[12:03] <Kamion> I prefer to leave sources.list.d for other packages and admin customisation
[12:03] <Kamion> zyga: oh, membership? yeah, I'll catch up tomorrow
[12:03] <Riddell> Kamion: great, thanks
[12:04] <Kamion> sorry I didn't do it over Christmas, but, well, I've been on holiday :)
[12:04] <zyga> Kamion: yes about membership, I'll mail you :)
[12:05] <Kamion> zyga: nah, don't bother, I've already got mail from someone (Lucas I think?) about it
[12:05] <zyga> :)
[12:08] <Kamion> if anyone knows "Keyes", please tell him/her that integration of "EasyUbuntu" in Dapper is not an appropriate topic for the community council
[12:11] <seth_k|lappy> I'll have robotgeek tell him, Kamion 
[12:14] <lucas> I'll ask him to consider going through REVU for easyubuntu
[12:19] <Burgundavia> jdub, you have any comments on  https://wiki.ubuntu.com/HelpingUbuntu
[12:19] <jdub> Burgundavia: can you mail me a ping? :)
[12:19] <Burgundavia> jdub, sure
[12:20] <jdub> thanks
[12:20] <jdub> i liked the easy page
[12:20] <jdub> and doc project news == very good
[12:21] <Burgundavia> that is mostly mdke and jerome. I can't claim any credit for that
[12:22] <Burgundavia> jdub, sent
[12:30] <Burgundavia> hmm http://www.nubuntu.org/
[12:32] <tseng> Burgundavia: "copyright 2005 canonical ltd"?
[12:33] <Burgundavia> is this an official subproject then?
[12:33] <tseng> unlikely
[12:34] <tseng> Tech Email:ownthebox.net@gmail.com
[12:34] <Burgundavia> http://en.wikipedia.org/wiki/NUbuntu
[12:35] <tseng> also notice the references to some off the wall irc network
[12:35] <tseng> he just ripped off the copyright notice for some reason
[12:35] <Burgundavia> it doesn't appear to be plone
[12:35] <tseng> he could have joined ubuntu hardened..
[12:35] <Burgundavia> people have this cathedral like view of distros sometimes
[12:36] <tseng> some people like to derive things unessecarily for some reason
[12:36] <Burgundavia> observe doc.gwos.org
[12:36] <slomo> tseng: maybe because they want to do something on their own? ;)
[12:36] <FireRabbit> tseng, its so they can change the graphics
[12:37] <FireRabbit> because you arent a hacker unless you use a dark theme
[12:37] <tseng> hah well i made a livecd of ubuntu with ethereal on it
[12:37] <tseng> i forgot to make my own website about it :(
[12:37] <FireRabbit> :(
[12:37] <tseng> dinner time, take care
[12:38] <FireRabbit> ttyl
[12:38] <FireRabbit> people sure waste no time putting stuff on wikipedia
[12:38] <FireRabbit> we dont need websites, we just need a wikipedia page
[01:29] <CarlFK> where is the right place to point out that the "list of files" on http://mirror.mcs.anl.gov/pub/ubuntu-iso/5.10/ is "worthless" (they are all truncated such that you can only tell install from live)
[01:29] <CarlFK> other mirrors are ok: http://mirror.yousendit.com/ubuntu/5.10/
[01:29] <HrdwrBoB> I would say the admin for the mirror.mcs.anl.gov site?
[01:30] <FireRabbit> what is the "opposite" of the "dput" command? - to remove a package from the repository
[01:55] <Riddell> ogra: do you know dsaa  Dino Solon A. Agcambot?
[02:29] <akurashy> is xorg 7 in dapper yet? :o
[02:31] <crimsun> no.
[02:37] <daniels> crimsun: well, basically yes
[02:37] <crimsun> daniels: aside from doc changes, correct?
[02:39] <lamont> jdong/Kamion: given-back packages are automatically retried.
[02:45] <psusi> SEJeff, are you around?  weren't you the one that had a page about "make it pretty I say!"?
[03:36] <SEJeff> psusi, ping
[03:36] <psusi> hey, sup?
[03:37] <SEJeff> psusi, You wanted that script I wrote to set up ubuntu?
[03:37] <SEJeff> psusi, Or did you just want the gconftool-2 commands to set up gnome to be a bit more sane?
[03:37] <psusi> didn't you have a page explaining the various ways to tweak things and make things look pretty?  I'd like to improve my fonts a bit I think
[03:38] <SEJeff> psusi, No, I wrote a script to set up a default ubuntu install and set things like ctrl alt del to gnome-system-monitor etc
[03:39] <SEJeff> psusi, It also stops and removes unneeded services, does some speed tweaks and whatnot
[03:39] <psusi> yea, I remember looking at it... I just thought I saw something about making fonts look nicer
[03:39] <psusi> now why the hell would this end up being zero?
[03:39] <psusi> pd->settings.size = be32_to_cpu(ti.fixed_packet_size) << 2;
[03:40] <psusi> when ti.fixed_packet_size is 0x80000000
[03:40] <SEJeff> exactly what I was thinking all along :)
[03:40] <SEJeff> I'm too tired to play hex. I just ran about 9 miles
[03:41] <psusi> lol
[03:46] <psusi> BAH! pd->settings.size is only an 8 bit type!
[03:47] <psusi> this is a load of gnarly code... how did it make it into the kernel?  sheesh.. and why do I get the feeling that I'm going to end up becoming it's maintainer?
[03:50] <SEJeff> psusi: And thats a bad thing?
[03:52] <psusi> I don't need that kind of responsibility ;)
[03:54] <psusi> bingo!  changed to a u32, and raised the compile time maximum packet size, and it appears to be correctly accessing this disc using 128 sector packets
[03:54] <SEJeff> psusi, Dump the patch on lkml and pretend you don't know anything about it after that. Be like the maintainer of ncpfs that never responds to emails about bugs
[03:54] <psusi> lol
[03:54] <psusi> that's what is going on here... the original developers of this stuff appear to have vanished
[03:54] <SEJeff> Do you have it working with the project utopia magic / black voodoo?
[03:54] <psusi> no new developments in over a year that I can see, and the mailing list is basically dead
[03:55] <psusi> project utopia?
[03:56] <SEJeff> Project Utopia is an umbrella project started by robert love. It includes things like the 2.6 kernel based hotplug infrastructure, gnome-volume-manager, hal, dbus, etc
[03:56] <SEJeff> It is the project that coordinates making hardware "just work TM" in linux
[03:56] <psusi> oh, yea... I made a hal fdi policy file to get it to auto mount
[03:57] <psusi> right now though, it tries to do this for all discs, not just ones formatted for packet mode... I need to fix it to detect that and act accordingly
[03:57] <SEJeff> psusi: rlove description of project utopia use cases http://kerneltrap.org/node/3450
[03:58] <SEJeff> psusi, Yes thats a problem. And does it impose any penalty on normal cdroms?
[03:58] <psusi> SEJeff, once it notices the difference, it won't change the operation of normal discs at all
[04:26] <jbailey> daniels: Around?
[04:26] <daniels> jbailey: represent
[04:26] <jbailey> =)
[04:26] <jbailey> daniels: re: 21779, it's likely not cdbs rebuilding configure, but that you have some patch that's left configure.ac with a more recent timestamp than configure.
[04:27] <jbailey> automake will then attempt to rebuild the whole mess.
[04:27] <daniels> jbailey: my 'why the christ would you do that' sentiment remains :)
[04:27] <jbailey> My usual solution to this when I'm patching configure.{ac,in} is to ALSO add AM_MAINTAINER_MODE so that unless --enable-maintainer-mode is added to the ./configure line, it will never regenerate the tools.
[04:28] <jbailey> daniels: Because the expectation is that if you've modified configure.{ac,in} more recently than configure, that you probably want it updated.
[04:28] <jbailey> Not cdbs, but usual autotools magic.
[04:28] <daniels> hm, I thought dbus already had AM_MAINTAINER_MODE in
[04:28] <daniels> i'll check it out anyway, though; thanks
[04:28] <jbailey> daniels: Np.  I'm around for 5-10m, if you want I can look at it now {for,with} you.
[04:29] <daniels> jbailey: ah, it's okay, thanks.  i think pitti and mvo are nominally the dus maintainers anyway ;)
[04:29] <daniels> i'm the first on the debian maintainer list -- i.e., least active
[04:30] <jbailey> =)
[04:30] <jdong> lamont: can you help me look at why some backports still don't have binaries then? comix is one example, 2.1-1~breezy1
[04:30] <jbailey> In other news, I haven't had a hang in a while on my X (r300, ppc64).  However, it now cmpletely fubar's the video after a while.
[04:30] <jbailey> Makes it look like an analog TV when you try watching the VCR on channel 4 instead of channel 3.
[04:31] <daniels> using xvideo, I assume
[04:31] <jbailey> The funny part is that I can tell when it's getting close to that time because the nautilus corruption stops.
[04:31] <daniels> if you could get a photo, that would be neat
[04:31] <daniels> heh!
[04:31] <jbailey> xvideo?
[04:31] <jbailey> No, just regular IRC session or whatever.
[04:31] <jbailey> But I can take a picture with my digital camera next time.
[04:31] <daniels> oh, so it screws up your whole thing
[04:31] <jsgotangco> good morning
[04:31] <daniels> sorry, I thought you meant when watching actual videos
[04:32] <jbailey> Even when I flip to a text console.
[04:32] <jbailey> Nonono
[04:32] <jbailey> The whole screen.
[04:32] <jbailey> But at least I can close x-chat and my email gracefully. =)
[04:32] <daniels> i get some interesting corruption every now and then on my r480 whereby you get random black lines roughly once every 380 scanlines, then they start multiplying, offset by a couple of lines each time
[04:32] <jbailey> Nothing like that, so far.
[04:32] <daniels> that's only with videos, though
[04:32] <daniels> so I just blame the video engine
[04:32] <jbailey> But the nautilus corruption that I usually see is that it likes to shift things one pixel or two up randomly.
[04:33] <daniels> hrm
[04:33] <jbailey> Like the text under an icon will shift up 1 or 2 pixels everytime I move my mouse over it.
[04:33] <jbailey> However, about 30-60 minutes before X pukes on me, all that corruption goes away.
[04:33] <jbailey> And Nautilus looks fine.
[04:33] <jbailey> Creepy eh? =)
[04:33] <daniels> do you have any effects on icons? like, underlining text or anything
[04:33] <daniels> i only have the folder sort of lighting up when I move over it
[04:34] <jbailey> No, it just lights up on hover.
[04:34] <daniels> intriguing
[04:35] <jbailey> A ghost of the text under the icon slowly crawls up with each highlight, though.
[04:35] <jbailey> I think I can get a screenshot if you want.
[04:35] <daniels> that would be pretty rad
[04:35] <daniels> this is on the powerpc, yeah?
[04:35] <jbailey> Yup
[04:36] <daniels> i think texture uploads there are just utterly screwed
[04:36] <jbailey> http://people.ubuntu.com/~jbailey/Capture.png
[04:36] <infinity> jdong: That didn't get auto-given-back due to a bug in sbuild that I've since fixed.
[04:37] <infinity> jdong: I'll do a mass-give-back in breezy-backports for you.
[04:37] <jbailey> daniels: Interesting things here are that everytime I exit or leave the icon, a new text is created.  So one for insert, one for removal.
[04:37] <jbailey> Also that you can see from the gnome panel at the bottom that the whole thing is already offset some.
[04:38] <lamont> jdong:   Section             : unknown
[04:38] <jbailey> And if you look at that Capture-2.png icon near the bottom, you can see that the corruption gets quite extreme after a while.
[04:38] <jbailey> And then suddenly it will all behave correctly.  No offsets, no corruption or anything.
[04:38] <lamont> jdong: --> comix needs to be added to the breezy-backports overrides file before it will successfully build (since it's not in main...)
[04:38] <jbailey> And a crash will happen soon after.
[04:38] <lamont> unknown gets lumped into main, knowing that it's probably wrong.
[04:42] <infinity> jbailey: Can it really be much worse than your usual screenshots?
