavahi-{core,common,client,glib,qt} will be pulled in as dependencies... and maybe we want avahi-utils / avahi-discover seeded but it's not needed...
[12:03] <pitti> seb128: to settle this finally, shall I try to build evo and e-d-s against libdb4.3 and check whether everything still works? (I strongly think it will, but it shoudl be tested)
[12:05] <pitti> seb128: or do you plan to merge with Debian anyway?
[12:06] <hunger> OK, bugs about all my mandb warnings (several X debs, perl, evo) are reported in malone.
[12:07] <seb128> pitti: no plan to merge with Debian, I wanted to speak with lool about that first, but if you want to give it a try you are welcome :)
[12:07] <pitti> seb128: there are no transaction logs in my ~/.evolution, just .db files; i. e. it seems that evo uses in-memory transactions only, which is fine
[12:08] <pitti> seb128: oh, evo b-deps gstreamer0.8 *sigh*
[12:09] <Mithrandir> elmo: please sync apr1.0 and apr-util1.0 from unstable, they aren't in Ubuntu yet, afaik.
[12:09] <hunger> pitti: May I nag you about my malone bugs #563, #6216 and #6257? (first is my cryptsetup init script, the latter two oneliners to pam_mount to deal with LUKS volumes more sanely).
[12:09] <pitti> hunger: nag yes, but I'm afraid I won't really have time to deal with this stuff
[12:10] <pitti> hunger: if they are assigned to me, then I won't forget at least :)
[12:10] <hunger> pitti: Is there anything I can do to help with these issues?
[12:10] <slomo_> seb128: it's only rhythmbox, gnomemeeting (in cvs), a patch for vino exists, gnome-games afaik in cvs too... no idea about kde stuff
[12:10] <hunger> pitti: Like send a patched deb or anything?
[12:10] <pitti> hunger: oh, if there are properly tested patches, I'll take a look :)
[12:11] <pitti> hunger: no, a debdiff or normal patch is fine; no debs please :)
[12:12] <seb128> pitti: that's for a stupid plugin, the Build-Depends will not stay for dapper don't worry
[12:12] <seb128> slomo_: k, I'll build gnomevfs with it
[12:12] <seb128> slomo_: rhythmbox/gst0.10 daap code doesn't work fine atm
[12:12] <hunger> pitti: One is a init-script I use since way before breezy, the others are not really patches but one-line changes to mount.crypt.
[12:12] <pitti> Riddell: is it possible to build libakokde-dev without jack? it's the only package that wants jack in main again
[12:13] <slomo_> seb128: ok, fine :)
[12:14] <tepsipakki> kamion: grub-install fails on daily dapper-netboot
[12:14] <pitti> Keybuk, Kamion: libsepol approved
[12:16] <seb128> infinity, lamont-away: could you give a retry to evolution-exchange build?
[12:18] <Kamion> Keybuk: any bzr branch for your pcmciautils, or shall I merge it by hand?
[12:19] <Kamion> pitti: and promoted (apart from sepol-utils, unless we actually need that)
[12:19] <Kamion> tepsipakki: fails how?
[12:21] <tepsipakki> kamion: hmm, because there's unmet dependencies.. so not really grub-installers fault?-)
[12:21] <tepsipakki> kamion: lang-support-en can't be installed so this fails
[12:21] <Kamion> tepsipakki: oh, ubuntu-desktop's just uninstallable at the moment *shrug*
[12:22] <tepsipakki> yes
[12:22] <Kamion> the single-stage installer causes this to show up at a different point from where it used to
[12:22] <Kamion> though I'm surprised you didn't get a failure in pkgsel
[12:22] <tepsipakki> where can I edit the seed? is it preseedable?
[12:30] <zyga> morning folks
[12:32] <ijuz> are the udeb's not unpacked by udpkg?
[12:33] <sivang> mornign zyga 
[12:35] <Riddell> pitti: done
[12:35] <pitti> Riddell: thank you
[12:48] <Nafallo> mjg59: ping
[01:03] <pitti> Keybuk: heh, so you liked the movie 'Chicago', too? :)
[01:10] <maswan> Keybuk: hmm.. wouldn't it be interesting with a bootchart:y thingie from login to default desktop ready?
[01:12] <martink> doko: any idea what went wrong with openoffice.org2-l10n?
[01:27] <mjg59> pitti: PAM configuration is through conffiles. One package shouldn't fuck with the ones produced by another package.
[01:28] <mjg59> pitti: So integration needs cooperation with libpam - see my mail to ubuntu-devel
[01:28] <mjg59> jk_work: With luck
[01:28] <mjg59> Nafallo: Hi
[01:28] <pitti> mjg59: ah, so you'll modify pam proper to use the new module?
[01:28] <pitti> ok
[01:29] <mjg59> pitti: Thanks for the comments - I'll fix and reupload
[01:30] <Nafallo> mjg59: hi! :-), I found something strange that I don't know if it's known... after hibernate/wakeup my brightness up and down keys works as both up and down (i.e. I press on it to its at max, then it goes to minimum and then back up to max etc).
[01:30] <mjg59> Nafallo: Hmm. Weird.
[01:31] <Nafallo> agreed ;-)
[01:31] <mjg59> Nafallo: Just the once, or always after wakeup?
[01:31] <Nafallo> reproducible aswell, only after wakeup.
[01:31] <mjg59> Sorry, I meant - once you've woken up, does it happen once and then start working properly?
[01:32] <Nafallo> hmm, works now
[01:32] <Nafallo> strange
[01:32] <mjg59> What sort of hardware?
[01:32] <Nafallo> http://www.magicalforest.se/darkelf/
[01:33] <Nafallo> ehm
[01:33] <Kamion> ijuz: mostly they're unpacked by udpkg, yes, apart from those in the initrd which are unpacked by dpkg on the build system
[01:33] <Nafallo> it's back
[01:33] <lucas> hi
[01:34] <Nafallo> mjg59: stranger yet... it worked when you asked me to test and now (few seconds after) it's reproducible again :-P
[01:34] <lucas> while trying to edit on the wiki: "/!\The authentication database is temporarily unavailable. Anonymous access only.
[01:34] <lucas> You are not allowed to edit this page."
[01:34] <Nafallo> mjg59: any idea what I can do to debug this? :-)
[01:34] <mjg59> Nafallo: Haha. Afraid not right now.
[01:35] <mjg59> I'll think about it
[01:35] <Kamion> lucas: the key word being "temporarily", indicating that you should do something else for a while during maintenance
[01:35] <lucas> ok, just wanting to make sure you were aware of it
[01:36] <Kamion> lucas: when you see that machine, it means that an admin is working on the machine, and therefore they're already aware of it unless they're sleepadmining
[01:36] <mdke> actually I'm not sure it is intentional, authentication is working on launchpad, but not on the wiki
[01:37] <mdke> was there an announce about it?
[01:37] <lucas> Kamion: note that the message doesn't say that the admins are working on it
[01:37] <Nafallo> mjg59: weirdest.bug.ever! I'll try to reboot :-P
[01:39] <pappan> Kamion: hi
[01:43] <pappan> Kamion: hi
[01:43] <maswan> dholbach: Should it be solved? I'm not sure I feel motivated enough to install flight-2 and upgrade it just to check that.
[01:44] <Kamion> pappan: hello?
[01:44] <dholbach> maswan: yes, i think it is - i thought you had an installation, where you could check, at hand
[01:44] <dholbach> maswan: nevermind, if you don't
[01:45] <maswan> dholbach: No, we played around a bit and I submitted a bunch of bugs I found, but now we have a FAI-installed breezy with a bunch of modifications and custom kernel
[01:45] <dholbach> right
[01:45] <dholbach> i'll keep the bug open and see if people still have that bug
[01:45] <maswan> dholbach: thanks. my collegue is still on vacation, so if it is important I could reinstall his workstation this week. :)
[01:45] <dholbach> nono, don't bother :)
[01:47] <maswan> dholbach: We mostly ran and tested flight-2 since that was the first distribution we found that installed on these machines, rather new sata chipset or something like that.
[01:48] <Lathiat> sata_sil24?
[01:48] <Lathiat> or another new one?
[01:50] <maswan> Lathiat:  [SiS]  5513 IDE chipset with some sata stuff, or something like that.
[01:51] <Lathiat> ah ok
[01:51] <maswan> Fujitsu-Siemens shipped it with a fedora core 4 dvd that didn't find a disk to install onto either. :)
[01:51] <Lathiat> haha
[01:51] <maswan> Since we asked for no windows
[01:52] <HiddenWolf> hah, you're kidding?
[01:52] <highvoltage> it would be nice if they sent an ubuntu dvd.
[01:52] <maswan> well, they don't actually support linux, so nothing preinstalled
[01:52] <maswan> highvoltage: well, suggesting them to bunlde flight-2 is kind of icky too
[01:53] <maswan> and support for the ide/sata stuff came with 2.6.15
[01:53] <highvoltage> maswan: better than fedora core :)
[01:53] <maswan> highvoltage: :)
[01:53] <maswan> Actually, I think you could tinker aroudn in bios and make it hide as an ide drive that you could install on
[01:53] <maswan> but then you couldn't access the cd/dvd-drive
[01:53] <highvoltage> yep, enable ide emulation in bios.
[01:54] <highvoltage> you should just emulate as another controller.
[01:54] <highvoltage> if you emulate your sata disks as hda and hdb, and your cdrom is on hda, then it wont work.
[01:54] <highvoltage> so you can just emulate your hard disks as a second or third controller, instead.
[01:59] <maswan> highvoltage: well, I didn't try it personally, just that it didn't work
[02:12] <Pygi> not that I haven't mentioned this for like 10 times (and no one replied :P) But I'll try once more :)
[02:12] <Pygi> Keyboard layout changing does NOT work :/
[02:12] <HiddenWolf> Pygi, if a bug is filed it'll be fixed.
[02:13] <Pygi> HiddenWolf: not sure if there's bug
[02:13] <Pygi> reported
[02:13] <HiddenWolf> then report one.
[02:13] <dholbach> there are bugs reported on it, it depends on what your issue is EXACTLY, please try to follow up on existing ones, if they match your problem
[02:13] <HiddenWolf> developers like bugs, so they can keep track of things, unlike things reported on irc.
[02:14] <Pygi> dholbach: k, I'll track the following bugs
[02:14] <Pygi> reported
[02:14] <dholbach> thanks
[02:20] <mdke> seb128, evolution-exchange is blocking ubuntu-desktop, is this known?
[02:21] <seb128> yep
[02:21] <Nafallo> dep-wait: builddadmin :-)
[02:21] <seb128> you can assume than un-installability on dapper for desktop are always know
[02:22] <seb128> no need to have a zillion on people pinging me by mail/bugzilla/IRC every time a soname change, thanks
[02:22] <seb128> :)
[02:22] <Nafallo> :-)
[02:22] <mdke> seb128, sorry
[02:23] <seb128> np, I know that people doesn't want to bother, but I've to reply to mails about that, saying the same things and different chans and to close some bugzilla bug already today
[02:23] <seb128> men, that's an unstable distro, stuff move, few a couple of hours :)
[02:23] <mdke> i should have looked on bugzilla
[02:24] <seb128> s/few/wait
[02:24] <mdke> i did wait 24 hours first tho ;) problem with searching bugzilla is that usually only open bugs come up, i guess
[02:24] <seb128> yep
[02:24] <mdke> anyway, i'll assume that uninstallability is known from now on
[02:24] <jdub> seb128: dude, you or dhbuild noticed that evo exchange is blocking u-d today?
[02:25] <mdke> lol
[02:25] <jdub> ;-)
[02:25] <Nafallo> :-)
[02:25] <dholbach> . o O { it must be REAL love }
[02:26] <seb128> mdke: in fact evolution-exchange waits for buildd admin to retry the build, nothing we can do about it
[02:28] <Nafallo> jdub: I rebooted with them :-)
[02:28] <jdub> Nafallo: you obviously survived. did your laptop?