[04:42] <jbailey> infinity: It was more the "make sure I don't actually have customer data on there this time"
[04:42] <daniels> jbailey: hm
[04:43] <daniels> jbailey: i've got a pretty decent idea what this is, but fixing it is non-trivial
[04:43] <daniels> jbailey: (as in, not going to happen without me shaking benh, which I'll do)
[04:44] <jbailey> daniels: Okay.  Is there anything I can do aside from shut up, sit down, and wait? =)
[04:45] <daniels> jbailey: try Option "XaaNoOffscreenPixmaps", maybe?
[04:46] <jbailey> daniels: Is that under         Identifier      "ATI Technologies, Inc. Radeon 9600 XT (RV350 AR)"
[04:46] <jbailey>  ?
[04:46] <daniels> yeah
[04:47] <daniels> sorry
[04:47] <jbailey> Bounce to check that
[04:49] <jbailey> daniels: Yes, that does it.
[04:53] <daniels> jbailey: bingo
[04:53] <daniels> jbailey: it won't be blazingly fast, but eh
[04:53] <jbailey> daniels: Cool.  If there's anythign you need me to test, just fire me an email.  I'll try to come online before I head to bed.
[04:55] <psusi> great... now it looks like the udf filesystem driver has problems...
[04:55] <daniels> jbailey: will do; don't expect anything very soon at all though -- i haven't got a powerpc with r3xx, and benh has been banging his head against various issues related to this for ages
[04:55] <psusi> it does not honor the uid,gid,umask params...
[04:56] <stratus> daniels, wasn't XaaNoOffscreenPixmaps block some optimizations? Isn't sidepanel option a better attempt?
[04:56] <daniels> stratus: well, it disables part of the acceleration, yes.  and what's 'sidepanel'?
[04:57] <daniels> jbailey: (also, exa might be differently broken for you; your call as to which is more tolerable.)
[04:57] <stratus> daniels, panelsize, sorry
[04:57] <stratus> daniels, i think  you already see but: https://bugs.freedesktop.org/show_bug.cgi?id=4294
[04:57] <jbailey> daniels: I'm not particularily worried about tollerable right now, this has been broken for sometime.  Mostly I want to help you get the most correct solution possible for release.
[04:58] <daniels> stratus: we never manually specify PanelSize
[04:58] <daniels> jbailey: cheers :) i'll ping you if I have anything more useful
[04:59] <jbailey> daniels: Thanks.  g'night!
[04:59] <daniels> jbailey: basically the swappers are pretty hopelessly fucked on r3xx
[05:00] <stratus> daniels, weird maybe Roger hit a different combination of options there
[05:01] <stratus> daniels, he's a DD so the packages aren't that different and i informed you a bug forwarded from debian bts
[05:02] <stratus> daniels, #318812 if you want to take a look
[05:02] <daniels> stratus: yeah, I've perused the bug
[05:03] <stratus> daniels, np thanks
[05:31] <psusi> boy, this is retarded... chown only accepts 16 bit uids?  but the kernel uses 32bits?
[05:37] <psusi> no.. looks like it does take 32 bit numbers... just won't accept 2^32-1
[05:37] <psusi> wtf?
[05:40] <lifeless> 0 is a valid uid, so what number represents an invalid one
[05:40] <desrt> no such thing
[05:41] <lifeless> well, twas a thought
[05:41] <desrt> well, maybe you're right :p
[05:41] <psusi> -1
[05:41] <psusi> chown(2) won't take -1 as a uid
[05:41] <psusi> very annoying
[05:41] <lifeless> desrt: an in 2's complement, -1 is ... :)
[05:41] <desrt> 0x[lots of fs] 
[05:42] <desrt> :)
[05:42] <lifeless> desrt: 2^bitwidth - 1
[05:42] <lifeless> i.e. yes :)
[05:42] <psusi> because it looks like the udf filesystem will not use the uid= mount option unless the file on disk is owned by -1... it's owned by root right now for some reason, and I can't chown it to -1 to get udf to use the mount option for the owning uid
[06:01] <jdub> apt-get CLI fat-finger usability issue: u and y are right next to each other on the keyboard ;-)
[06:02] <daniels> dist-upgrade?
[06:02] <daniels> er, dist-ypgrade
[06:03] <jdub> when using -u
[06:03] <jdub> -uy == eek!
[06:05] <SEJeff> jdub, *cough* dvorak
[06:08] <fabbione> morning guys
[06:12] <FireRabbit> hey
[06:14] <FireRabbit> i am maintaining my own apt repository .. and I use dput/debarchiver to add new packages to it... anyone know how to *remove* packages?
[06:21] <seth_k|lappy> FireRabbit, I just remove the files and then call dpkg-scanpackages, not sure if that will work for your setup
[06:21] <FireRabbit> hmm.. well dput is configured to run debarchiver with -x.. according to the man page that runs apt-ftparchiver not dpkg-scanpackages?
[06:22] <FireRabbit> i wonder if maybe i can just remove the files and call debarchiver
[06:22] <lifeless> I hate evo
[06:23] <lifeless> hung on a futex.
[06:25] <FireRabbit> seth_k|lappy, so there is no tool that will actually do the deleting for me?
[06:25] <seth_k|lappy> I'm unsure, FireRabbit.
[06:25] <FireRabbit> is there a debian devel channel i could ask in?
[06:28] <lifeless> #debian-devel
[06:28] <FireRabbit> oh.. didnt show up in the channel list
[06:28] <FireRabbit> thanks :)
[06:38] <psusi> YES!
[06:39] <psusi> I patched mkudffs and cdrwtool to set the uid and gid for the root to -1 and rebuilt the package
[06:39] <psusi> and it works... hehe
[06:49] <jdub> mdz: ping
[06:49] <ajmitch> elmo: please sync pngwriter, positron, albatross and vblade, dropping changes.
[06:49] <jdub> mighty vblade~
[06:49] <jdub> !
[06:51] <daniels> man
[06:51] <daniels> irssi CLI usability issue: ~ and ! are too close together
[06:53] <jdub> heh
[06:54] <jdub> lordy, xmltv is perl-a-go-go
[07:06] <fabbione> ajmitch: are you sure about vblade?
[07:06] <fabbione> ajmitch: it was on my todo list of pkgs to check for today
[07:07] <jdub> daniels: do you know anything about rage128 tv out support on ibooks?
[07:07] <daniels> jdub: 'not'
[07:08] <jdub> heh
[07:08] <ajmitch> fabbione: I was requesting the sync for stevenk, it wasn't an indepth look
[07:08] <fabbione> ajmitch: ok
[07:08] <daniels> jdub: i never really got atitvout to do anything useful for me
[07:08] <fabbione> i got tvout working
[07:08] <fabbione> but that was on i386
[07:08] <fabbione> i could test ppc tho
[07:08] <daniels> with r128, and ibook?
[07:08] <fabbione> daniels: is there any reason why we don't ship the theater modules from the ati driver?
[07:09] <daniels> fabbione: 'i forgot'
[07:09] <fabbione> daniels: ehehhe
[07:09] <jdub> theater?
[07:09] <daniels> i've got that fixed locally, when I bump to 1.0.0
[07:09] <fabbione> jdong: tvin/out
[07:09] <daniels> or whatever
[07:09] <daniels> jdub: yeah
[07:09] <jdub> oh :)
[07:09] <daniels> jdub: except it doesn't really work on anything except the aiws that no-one has, because gatos are, um, gatos
[07:09] <jdub> did gatos change names, or is this something else?
[07:10] <daniels> gatos merged into xorg
[07:10] <fabbione> not even all of it
[07:10] <fabbione> only a portion
[07:10] <daniels> which caused all sorts of havoc when we realised that it locked up most non-aiw cards, because obviously the only people who tested gatos were people with aiws
[07:10] <daniels> fabbione: well, all their ati.2 stuff, no?
[07:11] <fabbione> daniels: not sure 100%, the gatos site still reports some features as not merged/not working
[07:11] <fabbione> like tvout on some cards
[07:11] <fabbione> or tvin
[07:11] <fabbione> can't remember
[07:11] <fabbione> the only way i managed to get tvout was removing all the other monitors from the machine
[07:11] <fabbione> at that point tv was the only one
[07:11] <fabbione> detected and configured
[07:11] <fabbione> and that is working fin
[07:12] <fabbione> fine
[07:12] <fabbione> other setups are hitchy
[07:12] <fabbione> but right now i can only test ppc
[07:12] <fabbione> the i386 card is slammed in the hppa (by mistake) that's located underneath 3 layers of computers
[07:12] <daniels> afaict they only merged tv in because tv out is still really really sketchy
[07:13] <daniels> i used to care about such things, but I'd be lying if I said I did anymore
[07:13] <fabbione> actually... 4 layers
[07:13] <StevenK> fabbione: vblade looks like it can be synced - the package in Debian is based on what you did for Ubuntu.
[07:14] <jdub> StevenK: you still at ursys?
[07:14] <fabbione> StevenK: ok, i was only wondering if there are new upstream releases
[07:14] <StevenK> Yup.
[07:14] <StevenK> jdub: Still working there, and still at work actually.
[07:14] <jdub> how's it going there these days?
[07:14] <StevenK> Busy.
[07:15] <fabbione> StevenK: so did you check for new upstream?
[07:15] <StevenK> fabbione: I haven't. I will when I get home.
[07:15] <fabbione> StevenK: ok thanks
[07:25] <dholbach> good morning
[07:29] <jdub> i bet mythtv-common doesn't actually need pwgen and ttf-freefont
[07:29] <jdub> Keybuk!
[07:29] <jdub> dholbach!
[07:30] <dholbach> hellas jdub
[07:30] <jdub> Keybuk: usability note for init scripts and usplash -> lack of "ok" on any line makes neophyte users anxious
[07:33] <Keybuk> morning
[07:34] <Keybuk> jdub: yes, I would expect it would
[07:34] <Keybuk> good crimbo and new year?
[07:34] <Keybuk> jdub: if you're referring to the missing one in the udev scripts, I don't know why it doesn't print the ok ... there's the right call for it.  I suspect usplash bug
[07:35] <Keybuk> it doesn't print a fail either
[07:36] <infinity> Keybuk: usplash has a bit of a race, where if your init script is blindingly fast, you won't get your output.
[07:36] <infinity> Keybuk: Whether or not that's your issue, I dunno.
[07:36] <Keybuk> it's the opposite, that bit takes a few seconds
[07:36] <Keybuk> ~15s
[07:36] <infinity> In that case, it's not a usplash issue...
[07:36] <infinity> You're not backgrounding or something, and expecting to print the OK after other scripts have run? :)
[07:36] <jdub> Keybuk: doesn't it fork and return very quickly though?
[07:37] <Keybuk> nope
[07:37] <infinity> (If so, you should just "OK" on fork)
[07:37] <Keybuk> the point of that call is it's the program that waits for the backgrounded stuff to complete
[07:38] <infinity> (Are we all talking about the same script, by the way?)
[07:39] <Keybuk> I assume so
[07:39] <Keybuk> the udev startup script usually doesn't print an "ok" for the "Detecting and activating hardware" bit
[07:39] <Keybuk> never has
[07:39] <infinity> I'd have to reboot to see which one I don't get an OK on, but it'd bean a longstanding bug.
[07:39] <Keybuk> I've not bothered about it because it's cosmetic at best
[07:40] <Keybuk> funny thing is it does print it if you don't use usplash
[07:40] <Keybuk> so somewhere the log_end_msg bit doesn't reach usplash
[07:41] <Keybuk> but appears fine on the console
[07:41] <infinity> Hrm, that script is buggy in other ways.
[07:41] <jdub> networking does the same thing
[07:41] <Keybuk> it is?
[07:41] <Keybuk> jdub: networking doesn't anyway atm, gonna do that stuff at the sprint in Jan
[07:41] <infinity>  /etc/init.d/udev start shouldn't "fail" if udev is already started (say, on my running system)
[07:41] <jdub> Keybuk: yay
[07:41] <Keybuk> infinity: damn well should
[07:41] <Keybuk> you shouldn't try to start udev on a running system :p
[07:41] <infinity> Not according to both policy and the LSB.
[07:41] <Keybuk> I disagree with both
[07:42] <Keybuk> it does that by design
[07:42] <Keybuk> it's an rcS script, not a daemon
[07:43] <infinity> Then print "* udev already running... [ ok ] " and exit 0.
[07:43] <Keybuk> no ;)
[07:43] <Keybuk> that bit of policy only applies to rc2
[07:43] <Keybuk> it's fine exiting 1, in fact, it's better; it tells people they shouldn't have done that
[07:45] <Keybuk> "udev already running" could persuade people to do "udev stop" which will utterly fuck their system
[07:45] <infinity> But they can do anyway.