[02:28] <Nafallo> yepp :-)
[02:28] <jdub> lucky ;)
[02:28] <Nafallo> hehe
[02:28] <Nafallo> new kernel aswell :-P
[02:28] <Nafallo> and gdm ;-)
[02:29] <Nafallo> risky today :-)
[02:30] <doko> martink: no idea, it builds fine, when I build it by hand
[02:34] <jsgotangco> its pretty sweet
[02:36] <ijuz> Kamion: i wondered if there is something else than udpkg, one of my problems is that anna complains about wrong md5sum despite i have updated the Packages file
[02:57] <Kamion> ijuz: well anna's at a different layer from udpkg ...
[02:58] <Kamion> ijuz: perhaps you forgot to update Release with the new md5sum/sha1sum/size of Packages
[03:05] <ijuz> Kamion: what programm does generate the Release file? manually is a bit annoying
[03:06] <tseng> ijuz: man apt-ftparchive
[03:07] <ijuz> tseng: thank you
[03:07] <raphink> dpkg-scanpackages and dpkg-scansources
[03:07] <raphink> ijuz: 
[03:07] <raphink> you can use that too
[03:07] <raphink> dpkg-scanpackages /dev/null ./ | gzip -9 > Packages.gz
[03:07] <raphink> and same with sources
[03:07] <raphink> for small repos ;)
[03:08] <tseng> (he said release)
[03:08] <raphink> oh sorry
[03:12] <Keybuk> Kamion: it was such a trivial change that I figured it'd be easier for you to just take it by hand
[03:12] <Keybuk> pitti: heh, isn't the movie so much ... they're all lyrics from broadway/west end musicals
[03:13] <Keybuk> maswan: it'd be interesting to seb128, I guess, and easy to do; just put bootchart into the Xsession ... though you'd need a "stop" point for it
[03:13] <Keybuk> jdub: if it makes you feel better, I've been booting with them for weeks
[03:14] <Kamion> Keybuk: fair enough, already done
[03:15] <Keybuk> Kamion: I couldn't remember off the top of my head where your bzr branch for it was
[03:15] <Keybuk> one day, I suspect, launchpad will tell me :p
[03:16] <ijuz> does somebody know if bzip2 support in dpkg/apt really works?
[03:17] <Keybuk> ijuz: yes
[03:17] <Keybuk> we've been using it since hoary
[03:20] <ijuz> Keybuk: for Package files, but not for the deb packages themselfes
[03:20] <Keybuk> bzzzt!  wrong!  but you can try to double-your-money in the ready-money-round
[03:22] <ijuz> uhm, i decompressed all data.tar.gz from breezy with gzip, that was something like an indicator for me that they are compressed with gz ;)
[03:24] <Keybuk> *sigh* you so should have asked for the RETURN ticket from wrongland :)
[03:24] <Keybuk> descent scott% apt-cache policy language-pack-en
[03:24] <Keybuk>  *** 20051011 0
[03:25] <Keybuk>         500 http://archive.ubuntu.com breezy/main Packages
[03:25] <Keybuk> descent scott% aptitude download language-pack-en
[03:25] <Keybuk> descent scott% ar tf language-pack-en_20051011_all.deb
[03:25] <Keybuk> debian-binary
[03:25] <Keybuk> control.tar.gz
[03:25] <Keybuk> data.tar.bz2
[03:25] <Keybuk> *gasp*
[03:25] <Keybuk> bz2 ... well I never
[03:25] <Keybuk> :D
[03:25] <ijuz> odd
[03:25] <ijuz> but on the CD they are with .gz
[03:26] <Kamion> no, they aren't
[03:26] <Kamion> the CD takes packages verbatim from the archive; it doesn't recompress anything
[03:26] <maswan> seb128: so, do you want to do desktop bootcharting? in our experience, default desktop startup is almost as bad as bootup times
[03:27] <Keybuk> it's just really not your day, is it? :)
[03:27] <Keybuk> descent scott% cat /media/cdrom0/README.diskdefines
[03:27] <Keybuk> #define DISKNAME  Ubuntu 5.10 "Breezy Badger" - Release i386
[03:27] <Keybuk> descent scott% ar tf /media/cdrom0/pool/main/l/language-pack-en/language-pack-en_20051011_all.deb
[03:27] <Keybuk> debian-binary
[03:27] <Keybuk> control.tar.gz
[03:27] <Keybuk> data.tar.bz2
[03:27] <ijuz> 126751a2dc5528c2f9044d9e4ee36d61  ubuntu-5.10-install-i386.iso
[03:27] <ijuz> is that the right one?
[03:28] <ijuz> well, i have it from cdimage.ubuntulinux.org
[03:28] <Keybuk> Kamion: uh, where are the breezy CDs? :)  they're not on cdimage
[03:29] <Kamion> releases.u.c
[03:29] <Keybuk> duh, silly me
[03:29] <Keybuk> 126751a2dc5528c2f9044d9e4ee36d61  ubuntu-5.10-install-i386.iso
[03:29] <Keybuk> ijuz: yup, same CD I have
[03:29] <ijuz> ar tf /media/cdrom0/pool/main/g/gcc-4.0/gcc-4.0-base_4.0.1-4ubuntu9_i386.deb
[03:29] <ijuz> debian-binary
[03:29] <ijuz> control.tar.gz
[03:29] <ijuz> data.tar.gz
[03:29] <ijuz> most stuff is still gz
[03:29] <Keybuk> yes, that's nice
[03:30] <ijuz> BUT, that's excellent
[03:30] <Keybuk> that's because most stuff compresses better with gz
[03:30] <Kamion> cjwatson@little:~/cdimage/www/simple/breezy$ isoinfo -R -i ubuntu-5.10-install-i386.iso -x /pool/main/l/language-pack-en/language-pack-en_20051011_all.deb > ~/language-pack-en_20051011_all.deb
[03:30] <Keybuk> look at the package I showed you
[03:30] <Kamion> cjwatson@little:~/cdimage/www/simple/breezy$ ar tf ~/language-pack-en_20051011_all.deb
[03:30] <ijuz> that means even more soace saving
[03:30] <Kamion> debian-binary
[03:30] <Kamion> that's on the master cdimage system
[03:30] <Kamion> control.tar.gz
[03:30] <Kamion> data.tar.bz2
[03:30] <ijuz> s/soace/space
[03:31] <Keybuk> we hand-picked which debs are bz2-compressed
[03:31] <Keybuk> because for most of them, there's no point
[03:31] <seb128> maswan: "our" beeing? Sure it would be nice but I'm probably too busy to do that myself atm, but if you want to work on it you are welcome
[03:32] <seb128> maswan: dapper should be better regarding to desktop startup time, we changed quite a bunch of stuff (xrdb call, merged gconf, delay the startup of some stuff to after the login, etc)
[03:35] <Keybuk> descent scott% /bin/ls --block-size=4096 -ls gcc*
[03:35] <maswan> seb128: Well, just me and my collegue, spending time watching the startup time of various gnomey stuff after login before you get a usable desktop in flight-2
[03:35] <Keybuk> 44 -rw-r--r--  1 scott scott 44 2005-10-01 15:20 gcc-4.0-base_4.0.1-4ubuntu9_i386.deb
[03:35] <Keybuk> 44 -rw-r--r--  1 scott scott 44 2006-01-04 14:32 gcc-4.0-base_4.0.1-4ubuntu9_i386.deb-bz
[03:35] <Keybuk> ijuz: on most modern filesystems, those two files are the same size with gzip and bzip2 -- and bzip2 is significantly slower and more resource hungry
[03:35] <ijuz> there is no size
[03:35] <Keybuk> ijuz: the "44" bit
[03:35] <HiddenWolf> maswan, to be expected, some of what was done before login is now done after.
[03:36] <ijuz> Keybuk: what unit is that?
[03:36] <Keybuk> ijuz: when talking about archive or cd space saving, it's only useful to talk in terms of number of filesystem blocks
[03:36] <Keybuk> ijuz: filesystem blocks
[03:37] <ijuz> the base package is a bad example
[03:37] <ijuz> -r--r--r--  4 root root 5106876 2005-10-01 16:20 /cdrom/pool/main/g/gcc-4.0/libgcj6_4.0.1-4ubuntu9_i386.deb
[03:38] <ijuz> -rwxr-xr-x  1 ijuz ijuz 3644880 2006-01-03 21:38 libgcj6_4.0.1-4ubuntu9_i386.deb
[03:38] <ijuz> bytes
[03:38] <ijuz> bytes or blocks are the same in the end
[03:38] <maswan> HiddenWolf: Ok. Well, I don't care that strongly about it. We ended up not running the default desktop session anyway.
[03:38] <Keybuk> ijuz: dude, if you really think that, you need to stop this discussion now
[03:39] <ijuz> over n files it's the same
[03:39] <Keybuk> no its not
[03:39] <Kamion> it's the same on a CD image, but not in the archive
[03:39] <ijuz> ok, i take back "bytes or blocks are the same in the end"
[03:40] <Keybuk> ijuz: I contest your size finding for that file ...
[03:40] <Keybuk> 4988 -rw-r--r--  1 scott scott 5106876 2005-10-01 15:20 libgcj6_4.0.1-4ubuntu9_i386.deb
[03:40] <Keybuk> 4924 -rw-r--r--  1 scott scott 5040668 2006-01-04 14:40 libgcj6_4.0.1-4ubuntu9_i386.deb-bz
[03:40] <Keybuk> barely any difference
[03:40] <ijuz> for bzip2, maybe
[03:41] <Keybuk> where did you get your 3.4MB figure from?
[03:41] <Kamion> I think he's using 7zip or lzma or something
[03:41] <Kamion> and that this discussion is extremely confusing
[03:41] <ijuz> Kamion: yes, lzma
[03:41] <Keybuk> ijuz: we weren't discussing lzma, we were discussing bzip2
[03:42] <Keybuk> dpkg doesn't (currently) support lzma
[03:42] <ijuz> my dpkg does :)
[03:42] <Keybuk> I have the patches to do it here too
[03:42] <ijuz> i wondered about bzip2 because the question is if the cases in the different apt parts are working for bzip2
[03:43] <Keybuk> apt and dpkg are different things
[03:43] <Keybuk> apt only concerns itself with Packages files and the like
[03:43] <Keybuk> (well, except for the ftp archive bits, but in general that's true)
[03:43] <ijuz> well, apt-extracttemplate or something
[03:44] <Kamion> apt-extracttemplates only cares about the control.tar.gz member, not the data member
[03:45] <Keybuk> ah yes, of course
[03:45] <ijuz> you are right
[03:45] <ijuz> it outputs only a warning
[03:45] <Kamion> the only thing in apt I can see that cares about the data member is apt-ftparchive's contents generation
[03:45] <ijuz> i'm a bit confused today, sorry
[03:46] <Kamion> there's a .deb extractor in apt-inst but as far as I can see it's unused by apt itself, although it might be used by other users of libapt
[03:46] <Keybuk> jennifer, maybe
[03:47] <Kamion> yes
[03:47] <mvo> python-apt proivdes some additional support for the data extraction
[03:48] <Diziet> dholbach: I wanted to talk to you briefly about your change to ffox's debian/rules.  Your comment says `commented out, broke on everything but i386'.  Can you say in what way ?
[03:49] <ijuz> i just wondered about lzma in ubuntu because space should be soon a contraint with the same package set for 1 CD
[03:50] <Keybuk> ijuz: we've briefly looked into it, it needs a thorough investigation though
[03:50] <Keybuk> in particular, how much space is gained broken down into package groups
[03:50] <Keybuk> difference in resources (cpu, memory, time, etc.) to unpack and pack
[03:50] <Keybuk> etc.
[03:50] <Keybuk> for dapper we still have wiggle-room on the CD, though
[03:50] <Keybuk> there's still some WinFOSS and language-packs that can be tossed :p
[03:51] <dholbach> Diziet: it didn't build, let me try to find the buildlog
[03:51] <Mithrandir> ijuz: I'm working on squashfs, possibly with lzma for the live cd.  It saves quite a bit, but needs unionfs, which is problematic on ppc
[03:52] <ijuz> uhm, i did that 1 year ago for knoppix
[03:53] <Mithrandir> ijuz: there aren't any debian packages of mksquasfs-lzma that I could find at least.  Anyway, it's not something we just jump at, we need to investigate it a bit
[03:53] <ijuz> for a live-cd it is a slightly speed problem on slow boxes
[03:53] <dholbach> Diziet: http://people.ubuntu.com/~lamont/buildLogs/f/firefox/1.4.99+1.5rc3.dfsg-1ubuntu6/
[03:54] <dholbach> Diziet: it only built on i386
[03:54] <ijuz> Mithrandir: but i don't get the unionfs requierement
[03:54] <Mithrandir> ijuz: oh?  How else would you make the CD writeable?
[03:54] <Mithrandir> ijuz: you can do apt-get install $whatever on the ubuntu live cd.
[03:54] <Mithrandir> +s
[03:55] <seb128> Diziet: I did change the -X"*nspr*" to -Xnspr (same for some other stuff than nspr too), I think I mentionned the change on a malone bug
[03:55] <ijuz> Mithrandir: sure, but don't you need that too for what you are using now?
[03:55] <Kamion> ijuz: no, with cloop you can use device-mapper snapshots
[03:55] <Mithrandir> ijuz: no, devmapper with a writable snapshot doesn't need it.
[03:55] <Kamion> but they don't work with squashfs because squashfs is a fundamentally read-only fs
[03:55] <ijuz> oh, didn't know this
[03:57] <Mithrandir> ijuz: so our current live cd is ext2+cloop with devmapper writable snapshots, while squashfs would be squashfs+unionfs
[03:58] <ijuz> yes, sure, have you allready tried using squashfs?
[03:58] <Mithrandir> ijuz: yes, non-lzma.
[03:59] <ijuz> i think one of the problems is that squashfs still has no metadata about the used compressor
[04:00] <ijuz> Mithrandir: 2.2r2?
[04:01] <Mithrandir> yes
[04:02] <elmo> mmk, so I installed a laptop without network, have no put a wireless PCMCIA card in, but gnome refuses to see it in the panel - how do I beat some sense into it?
[04:02] <elmo> s/no/now/
[04:02] <mvo> Kamion: ok if I change notification-daemon -> notify-daemon in the desktop seed?
[04:02] <Kamion> mvo: yes
[04:04] <Diziet> dholbash: Yes, but why did what you do fix it ?
[04:04] <jdub> elmo: add the network status applet? switch the network interface in the status applet prefs?
[04:04] <Diziet> Err, I mean:
[04:04] <Diziet> dholbach: Yes, but why did what you did fix it ?
[04:05] <dholbach> Diziet: that's a very good question
[04:05] <dholbach> Diziet: i'm afraid, I have no answer :-(
[04:05] <Diziet> I'm talking about the one where you unwrapped the seddery.
[04:06] <dholbach> Yeah, i don't know. It was a matter of trial and failure. :(
[04:06] <Diziet> You didn't _change_ the seddery, did you ?  You just removed various instances of   tab tab \ lf tab tab  .
[04:06] <Diziet> It's a bit hard to tell because of the formatting change :-).
[04:06] <Diziet> I should thank you from clearing up my mess, though, so thanks.
[04:07] <dholbach> No problem, I always hoped I didn't produce more mess. :-)
[04:07] <Diziet> Do non-i386's have different /bin/sh, or different make ?
[04:07] <Diziet> I know this seems optimistic, but I want to understand it so that I can fix the root cause (whether that's in my head or somewhere else ...)
[04:07] <Kamion> make had a related change recently
[04:08] <Kamion> I think it may have been reverted, but can't remember exactly; check the recent changelogs
[04:08] <Kamion> I know Joey had to make a similar unwrapping change in debconf (in fact I think he moved the seddery complete with backslash-newlines into a define)
[04:08] <Diziet> Joy.
[04:08] <Diziet> And that fixed it ?
[04:08] <Kamion> yes
[04:08] <Diziet> Cripes.
[04:09] <Kamion> perhaps i386 built firefox with a different version of make due to timing
[04:09] <Kamion> the top of the build log should say
[04:09] <Diziet> I think I'll try a test build with dholbach's change reverted and see.
[04:10] <psusi> does this lzma squashfs use a larger block size than the normal squashfs?  imho, using a larger block size has far more impact on compression than lz vs lzma
[04:10] <Kamion> bah, no, the build log doesn't list that
[04:10] <Kamion> lamont-away: it'd be nice if make were listed among "Toolchain package versions"
[04:10] <Diziet> Lots of the stuff in make is completely crackheaded.   Recently I discovered that  $(shell $(variable))  after  define variable  splits variable up by newlines into separate arguments and appends that argument list to  sh -c  .
[04:11] <psusi> iirc, squashfs defaults to a 4 KB block size... you can get significantly better compression by increasing that to 64k or more, but it has more overhead for small random io
[04:11] <Mithrandir> psusi: it defaults to 64k in my tests.
[04:12] <psusi> Mithrandir: so both normal and lzma you have set to use 64k?
[04:12] <psusi> and lzma is significantly better compression with equal block size?
[04:13] <Mithrandir> yes
[04:13] <psusi> wow
[04:14] <seb128> Diziet: I did that change to firefox: https://launchpad.net/distros/ubuntu/+source/firefox/+bug/5999 (the formatting of the patch is screwed by launchpad)
[04:15] <Diziet> seb128: Right.
[04:15] <ijuz> Mithrandir: when are you going to build a new live cd? :)
[04:15] <Diziet> I RTFM'd dh_install earlier and you were right.
[04:15] <Diziet> Although I don't understand why it ever worked.
[04:15] <Diziet> Maybe dh_install has changed its behaviour to match the docs ?
[04:16] <seb128> not sure
[04:16] <seb128> it was doing "find mozilla-firefox ! \( -regex .\*\*nspr\*.\* \)" with the '-X*nspr*'
[04:16] <Mithrandir> ijuz: we need to change the build infrastructure a little for squashfs, but I'm hoping that we'll have something ready early next week
[04:17] <ijuz> Mithrandir: ok
[04:18] <lamont-away> Kamion: anything besides make?
[04:18] <lamont-away> @toolchain_regex = ( 'binutils$', 'gcc-[\d.] +$', 'g\+\+-[\d.] +$', 'libstdc\+\+', 'libc[\d.] +-dev$', 'linux-kernel-headers$', 'dpkg-dev', 'make' );
[04:18] <Kamion> *could* be related to the change in debhelper 4.9.4 I suppose ...
[04:18] <Diziet> Well, I can't be bothered tracking that one down :-).
[04:18] <Diziet> Since it's all fine and dandy now.
[04:18] <Kamion> lamont-away: nothing else comes to mind right now
[04:19] <lamont-away> Kamion: it'll be in the next round of updates, as yet unscheduled
[04:19] <Kamion> thans
[04:19] <Kamion> +k
[04:21] <Kamion> can I call gtk.main() from within a gtk signal handler?
[04:22] <Kamion> hmm, I suppose gtk will block until the signal handler finishes
[04:23] <mvo> Kamion: what is your use-case/goal?
[04:24] <Kamion> mvo: I'm trying to start up a debconf coprocess from a GtkNotebook switch_page handler to drive the logic of the new page
[04:25] <Kamion> mvo: when that coprocess gets an INPUT command, it needs to wait for more gtk events to find out whether to back up or not
[04:25] <Kamion> the problem is that when I try to do that, the new page doesn't get drawn
[04:26] <mvo> Kamion: would a  "while(gtk_events_pending()) do gtk_main_iteration(); " help you there?
[04:26] <zyga> Kamion: coprocess? 
[04:26] <Kamion> zyga: too complicated to explain right now
[04:26] <mvo> (or the python equivalent)
[04:26] <Kamion> mvo: I don't see how, since gtk.main() doesn't get around to drawing the page so I'd guess repeated iterations won't either
[04:27] <Kamion> and inside debconffilter, I need to actually wait for user interaction, not just draw the page
[04:29] <Kamion> hmm, perhaps I can do gtk.main_quit() in the switch_page handler and start up debconf outside that
[04:31] <zyga> Kamion: all sounds fishy, while I don't know the problem I'd opt for a longish clean solution
[04:32] <sbalneav> Hello all!  Can anyone point me at some docs for adding launchers to the panel in Gnome 2.12 in some scriptable way?  The latest Gnome Administrator's guide is from 2.6, and it seems that most of the gconftool-2 keys have changed about.
[04:32] <elmo> jdub: only 'lo' appears in the properties of the network status applet, that's the problem
[04:33] <Kamion> zyga: I'm not sure what you mean
[04:33] <tseng> elmo: if the other interfaces are down they dont show in the drop down
[04:33] <elmo> tseng: eth1 is up, and in use
[04:35] <Kamion> zyga: it's similar to oem-config; I'm basically filtering the debconf protocol in order to implement a pretty interface on top of questions asked by debconf (specifically, UbuntuExpress calling d-i code)
[04:35] <zyga> Kamion: well if you need gtk.main() and main_quit() voodoo around the app needs redesign IMHO
[04:35] <zyga> Kamion: hmm, cannot you spawn a subprocess to get the input and report back?
[04:35] <zyga> I don't know debconf internals but it seems to be ask-and-wait, right?
[04:35] <Kamion> zyga: no, because the subprocess won't exit
[04:36] <zyga> Kamion: why not
[04:36] <Kamion> the flow is that the debconf confmodule (the script running under debconf) calls the INPUT command, and at that point the user interface needs to do stuff
[04:36] <zyga> Kamion: subprocess like simple text dialog for example
[04:36] <Kamion> and then report back to debconf whether to go back or forward
[04:37] <zyga> Kamion: but why cannot that input command simply spawn a UI process, wait for it to exit and report back?
[04:37] <Kamion> zyga: because then I can't keep the UI for the whole thing within a single unified window
[04:38] <psusi> pitti: you around?
[04:38] <Kamion> and several questions from debconf may correspond to a single piece of UI
[04:38] <elmo> partman (68/76ubuntu1+1): in main - skipping.
[04:38] <zyga> Kamion: ah, you need plug into existing window
[04:38] <elmo> baseconfig-udeb (1.04/1.04): in main - skipping.
[04:38] <pitti> psusi: yes
[04:38] <zyga> Kamion: you can spawn ui daemon that can show and hide windows
[04:38] <elmo> Kamion: any of those safe for removal yet?
[04:38] <psusi> pitti: you seen the spec page I made?  https://wiki.ubuntu.com/PacketCD
[04:38] <zyga> Kamion: and use something like DBUS to talk to
[04:38] <Kamion> elmo: partman is different and needs to stay permanently; baseconfig-udeb can go
[04:38] <zyga> this would really rock btw
[04:38] <Kamion> zyga: ok, you just got way beyond the complexity I'm prepared to deal with in the time available. maybe later.
[04:38] <pitti> psusi: wow, no, I didn't
[04:39] <elmo> Kamion: ok, thanks
[04:39] <psusi> pitti: I'm looking for ideas on how to make the hal fdi policy look up the type field from the track info on the disc and key on that... think that's possible? 
[04:39] <zyga> org.debian.DConf.TextInput :-)
[04:39] <zyga> Kamion: I see :)
[04:39] <zyga> (then various daemons/services) could implement that interface and spawn text based ui or gui :)
[04:39] <Kamion> zyga: I don't disagree that something involving dbus would be the right solution long-term, but I just don't have time
[04:40] <sivang> dbus in Debconf?? ;-)
[04:40] <lamont-away> mdz: ping
[04:40] <mvo> Kamion: you need to interact with the user at this stage (when switch_page is called)? or "only" with gtk and debconf?
[04:40] <Keybuk> \o/
[04:40] <Keybuk> SCORE!
[04:41] <zyga> sivang: no, dbus in one of debconf interfaces
[04:41] <Keybuk> it seems that ifup --allow=auto works already
[04:41] <Kamion> mvo: not at that particular stage, I just need to kick off debconf
[04:41] <Kamion> interaction will be later
[04:41] <Kamion> I'm trying turning the code upside-down at the moment
[04:41] <psusi> anyone know how to kick the wiki and make it treat attachments as text/plain so they can be properly viewed?
[04:43] <mdke> psusi, what extension did you give it?
[04:43] <psusi> mdke: .fdi
[04:44] <mdke> maybe the wiki doesn't recognise that
[04:44] <psusi> you're not supposed to key on the name though, it should allow you to specify the mime type, or auto detect it based on the contents.. but there doesn't seem to be a way to specify the mime type
[04:44] <mdke> no, you can't specify the mime type
[04:44] <psusi> damn wiki!