[07:46] <infinity> And they may do, right after the "failed" start.
[07:46] <Keybuk> sure, but there's absolutely no reason for the fully message
[07:46] <Keybuk> and those "use log_* for fluffy messages" are bugs too, imo
[07:46] <Keybuk> "echo" is sufficient if you want it printed on the console
[07:47] <infinity> If you consider it fluff, then exit with no text if it's running already. :)
[07:47] <infinity> I was just suggesting changing the text if it's running, not adding more.
[07:49] <Keybuk> if you can find an actual bug or problem with it behaving like it does, I'll consider it ;)  but I think it's fine as it is
[07:49] <Keybuk> it's a startup script, not a daemon start/stop init script
[07:50] <infinity> (Then why does it have a restart target that appears to do something useful)?
[07:51] <Keybuk> because I was trying to put useful things in there :p
[07:52] <Keybuk> amusing, that restart target has a bug
[07:54] <Mez> bit early for you isnt it Keybuk ?
[07:55] <Mez> or are you in a random country again ?
[07:55] <dholbach> "the early bird..."
[07:56] <Keybuk> heh, nope; I like waking up before sunrise
[07:56] <Keybuk> and gonna try and do a split day this year; up early so I can catch the aussies, have a long lunch at the gym, then start again in the afternoon to catch the yanks :)
[07:57] <Mez> Keybuk - lol - fair enoug h :D btw - did you see my patch to signkey.pl (and we still have to meet up for a beer at some point!)
[07:57] <Keybuk> and he said we'd still be friends, and that it wouldn't be awkward
[07:58] <jdub> haw haw haw
[07:58] <Mez> wow - people are actually here :D
[07:58] <Mez> oh, and jdub: unping :P
[07:58] <Keybuk> Mez: I can see it in my INBOX.  I haven't read e-mail for a few weeks now, so I have a *lot* of it to go through today
[07:58] <Keybuk> in fact, expect me not to finish dealing with e-mail until the end of the week
[07:58] <Mez> lol - basically I patched your script to clean signed keys :D tis shiny :D lol :D
[07:59] <Keybuk> fair enough, publish away
[07:59] <Keybuk> that script's MIT I think
[07:59] <Mez> yeah it is - lol - I just sent you a copy of the patch (though I've modified since then0
[07:59] <Mez> lol :D 
[08:00] <Mez> so 1) it doesnt export all the sigs accidentally and 2) so it checks for pgp-clean :D lol - I'm only publishing the patch anyways
[08:00] <Keybuk> I had objections to the whole idea of sending "clean" PGP signatures, but I can't remember what they are
[08:01] <Mez> lol - fair enough - I think that was your excuse at UBZ too :D
[08:01] <Mez> lol
[08:01] <Mez> so - you up for grabbing a beer at some point seeing as you're in england now ?
[08:01] <Keybuk> sure
[08:01] <Keybuk> I might actually try and get to a WolvesLUG again this month
[08:02] <Mez> lol
[08:02] <Keybuk> you should, it's fine; there's lots of beer, curry and aq
[08:02] <Mez> Keybuk - well - I have this weekend off :D lol
[08:02] <jdub> AQ ATTAQ!
[08:03] <Mez> but if you ever fancy bringing the LUG to a casino - head to dudley :D and you can see me at work
[08:03] <Mez> and Aq... is... hmm - sorta scary :D lol
[08:04] <Mez> I'm assuming you'll be at LRL this year ?
[08:05] <Mez> brb
[08:05] <Treenaks> I will be :)
[08:05] <Mez> have to go kill the *********** that's downloading crap over the network and blocking everyones net connection :D
[08:05] <Mez> Treenaks, I know :D lol
[08:05] <Treenaks> Jono asked me to describe the crack that will be going into dapper+1 ;)
[08:05] <jdub> ( | )
[08:05] <jdub>   ^ dapper+1 crack
[08:06] <Keybuk> yeah, most likely; if it isn't at a time I'm out of the country
[08:06] <jdub> Treenaks: MULTIARCH!
[08:06] <Keybuk> dholbach: oh, I came up with a motto for you guys btw ... "Magister mundi sum" :p
[08:06] <dholbach> Keybuk: you think we're that elite? :-p
[08:06] <Keybuk> you guys rock the world
[08:06] <ajmitch> 'you guys'?
[08:06] <Keybuk> motu
[08:06] <ajmitch> ah
[08:07] <jdub> ajmitch: the german football team.
[08:07] <jdub> *cough*
[08:07] <dholbach> hahahaha
[08:07] <Keybuk> jdub: dude, not at this time in the morning ;)
[08:15] <Mez> Keybuk: when's hct planning on being released for public usage
[08:17] <Keybuk> not for a while
[08:17] <jdub> really?
[08:17] <Keybuk> it needs bzr to be stable first
[08:18] <Keybuk> yeah, too many moving pieces, and people keep getting the fingers caught in the cogs
[08:18] <Keybuk> the last decision was that we'd stall it until dapper+1 when bzr and lp are a more stable base for it
[08:18] <Mez> :(
[08:18] <jdub> ahr
[08:19] <Mez> I was looking forward to it too :D
[08:19] <Mez> lol
[08:19] <Keybuk> there's not a huge amount of hct development work left, just lots of dot joining 
[08:20] <Keybuk> but we're gonna wait for the dots to be in the right places first :)
[08:20] <jdub> at least the dots aren't being shot by fully automatic revolving shotgun turrets
[08:20] <jdub> now
[08:21] <Mez> jdub: why not - that'd be cool
[08:23] <Keybuk> Mez: it depends where you're standing
[08:23] <Burgundavia> what is is about ubuntu-devel that screams "Please take my bug report now!"
[08:24] <Mez> Keybuk, behind the guns - of course - or in the targeting seat *winds the cogs*
[08:24] <jamesh> Burgundavia: it is a support escalation service, right?
[08:24] <ajmitch> jamesh: no, it's somewhere to complain about how thinly spread we all are
[08:25] <Burgundavia> jamesh, I am talking about the random community members who tell us about problems
[08:25] <ajmitch> Burgundavia: in excruciating detail, at times :)
[08:26] <Burgundavia> yes
[08:26] <Burgundavia> I appreciate their enthusiasm, just not how it is directed
[08:26] <Burgundavia> I noticed that on Fedora-devel they don't shut those people down and it seems that is more like -support
[08:27] <jdub> that's because there's nothing else to do there ;-)
[08:27] <Burgundavia> ouch
[08:27] <dholbach> what happened to the amd64 buildd?
[08:27] <ajmitch> some of those threads will have to stop soon if -devel will be readable
[08:27] <jdub> we should make sure our responses always include "this might be more appropriate on ... list"
[08:27] <jamesh> Burgundavia: I realise that it isn't ubuntu-support, but the developers read ubuntu-devel, so they are more likely to see and respond to my problem if I ask on that list
[08:28] <jamesh> Burgundavia: (and naturally, my problem is more important than the other problems reported through the standard channels)
[08:28] <Burgundavia> jamesh, indeed
[08:28] <jdub> it's *okay* to say "we have a legitimate difference of opinion - we appreciate your feedback and contribution"
[08:28] <infinity> dholbach: More detail than "what happened" would be nice. :)
[08:29] <infinity> dholbach: If you're referring to that crazy glibc error, a mass give-back will be happening later today to retry everything, once some dust has settled.
[08:29] <dholbach> infinity: ok, in the evolution build log, python2.4 doesn't seem to be installable because of some "inconsistency detected by ld.so"
[08:29] <infinity> dholbach: Yeah, that's fixed.  FWIW, all the other arches would have the same problem right now, if they weren't all shut down. :)
[08:29] <dholbach> i'll do the package updates on i386 then ;)
[08:30] <infinity> amd64 is the only arch building, while I sort some holiday fallout.
[08:31] <Mez> infinity - any idea why I'm not getting email from katie? (and are you the person to talk to ?)
[08:31] <infinity> Mez: No and no.  But which mail are you expecting to see?
[08:31] <Keybuk> I _almost_ have all three architectures now \o/
[08:31] <Mez> infinity - anything - NEW REJECTED, ACCEPTEd
[08:31] <Mez> not getting ANYTHING from katie :D havent done since september
[08:31] <infinity> Mez: For sources or binaries?
[08:31] <Mez> Sources
[08:31] <infinity> Mez: You won't get mail for binaries ever.
[08:32] <Mez> inifinty - I've never uploaded binaries :d
[08:32] <infinity> Mez: If it's for sources, then you're probably uploading with an email address that isn't in the whitelist.
[08:32] <Mez> infinity, - mez@ubuntu.com
[08:32] <dholbach> Mez: elmo will know who's in the whitelist
[08:32] <infinity> Mez: Oh.  *@ubuntu.com is whitelisted, AFAIK.
[08:32] <Mez> which should be auto-whitelisted - should it not?
[08:32] <ajmitch> Mez: overaggressive spamfilters somewhere along the line, perhaps
[08:33] <Mez> ajmitch: hmm - i'll chekc i have spamassasin off - but - hmm .... it worked before :P lol - in septemerb and i havent changed anything
[08:33] <Mez> though if spamassasins on - I can guess the rules were auto-updated
[08:34] <Mez> Spam Assassin is currently: disabled 
[08:34] <Mez> hmm
[08:34] <ajmitch> not just on your local box, I mean
[08:35] <Mez> ajmitch, no, it should goto the junk folder
[08:35] <Mez> but theres nothing from katie in the junk folder
[08:36] <Mez> nor in any other folder
[08:37] <Mez> it's very annoying
[08:37] <Mez> specially when i dont get rejects :D so i dont know if it's been rejected or NEW'd
[08:38] <Mez> I have emailed elmo - but i guess he's in the same situation as Keybuk regarding email
[08:38] <infinity> Most of us are, yes.
[08:38] <Keybuk> I shouldn't expect elmo is up yet either ;)
[08:39] <Keybuk> it's about 7 hours too early for him
[08:39] <Mez> I emailed him 3 days ago :D
[08:39] <Keybuk> this is the first year I've had so much e-mail
[08:39] <Keybuk> I've taken a new policy of not going near computers when I'm not "at work"
[08:39] <Keybuk> so they've all been turned off all holiday
[08:40] <infinity> Yeah, I did the same.  I'm now regretting it.
[08:40] <Mez> lol
[08:42] <Keybuk> I'm rather happy with it, I feel like I've actually had a holiday
[08:42] <infinity> Well, yes, that bit was nice.  The consequence this morning was not.
[08:43] <jsgotangco> lol
[08:43] <Keybuk> heh, I once decided on returning from a 3-week holiday to delete all my e-mail on the basis that anything important would get nagged again :p
[08:43] <infinity> It's like some sort of horrible electronic hangover.
[08:43] <Keybuk> this didn't work so well
[08:43] <jdub> infinity: i find that makes it really hard to feel rested - as soon as you get back, you have the stress of so much in one hit
[08:43] <Keybuk> jdub: yeah, I'm going to do a folder a day or something, so it's not one hit
[08:44] <Mez> I forgot to suspend my mailing lsits while I was away - came back to 10,000 new emails
[08:44] <jsgotangco> i still like taking small doses everyday
[08:44] <Mez> :()
[08:44] <Mez> :( *
[08:44] <Keybuk> fortunately there isn't actually a huge amount anyway, certainly a lot less than any other two weeks in the year
[08:45] <pitti> Good morning
[08:45] <ajmitch> morning pitti 
[08:45] <Keybuk> morning pitti
[08:45] <pitti> hi ajmitch 
[08:47] <Mez> morning pitti :D
[08:47] <Mez> (everyone else was doing it - I just wanted to be popular)
[08:47] <pitti> hi Mez 
[08:48] <Keybuk> you can get a cream for that
[08:48] <Mez> Keybuk - where from - I've been looking for ages ... cant find any
[08:49] <ajmitch> Keybuk: do you have the MoM source somewhere?