[04:45] <mdke> dig around at http://moinmoin.wikiwikiweb.de
[04:45] <mdke> you'll find a soltion or a bug, or you can file a bug yourself
[04:45] <HiddenWolf> mdke, that must be the weirdest url I ever saw. :)
[04:46] <psusi> lol.. yea
[04:46] <mdke> ok...
[04:46] <ijuz> how retro, breezy automake is still default to 1.4
[04:47] <Keybuk> same as Debian
[04:47] <ijuz> yes, still retro :)
[04:47] <Keybuk> the maintainer likes it that way
[04:49] <mdz> lamont-away: pong
[04:50] <lamont-away> mdz: wondering if mysql 5.0 is supposed to land in dapper, or if I should switch back to libmysqlclient14-dev for a package or 2.
[04:51] <Diziet> Aaargh!  Why do we throw stuff out of our archive so quickly ?
[04:51] <ijuz> Mithrandir: you may be interested in this: http://christian-leber.de/~ijuz/advancecompsquash-2.2r2.tar.gz
[04:51] <Diziet> Is there some magic history site somewhere I just don't know about ?
[04:51] <fabbione> Diziet: it's fun!
[04:52] <tseng> Diziet: http://snapshot.archive.ubuntu.com/ubuntu/ like this?
[04:52] <lamont-away> Diziet: morgue.ubuntu.com
[04:52] <lamont-away> tseng: there's a snapshot?  kewl;
[04:52] <lamont-away> morgue is kinda unindexed, which can annoy would-be users...
[04:52] <Kamion> lamont-away: except that hasn't actually been updated since April
[04:52] <lamont-away> Kamion: morgue? oh, yeah. that.
[04:52] <Diziet> snapshot.archive contains what exactly ?
[04:53] <Diziet> For example, it contains only firefox ubuntu4 and 9 and not 6.
[04:53] <Kamion> err ... snapshot.archive.ubuntu.com is just one of the *.archive.ubuntu.com wildcard DNS, AFAIK
[04:53] <ijuz> Mithrandir: it is a mksquashfs that uses the gzip compression of 7zip and compresses a bit (2-5%) better than the normal zlib while staying completly compatible, you don't have to change anything else
[04:54] <tseng> hm i was sure there was a snapshot at one point
[04:54] <Kamion> I suspect you'll find it's identical to archive
[04:54] <tseng> (maybe not here)
[04:54] <Kamion> Diziet: if there's a particular version you want, I can fish it off jackass for you
[04:54] <Diziet> Never mind.
[04:54] <Diziet> /dev/vg_davenant/lv_mirror 180G  137G   43G  76% /export/mirror
[04:54] <Diziet> I'll just set up magicmirror to keep me a some versions.
[04:55] <Diziet> Shame I'll still miss some because I don't want to run the mirror more than about once per day.
[04:55] <Diziet> -1s/me a/me/
[04:55] <psusi> why not more than once a day?
[04:55] <Diziet> I don't know exactly what I want.  I was hoping to grobble around in pool to find it.
[04:56] <Kamion> Diziet: dead packages stay in the archive for a day anyway
[04:56] <Kamion>   StayOfExecution 86400;
[04:56] <Diziet> psusi: magicmirror needs some work to improve the performance during the `update mirror target area' step.
[04:56] <psusi> can you sync the archive with rsync?  then you can use rsync snapshots to keep previous versions
[04:56] <Diziet> kamion: Oh, good.
[04:56] <psusi> ahhh
[04:57] <Kamion> /srv/ftp.no-name-yet.com/morgue/rhona/2005-12-21/firefox_1.4.99+1.5rc3.dfsg-1ubuntu6.diff.gz
[04:57] <Kamion> /srv/ftp.no-name-yet.com/morgue/rhona/2005-12-23/firefox_1.4.99+1.5rc3.dfsg-1ubuntu7.diff.gz
[04:57] <Kamion> /srv/ftp.no-name-yet.com/morgue/rhona/2005-12-24/firefox_1.4.99+1.5rc3.dfsg-1ubuntu8.diff.gz
[04:57] <Kamion> etc.
[04:57] <Diziet> The following URL could not be retrieved: ftp://ftp.no-name-yet.com/morgue
[04:57] <Diziet> Squid sent the following FTP command:
[04:57] <Diziet> RETR morgue
[04:58] <Diziet> and then received this reply
[04:58] <Diziet> Err, with a /, I get
[04:58] <Kamion> it's not accessible, but as I say I can fish stuff out
[04:58] <Diziet> CWD morgue
[04:58] <Keybuk> heh, why do our servers STILL call the archive "ftp.no-name-yet.com" :)
[04:58] <Diziet> Oh, right.
[04:58] <Diziet> Why can't it be accessible ?
[04:58] <Kamion> probably in case people mirror it and cane our datacentre's bandwidth
[04:58] <Keybuk> diz: http://morgue.ubuntu.com/
[04:59] <lamont-away> Keybuk: because changing it is a royal pain.
[04:59] <Kamion> Keybuk: out of date
[04:59] <Keybuk> oh, is it?
[04:59] <Keybuk> bah
[04:59] <lamont-away> Keybuk: ENOSPC
[04:59] <Kamion> it used to be mirrored there; IIRC rookery ran out of space
[04:59] <lamont-away> Kamion: that's my memory too
[04:59] <Kamion> the morgue is HUGE
[04:59] <lamont-away> Kamion: well, of course it is... all those dead bodies piled up...
[04:59] <Kamion> 191823204       /srv/ftp.no-name-yet.com/morgue/
[05:00] <Kamion> hmm, not as big as I thought, but still non-trivial to push around
[05:00] <Diziet> What's our mirror turnover like ?  Eg, if I kept a copy of every day for the past month, would I have to buy another disk ?
[05:00] <Diziet> (Files the same in all copies are hardlinked.)
[05:00] <Kamion> for comparison, the archive proper is 144073008       /srv/ftp.no-name-yet.com/ftp
[05:00] <Diziet> Oh, that's helpful.  And morgue goes back forever ?
[05:01] <lamont-away> Kamion: how pre-warty-release does the morgue go? day 1?
[05:01] <Kamion> Diziet,lamont-away: I'm not sure actually. I don't *think* elmo's ever purged it.
[05:02] <Kamion> I'm just trying to drive the stats program now
[05:02] <lamont-away> Kamion: hrm...
[05:02] <lamont-away> I suppose it's possible that it only dates back to oxford....
[05:02] <lamont-away> archive events can make for ugly morgues
[05:02] <psusi> I was under the impression that the morge only contains packages that have been removed from the archive, not each superceeded version
[05:03] <lamont-away> psusi: they get removed from the arvhive when superseded
[05:03] <Kamion> psusi: nope
[05:03] <lamont-away> --> morgue
[05:03] <psusi> hrm.... what's that url then? ;)
[05:03] <lamont-away> morgue == everything that was ever in the archive, but isn't now.
[05:03] <Kamion> psusi: people have abused the word "morgue" in a confusing way, which gave you that impression
[05:03] <Kamion> that's why I wish people would just say "remove" when they mean "remove from the archive" rather than inventing new inaccurate verbs
[05:04] <lamont-away> psusi: morgue.ubuntu.com is frozen as of ~april 2005, which kinda limits its usefulness these days
[05:04] <psusi> yea... so is there an active usefull morgue?
[05:04] <lamont-away> psusi: not that I know of that is publicly reachable
[05:04] <Kamion> 20060101 570 533
[05:04] <Kamion> 20060102 499 855
[05:04] <Kamion> 20060103 697 1121
[05:04] <Kamion> 20060104 307 1196
[05:05] <Kamion> ^-- date, packages, size in megabytes
[05:05] <Kamion> that's the amount installed into the archive
[05:05] <psusi> total, or new on that day?
[05:05] <Kamion> total
[05:06] <Kamion> I mean, each time a package gets installed, the size gets increased
[05:06] <psusi> hrm.... only 1 gig as of today?
[05:06] <Keybuk> no, that means 1 gig got added today
[05:06] <Kamion> neither
[05:06] <Keybuk> bah
[05:06] <psusi> 1 gig worth of new packages today?  holy shit!
[05:06] <Keybuk> 1 gig got removed today
[05:06] <Kamion> no, STOP
[05:06] <Kamion> 1GB of packages got installed by kelly, the program that installs .debs, source packages, and so on into the pool
[05:07] <Kamion> it doesn't mean net increase or decrease of the total size of the pool
[05:07] <Kamion> (because removals aren't counted there)
[05:07] <psusi> what's the pool?  and how much is in the archive?
[05:07] <Kamion> psusi: http://archive.ubuntu.com/ubuntu/pool/
[05:07] <Kamion> those figures are purely turnover
[05:07] <Keybuk> the pool is what the developers piss in
[05:07] <Keybuk> uh, I mean, ... :p
[05:07] <Kamion> psusi: 16:00 < Kamion> for comparison, the archive proper is 144073008       /srv/ftp.no-name-yet.com/ftp
[05:07] <psusi> lol
[05:08] <mdz> mvo: what's the reason for notification-daemon->notify-daemon?
[05:08] <psusi> that is kb?  so the archive is ~144 gb?
[05:08] <Kamion> psusi: yes
[05:08] <psusi> wow... that's a bit larger than I expected
[05:09] <pitti> hi mdz
[05:09] <lamont-away> psusi: and smaller than I expected by a shade
[05:09] <mvo> mdz: upstream changed the name
[05:09] <psusi> oh, that's for all releases, not just dapper right?
[05:09] <Kamion> Diziet: taking 2005-11 as a reasonably normal month when we weren't all off on holiday, total turnover was 28482KB, so you shouldn't have to buy a new disk
[05:10] <Kamion> psusi: yes
[05:10] <psusi> ahh... ok... 
[05:10] <Kamion> er, 28482MB, duh
[05:10] <psusi> 28 gigs of new packages in a month?  wow
[05:11] <lamont-away> Kamion: it's those wonderful weeks when kernel, toolchain, and oo.o all upload twice. :0)
[05:11] <lamont-away> psusi: those 28GB were probably accompanied by 27+GB moving to the morgue.
[05:11] <Kamion> Diziet: if you skip some architectures, it'll of course be considerably less
[05:12] <psusi> lamont-away: but the morgue is no longer active?
[05:12] <lamont-away> Kamion: those are jackass numbers?
[05:12] <Kamion> psusi: morgue.ubuntu.com isn't (the mirror), but the master morgue on jackass certainly is
[05:12] <lamont-away> psusi: the visible morgue is not current.
[05:12] <Kamion> lamont-away: yeah, from /srv/ftp.no-name-yet.com/log/
[05:12] <lamont-away> ok.
[05:12] <xhaker> hey.. somehow.. a new version of initscripts showed up for upgrade.. i don't even see the source package for that on dapper-changes and it breaks the shutdown procedure
[05:12] <lamont-away> so those include all 6 architectures
[05:12] <Kamion> and a hacked copy of saffron
[05:12] <psusi> ohh... why is the active morgue not publically accessible?
[05:12] <Kamion> psusi: we just did this
[05:12] <lamont-away> psusi: ran out of disk space.
[05:13] <psusi> ohh... so will be be public at some point?
[05:13] <Kamion> I think we've just encountered the middle of this conversation.
[05:13] <lamont-away> heh
[05:13] <ogra> set a mark :)
[05:13] <psusi> lol
[05:14] <lamont-away> although I don't think anyone has committed to fixing the disk space and shlepping all the data to a reachable place, it would be nice...
[05:15] <mdz> pitti: morning
[05:15] <trulux> pitti: hey
[05:15] <mdz> lamont-away: I have a note from the devel meeting saying that infinity wanted to discuss that with me (mysql 5.0)
[05:15] <lamont-away> ok
[05:16] <lamont-away> meantime, I'll just (a) make it say libmysqlclient15-dev|libmysqlclient14-dev, and (b) hug Kamion for making that work for me back in oxford.
[05:16] <pitti> hi trulux 
[05:16] <lamont-away> then a simple rebuild is all that's needed when 5.0 is there
[05:17] <Keybuk> ROFL @ The "Ben got a PowerBook for Christmas" Release.
[05:17] <Keybuk> BenC got a powerbook, sabdfl got a Mac, I got a powerbook, did anyone NOT get an Apple over xmas?
[05:17] <ogra> *giggle*
[05:18] <Nafallo> me! :-/
[05:18] <Nafallo> I got a BED :-P
[05:18] <Kamion> lamont-away: s/Oxford/Mataro/ IIRC :)
[05:19] <lamont-away> Kamion: 'twas orginal ogre-model --> oxford (pre-warty)
[05:19] <psusi> I got scuba diving lessons ;)
[05:20] <Nafallo> woohoo Keybuk rocks! :-D
[05:20] <mdx> I do?
[05:21] <Nafallo> mdx: no, not you. Keybuk :-).
[05:21] <lamont-away> Kamion: hrmpf.
[05:21] <Nafallo> now you do! :-)
[05:21] <Nafallo> atleast if that ntpdate-thingie works :-)
[05:21] <lamont-away> patch was committed 8 dec 2004.  damn.  and here I could have sworn it was oxford
[05:22] <Kamion> lamont-away: was a bug you discovered in the ogre model and commented on in the quiet room in Mataro loudly enough for me to offer to fix it. :)
[05:22] <Keybuk> works for me
[05:22] <lamont-away> Keybuk: my daughter got an iPod, does that count?
[05:23] <Kamion> having a highly visual memory is handy sometimes - I remember the way the room looked and can mentally zoom out until I see the Cristal ballroom and thereby remember which conference it was. :)
[05:23] <lamont-away> Kamion: ah, yes.  although I seem to recall some mild begging was involved before you offered to shut me up.
[05:23] <ijuz> Nafallo: is ntpdate now less annoying when net doesn't work?
[05:23] <xhaker> ijuz, it should
[05:23] <Nafallo> ijuz: init.d -> ifup.d :-)
[05:24] <xhaker> Nafallo, init.d/rc file has a bug
[05:24] <xhaker> "bug"
[05:24] <Keybuk> xhaker: it does?
[05:24] <ijuz> Nafallo: excellent, i really should install dapper on my laptop, so that i can retry pppoeconf
[05:25] <xhaker> Keybuk, first_step is not set for runtime.... can't remeber
[05:25] <xhaker> i'll check
[05:25] <elmo> Kamion/etc.: the morgues actually migrating to hutte now, I finally did run out of space with dapper
[05:25] <elmo> jackass only has the last 8 months or so
[05:25] <Kamion> ah, that would explain it ...
[05:25] <xhaker> Keybuk, had to fix it to be able to shutdown the machine cleanly
[05:25] <elmo> pending my copious freetime, I'll make tghe complete copy on hutte public
[05:26] <Keybuk> xhaker: could be, can you file it ... I wasn't my most awake when I wrote that stuff
[05:27] <Kamion> I think this is a good sign; espresso just showed up its first bug in d-i
[05:27] <xhaker> Keybuk, ok.. i'll attach a patch hihi
[05:27] <xhaker> single line tho
[05:29] <Nafallo> hmm, that bittorrent tracker thing is ugly on shutdown :-P
[05:30] <Riddell> pitti: do you know if any other distro uses cups 1.2?
[05:30] <pitti> Riddell: I don't know for sure, but I doubt it
[05:31] <Riddell> pitti: guess I have to port KDE to cups 1.2 myself then
[05:32] <pitti> Riddell: oh?
[05:32] <Nafallo> pitti: oh! that cups thing always have to wait 5 seconds and then gets killed on shutdown :-P
[05:32] <pitti> Riddell: gnome-cups worked fine with 1.2
[05:32] <dmk> Riddell: there has been problem with cups 1.2 and kde in fedora
[05:32] <xhaker> Keybuk, what package is it in? binary please
[05:32] <Keybuk> PostgreSQL is my shutdown hate
[05:32] <Keybuk> xhaker: sysvinit
[05:32] <pitti> Nafallo: cupsd is incredibly hard to kill :(
[05:32] <Riddell> pitti: it works fine as long as you have a printer setup but if you start with a fresh install and no printers set up it doesn't connect
[05:32] <psusi> pitti: ohh.. say... I've noticed people complaining lately about removable media, most notably, usb memory sticks, not being properly unmounted due to the icon vanishing right away when they choose eject, but the syncing continues in the background
[05:32] <pitti> Keybuk: ?
[05:33] <Keybuk> xhaker: actually, binarywise might be sysv-rc
[05:33] <Nafallo> hmm
[05:33] <Keybuk> pitti: if postgresql isn't running, but there's a pid file lying around, it waits a metric week before letting the shutdown finish
[05:33] <psusi> pitti: have any thoughts as to how to fix that?  I'm thinking some kind of progress indicator during the sync would be nice... but I don't think the kernel provides any means of getting that kind of info
[05:33] <pitti> psusi: https://bugzilla.ubuntu.com/show_bug.cgi?id=13169
[05:33] <pitti> psusi: I forwarded this to upstream, and  that's in fact what they will do
[05:34] <pitti> psusi: http://bugzilla.gnome.org/show_bug.cgi?id=313639
[05:34] <Nafallo> food now though :-P
[05:34] <pitti> Keybuk: well, that shouldn't be the common case :)
[05:34] <pitti> Keybuk: but I'll look at it
[05:34] <pitti> Keybuk: can you please file a Debian bug?
[05:35] <pitti> Keybuk: (postgresql-common)
[05:35] <Keybuk> bah, filing debian bugs is such hard work
[05:35] <psusi> I htink it is caused by the kernel removing the entry from /proc/mounts before the sync is done... I'm not sure if that can be considered a bug in the kernel or not
[05:35] <pitti> Keybuk: hm? it's so much easier than filing an Ubuntu bug, at least for me
[05:35] <pitti> Keybuk: s/easier/faster/
[05:35] <Keybuk> reportbug sulks because I don't have an MTA
[05:36] <pitti> Keybuk: I don't use reportbug any more, I just write a mail to submit@b.d.o
[05:36] <Keybuk> then I have to get the dev-ref out to remember the silly fields you need for that now <g>
[05:36] <pitti> Package: postgresql-common\nVersion: 38, that's enough for me
[05:36] <zakame> gaah
[05:36] <pitti> Keybuk: Severity: wishlist if you want, but I'll care for severity inflation myself usually :)
[05:37] <zakame> pitti: out of curiousity, how do you handle X-Debbugs-CC: ?
[05:37] <Keybuk> right, I think that's one filed
[05:37] <pitti> zakame: I don't explicitly set it, I rely on submit@ to DTRT
[05:38] <jm_> where can I download the Flight CD 2 ? Thanks
[05:39] <Riddell> pitti: how come we're using cups 1.2 when it's not had any sort of release yet?
[05:39] <Riddell> jm_: cdimage.ubuntu.com
[05:39] <jm_> thanks Riddell 
[05:39] <Riddell> releases/dapper/flight-2 or something like that
[05:39] <pitti> Riddell: it was promised for December...
[05:39] <Riddell> ah :)
[05:40] <zakame> pitti: ah :)
[05:40] <pitti> Riddell: it behaves much better with USB printers, that's why we wanted it in the Printer spec
[05:41] <psusi> pitti: I think a simple workaround to the problem would be to first remount the filesystem as r/o, then umount it.. that would keep the icon there while the dirty pages are flushed.. might be worth considering if a better solution is taking too long
[05:42] <pitti> psusi: unmount waits until sync is complete, too
[05:43] <pitti> psusi: I think the problem is rather that g-vfs calls pumount asynchronously and doesn't wait until it's finished before removing the drive
[05:43] <psusi> pitti: umount does, but the problem is that the kernel immediately removes the entry from the list in /proc, causing gnome-vfs to remove the icon... I believe... doing a remount first will keep the icon until the dirty pages are flushed, then it's safe to umount
[05:44] <psusi> I don't htink so because you can umount it from the command line and the icon goes away immediately... I believe it is in response to what the kernel lists in /proc/mounts
[05:44] <psusi> even though the umount blocks for some time
[05:44] <pitti> oh, right, I see
[05:45] <Keybuk> whoah
[05:45] <Keybuk> this libsane stuff is total crack
[05:45] <Keybuk> it has a script to convert a usermap into a rules file
[05:45] <Keybuk> WHY WHY WHY
[05:45] <Diziet> Maybe I need an even more sophisticated fs.  cp -al on my local mirror seems to take quite a while.
[05:45] <psusi> I'm not sure if it can be considered a kernel bug or not... that it removes the mount from /proc before it is done flushing
[05:45] <psusi> but I'm fairly sure that is what causes the icon to go away before the umount completes
[05:47] <mdz> mvo: I saw a note in JaneW's status report that you wanted to talk with me about the upgrade tool
[05:48] <mvo> mdz: can I come back to you about it in ~1h? I need to look at my notes what it was about first
[05:49] <mdz> mvo: ok
[05:50] <xhaker> Keybuk, http://bugzilla.ubuntu.com/show_bug.cgi?id=21899  ;)
[05:55] <Mithrandir> ijuz: I'll take a look at it
[05:55] <psusi> another thing I've found very annoying about the kernel mount mechanics is that there is no umount --i-really-mean-force-it-right-now-and-damn-the-dirty-buffers option
[05:56] <psusi> it's really annoying when there's a problem writing to the disk and you can't umount the damn thing because the kernel keeps trying to flush the dirty buffers and failing
[05:56] <ogra_ibook> umount -l isnt enough for you ? 
[05:57] <psusi> no... that just makes it not block the umount program
[05:57] <psusi> you still can't eject the disc from the drive ;)
[05:57] <psusi> and your syslog grows to 50 MB from all the error messages
[06:02] <mdz> doko: are you aware of this?  trying to overwrite `/usr/lib/openoffice2/help/en/swriter.idx/DICTIONARY', which is also in package openoffice.org2-help-en-us
[06:07] <doko> mdz: yes, working on updated packages
[06:08] <mdz> doko: ok, thanks
[06:09] <mdz> seb128: I find the on-the-fly spell checking in xchat-gnome to be very distracting; is it a good idea to have it enabled by default?
[06:09] <Kamion> woo
[06:10] <Diziet> Can I do a quick survey ?
[06:10] <seb128> mdz: probably not, especially that the list of language is not configurable from the interface and it underline every english word on a french installation :) I'll speak with upstream about that
[06:11] <Diziet> update-alternatives --display x-www-browser   who sees it showing up as manual, set to /usr/bin/mozilla-firefox ?
[06:11] <Diziet> And who sees it as auto, set to /usr/bin/firefox (which is correct) ?
[06:11] <mdz> seb128: agreed, thanks
[06:11] <mdz> Kamion: nice!
[06:11] <seb128> np
[06:12] <mdz> Diziet: mine was manual until I fixed it earlier in dapper (after it broke everything)
[06:12] <xhaker> mdz, will it be in main? xchat-gnome?
[06:12] <seb128> mine is set manual/galeon ...
[06:12] <mdz> I've been using it instead of xchat for months now, and I think it's ready.  seb128?
[06:13] <seb128> xhaker: waiting on pitti to review it for promotion
[06:13] <Diziet> Hrm.  Earlier in dapper ?  How early ?  My dapper chroot here doesn't have it as manual but I probably haven't booted it since dapper.  The testbed dapper machine has manual and is broken.
[06:13] <seb128> mdz: yeah, it just wait for pitti to review it for main promotion
[06:13] <xhaker> mdz, i've said so before here in this same channel.. they don't listen to me.. hehe
[06:13] <pitti> seb128, mdz: it's approved since last year, just needs to be germinated :)
[06:13] <seb128> hum, weird
[06:14] <mdz> Dec 06 09:37:57 <mdz_>   link currently points to /usr/bin/mozilla-firefox
[06:14] <mdz> Dec 06 09:37:57 <mdz_>  /usr/bin/firefox - priority 70
[06:14] <mdz> Diziet: ^^
[06:14] <pitti> seb128, mdz: xchat-gnome still sucks as long as it can't split windows, but I'd only keep one, not both
[06:14] <Diziet> Hrm.  December.
[06:14] <seb128> pitti: did you review 0.7 or 0.8?
[06:14] <mdz> seb128: feel free to make the seed change
[06:14] <seb128> mdz: will do
[06:14] <seb128> that and avahi :)
[06:14] <pitti> seb128: dunno, whatever was current in the end of december
[06:15] <pitti> seb128: at Dec 13 to be exact
[06:16] <Diziet> I think I'll just have to do a clean install (eg, breezy, dapper upgrade) and see if and when it breaks.