[08:49] <Keybuk> yes, I do
[08:49] <ajmitch> somewhere publically available & published, that is
[08:49] <Keybuk> no
[08:50] <Keybuk> I'll prod Mark again, I mailed him a while back asking for permission to open it
[08:50] <Keybuk> it probably vanished in the million-mails-a-day of his INBOX
[08:51] <Burgundavia> Keybuk, email Claire. It probably has a higher chance of actually getting to him
[08:52] <jdub> Burgundavia: you just sent a reply to ubuntu-users and ubuntu-users-owner
[08:53] <Burgundavia> jdub, hmm, ouch
[08:53] <Burgundavia> jdub, I just sent to -owner
[08:54] <Mez> jdub- btw - were you getting bounces from my email address for dapper-changes ?
[08:54] <sivang> morning everybody!
[08:54] <jdub> sorry, meant s/and/to/
[08:54] <jdub> Mez: no idea
[08:54] <Mez> lol - fair enough :
[08:54] <Keybuk> probably the main problem is that the code isn't really "releasable" quality
[08:54] <Keybuk> it's one huge uncommented python script
[08:54] <Burgundavia> jdub, it came in through -owner and I responded to both him and -owner
[08:55] <jdub> Burgundavia: oh
[08:55] <Keybuk> with dependencies on hacky scraps of python written for other purposes and culted in
[08:55] <Mez> Keybuk: uncommented huge scripts are always the most fun to break though
[08:55] <Burgundavia> jdub, to prevent all of us responding with the same thing
[08:55] <ajmitch> Keybuk: I don't mind, I just wanted to look at it & pick out anything useful
[08:55] <Keybuk> though I guess that didn't stop us releasing Germinate, which was in much the same style <g>
[08:55] <ajmitch> others might mind though
[08:55] <jdub> Keybuk: at least Kamion writes good shitty code ;)
[08:55] <tepsipakki> is my dapper the only one which is having interactivity problems.. the system feels quite sluggish, even though the hardware isn't (2.2GHz celeron, 512MB)?
[08:56] <Mez> tepsipakki, -> support = #ubunut
[08:56] <Mez> s/ubunut/ubuntu/
[08:56] <jdub> UBUNUT!
[08:56] <jdub> we own that domain
[08:56] <Keybuk> jdub: heh, it was still largely my code for the first release
[08:56] <jdub> we should do something rad with it
[08:56] <jdub> LIES!
[08:56] <Keybuk> jdub: Ubuntu fan page?  Let people post details of themselves and how they use Ubuntu and why they love it so much?
[08:57] <tepsipakki> mez: just noticed last night that someone of you devels had some problems too.. but anyway
[08:57] <Keybuk> you can have badges, "I'm an Ubunut!"
[08:57] <jdub> Keybuk: yeah, was thinking along those lines
[08:57] <Mez> Keybuk - I was thinking along those lines-  but I think we might get scary people
[08:57] <jdub> Keybuk: maybe a blog service ;-)
[08:57] <Keybuk> scary people are the most fun
[08:57] <Mez> Keybuk - you'd know ;)
[08:57] <Mez> yeah you ae
[08:57] <Mez> specially with a microphone in your hand :d
[08:57] <Mez> :-"
[08:58] <Keybuk> a friend accused my new hair cut of being terrifying :-/
[08:58] <jdub> SSHOT!
[08:58] <jdub> SSHOT!
[08:58] <Mez> Keybuk, been cut since UBZ ?
[08:58] <sivang> Keybuk: yeah, any changes to it since the hair color ? :)
[08:58] <Keybuk> yeh, I have a mohawk now :p  (I was drunk, it seemed like a good idea, and now I kinda like it :p)
[08:58] <sivang> hehe
[08:59] <Mez> mohawks are cool :D
[09:02] <jdub> is there a useful CLI tool for changing hostname post-install?
[09:02] <Keybuk> jdub: vi /etc/hostname? :p
[09:03] <jdub> other instances of hostname though
[09:03] <jdub> though i guess we use localhost.localdomain
[09:03] <jdub> man, under vmware, breezy installs lilo
[09:03] <jsgotangco> huh?
[09:03] <jdub> don't cry, baby jesus!
[09:05] <Keybuk> where else would we embed the hostname?
[09:05] <Keybuk> we don't ship an MTA anymore, which is the usual suspect
[09:06] <Keybuk> /etc/hostname and /etc/hosts look like the only two places
[09:07] <Mithrandir> hi Scott
[09:07] <infinity> Tollef!
[09:08] <Mithrandir> good morning, Adam :-)
[09:08] <Mithrandir> 'sup?
[09:08] <infinity> Not much, just rejoicing.  No computer for a week meant no Tollef for a week, which is clearly a bad thing.
[09:08] <infinity> Or something.
[09:08] <infinity> Or, I'm buttering you up, so I can reassign some bugs to you.
[09:08] <infinity> YOU DECIDE.
[09:08] <pitti> hi infinity, happy new year
[09:08] <Mithrandir> yeah, I've been meaning to ask you when we can get squashfs builds going and such.
[09:09] <Mithrandir> oh, what kind of bugs?
[09:09] <infinity> Oh, I'll surprise you. :)
[09:10] <infinity> And yes, squashfs sounds like fun.  Should we make a date to attack that Fridayish?  (post-meeting, post-holiday-catchup)
[09:10] <Mithrandir> sounds fine with me.
[09:10] <infinity> pitti: You too.
[09:10] <Mithrandir> hopefully, jbailey will have the squashfs stuff in klibc by then as well, so we don't just end up breaking the CD.
[09:11] <infinity> No one's stopping you from uploading with your patch.
[09:11] <infinity> (hint, hint)
[09:11] <Mithrandir> I thought jbailey did, and it FTBFS-ed on ppc due to lkh breakage?
[09:11] <infinity> I have an odd feeling any distro time jbailey has to spend will be spent tracking glibc 2.3.6 bugs for the next 2 weeks.
[09:12] <infinity> Oh, that's right.  Headers jumping around randomly, world confused.  I remember now.
[09:15] <Keybuk> hey Tollef!  Happy new year, etc.
[09:16] <sivang> happy new year infinity 
[09:17] <dholbach> hellas mvo!
[09:18] <Mithrandir> Keybuk: what do you think a sensible way to set up the network on the live cd would be, now that we don't do it in the initramfs any more?
[09:18] <Keybuk> we don't?
[09:18] <mvo> hey dholbach 
[09:19] <Mithrandir> Keybuk: no, the initramfs just does the minimal stuff needed to boot the system now.  simplifiedlivecd.
[09:19] <Keybuk> can we rewind a bit, I've missed the context of this :)
[09:19] <Keybuk> go gently with me
[09:20] <Mithrandir> currently, the network interfaces on the live cd are completely unconfigured.  Including lo not being upped.
[09:20] <infinity> The old initrd was a cut down d-i instance that did most everything d-i does.
[09:20] <infinity> Now it's just a regular old initramfs.
[09:21] <Keybuk> Mithrandir: you're not running S40networking?
[09:21] <infinity> I'm sure he is.  Of course, it doesn't work currently.
[09:21] <Mithrandir> Keybuk: yes, but that doesn't do anything if you don't put anything in /etc/network/interfaces, does it?
[09:22] <Keybuk> Mithrandir: true :p
[09:22] <infinity> I've intentionally not fixed my interfaces(5) file to track when that bug gets fixed. :)
[09:22] <Keybuk> infinity: yeah, that's deliberate; I'm going to work on that at the sprint
[09:22] <infinity> Mithrandir: You could write an interfaces file if you're booting live... No reason why casper can't have an init script.
[09:23] <infinity> But that breaks for people who want a static setup.
[09:23] <Keybuk> in theeeeory, "ifup lo" will be handled by a separate early init script without relying on /etc/network/interfaces
[09:24] <infinity> The real answer is "Gee, wasn't NetworkManager supposed to fix all of this?", but I still don't know if anyone has the round tuits to solve the "NM hates our way of doing things, and we hate it" issues.
[09:24] <infinity> (where "all of this" doesn't include "lo", obviously)
[09:24] <Keybuk> infinity: NetworkManager hates my Atheros
[09:24] <Keybuk> I had some keen guy looking into it, but haven't heard back from him
[09:24] <Keybuk> I suspect he hit the "oh gods, this is HARD" problem
[09:25] <Mez> Keybuk, NM decided to cease to exist for me :D
[09:26] <infinity> Keybuk: Well, I still have other fundamental issues with it.  Like, it needs a way to configure the "on boot" network settings, for instance (if you can become root).
[09:26] <StevenK> fabbione: Yes, vblade 10 is out upstream.
[09:26] <Mithrandir> it also needs to stop writing invalid resolv.conf files.
[09:27] <infinity> So you don't log in, configure the network, log out, have Mom log in, and Mom can't get to the internet, cause her NM isn't configured the same as yours (or at all)
[09:27] <Keybuk> infinity: agree
[09:27] <fabbione> StevenK: ok, do you want to take care of it or should i?
[09:27] <fabbione> StevenK: both ways i don't mind.. it's a small package
[09:28] <StevenK> fabbione: I was just going to copy the debian dir from 5-0ubuntu2 and update the changelog.
[09:28] <fabbione> ok whatever :)
[09:39] <StevenK> fabbione: Does this mean you'll upload it for me? :-)
[09:41] <fabbione> StevenK: whatever.. 
[09:41] <Kamion> Keybuk: at least the installer no longer breaks messily due to no network plugging on reboot ...
[09:42] <Kamion> so I can stop going "argh" about that one
[09:42] <Keybuk> Kamion: morning!
[09:42] <Keybuk> happy new year
[09:42] <Keybuk> good crimbo, etc.
[09:42] <Keybuk> ?
[09:43] <Keybuk> Kamion: in theeeory, we probably want netcfg to write auto lines again, not whatever mess is popular this week; and that'd solve the problem temporarily
[09:43] <StevenK> fabbione: http://wedontsleep.org/~steven/vblade
[09:44] <fabbione> StevenK: ok thanks
[09:44] <fabbione> StevenK: don't you have universe upload privileges?
[09:44] <fabbione> i thought you went trough the process
[09:44] <StevenK> fabbione: Nope.
[09:45] <StevenK> I'm a member, but I haven't attended a TC meeting yet.
[09:45] <Kamion> Keybuk: hmm, just want to make sure that the transition from all the various different things netcfg has ever written out will have a chance of working
[09:45] <Kamion> will introducing another one be a problem? :)
[09:45] <fabbione> StevenK: ah ok
[09:45] <Keybuk> this is why I haven't done it yet;  gonna sort networking out at the sprint when we're all together
[09:46] <infinity> Do you really want to be in hitting distance when this is sorted out?
[09:47] <Keybuk> yes, I'm into that kind of thing
[09:47] <Keybuk> I'll even bring the restraints if you like ;)
[09:47] <infinity> It's more fun to watch you run and trip.
[09:48] <infinity> Restraints are for lazy people.
[09:49] <Keybuk> hmm, "I'm not a bottom, I'm just lazy" ... that's exactly what a friend of mine says
[10:02] <jdub> Keybuk: hosts, motd (which is regenned though)
[10:05] <Mithrandir> Keybuk: are the restraints to prevent you from running away or us killing you completely?
[10:07] <Keybuk> yes :)
[10:13] <fabbione> StevenK: uploaded
[10:13] <fabbione> StevenK: did you also check aoetools?
[10:32] <Keybuk> \o/ only 10,000 e-mails to go
[10:32] <Keybuk> and damn, I _wish_ that number was an exaggeration :'(
[10:36] <maswan> Keybuk: been on vacation?
[10:37] <pitti> Keybuk: a tiny modification to your .procmailrc before holidays would have cured that :)
[10:39] <Keybuk> maswan: xmas, new year, etc.
[10:39] <Keybuk> pitti: heh, I don't have such a thing ... and people get upset if you /dev/null their mail
[10:39] <pitti> Keybuk: what do you use instead?
[10:40] <pitti> Keybuk: I know, procmail is kind of old skool, but I was grown up with it...
[10:40] <Keybuk> subscribe to mailing lists as different addresses
[10:40] <maswan> Keybuk: just skip on those pesky mailing lists? :)
[10:40] <Keybuk> and then just have that address delivery straight into the right folder
[10:40] <pitti> hmm
[10:41] <Keybuk> ie. scott-canonical-ubuntu-devel@netsplit.com -> ./Maildir/.Lists.canonical.ubuntu.devel/
[10:41] <pitti> Keybuk: I have such fancy things like 'only show me new unassigned bugs', that's certainly not possible just with addresses
[10:41] <pitti> or filtering out MS office attachments
[10:42] <Keybuk> right, I have a Python filter program to do that kind of thing in the mail delivery rules
[10:52] <sivang> pitti: care to discuss those procmail improvements over a cup of tea? I think I need some of this cure :-)
[10:53] <pitti> sivang: feel free to spy on /home/martin/.procmailrc on my server ;)
[10:54] <sivang> pitti: will do , thanks
[10:58] <Treenaks> pitti: Is mozilla-firefox-locale-nl-nl known broken?