[06:17] <mdz> Keybuk: what's the correct first_step value in rc for runlevels 0 and 6, so I can fix it locally?
[06:17] <seb128> pitti: oh, I said to wait on 0.8, which is from Dec 21 which is probably much better (they stopped to ship the xchat sources for text, gtk, etc ... they just have the common folder from it now)
[06:17] <pitti> seb128: well, doesn't make much of a difference, review-wise
[06:17] <seb128> pitti: anyway, that does the trick too, new one is just better, thanks :)
[06:18] <pitti> seb128: I didn't audit the source for this one
[06:18] <seb128> k
[06:18] <mdz> xhaker: it's something we've discussed for a long time; it's just been a matter of choosing the right time to switch
[06:19] <mdz> pitti: split windows?
[06:19] <pitti> mdz: watch two channels at the same time
[06:19] <xhaker> pitti, xchat allows that?
[06:19] <pitti> mdz: in xchat I can split off a channel window
[06:19] <pitti> xhaker: for ages, yes
[06:19] <pitti> it's a feature I love
[06:19] <mdz> pitti: oh, the attach/detach tab thing
[06:19] <pitti> how else can you follow two channels at once?
[06:20] <pitti> mdz: right
[06:20] <seb128> I just switch between chans :)
[06:20] <pitti> every minute?
[06:20] <mdz> me too
[06:20] <mdz> no, just when it's highlighted
[06:20] <pitti> that's too distracting for me
[06:20] <mdz> so every 5 minutes approximately
[06:20] <pitti> #u-d{esktop,devel} highlight every 10 seconds...
[06:20] <seb128> pitti: when I'm busy I switch when somebody speaks to me
[06:20] <pitti> hm
[06:21] <xhaker> pitti, but it goes to the taskbar
[06:21] <pitti> maybe I need to change my working style then, I just got used to the feature
[06:21] <seb128> otherwise I read the log every few and then
[06:21] <xhaker> i thought you were talkink about something like split view.. like in IDE's
[06:21] <seb128> I have like 20 chans anyway, I could not get them all at screen
[06:22] <pitti> seb128: http://people.ubuntu.com/~pitti/shots/crashrep-gui.png <- that's how I do it
[06:22] <pitti> seb128: (this shot has a different purpose, but shows off xchat as well)
[06:22] <seb128> pitti: anyway: http://bugzilla.gnome.org/show_bug.cgi?id=323486
[06:23] <Kamion> mm, I'm with seb on this one
[06:23] <xhaker> they should do it then
[06:24] <xhaker> i like Guillaume option
[06:24] <pitti> me too, sounds nice
[06:25] <xhaker> like in eclipse, or any word processor
[06:25] <pitti> or vim's :split
[06:25] <jm_> bye
[06:25] <xhaker> hehe.. i should really try vim more
[06:26] <pitti> xhaker: /me <3 vim 
[06:26] <pitti> xhaker: /me <3 vim 
[06:26] <xhaker> oh crap
[06:27] <xhaker> how dows one reattach a channel window?
[06:27] <xhaker> i just parted the channel trying to do that :P
[06:27] <Riddell> Keybuk: what's the status of network manager in ubuntu?
[06:28] <Keybuk> universe
[06:28] <Keybuk> I could explain now, but it'd ruin the surprise for the status meeting
[06:29] <xhaker> surprise surprise
[06:29] <seb128> hum
[06:29] <seb128> $ bzr get sftp://chinstrap.ubuntu.com/home/warthogs/archives/seeds.ubuntu.com/ubuntu/seeds/dapper
[06:29] <seb128> bzr: ERROR: Not a branch: sftp://chinstrap.ubuntu.com/home/warthogs/archives/seeds.ubuntu.com/ubuntu/seeds/dapper
[06:29] <Kamion> I think it's a bzr bug, best ask on #bzr
[06:30] <seb128> k, thanks
[06:30] <Kamion> I just haven't upgraded my currently-working daily build of bzr from a month or two ago ...
[06:30] <Kamion> what do you need to change?
[06:30] <pitti> Keybuk: you want to motivate us to get up at 3 am? :)
[06:30] <seb128> right, works with the dapper version
[06:31] <seb128> s/xchat/xchat-gnome
[06:31] <pitti> seb128, Kamion: no, it's actually a bug fix
[06:31] <seb128> and list avahi-daemon
[06:31] <Kamion> oh, if you've got a checkout with the dapper version then go ahead
[06:31] <pitti> seb128, Kamion: try with ... .com/%2Fhome/warthogs
[06:31] <seb128> yeah, seems to be fine with dapper version
[06:32] <seb128> pitti: hum, why?
[06:32] <pitti> seb128: the sftp:// specification says that the URL is relative to the home dir
[06:32] <pitti> seb128: so sftp://host/foo refers to ~/foo, not to /foo
[06:32] <seb128> ah
[06:32] <pitti> seb128: older bzr versions didn't respect that
[06:32] <ajmitch> which really does get annoying
[06:33] <Mithrandir> you don't have to escape the /, though.  sftp://host//wherever should work
[06:33] <jbailey> Yup, just double the /
[06:33] <pitti> Mithrandir: maybe, I just followed the advice in the bug report
[06:33] <jbailey> (if you're using my nightly snapshots)
[06:33] <pitti> but so much the better
[06:33] <Kamion> that's a little less horrible at least
[06:33] <Mithrandir> it's utterly insane.
[06:34] <Mithrandir> or arcane
[06:34] <jbailey> But who is to blame?
[06:34] <jbailey> mpool is his name!
[06:34] <jbailey> (sorry, I'll stop the rap now)
[06:35] <Mithrandir> he wrote the sftp spec?
[06:35] <pitti> http://bugzilla.gnome.org/show_bug.cgi?id=322394, btw
[06:35] <jbailey> AFAIK, the IETF spec wasn't actually approved
[06:35] <jbailey> It's just better than nothing.
[06:35] <pitti> it's not a bzr bug, but that's where jamesh pointed to
[06:36] <pitti> http://www.ietf.org/internet-drafts/draft-ietf-secsh-scp-sftp-ssh-uri-03.txt <- spec
[06:36] <Kamion> None of the secsh specs are RFCs yet
[06:37] <Kamion> although that one isn't even in the editor's queue, true
[06:48] <seb128> dholbach: you said you would make a promotion page for libsexy no? xchat-gnome builds with it atm
[06:48] <jbailey> libsexy?
[06:48] <dholbach> seb128: i can do that, yes
[06:48] <dholbach> jbailey: "doing naughty things to widgets" :)
[06:48] <jbailey> lol
[06:49] <seb128> dholbach: thanks
[06:49] <stratus> btw libsexy is cool!
[06:49] <stratus> oh, there's sexy-python too
[06:50] <Diziet> dholbach: Your symlinks for /usr/include/mozilla-firefox:
[06:50] <Kamion> pitti: libwnck
[06:50] <Diziet> Firstly, you didn't fix up the transition yet, did you ?  I mean, where you can end up with both directories because of dpkg/symlink transition issues ?
[06:50] <Diziet> Secondly, these links are needed for the embedders, not for the libnss/libnspr users, right ?
[06:50] <pitti> I think libaqbanking-plugins-libgwenhywfar17 is my most favourite package name, though
[06:51] <seb128> rofl, that's a real package
[06:51] <Mithrandir> pitti: looks like somebody who has taken my advice of using pwgen-generated names to heart.
[06:51] <pitti> I couldn't type that without tab completion...
[06:51] <ijuz> Keybuk: i tried it, on a 2 Ghz Athlon XP the decompression of the packages on the breezy cd (compressed with lzma) took 107 sec that results in about 15 MB/s of decompressed data; memory usage is sligtly over 8 MB when you compress with 8 MB dictionary
[06:53] <dholbach> Diziet: I didn't think of the transition issue, when i added the symlink. Kamion told me afterwards. About your second point: I don't know - aren't embedders libns{s,pr} users? :-)
[06:54] <dholbach> Diziet: Somebody, I think it was Kamion, gave me a small script for the transition, let me try to find it and send it to you in a query.
[06:54] <Diziet> dh: Yes, embedders are libns{s,pr} users.
[06:55] <Diziet> But libnss/nspr users just look in the .pc file which has the /usr/include/mozilla path and should just work.
[06:55] <Diziet> It's the embedders (and the firefox binary itself) which look in /usr/lib/mozilla-firefox.
[06:55] <Keybuk> pitti: if I'm getting up, you are
[06:55] <Diziet> (It's a bit schizoid because the ffox build system wasn't meant to do this.)
[06:55] <Kamion> mdz: I take it you want to have that talk after the management meeting now?
[06:55] <Diziet> transition script> 'sok, I can roll my own.
[06:56] <mdz> Kamion: yeah
[06:56] <pitti> Keybuk: I'm already afraid enough of JaneW's whip :)
[06:56] <Diziet> There's an attempt in the bug report even :-).
[06:56] <mdz> Kamion: assuming you'll be around
[06:57] <Kamion> yeah, although I'm having dinner soonish
[06:57] <Diziet> OK, well, I'll put it in the firefox-dev package and see what breaks.
[06:59] <dieman> heh
[06:59] <dieman> i had one user get cranky about us going to UTF8
[06:59] <mdz> Keybuk: ok, I'm guessing 50
[06:59] <dieman> grep is slower (at least in hoary)
[06:59] <dieman> need to check if newer versions are appreciably faster.
[06:59] <pitti> Riddell: still here?
[07:00] <Keybuk> mdz: 50?
[07:00] <Keybuk> or do I need to take your tab key away? :P
[07:00] <mdz> Keybuk: <mdz> Keybuk: what's the correct first_step value in rc for runlevels 0 and 6, so I can fix it locally?
[07:00] <Keybuk> ahh
[07:01] <Keybuk> sorry
[07:01] <Keybuk> I have no idea why x-chat didn't hilight that, strange
[07:01] <Keybuk> 0 I would guess
[07:02] <Riddell> pitti: yo
[07:02] <Kamion> dieman: I think that got fixed eventually upstream, but it took a while
[07:02] <pitti> Riddell: I'm currently reviewing the patch, and I will do the patching and exploit testing tomorrow
[07:03] <Kamion> grep was using the libc multibyte functions, which were a bit too generic and slow for what it was doing; Markus Kuhn mentions the issue on one of his Unicode pages
[07:03] <Riddell> pitti: ok, need me to do anything?
[07:03] <pitti> Riddell: however, since this is really not funny any more, maybe you could build koffice and kpdf against poppler in dapper?
[07:03] <Kamion> http://www.cl.cam.ac.uk/~mgk25/unicode.html#mod
[07:03] <pitti> Riddell: it's really pretty easy, you just have to tickle the build systems a bit; it was pretty easy for tetex-bin and xpdf-utils
[07:03] <Riddell> hmm
[07:04] <pitti> Riddell: using /usr/include/poppler and linking against -lpoppler is all it should take
[07:04] <Riddell> if we get libgs.so in we can use okular which is the kpdf replacement that uses poppler
[07:04] <pitti> Riddell: well, maybe do some small adaptions (s/GString/GooString/)
[07:04] <Kamion> oh, crap, it's tonight's meeting that's 2am
[07:04] <martink> kpdf has some enhancements to the Xpdf code that are not in poppler yet (or so its author says)
[07:04] <pitti> Riddell: I leave that to you :) I Just wanted to point out the possibility
[07:05] <Kamion> can somebody send an ubuntu-devel-announce@ reminder please? it's a bit late, but might help
[07:05] <Diziet> *snort*
[07:05] <pitti> JaneW pointed to it in her last report, but of course another one shouldn't hurt
[07:06] <dieman> Kamion: yeah, im reading the bug, #1148
[07:06] <dieman> Kamion: looks like it was fixed, reverted, then fixed again
[07:07] <dieman> Kamion: appears dapper has the version with the patches
[07:08] <pitti> Kamion: sent
[07:08] <Kamion> pitti: thanks
[07:08] <Kamion> (approved)
[07:11] <pitti> Riddell: btw, http://scary.beasts.org/security/b0dfca810501f2da/CESA-2005-003.txt has some shiny exploits for testing
[07:12] <Kamion> Diziet: should #1148 be closed?
[07:13] <Diziet> I don't know; I haven't tested on our end.
[07:13] <pitti> mdz: could you please take a short look at the status whiteboard of https://launchpad.net/distros/ubuntu/+spec/gstreamer-audio-backend ?
[07:13] <Diziet> I don't know of any reason why we would disagree with Debian.
[07:13] <pitti> mdz: everything is implemented from our side, and the missing upstream bit doesn't actually conflict with our changes; would you be fine if I set this to implemented?
[07:14] <mdz> pitti: meeting now, but will take a look later
[07:14] <pitti> oh, sorry
[07:21] <Kamion> Mithrandir: is there a way I can get at the unmodified CD filesystem from the current live CD?
[07:22] <Mithrandir> Kamion: I could arrange that.
[07:22] <Kamion> please - need it for u-e
[07:22] <Kamion> (per https://wiki.ubuntu.com/UbuntuExpress/CopyFileSystem
[07:22] <Kamion> )
[07:22] <Mithrandir> sure, I'll fix that.
[07:28] <seb128> pitti: hum? distro meeting at 2am? previous one was at 2pm no?
[07:29] <pitti> seb128: previous one was 1400, but the 2000 was skipped due to holidays
[07:29] <seb128> oh, skipped, not shifted
[07:29] <seb128> grumpf
[07:30] <seb128> are you sure? :)
[07:30] <pitti> seb128: JaneW wrote the date in her last report
[07:30] <pitti> seb128: I'd prefer 2000, too, believe me
[07:30] <seb128> bah
[07:30] <ogra_ibook> sigh
[07:30] <seb128> k, I'll stop working now for a bit so, or I'll no keep until 3am
[07:31] <pitti> seb128: I will sleep before :)
[07:31] <seb128> I don't want to try that
[07:32] <seb128> waking up after 2 hours of sleep ... not for me
[07:35] <psusi> pitti: do you think it would be rather easy to patch pumount to remount r/o first, then really unmount?  or would that not be so simple?
[07:38] <pitti> psusi: that's what I had in mind when we do that
[07:39] <psusi> cool... the progress bar would be nice ( especially if it were actually accurate ) but the remount should be a much simpler fix... if it is a simple bug fix, can't it even be back ported to breezy?
[07:40] <pitti> psusi: let's see, maybe the gnome guys will figure out a nice progress bar or sth
[07:40] <pitti> psusi: I doubt that it will be backported, but for dapper it's an easy alternative, right
[07:40] <psusi> well, like the author of gumount said, the kernel doesn't export enough information
[07:40] <pitti> right, progress bar is probably hard
[07:40] <pitti> but at least a warning dialog would be sufficient
[07:41] <psusi> hell, as long as the icon doesn't vanish until the flush is done, that would keep people from yanking the drives early and corrupting their data
[07:41] <pitti> (well, more of a 'monolog' :) )
[07:41] <psusi> though they might wonder WTF it isn't going away when they chose eject
[07:42] <ogra_ibook> pitti, make it a game ... to keep the user busy until he can unplug ;)
[07:42] <psusi> lol
[07:42] <zul> ooh tetris would be my vote
[07:42] <zakame> haha
[07:43] <psusi> I say pong
[07:43] <marcin_> marcin: jestes ?
[07:43] <ogra_ibook> if you integrate pmount with n-m, you could make a netris session :)
[07:47] <pitti> psusi: yes, you could shoot the dirty blocks into the USB device, and you'll lose all data that miss the target in the game :)
[07:47] <marcin_> hmm
[07:47] <marcin_> nvm
[07:48] <psusi> roflmfao
[07:48] <marcin_> :D
[07:48] <marcin_> freenode is beautifull :D
[07:48] <psusi> I used to know a crazy programmer who said the only game worth playing was one where it blew away random clusters on your hard disk as you took damage ;)
, I hope?
[07:48] <marcin> :>
[07:48] <marcin> :P
[07:48] <marcin> bye
[07:53] <dholbach> seb128: report written
[08:00] <Keybuk> right, dinner time I think
[08:00] <Keybuk> see everyone at OHMYGOD am
[08:01] <wasabi> There a trick to getting vmware compiling on 2.6.15-10?
[08:25] <HiddenWolf> jdub, is there any way that the fridge can show where it links to instead of those node links?
[08:25] <HiddenWolf> jdub, call me paranoid, but I'd like to know what I'm clicking on.
[08:27] <Treenaks> HiddenWolf: http://www.w3.org/Provider/Style/URI you mean? :)
[08:28] <HiddenWolf> Treenaks, just that when I see a link pointing to fride.ubuntu.com/bla, I expect to end up at f.u.c and not at eweek. If it points to eweek, I want to know that before I click on it.
[08:28] <xhaker> new libsane package tryed to do some wierd shit
[08:30] <xhaker> it's trying to remove /etc/hotplug dir
[08:30] <xhaker> that's not a good idea
[08:30] <xhaker> it does fail.. it's a rmdir
[08:30] <ogra_ibook> nope, since this dir shouldnt exist on dapper
[08:31] <xhaker> blaccklist.d is in ther
[08:31] <ogra_ibook> dpkg -S /etc/hotplug/blaccklist.d ?
[08:31] <ogra_ibook> -c
[08:32] <Kamion> xhaker: it was probably just the last package on your system owning /etc/hotplug, so dpkg tried to remove it; nothing to worry about
[08:33] <Kamion> note that stuff created by maintainer scripts is not generally known to dpkg
[08:36] <ogra_ibook> ogra@aleph:~$ dpkg -S /etc/hotplug/blacklist.d
[08:36] <ogra_ibook> hotplug, alsa-base, linux-sound-base: /etc/hotplug/blacklist.d
[08:36] <ogra_ibook> mine is cleaver :)
[08:36] <ogra_ibook> -a
[08:37] <Kamion> oh, what do you know, it is the maintainer script doing it
[08:37] <Kamion> it doesn't matter though, it's || true'd and it doesn't try to force it
[08:38] <Kamion> it should really use rmdir --ignore-fail-on-non-empty though
[08:38] <Kamion> (which would suppress the error message too)
[08:41] <Kamion> I'll go clean that up
[08:44] <Kamion> (done)
[08:54] <xhaker> ;)
[08:58] <mhz> fabbione: ping
[09:38] <fabbione> mhz: pong
[09:47] <lamont> hrm... /me will miss the distro-team meeting tonight
[09:48] <fabbione> OH CRAP
[09:48] <fabbione> it's already THAT thursday of the month!
[09:48] <pitti> fabbione: yes :(
[09:48] <ogra> yup
[09:48] <fabbione> pitti: why #ubuntu-devel????
[09:49] <fabbione> we always did it in #ubuntu-meeting
[09:49] <Burgwork> fabbione, likely a typo
[09:49] <Burgwork> jdub, ping
[09:49] <pitti> fabbione: oh, sorry, that was a typo
[09:49] <fabbione> eheh
[09:53] <fabbione> elmo, Kamion: would you be so nice to NEW the sparc kernel when  you have a minute? so i can kick l-r-m
[09:53] <elmo> fabbione: it's not in NEW
[09:54] <elmo> does anyone know if it's safe to take home from an amd64 box and copy it across to an i386 one?  i.e. any common apps which store stuff in 32/64 sensitive format in ~/
[09:55] <fabbione> elmo: thanks... i missed the mail of ACCEPT :(
[09:55] <fabbione> elmo: it should be. i share my /home with all our 7 arches..
[09:55] <fabbione> make that 6..
[09:56] <fabbione> m68k doesn't count :)
[09:56] <elmo> fabbione: any 64 vs. 32 tho?
[09:56] <elmo> I suppose endian switching should be guarantee enough
[09:56] <fabbione> elmo: amd64/hppa/i386/sparc64/ia64/ppc
[09:57] <fabbione> i use all of them..
[09:57] <fabbione> same /home
[09:57] <elmo> fabbione: k, thanks
[09:57] <fabbione> i didn't notice anything...
[09:57] <fabbione> ... hmm.. hold on.. where is my GPG key???
[09:57] <fabbione> :P
[09:57] <maswan> work has $HOME in afs, shared between amd64, i386, AIX.. hmm.. no sparc/solaris anymore though
[09:58] <maswan> I'd consider it a fairly big bug to have things break because you switch arch
[09:59] <elmo> yeah, I'd assume it would work, but I've never tried it before, esp. on a desktop, so I thought I'd ask
[09:59] <fabbione> elmo: are you just tarring up the data? or moving the disk?
[10:00] <elmo> rsyncing from backup
[10:00] <elmo> I'm trashing an amd64 laptop, reinstalling it as i386
[10:00] <fabbione> nah no problem than
[10:01] <fabbione> elmo: why?
[10:02] <ogra> oh, have fun with the fans then :)
[10:02] <fabbione> elmo: do you have userland or kernel problems?
[10:03] <elmo> fabbione: userland, need proprietary crap to work
[10:03] <fabbione> elmo: as proprietary drivers for the kernel i assume
[10:03] <elmo> no, proprietary userland
[10:03] <fabbione> ah
[10:03] <fabbione> ok
[10:03] <fabbione> elmo: than let me give you a suggestion
[10:04] <fabbione> install whatever release you want with standard kernel
[10:04] <fabbione> grab the the proper amd64 kernel .deb
[10:04] <fabbione> dpkg -x blabla.deb .
[10:04] <fabbione> dpkg -e blabla.deb
[10:04] <fabbione> in DEBIAN/* there are only 3 or 4 references to amd64
[10:04] <fabbione> that changed to i386
[10:05] <fabbione> + a dpkg-deb -b
[10:05] <fabbione> will give you a nice amd64 kernel for an i386 userland .deb :)
[10:05] <fabbione> elmo: sounds crack, but it works perfectly
[10:05] <fabbione> and you get all the hw support from amd64 with 32 bit userland
[10:06] <ogra> oh, thats nice to know ...
[10:06] <maswan> or just not bother about it, sounds unlikely that you need 64-bit adressing on a laptop
[10:06] <elmo> fabbione: hmm, nice idea, but I'm going for maximum stability + simplicity, as this laptop's going on the road with a new-to-linux-user and needs to Just Work
[10:06] <sivang> elmo: is that a light weight laptop?
[10:06] <maswan> and the rest of it is just performance
[10:06] <ogra> my fans always freakout if i use a 32bit os on this amd64 lappie
[10:06] <fabbione> elmo: ok.. in that case it makes sense, be sure the i386 kernel can control fans and so on..
[10:07] <fabbione> maswan: well. not only. there are some chipsets that are not supported by x86
[10:07] <fabbione> the one that are specific for x86_64 only
[10:07] <fabbione> like dunno.. thermal control?
[10:07] <ogra> yup
[10:07] <fabbione> cpu_freq?
[10:07] <fabbione> otherwise the battery on that laptop will last approx 10 minutes
[10:08] <fabbione> elmo: i think we will ship such a kernel in universe.. BenC and I didn't agree yet on how to build it :)
[10:08] <ogra> mine hovers 1m above the ground on i386 kernel... i never tried it on byttery
[10:08] <spacey_ki> :)
[10:08] <spacey_ki> quite impressive with that weight :)
[10:09] <ogra> heh
[10:09] <elmo> no, it's a compaq
[10:09] <elmo> presario v2200 or something
[10:09] <maswan> fabbione: oh, I didn't know that
[10:09] <ogra> ah, that leaves some hope
[10:09] <elmo> I'll check the fans etc. work once it's installed
[10:09] <fabbione> maswan: ehehe
[10:13] <elmo> wow, that's impressively confused
[10:13] <elmo> the wireless stack thinks the pcmcia card is both eth1 _and_ eth2
[10:16] <dilinger> heh
[10:16] <dilinger> someone else saw this earlier
 08:22:23.908402 00:30:48:51:88:c3 > Broadcast, ethertype ARP (0x0806), length 42: arp who-has 209.123.2.10 tell 209.123.2.10
[10:16] <dilinger> let's hear it for confused kernels
[10:17] <fabbione> elmo: is that an aironet/cisco card?