[10:58] <Treenaks> pitti: or should I file a bug?
[10:58] <pitti> Treenaks: all m-locale packages seem to be ATM
[10:59] <Treenaks> pitti: ok, so it's known :)
[10:59] <pitti> Treenaks: yes, there are a lot of bugs already
[10:59] <Treenaks> (also, firefox itself seems to be at 1.5rc3 while 1.5-final has been out for ages, but that might be on purpose)
[11:01] <jsgotangco> heya sabdfl happy new year
[11:01] <sabdfl> happy 2006 to you too!
[11:01] <pitti> Hi sabdfl, how are you? happy new year
[11:01] <sabdfl> managed to survive the slopes
[11:01] <dholbach> happy new year, sabdfl :)
[11:01] <sabdfl> 57 inches of snow in 7 days, it was a playground of note
[11:01] <pitti> sabdfl: downhill skiing?
[11:01] <sabdfl> snowboarding
[11:01] <sabdfl> how's everyone here?
[11:01] <pitti> ah, cool; I always enjoyed that
[11:01] <jordi> sabdfl: hey
[11:02] <jsgotangco> colorado
[11:02] <jordi> sabdfl: so you're back in one piece: )
[11:02] <sivang> happy new year sabdfl !
[11:02] <sabdfl> jordi: no amputations this season
[11:04] <seb128> Happe new year sabdfl !
[11:04] <ajmitch> hi sabdfl 
[11:04] <sabdfl> hey seb128, ajmitch
[11:05] <jordi> sabdfl: good!
[11:11] <fabbione> hey sabdfl !
[11:13] <mdke> is there a Community Council meet today, or TB, or nothing?
[11:13] <Keybuk> TB will be next week
[11:14] <Keybuk> I updated the agenda not 5 minutes ago
[11:14] <mdke> ah, sorry yeah just saw that
[11:14] <mdke> so CC today then I suppose
[11:18] <doko> seb128: hmm, bad upgrade time? dist-upgrade wants to remove gnome-panel ...
[11:18] <seb128> doko: what else does it want to remove? what arch?
[11:20] <doko> seb128: amd64, others to remove:   base-config evolution evolution-exchange evolution-plugins gnome-applets
[11:20] <doko>   gnome-applets-data gnome-panel libcamel1.2-6
[11:20] <seb128> that's amd64 lagging behind, yeah just wait for a buildd retry
[11:21] <mdke> Kamion, is CC today?
[11:25] <Kamion> mdke: think we're postponing it but I'm honestly too behind to know
[11:26] <Kamion> it's my first day officially back, give me a break :)
[11:26] <jsgotangco> lol
[11:26] <mdke> Kamion, sure thing, didn't mean to bother ya
[11:30] <Kamion> ah, sabdfl says in mail it's on the 10th, and to postpone TB by a week
[11:30] <Kamion> Keybuk: ^--
[11:30] <zyga> monring :--)
[11:30] <doko> infinity: what is the plan with db4.x? stick with 4.3 or go to 4.4?
[11:31] <zyga> Kamion: should I ping you about the three missing members later?
[11:31] <Kamion> zyga: no
[11:31] <Kamion> leave me alone and I'll do it :)
[11:31] <zyga> great :-)
[11:32] <Keybuk> Kamion: heh, confusing ;)  but fair enough, will update the TB time
[11:32] <Keybuk> done
[11:35] <mdke> Kamion, what time on the 10th? (I'll update the wiki page)
[11:38] <Kamion> mdke: no idea
[11:39] <Kamion> oh, yes I do, 1500 UTC I think, but Seveas has mail about that and has been asked to update the agenda already, so just leave it :)
[11:39] <mdke> alright, thanks
[11:39] <Kamion> zyga: I have no record of you being approved; mako didn't say yes
[11:39] <Kamion> we had two +1s out of a required three
[11:41] <Kamion> sabdfl: ok with zyga for membership? we had an incomplete discussion in http://people.ubuntu.com/~fabbione/irclogs/ubuntu-meeting-2005-12-06.html that had elmo and I both saying yes
[11:41] <sabdfl> +1 on zyga
[11:41] <sabdfl> welcome aboard :-)
[11:41] <Kamion> thanks
[11:41] <jdub> morning sabdfl, good break?
[11:47] <zyga> sabdfl: thank you :-)
[11:48] <sabdfl> jdub: yes thanks. you and the missus get some r&r time in?
[11:48] <ajmitch> zyga: well done :)
[11:50] <zyga> :-)
[11:51] <sivang> welcome aboard zyga 
[12:08] <Seveas> lol, very short ad hoc CC meeting :)
[12:18] <ajmitch> elmo: please sync python-setuptools from sid, dropping changes thanks
[12:24] <mjg59> Keybuk: Around?
[12:34] <jdub> sabdfl: yep, hot and sweaty it was.
[12:54] <infinity> doko: I wouldn't mind a switch to 4.4, but it needs some thinking.  Maybe we should talk about it at the sprint.  That'd an ideal time to discuss mass uploading to switch, if we want to.
[12:57] <ogra_ibook> infinity !
[12:57] <ogra_ibook> happy new year 
[12:57] <infinity> To you too.
[12:57] <Mithrandir> infinity: any chance I could get a new livefs build with user-setup added to the image?
[12:58] <Mithrandir> infinity: or should I just do it myself?
[12:59] <ogra_ibook> infinity, i have a small attack waiting for the initramfs maintainer ... could you have a look at http://bugzilla.ubuntu.com/attachment.cgi?id=5494 if thats ok ? i need it to have images small enough to netboot ppc
[12:59] <ogra_ibook> (for ltsp that is)
[01:01] <infinity> ogra_ibook: Why add the raid and i2o_block modules if it's for booting without block devices?
[01:02] <infinity> Or usb_storage..
[01:02] <ajmitch> elmo: please sync pyxmms-remote from sid, dropping changes.
[01:02] <ogra_ibook> infinity, thats what i wanted to hear :) i wasnt sure if we'd need them for local device support at some point, but udev should care anyway
[01:02] <fabbione> ogra_ibook: that patch is no good.. the problem is not initramfs
[01:02] <fabbione> ogra_ibook: the problems are ppc kernel udebs too big
[01:02] <ogra_ibook> fabbione, i dont need any block device support in initramfs for netbooting
[01:03] <StevenK> fabbione: aoetools 8-0ubuntu1 building.
[01:03] <ogra_ibook> fabbione, its not only the yaboot restriction
[01:03] <fabbione> that's something has been discussed already with Kamion and we need to solve it in a more general way
[01:03] <fabbione> StevenK: cool
[01:03] <ogra_ibook> fabbione, i still need a smaller initramfs for 32MB clients
[01:03] <fabbione> ogra_ibook: meh ok
[01:03] <ogra_ibook> so dropping cblockdev support is the best imho
[01:04] <ogra_ibook> since udev will load whats needed post boot ...
[01:04] <infinity> fabbione: Regardless of the other PPC/yaboot issues, a "noblock" mode for initramfs makes sense to me.
[01:04] <infinity> (as does a "no network" mode, for local-only booting)
[01:05] <ogra_ibook> we could also call it "netboot" 
[01:05] <ogra_ibook> i'm not sure about the "noblock" name
[01:05] <infinity> Meh, it'll get a good name when I integrate it.
[01:05] <fabbione> infinity: i am not 100% sure you can kill everything... 
[01:05] <fabbione> but well
[01:05] <infinity> Don't worry about that.
[01:05] <ogra_ibook> oki :-D
[01:06] <infinity> fabbione: initramfs doesn't do anything that the boot process doesn't duplicate again later.  If all we need to boot from a network is a NIC (which is generally true, give or take), then we can leave the rest for runlevel S.
[01:07] <fabbione> infinity: right
[01:07] <fabbione> make sense
[01:08] <StevenK> fabbione: http://wedontsleep.org/~steven/aoetools
[01:08] <Mithrandir> infinity: also, any thoughts on usplash-in-initramfs?
[01:08] <fabbione> StevenK: i am on it, thanks
[01:10] <fabbione> StevenK: uploading
[01:10] <fabbione> Successfully uploaded packages.
[01:11] <infinity> Mithrandir: Doesn't require much thought to do the dirty hack I originally spec'd... If you have designs for something more elegant, I'm all ears.
[01:11] <Mithrandir> infinity: what's your dirty hack?
[01:13] <infinity> Mithrandir: If you don't already know, you don't want to.
[01:13] <infinity> Mithrandir: Suffice it to say, it'll work fine, I just need to upload it.
[01:14] <infinity> Mithrandir: Perhaps around the same time we have that squashfs date, if I don't get to it tomorrow.
[01:14] <infinity> (Playing catch-up with buildd headaches could monopolize tomorrow)
[01:14] <Mithrandir> infinity: 'k..
[01:14] <Mithrandir> infinity: bah, silly buildds.
[01:14] <Mithrandir> whack them with a hammer. 
[01:15] <maswan> Mithrandir: you mean like FlerpIT.com Security Solutions? http://www.acc.umu.se/images/archive/20010821-FlerpIT.com-Security/
[01:16] <Mithrandir> maswan: yeah, sometihng like that
[01:22] <Kamion> fabbione: the stuff that's been discussed with me is entirely separate
[01:22] <Kamion> fabbione: we've talked about d-i; ogra's talking about the LTSP initramfs, which has nothing to do with udebs. It should never have been in the same bug in the first place
[01:22] <fabbione> Kamion: ok
[01:23] <ogra_ibook> fabbione, thats why i only posted the link to the attachment, not to the bug ;)
[01:40] <zyga> what's with current dapper kernels, why cpu frequency scaling is not active?
[01:40] <Nafallo> zyga: is here
[01:40] <zyga> Nafallo: hmm, doesn't work on k7
[01:40] <Nafallo> I have k8 :-)
[01:44] <mjg59> zyga: Will be fixed in the next release
[01:44] <zyga> mjg59: thank you :)
[01:44] <mjg59> It was disabled in SMP kernels, even if there was only one processor
[01:45] <mjg59> And we only ship SMP kernels now
[01:45] <zyga> oh, so will there be smp kernel with cpu freq scaling or non-smp kernels too?
[01:46] <infinity> Erm, frequency scaling is still working for me.  It just doesn't show up correctly in /proc/cpuinfo.
[01:46] <infinity> The /sys tree still shows the cur_freq going up and down with load.
[01:46] <zyga> infinity: I'll check
[01:47] <infinity> sudo cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq
[01:47] <zyga> infinity: I have no cpufreq in my sys
[01:47] <zyga> so it seems disabled
[01:47] <infinity> Shows at 800MHz right now.  If I pump my load up, it jumps up to 2.0 GHz.  Then scales back down slowly to 800 over time.
[01:48] <Lathiat> zyga: is powernowd running?
[01:48] <zyga> no
[01:49] <infinity> Yeah, it won't be if the /sys tree isn't there (ie: If there's no kernel support)
[01:49] <zyga> it said: /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq: No such file or directory
[01:49] <infinity> So maybe the k7 kernel is wonky, compared to the 686 kernel (which I'm using)
[01:57] <mjg59> infinity: On a K7?
[01:57] <mjg59> The issue is K7 specific
[01:57] <mjg59> All other cpufreq should be working
[02:03] <infinity> mjg59: Ahh, check.  I'm on a Pentium-M.
[02:05] <slomo> elmo: please sync ipod-sharp from debian/experimental, thanks :)
[02:06] <slomo> infinity, lamont: please give-back njb-sharp on ia64/amd64 and remove gnunet from dep-wait... thanks :)
[02:07] <infinity> slomo: There will be a mass give-back in about 12 hours to deal with holiday fallout.  Your still will end up in that batch. :)
[02:10] <slomo> infinity: ok, thanks... btw, did you already read tseng's mail about the mono breakage on ppc? any ideas on this?
[02:17] <jbailey> mjg59: Does cpufreq not go into SMP kernels because it's two hard to keep the two CPUs in sync?
[02:18] <mjg59> jbailey: "too"? :)
[02:18] <lamont> slomo: please fix njb-sharp and re-upload
[02:18] <jdub> haw haw
[02:18] <jbailey> mjg59: Right.  Sucks to be an audio thinker sometimes. =)
[02:18] <lamont> it Build-Depends: cdbs (and probably others), not build-depends-indep.
[02:18] <mjg59> jbailey: Not all SMP architectures handle CPUs with different clock speeds
[02:18] <mjg59> I think K8 and P4 can manage it
[02:18] <jbailey> Ah, cool.
[02:19] <Mithrandir> mjg59: sparc too, but not under linux.
[02:19] <jbailey> I was wondering a bit last week if there was any way of having my boxes consume less power.  Part of why I was playing with suspend on my desktop machines.
[02:21] <slomo> lamont: uh, sorry... uploaded fixed version
[02:21] <lamont> that would be the source of the "mysterious" build failure
[02:22] <Mithrandir> infinity: any reason why you're not on #d-apache?
[02:22] <lamont> depwaits on libmysqlclient15-dev cleared, although I don't see it anywhere in the archive at a quick glance....
[02:22] <lamont> jbailey: smaller pixels. :-)
[02:24] <jbailey> lamont: Well, white on black text *must* consume less power, right?
[02:24] <lamont> red on black would only use 33% as many pixels, I'd think.....
[02:24] <pitti> jbailey: depends on whether an enabled or disabled crystal consumes more, but I think an enabled one uses more
[02:25] <mjg59> Black on white ought to take less power on a TFT
[02:25] <lamont> winona suggests removing files to make the laptop lighter, too. :-)
[02:25] <pitti> jbailey: so, surprisingly, black on white should be more efficient
[02:25] <jbailey> Hah, funny. =)
[02:25] <jdub> mjg59: and better contrast
[02:25] <mjg59> pitti: Any chance you can take a quick look at libpam-foregound for security purposes?
[02:25] <pitti> lamont: right, mysql-dfsg-5.0 is still not synced to Ubuntu (which builds mysqlclient15)
[02:25] <mjg59> jdub: Hush
[02:26] <pitti> mjg59: could you please mail me a link to the diff? I'd like to do it today, but not right now
[02:26] <lamont> pitti: is it going to be soon-ish?  (like before uvf?)
[02:26] <pitti> lamont: I seriously hope so - we actually planned to have 5.0 as default and only version in dapper main
[02:27] <pitti> I just wonder why it doesn't autosync
[02:28] <pitti> lamont: but it is in Debian for months...
[02:28] <slomo> lamont: it depends on libmysqlclient14-dev now until we get 15... the same was already done for a few other packages
[02:28] <ogra_ibook> hi
[02:28] <ogra_ibook> oops, ECHAN
[02:28] <pitti> ogra_ibook: it's not entirely OT to say 'hi' here (at least I hope so :) )
[02:29] <ogra_ibook> heh
[02:29] <ogra_ibook> hi pitti :)
[02:30] <infinity> pitti : I'll be sorting out MySQL 5.0 early next week, I believe.
[02:31] <infinity> Mithrandir: Not on a lot of channels until I get home from the tropics.  Right now, I'm just in the minimum required channels to work.
[02:31] <Mithrandir> infinity: ah, ok, I didn't know you were still on a beach.
[02:41] <seb128> ogra: do you know about http://bugzilla.ubuntu.com/show_bug.cgi?id=21795 ?
[02:42] <seb128> ogra: the guy mailed me, he wants to switch 60 school computer to Ubuntu but is blocked because of that
[02:43] <tepsipakki> hm, /proc/meminfo output is incomplete, which is probably why "top" doesn't show memory usage
[02:43] <ogra_ibook> seb128, we dont support ltsp 4.1 at all
[02:44] <ogra_ibook> seb128, our supported ltsp implementation doesnt use xdmcp ...
[02:44] <seb128> ogra_ibook: could you reply on the bug please?
[02:44] <ogra_ibook> yup, will do
[02:44] <seb128> I don't know much about ltsp
[02:44] <seb128> thanks
[02:45] <ogra_ibook> i'll also reassign it if you dont mind loosing a bug :)
[02:45] <Mithrandir> uhm, how do I rename a page on the wiki=
[02:45] <Mithrandir> s/=/?/
[02:45] <maswan> What's the correct way of pointing out meta-bugs in the wiki? ("Otherwise, if "Rename to" is left blank, the original filename will be used." but it is "save as", not "rename to")
[02:45] <ogra_ibook> Mithrandir, look at the pulldown menu
[02:46] <Mithrandir> ogra_ibook: thx
[02:49] <seb128> ogra_ibook: not at all, thanks :)
[02:49] <tepsipakki> disregard my top-problem, the problem was in my head
[02:51] <maswan> Keybuk: there you go, while I had flight2 installed for a daniels-bug, I gave you a bootchart too
[02:58] <jdub> BenC: hrm, now i remember what i really wanted to ask you about
[02:59] <jdub> BenC: i noticed on my MIL's IBM T42 that the new -10 ABI breezy kernel (security update) was hanging in difficult to accurately reproduce ways, while -9 wasn't.
[03:00] <BenC> jdub: 20771
[03:00] <BenC> jdub: CC yourself and join the testing effort :)
[03:00] <jdub> oh
[03:01] <jdub> weird thing is, pia has the same machine, is running -10, and afaik hasn't had any crashies
[03:01] <BenC> it's a very odd problem...fabbione even went back and recompiled -9.23, and people were still seeing problems with that
[03:01] <BenC> it's like breezy suddenly went broke for no reason, and very randomly
[03:02] <fabbione> i blame binutils
[03:02] <BenC> yeah, I think binutils will take the rap for this one too
[03:02] <jdub> hrm
[03:02] <fabbione> i need to ask elmo to dig the old binary package for me
[03:02] <fabbione> and try to build with that one
[03:02] <jdub> i'll find out if pipka's had problems wiht hers
[03:03] <jdub> then i'll have a machine toplay with locally
[03:03] <Mithrandir> elmo: please sync ttf-unfonts, overriding ubuntu changes is ok.
[03:07] <zul> fabbione: i blame <deity>
[03:08] <fabbione> zul: heheh
[03:08] <zul> hi fabbione btw :)
[03:08] <fabbione> yo
[03:09] <zakame> evening devs :)
[03:10] <rob^^^> howdy
[03:22] <mjg59> pitti: Any idea how to deal with the MMC card problem? pmount (and, presumably, hal) seem to want the removable flag to be set in order to automount stuff, but the kernel people disagree
[03:23] <pitti> mjg59: hm, a device must either be hotpluggable, or removable for that
[03:23] <pitti> but MMC cards being removable sounds right to me? because they actually are
[03:23] <pitti> and the reader itself isn't hotpluggable
[03:23] <mjg59> pitti: The reader is a PCI device
[03:23] <pitti> yes, internal readers are quite common
[03:24] <pitti> but e. g. cdroms are 'removable', so we can pmount them
[03:24] <pitti> why don't the kernel folks treat MMC cards as removable?
[03:24] <elmo> Kamion: ?
[03:24] <Kamion> elmo: hi
[03:25] <mjg59> pitti: "GENHD_FL_REMOVABLE should be unset for removable block devices with permanent media"
[03:25] <mjg59> pitti: In the CD-ROM case, the block device is always there but media appears and goes away
[03:25] <mjg59> pitti: But in the MMC case, the block device only appears when the card is inserted
[03:25] <pitti> hm, I see
[03:25] <pitti> is there anything else in sysfs that tells apart a MMC device from a normal HD?
[03:26] <elmo> Kamion: what's a good dapper CD to use?  I've got a laptop where I need wireless and breezy isn't giving me no love
[03:26] <elmo> ISTR you said latest would be "fun"
[03:27] <Kamion> elmo: fun in a good way
[03:27] <Kamion> yesterday's was fine when I tested it
[03:27] <elmo> ok
[03:27] <Kamion> today's should be OK too, I think
[03:28] <mjg59> pitti: Not that I can see
[03:28] <Kamion> er, apart from being uninstallable
[03:28] <Kamion> yeah, try yesterday's
[03:28] <pitti> mjg59: *scratching head*
[03:28] <mjg59> pitti: This may have to be something that's dealt with upstream
[03:29] <pitti> mjg59: maybe they can add another attribute then; 'hotpluggable' would be nice
[03:29] <pitti> mjg59: right now, pmount defines hotpluggable as 'is attached to an USB or FireWire bus'
[03:29] <mjg59> pitti: Fancy taking this up with the kernel people, then? :)
[03:29] <pitti> yes
[03:30] <mjg59> What's the easiest way to give you the sysfs contents?
[03:31] <elmo> also, anyone know what I have to do to get an i915 to use the external monitor?
[03:31] <mjg59> elmo: Video switch doesn't work?
[03:31] <elmo> mjg59: wassat?
[03:32] <pitti> mjg59: looking in /sys/block/sda should be fine
[03:32] <mjg59> elmo: Is there no hotkey on the keyboard to trigger switching?
[03:32] <mjg59> pitti: Uh, it's not a SCSI device
[03:32] <pitti> yes, /sys/block/hdc/, or whatever
[03:32] <mjg59> pitti: Tarred up?
[03:33] <pitti> mjg59: if there is anything interesting in it, sure :)
[03:34] <mjg59> pitti: Ok, tar works really badly on sysfs
[03:34] <mjg59> pitti: Right, sorted. Shall I just attach this to the bug?
[03:35] <pitti> mjg59: that would be nice; which bug is it in particular/
[03:35] <pitti> ?
[03:35] <mjg59> #21767
[03:35] <elmo> mjg59: hmm, nope, doesn't seem to
[03:35] <mjg59> elmo: Right. Hm.
[03:36] <mjg59> elmo: i810switch might do it
[03:37] <elmo> ah, right, that's the name, I'll try that, thanks
[03:37] <mjg59> elmo: Alternatively, there's a patch to add support to 1855crt at http://sourceforge.net/tracker/?atid=647739&group_id=107484&func=browse
[03:38] <janimo> elmo, any news on the xubuntu-docs still in NEW vs GPL licensing issue?
[03:42] <janimo> do NEW (source) packages in debian require explicit sync requests or they come in automatically
[03:43] <Keybuk> mjg59: you ping'd?
[03:43] <mjg59> Keybuk: Can we chat briefly about NetworkManager at some point?
[03:43] <Keybuk> sure
[03:43] <jsgotangco> yay!
[03:43] <zakame> janimo: hm seems they do come in automatically
[03:43] <janimo> zakame, ok thanks
[03:44] <mjg59> Keybuk: Am I right in thinking that the major sticking point now is what it does on startup?
[03:44] <Keybuk> no
[03:44] <mjg59> Keybuk: Ah. What was the issue, then?
[03:44] <Keybuk> the major sticking point is that it doesn't support or work with one of the most common wifi cards
[03:44] <elmo> mjg59: oh, wow, shiny
[03:44] <mjg59> Keybuk: That being?
[03:44] <Keybuk> Atheros
[03:44] <mjg59> Uhm. It's bloody well meant to.
[03:44] <Keybuk> nope
[03:44] <elmo> mjg59: i810switch "works" in as much as it turns the monitor on, but all that's displayed is an ever-brithening blue blob
[03:45] <Keybuk> long flamewars about it
[03:45] <Keybuk> basically it comes down to an interpretation of the standard
[03:45] <Keybuk> there's a "scan for new networks" ioctl that nm issues every 30s
[03:45] <Keybuk> the ipws take this to just mean "update your table of nearby networks and tell me when done"
[03:45] <Keybuk> the atheros takes it to mean "Bored of this network, deconfigure the card and find me a better one"
[03:46] <mjg59> Keybuk: I'm aware that that /has/ been the case, but I didn't believe it was currently the case
[03:46] <mjg59> Given that Novell are using it, and a lot of Novell people have Atheros hardware...
[03:46] <Keybuk> was the case last time I checked (at UBZ)
[03:47] <Keybuk> there may be a new fixed version, Christian is going to upload that to universe and I'll play with it and see what happens
[03:47] <Keybuk> the symptom is quite obvious, every 30s your network connection drops
[03:47] <Keybuk> and if you (like me at home) have a stronger network that's not yours, your laptop jumps to that one instead
[03:47] <mjg59> Right. Give me a second and I'll test it here.