[10:18] <psusi> rofl
[10:18] <psusi> ohh, that's probably an anti collision check
[10:19] <elmo> fabbione: yeah it is
[10:20] <fabbione> elmo: i was that problem only with the airo pci version of the card and only with .15. I didn't have time to bisect the kernel, but unload and reload the module seems to do the trick
[10:20] <fabbione> and it doesn't even happen all the time
[10:20] <elmo> it didn't happen with an amd64 kernel :(
[10:20] <fabbione> s/was/saw
[10:20] <fabbione> yes.. i386 only it seems
[10:20] <fabbione> ppc doesn't do that either
[10:21] <elmo> yeah, this is the cisco from my powerbook, and I never saw it there
[10:21] <fabbione> elmo: i guess that one is .12, isn't it?
[10:22] <ogra> for a newbie on the road ? 
[10:22] <fabbione> ogra: i think that laptop is new enough that needs .15
[10:23] <ogra> but that breaks a lot with breezy ...
[10:24] <maswan> it's just a sign that the release cycle is way too long, it should be a new release every 3 weeks!
[10:24] <fabbione> ahah
[10:25] <fabbione> actually releasing every month or two is almost doable, if you can live without new features and no new crack :P
[10:25] <ogra> 3 weeks is fine, then the freezes would only be 2 weeks max :)
[10:26] <lifeless> moin
[10:26] <maswan> fabbione: actually, I'd love to see a breezy backport of 2.6.15. we currently run a hacked-together kernel.org on our workstations, but it lacks some stuff, like alsa. :)
[10:26] <maswan> fabbione: flight-2 wasn't quite up to it for my main workstation at work. :)
[10:27] <fabbione> maswan: you can't.. you will need too much from dapper to make it worth a backport
[10:27] <maswan> fabbione: yeah, I know
[10:27] <maswan> fabbione: that's why it's a kernel.org and not dapper kernel
[10:27] <fabbione> ehm...
[10:27] <fabbione> well whatever you say :)
[10:28] <maswan> "it's" - the thing we run now
[10:28] <maswan> would have been neat if there was an ubuntu 2.6.15 to run though
[10:28] <maswan> but I'll guess we'll just wait until dapper releases
[10:29] <fabbione> iirc the udev requirements between ubuntu and kernel.org are the same
[10:29] <maswan> ah
[10:29] <sivang> anyway, I'm off - night everyone
[10:29] <maswan> fabbione: well, it works mostly fine for us as it is
[10:32] <maswan> fabbione: just saying that it was rather annoying finding breezy and all other relaesed distributions uninstallable. :)
[10:32] <fabbione> maswan: you lost me...
[10:33] <fabbione> maswan: what do you mean other released distros uninstallable?
[10:33] <fabbione> what workstations do you have?
[10:33] <Burgwork> how do I debug gnome-session hanging?
[10:33] <maswan> fabbione: fujitsu-siemens athlon64 x2 with a new enough sis chipset to need 2.6.15 to find the sata disk
[10:34] <psusi> can mutt generate mbox indexes, or does it have to read and parse the entire thing every time like pine?  that doesn't work too well on 50 MB mboxes
[10:34] <fabbione> maswan: ahhhhhh ok
[10:34] <maswan> fabbione: the bundled fedora core 4 dvd didn't install either. :)
[10:34] <mdke> psusi, #ubuntu can help
[10:34] <fabbione> maswan: that's something similar to what i had to do to install my ppc
[10:35] <psusi> heh... sivang mentioned he uses mutt... figure'd pose the question... I don't think it can, and i"m happy with thunderbird... was just curious
[10:35] <maswan> fabbione: we just hacked together something out of kernel.org sources though, and it works. lacks some stuff, like alsa sound and possibly hal/dbus stuff. and splashy thingies. but it works. :)
[10:35] <ogra> Burgwork, easiest is probably to run a failsafe terminal session and run gnome-session manually
[10:36] <ogra> Burgwork, or run it with gdb if you have a -dbg package around
[10:36] <fabbione> maswan: you could try installing .15 from dapper on one of them.
[10:37] <maswan> fabbione: We have installed flight-2 on them. Installing the dapper kernel seemed to want to pull in a bit too much from dapper though.
[10:37] <maswan> fabbione: we have test-installed flight-2 I mean
[10:37] <fabbione> ok
[10:46] <elmo> fabbione: fyi, that tricked didn't work too great for this laptop, pcmcia doesn't come up at all
[10:46] <elmo> or at least airo_cs doesn't work
[10:46] <elmo> and I've got "APIC error on CPU0" in the logs 
[10:46] <fabbione> hold on
[10:47] <fabbione> cat /proc/cmdline 
[10:47] <fabbione> root=/dev/mapper/Ubuntu-root ro quiet splash noapic
[10:47] <fabbione> try adding the noapic to the boot options
[10:47] <fabbione> or disable it from the BIOS if you can
[10:47] <fabbione> apic makes baby jesus cry
[10:48] <elmo> this laptop is making me cry
[10:48] <fabbione> do you also have airo loaded?
[10:48] <fabbione> or only airo_cs ?
[10:48] <ogra> i thought thats only the liar apic or lapic :)
[10:48] <fabbione> i know mine is airo..
[10:48] <fabbione> the module i unload and reload
[10:50] <elmo> fabbione: ok, got rid of the APIC errors with no apic,but still no cisco love
[10:50] <elmo> modprobe-ing airo or airo_cs runs without  error, it just doesn't find the card
[10:50] <fabbione> oh
[10:51] <fabbione> did you try to stop and start pcmcia from init.d?
[10:52] <elmo> that generated a bunch of ioctl32(cadmgr:7557): Unknown cmd fd(3) cmd(c0146402){00} arg(0806f388 on /var/run/cm-7557-1 (deleted) in dmesg
[10:52] <elmo> I think/guess 32-bit cardmgr doesn't work with 64-bit kernel (anymore?)
[11:06] <fabbione> ehm
[11:06] <fabbione> i thought you were using a 32bit kernel?
[11:06] <ogra> he tried your suggestion
[11:07] <fabbione> elmo: it's possible that compat_ioctl are broken
[11:07] <fabbione> there are not that many people using that setup
[11:07] <fabbione> i guess you win the biscuit for finding one of them
[11:11] <fabbione> elmo fabbione: fyi, that tricked didn't work too <- i guess this one was referred to amd64->i386
[11:11] <fabbione> anyway
[11:11] <ogra> thats how i understood him 
[11:11] <fabbione> time to boot into the new kernel
[11:12] <fabbione> brb (hopefully)
[11:12] <elmo> ok, not feeling the ubuntu love on turion laptops, it has to be said
[11:13] <ogra> cant you run the proprietary app in a 32bit chroot on the 64bit os ? 
[11:14] <ogra> since i guess the amd64 version runs fine
[11:31] <elmo> argh, how do I convince gnome to use run a damn shell script and not throw up a confusing prompt?
[11:31] <ogra> you reate a launcher to launch the script
[11:31] <ogra> *create
[11:32] <elmo> how do I do that?
[11:32] <ogra> right click on the desktop
[11:33] <seb128> elmo: you mean the "show" "launch" ... dialog?
[11:34] <seb128> there is a nautilus preference to always open them as text or always run them
[11:34] <elmo> seb128: ah, thanks
[11:34] <ogra> ah, yes, that too
[11:34] <elmo> I'll try a launcher tho, that sounds fun
[11:36] <lifeless> seb128: an actual preference? or a gconf hidden flag ?
[11:36] <seb128> lifeless: UI preference
[11:36] <ogra> lifeless, actual preference iirc ...
[11:36] <lifeless> wow
[11:37] <lifeless> sabdfl: btw LOVED your warthogs mail ;)
[11:38] <fabbione> this wasn't such a great jump
[11:38] <ogra> :(
[11:39] <fabbione> cups doesn't stop properly
[11:39] <fabbione> there is a typo somewhere in the init scripts at shutdown
[11:39] <fabbione> that gives a syntax error on /etc/init.d/rc
[11:40] <fabbione> reboot asked me for a root/maintain passwd after that error (i assume they are related)
[11:40] <fabbione> (line 210 of that file)
[11:40] <fabbione> at boot we have "Setting system clock" twice
[11:40] <fabbione> a postfix reload error
[11:41] <fabbione> and the worst thing is the new gdm logout screen that seems to be written by m$ :P
[11:41] <ogra> oh thats in already ? 
[11:41] <fabbione> imho it's horrible
[11:42] <fabbione> it did really scare me
[11:42] <ogra> i havent upgraded for some days
[11:42] <fabbione> this is a 30 minutes ago dapper
[11:42] <lamont> fabbione: hwclockfirst.sh and hwclock.sh
[11:42] <fabbione> lamont: is that wanted?
[11:42] <lamont> I suppose I could make the messages slightly different
[11:42] <lamont> iz necessary
[11:42] <fabbione> ok
[11:42] <lamont> it's always been there, but I cleaned up a bunch of bugs by merging the files
[11:43] <fabbione> i think it would be a good idea to see 2 slightly different messages
[11:43] <lamont> so now hwclockfirst.sh is a sed command to generate it from the other
[11:43] <lamont> fabbione: ok.  file a bug in bz and assign it to me?  (or at least make me a cc)
[11:43] <fabbione> lamont: sure :)
[11:43] <fabbione> it's nothing urgent
[11:43] <lamont> what was the postfix reload error?
[11:44] <fabbione> for people used to read what happens at boot, it seems an error
[11:44] <fabbione> lamont: dunno... it's masked by the boot thingy and lsb
[11:44] <fabbione> it happens immediatly after network comes up
[11:44] <fabbione> so i *think* it's that script you put in /etc/network/if-up.d or something
[11:45] <fabbione> i assume the error is due to the fact that postfix isn't running et
[11:45] <fabbione> yet
[11:45] <fabbione> and it attempts to stop/start it again
[11:45] <lamont> fabbione: actually, the old hwclockfirst.sh didn't print anything
[11:45] <lamont> ew
[11:45] <lamont> I suppose I'll have to track that down
[11:52] <elmo> holy COW this is annoying
[11:55] <elmo> is there anythingmore I have to do in gnome to get it to change the keyboard, beyond selecting the layout in keyboard preferences?
[11:55] <elmo> maybe I'm going insane, but I'm sure this worked for amd64 and is no longer working on the same machine in x86 mode
[11:55] <seb128> what is not working?
[11:56] <elmo> it's stuck in UK mode, and I want it in Canadian French
[11:56] <elmo> (seriously :P)
[11:56] <seb128> did you select different layouts or only one?
[11:56] <elmo> I've still got UK in there, it's just not marked as default and is at the bottom
[11:56] <ogra> and did you use the anguage selector ? 
[11:56] <elmo> I'll try removing it entirely
[11:56] <seb128> it's xorg bog
[11:56] <ogra> *language
[11:56] <elmo> ogra: no! I only want the keyboard changed
[11:56] <elmo> seb128: oh, really?
[11:56] <seb128> yep
[11:56] <elmo> is there a workaround?
[11:57] <seb128> do you use breezy or dapper?
[11:57] <seb128> having 1 layout only is a workaround
[11:57] <elmo> seb128: breezy
[11:57] <elmo> ok, 1 layout works
[11:57] <elmo> seb128: thanks a lot
[11:57] <seb128> np
[11:58] <seb128> that's supposed to be fixed with xkeyboard-config from breezy-updates, daniels fixed that after the 5.10 freeze
[11:59] <elmo> hmm, it doesn't seem to fixed, and I've got breezy-updates installed
[12:00] <seb128> does "setxkbmap -layout 'ca,gb' -model pc105 -option '' | xkbcomp - :0.0" returns some error after the list of warnings?
[12:00] <elmo> oh dear god, it's STILL in UK mode
[12:00] <seb128> pc105 to replace by your pc10n variant
[12:00] <jdub> fabbione: you're famous! :)
[12:00] <jdub> elmo: quebecistani :)
[12:00] <fabbione> jdub: ?
[12:01] <jdub> fabbione: your interview -> FRIDGE TIME!
[12:01] <ogra> i'd still check the input methods column in the language selector .... :)
[12:01] <fabbione> jdub: dude HALT!
[12:01] <fabbione> jdub: only if you add one NOTE
[12:01] <jdub> yeah?
[12:01] <fabbione> jdub: it's not my ITAGLISH in there
[12:01] <fabbione> i was asked to answer in italian
[12:01] <jdub> hahaha
[12:01] <fabbione> somebody did translate it
[12:01] <jdub> ok, will do
[12:01] <fabbione> and missed some stuff