[03:47] <elmo> janimo: sec
[03:48] <whiprush> I don't see that here. nm and atheros.
[03:48] <Keybuk> run wavemon, every 30s or so the signal will drop out totally
[03:48] <mjg59> elmo: Ok, that sounds like it's powering up the crtc but not changing the display pipes
[03:48] <mjg59> elmo: Try i855crt with that patch?
[03:49] <elmo> mjg59: yeah, already compiling it
[03:56] <Mithrandir> would adding "auto %s\niface %s inet dhcp" for each network interface found be considered rude?  (In the live CD, that is)
[03:56] <mjg59> Mithrandir: Given that it'll block, yes
[03:56] <elmo> mjg59: meh, doesn't work (same problem with rawpipe as with i810switch, and SNAFU screen with a specific res), oh well - thanks anyway
[03:57] <Keybuk> mjg59: not for too much longer
[03:57] <Mithrandir> mjg59: I can blame Scott for that and he'll have an incentive to fix it. :-)
[03:57] <mjg59> elmo: Only other real option is to set up mirroring at all times in xorg.conf
[03:58] <elmo> mjg59: if I do that, does it work without the monitor?
[03:58] <mjg59> elmo: Yes
[03:58] <mjg59> Power consumption will be slightly higher, though
[03:58] <elmo> ah, ok, I guess I can try that
[04:03] <Keybuk> mjg59: just tried nm 0.5.1-0ubuntu6 (current dapper) and it does exactly the same thing
[04:03] <Keybuk> leaps off my network and joins my neighbour's whenever told to scan
[04:05] <mdz> jdub: pong
[04:07] <Keybuk> mjg59: in fact, it's every 13s now
[04:08] <mjg59> Hlaghlkjlkjafhalgh
[04:08] <pitti> mjg59: bless you
[04:12] <lucas> hi
[04:12] <lucas> I've a problem with the ubuntu wiki
[04:12] <lucas> I can login correctly using my LP id
[04:12] <lucas> but when I want to update my preferences on https://wiki.ubuntu.com/UserPreferences
[04:12] <lucas> it says "Passwords don't match"
[04:12] <lucas> can somebody confirm the problem ?
[04:15] <elmo> aww, crap it's a bcm4318
[04:15] <elmo> is the reverse engineered driver usable/useful in dapper yet?
[04:17] <tseng> i think mjg59 got it to associate but not transmit
[04:17] <elmo> what's the fallback, beyond a wireless PCMCIA card?  ndis?
[04:18] <Keybuk> ndis is reasonably well supported at the moment, if you do the voodoo it should work with udev
[04:19] <Keybuk> of course, your machine will crash an hour or so later, but hey, that's ndis for you :p
[04:20] <mjg59> elmo: Latest kernel should just about drive it
[04:20] <mjg59> Though "latest kernel" may be what's in git rather than what's in the archive
[04:21] <mjg59> Keybuk: We might want to look at switching to the madwifi-ng crap
[04:21] <Keybuk> Linus may have been too busy flaming people to put it in the last rc? ;)
[04:21] <Keybuk> mjg59: that allegedly fixes it, because it works the same way the ipws do
[04:22] <mjg59> Keybuk: Uh, our kernel. It's not in 2.6.15 stock.
[04:22] <Keybuk> oh, sorry
[04:22] <Keybuk> wrong git
[04:22] <hno73> lucas: It seems to work if you blank the password
[04:23] <seb128> Kamion: could you promote the new evolution-data-server binaries (soname changes): libcamel1.2-7 libexchange-storage1.2-1? evolution needs them to build
[04:23] <hno73> (though there is clearly a usability issue there)
[04:23] <mjg59> Argh christ the madwifi HAL comes from their Windows drivers
[04:23] <lucas> hno73: oh true, thanks
[04:24] <Keybuk> you didn't know that?
[04:25] <Keybuk> madwifi-ng is not the solution you're looking for, anyway
[04:25] <Keybuk> that's just version 2 of the existing crap
[04:26] <mjg59> Keybuk: I knew it was closed, I didn't know it was Windows derived
[04:26] <Keybuk> ath-driver is the one that "works"
[04:26] <Keybuk> http://www.ath-driver.org/
[04:26] <mjg59> Well, except for not working
[04:26] <mjg59> (so far)
[04:26] <Keybuk> I got it to associate to my network
[04:26] <mjg59> Ooh
[04:26] <Keybuk> not much else, but that's a start ;)
[04:26] <mjg59> Heh
[04:27] <mjg59> I'd be surprised if it's good enough for Dapper
[04:27] <mjg59> Whereas madwifi-ng is realistic
[04:27] <Keybuk> afaik madwifi-ng has exactly the same problem
[04:27] <mjg59> Keybuk: It has support for background scanning
[04:27] <Keybuk> oh, it does?  in the same way as the ipw implements it?
[04:27] <mjg59> No idea
[04:28] <mjg59> http://www.madwifi.org/wiki/ChipsetFeatures/BackgroundScanning
[04:28] <mjg59> "A cool feature whereby the driver scans available channels without interrupting a current AP association.
[04:28] <mjg59> Not at all an issue
[04:28] <Keybuk> right
[04:28] <mjg59> (cough)
[04:29] <Keybuk> wonder if they've bothered to implement Ad-Hoc yet
[04:30] <mjg59> Nngh and it still has its own 80211 stack
[04:31] <Keybuk> yeah, it has a port of the FreeBSD one, iirc
[04:31] <Lathiat> mm, sata_sil24, cooler ntfs write and ivp6 stateful connection tracking in 2.6.15
[04:31] <Keybuk> because what the kernel needs is yet another 802.11 stack
[04:31] <Keybuk> didn't I read somewhere about YET ANOTHER one being written (in addition to the madwifi and ipw ones?)
[04:32] <zyga> hmmm
[04:32] <zyga> why are they all duplicating work
[04:33] <Kamion> seb128: done
[04:34] <Keybuk> zyga: the "standard" 802.11 stack was written by the Intel guys for the IPW cards; and they're very resistant to changing it to accomodate other models afaiui
[04:34] <zyga> Keybuk: eh, the power of egos
[04:34] <jdub> 80211 stacks are the new irc client
[04:35] <mjg59> Keybuk: The ipw one doesn't have terribly complete support for managing softmac type stuff
[04:35] <zyga> Keybuk: are they all GPL'd?
[04:35] <Keybuk> ah, it was the free broadcom one
[04:35] <Keybuk> zyga: they're all GPL-compatible
[04:35] <zyga> at least that's good
[04:35] <zyga> I hope that in 5 years wifi will just work...
[04:35] <mjg59> The broadcom driver has been ported to two stacks (softmac, which is built on top of the ipw one, and dscape, which is complete but not terribly well integrated into the kernel yet)
[04:37] <Keybuk> *sigh*, etc.
[04:38] <seb128> Kamion: thanks
[04:39] <mjg59> Keybuk: The new madwifi stuff claims to support adhoc
[04:39] <elmo> Kamion: hum, I think the network stuff has regressed
[04:40] <elmo> Kamion: I'm trying to install today's i386 install, and I selected 'don't configure network at this time', and now I'm stuckk in a loop at the "http proxy" prompt ("choose a mirror of the Ubuntu archive")
[04:40] <Keybuk> mjg59: the old stuff claims to support it too, and then in very small letters adds "but it doesn't work, so you should just use hostap mode instead"
[04:40] <Keybuk> maybe they've fixed it in -ng, but I didn't see it on the NgFeatures page
[04:40] <Kamion> elmo: yeah, known, I mentioned it in the Flight 2 release notes
[04:41] <mjg59> I don't think that's a complete changelog
[04:41] <elmo> Kamion: ah, ok
[04:41] <elmo> Kamion: sorry
[04:41] <Kamion> no problem, I need to fix it RSN
[04:41] <Keybuk> I think we should all go back to 10base-2
[04:42] <dilinger> * Add a new command "play" to play an audio file on PC.
[04:42] <dilinger> grub2 scares me
[04:43] <jdub> you mean, we can have the ubuntu beat at grub now? elite!
[04:43] <jbailey> dilinger: Someone was joking the other day about porting glibc to the grub2 OS. =)
[04:43] <Mithrandir> jbailey: what about emacs?
[04:43] <dilinger> jbailey: and you haven't gotten on that yet? ;p
[04:44] <Mithrandir> a bootloader just needs a mail reader in the same way a fish needs a bike.
[04:44] <jbailey> dilinger: Grub 2 won't load an elf64 kernel yet.
[04:44] <jsgotangco> good night
[04:44] <jbailey> dilinger: So I can't test it on my main machine. =)
[04:44] <jbailey> Mithrandir: I'm not an emacs lover.  But I wouldn't be surprised if someone hadn't done it already. =)
[04:44] <tenco> hi
[04:45] <dilinger> http://lists.gnu.org/archive/html/grub-devel/2005-11/msg00032.html
[04:45] <tenco> iam currently trying to make breezy boot faster on an old 400mhz desktop
[04:45] <tenco> and i just looked at /etc/modprobe.d
[04:45] <jbailey> dilinger: Oh, that's just the generic speaker.
[04:45] <jbailey> dilinger: I helped Marcus write that code in the Hurd. =)
[04:46] <tenco> IMHO thats pretty evil bruteforce
[04:46] <Keybuk> tenco: hmm?
[04:46] <jdub> dilinger: um. 'mod'. holy crap.
[04:46] <Keybuk> tenco: what about /etc/modprobe.d concerned you?
[04:46] <tenco> especially alsa-base
[04:46] <mjg59> Hmm. Actually, the openhal code looks like it possibly supports a sensible amount of hardware now
[04:46] <dilinger> jbailey: yea, i noticed the comment that it was stolen from hurd
[04:46] <tenco> Keybuk: in alsa-base e.g. are lots of "install <module>" lines
[04:47] <Keybuk> tenco: do you know what the "install" keyword in a modprobe configuration file does?
[04:47] <tenco> Keybuk: so "Calculating module dependencies..." takes a lot of time
[04:47] <Keybuk> *blink*
[04:47] <Keybuk> no, that isn't why
[04:47] <tenco> trying to install that module?
[04:48] <Keybuk> no, that's not what it does
[04:48] <tenco> what else?
[04:48] <Keybuk> I suggest you and the module-init-tools documentation become friends :)
[04:48] <tenco> jup :-)
[04:49] <tenco> anyway, calc the mod deps takes a lot of time
[04:49] <Keybuk> yes, it does
[04:49] <Keybuk> which filesystem are you using?
[04:50] <mjg59> Keybuk: If you've got a chance, can you take a look at http://pdos.csail.mit.edu/~jbicket/openhal/ at some stage?
[04:50] <tenco> ext3
[04:50] <Keybuk> tenco: ok, that's interesting; usually it's the more ricer ones that tend to be slow
[04:50] <Keybuk> basically it's just taking that long to stat a directory tree of 2,000 nodes
[04:51] <Keybuk> (/lib/modules/$ver)
[04:51] <tenco> o_O
[04:51] <Keybuk> this is rather irrelevant though
[04:51] <Keybuk> we don't do that during boot in dapper
[04:52] <tenco> a simple file would be faster to parse...
[04:52] <tenco> text file
[04:52] <Keybuk> uh
[04:52] <Keybuk> dude, seriously, go read documentation
[04:52] <Keybuk> all that step is doing is creating the "simple text file" you want
[04:52] <tenco> every boot?
[04:53] <Keybuk> yes
[04:53] <Keybuk> like I said, we don't do that in dapper anymore
[04:53] <tenco> ok
[04:53] <mdz> Keybuk: what's behind the issue of network interfaces not being brought up at boot?  I saw this in a test install yesterday
[04:53] <Keybuk> mdz: I've not written anything to bring them up at boot yet
[04:53] <Keybuk> mostly
[04:54] <mdz> I see
[04:54] <Keybuk> need to port over the old hotplug net.agent to udev, use the ifrename support built-in to udev, and sort out the mess with n-m too
[04:54] <Keybuk> is my plan for the sprint when everyone's together
[04:54] <mdz> we should do something about it before then, though
[04:54] <Keybuk> yeah, the answer is add an "auto" line to /etc/network/interfaces
[04:55] <Kamion> should the allow-hotplug line stay?
[04:55] <Keybuk> I have no idea what "allow-hotplug" is
[04:55] <Keybuk> that turned up recently
[04:55] <mdz> me either
[04:55] <Keybuk> it seems like aj-crack to me
[04:55] <Kamion> no
[04:55] <Kamion> Thomas Hood suggested it
[04:55] <Kamion> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=340935
[04:56] <Keybuk> if all hardware is brought up by hotplug, then either it or the "auto" line are meaningless
[04:56] <mdz> doko: have you had a chance to look at this patch from Martin?
[04:57] <tenco> Keybuk: do you think its easy to transfer the dapper mechanism to breezy?
[04:57] <Kamion> 15:44:08 [I]  Log message: configuring ssh after installation
[04:57] <Kamion> 15:44:08 [I]  /home/cjwatson/src/ubuntu/espresso/tailor $ svn update --revision 1480 .
[04:57] <Kamion> I'm so glad I've branched from that code and am not obliged to follow it
[04:58] <Kamion> (that change does 'dpkg-reconfigure ssh' after installation. What's the point?)
[04:58] <mdz> isn't ssh just a metapackage now?
[04:58] <Kamion> yes
[04:59] <Kamion> reconfiguring it does nothing
[05:00] <Kamion> oh, great, and they've cloned-and-hacked grub-installer straight in
[05:01] <Keybuk> Kamion: was the log message "merry christmas *hic*" ?
[05:05] <Kamion> Keybuk: I've made that change in netcfg for now, although I'm still not convinced I really understand what's going on
[05:05] <jdub> mdz: from your past experience, how grotesquely conflated are mythtv's backend and frontends? do you think it would be feasible/easyish to develop an alternative frontend without a huge amount of work?
[05:06] <tseng> jdub: they are totally seperate in theory
[05:06] <tseng> they can happily live on seperate systems
[05:06] <Keybuk> Kamion: me neither, I'm not entirely sure where we're going either
[05:06] <jdub> yeah, that's different to conflated codebases though :-)
[05:06] <Keybuk> my new-world-order-tom-tom is stuck in john cleese and making random comments about its mother
[05:06] <jdub> the 'protocol' ends up being the database, which is a saving grace to a certain extent
[05:07] <tseng> thats true, but there is more than just the db
[05:07] <jdub> nfs? :)
[05:07] <tseng> no
[05:07] <jdub> and all the xml bits
[05:07] <mdz> jdub: with or without basing it on the existing code?
[05:07] <Kamion> Keybuk: yes, dear
[05:07] <jdub> mdz: without
[05:07] <tseng> you need to connect to the server to stream the ring buffer frmo live tv capture
[05:07] <jdub> tseng: ahr
[05:08] <tseng> i think it also wants to stream saved captures over the client/server, even though you can do nfs
[05:08] <mdz> jdub: it would be a lot of work to forego the library, but the protocol would support it
[05:08] <tenco> Keybuk: depmod --quick is taking that long?
[05:10] <Kamion> BenC: I've checked in image[macrisc4]  stuff to debian-cd now; it'll be in tomorrow's daily builds
[05:11] <fabbione> hey mdz
[05:12] <Keybuk> tenco: yes
[05:12] <Keybuk> about 12-15s if memory serves
[05:13] <Keybuk> of course, the second time you do it, it takes nanoseconds
[05:13] <mdz> fabbione: good morning
[05:13] <Keybuk> which all the clever kids know is a clue that it's a filesystem issue
[05:13] <mdz> Keybuk: if we don't know where we're going with this yet, now is the time to talk it out and decide ;-)
[05:14] <Keybuk> mdz: perhaps
[05:14] <tenco> Keybuk: thanks. turning that off should be save if either dpkg/apt handles an initial depmod on installing a kernel-package or i do that manually myself, i guess
[05:14] <Keybuk> tenco: I think we had to add those
[05:15] <tenco> Keybuk: thanks so far :-)
[05:15] <Keybuk> so there's a bunch of "network-like" things we have to bring up during boot
[05:15] <mdz> how do we get away with not running depmod for l-r-m?
[05:16] <mdz> or is it broken and nobody's noticed yet?
[05:16] <Keybuk> mdz: we depmod when we install the package, and expect the dependencies to not change <g>
[05:16] <mdz> Keybuk: the modules for l-r-m don't exist until boot time
[05:16] <Keybuk> they exist at install time too
[05:16] <mdz> hmmm
[05:17] <Keybuk> the postinst runs lrm-manager, which calls depmod
[05:17] <mdz> I guess it's ok unless someone unmounts the tmpfs and installs a new kernel
[05:17] <Keybuk> why would that matter?
[05:17] <Keybuk> different tmpfs per kernel version
[05:17] <mdz> because that would remove the entries from modules.dep
[05:17] <mdz> and upgrading the kernel would run depmod
[05:18] <Keybuk> and there are a whole gangbang of problems about removing or upgrading your running kernel anyway
[05:18] <BenC> Kamion: sweet, thanks
[05:18] <mdz> Keybuk: so about these network-like things
[05:18] <Keybuk> if someone installs kernel A-1, and the appropriate lrm for it, then unmounts the tmpfs, and then upgrades to A-2 ...
[05:18] <Keybuk> then yes, it wouldn't work
[05:19] <Keybuk> but I think that comes under a severe case of "don't put your gun at your feet"
[05:19] <Keybuk> right, networking
[05:19] <Keybuk> so we have to bring up:
[05:19] <Keybuk> 1) lo
[05:19] <Keybuk> 2) devices for which we find the hardware, load modules, etc.
[05:20] <Keybuk> 3) static magic devices like ppp, tunnels, etc.
[05:20] <mdz> for 2), do we need to make any distinction between whether the user is using a modular kernel or an elmo kernel?
[05:20] <Keybuk> mdz: no, fortunately we don't; they look the same to us
[05:20] <mdz> cool
[05:21] <Keybuk> 2) can be summed up as "everything the kernel has hardware for"
[05:21] <Keybuk> if the kernel doesn't know about it, you clearly have a Square One issue
[05:21] <Keybuk> except for the stuff in 
[05:21] <Keybuk> 3) which is the stuff there's no hardware for
[05:21] <mdz> for 1), we currently have an init script which configures it based on /etc/network/interfaces.  any reason to change that?
[05:22] <Kamion> you sure about listing ppp under 3)?
[05:22] <Keybuk> I'd rather like to fix the "eth0:2 isn't brought up when eth0 is plugged" bugs too
[05:22] <Keybuk> mdz: we do?  when did that show up?
[05:22] <Kamion> I suppose it's there sometimes, e.g. for serial modems
[05:22] <mdz> Keybuk: we do 'auto lo' and 'ifup -a' still, no?
[05:22] <Keybuk> Kamion: dunno, I've seen people with ppp listed in eni
[05:22] <Keybuk> mdz: currently, yes ... but that's far too late
[05:23] <Keybuk> we need to "ifup lo" before we do just about everything, including starting uedv
[05:23] <Keybuk> which means the "auto lo" won't mean a thing, because there's (currently) no "ifup --auto-only lo" type syntax
[05:24] <mdz> Keybuk: so how about we configure lo sanely in early userspace, and if the user does something custom via /etc/network/interfaces, that takes effect later in the boot/
[05:24] <Keybuk> mdz: ifupdown doesn't like reconfiguring interfaces :-/
[05:25] <hyperactivecrond> what does dapper use for the menu on boot from the install cd?
[05:25] <Keybuk> we'd have to ifdown it
[05:25] <Keybuk> my idea was to do something like:
[05:25] <Keybuk> - add the --auto-only flag to ifup
[05:25] <tseng> hyperactivecrond: gfxboot.
[05:25] <mdz> Keybuk: s/configure lo sanely/& without ifupdown/
[05:25] <Keybuk> - "ifup --auto-only lo" early in userspace
[05:25] <hyperactivecrond> thx tseng
[05:25] <mdz> ifconfig lo inet 127.0.0.1 # TYVM HAND
[05:25] <Keybuk> - call "ifup --auto-only $DEVNAME" in udev when network hardware is detected
[05:25] <Keybuk> and then deal with 3) somehow
[05:26] <mdz> if the user wants to add additional IP addresses to lo or something, it's fine for that to take effect later on
[05:26] <Keybuk> probably just with "ifup -a"
[05:26] <Keybuk> as that takes into account anything already up
[05:26] <Keybuk> afaict
[05:26] <doko> mdz: test build done, uploaded, waiting for the buildd (approx. 20h)
[05:27] <fabbione> doko: does the new OOo2 fix ppc build?
[05:28] <mdz> Keybuk: why bother with ifup in early userspace just for configuring loopback?
[05:28] <Mithrandir> mdz: would we want an integrity check of the live cd as well as of the install cd?
[05:28] <mdz> Mithrandir: yes; it'd be implemented identically
[05:28] <doko> fabbione: no, not yet. it looks like a gij/gcj issue, which does not occurr in breezy. I'll come to it again after a gcj-4.0 update
[05:28] <Keybuk> mdz: tbh, I never understood why lo is even in /etc/network/interfaces
[05:28] <fabbione> doko: ok. do you want me to build on sparc or should i wait?
[05:29] <Mithrandir> mdz: no, it wouldn't, unless you want a full d-i on the live cd as well?
[05:29] <Keybuk> we could always drop it from there, don't write it in netcfg, and just bring it up as you suggest
[05:29] <Keybuk> no sane person changes its details, after all
[05:30] <mdz> Mithrandir: the udeb boils down to md5sum -c
[05:30] <fabbione> Keybuk: remember that lo can be something different than 127.0.0.1/8 and that the network is routable...
[05:30] <mdz> I guess we'd need a separate progress bar
[05:31] <Kamion> we don't have any of the d-i infrastructure on the live CD any more
[05:31] <doko> fabbione: I think it can wait
[05:31] <fabbione> doko: ok
[05:31] <Kamion> I suppose I can resurrect it if I have to
[05:31] <mdz> Mithrandir: speaking of which, make sure re-adding usplash support is on your list for casper
[05:32] <Kamion> but it was a significant space saving
[05:32] <mdz> Keybuk: it seems reasonable to have it there so that users can add custom configuration to it
[05:32] <mdz> Kamion: I think it's better to leave it out
[05:32] <Keybuk> mdz: but then that wouldn't work
[05:32] <Keybuk> cf.
[05:32] <Keybuk> # ifdown lo
[05:32] <Keybuk> # ifconfig lo down
[05:33] <Keybuk> # ifconfig lo up 10.0.0.1
[05:33] <Keybuk> # ifup lo
[05:33] <Keybuk> SIOCSIFADDR: File exists
[05:33] <Keybuk> Failed to bring up lo.
[05:33] <mdz> ugh
[05:33] <Keybuk> THE SKY IS FALLING.
[05:33] <Keybuk> type thing
[05:33] <mdz> that's surely a bug
[05:33] <Kamion> mdz: I agree - but that does mean we can't reuse the integrity check
[05:33] <Kamion> perhaps we can have a simple switch in casper that does md5sum -c /cdrom/md5sum.txt or whatever it is
[05:33] <Mithrandir> mdz: it's already there.  Or should be, at least.
[05:34] <Keybuk> mdz: perhaps, but I get scared whenever I look at the ifupdown source code
[05:34] <Keybuk> it makes me want to cry
[05:34] <mdz> Kamion: yes, I meant more that the guts would be the same
[05:34] <Kamion> ah, ok
[05:34] <mdz> Keybuk: from what I see it should just be doing "ifconfig <iface> 127.0.0.1 up"
[05:34] <mdz> mizar:[/tmp/ifupdown-0.6.7ubuntu2]  sudo ifdown lo
[05:35] <mdz> mizar:[/tmp/ifupdown-0.6.7ubuntu2]  sudo ifconfig lo down
[05:35] <mdz> mizar:[/tmp/ifupdown-0.6.7ubuntu2]  sudo ifconfig lo 127.0.0.2 up
[05:35] <mdz> mizar:[/tmp/ifupdown-0.6.7ubuntu2]  sudo ifup lo
[05:35] <mdz> wfm
[05:35] <Keybuk> mdz: ifconfig it to something outside of the range ifup wants for it
[05:35] <Mithrandir> Kamion: ENOMD5SUM.TXT, but yeah, we would.  I'm pondering what we should do about the huge filesystem.cloop file.. I might end up writing something which reads chunks of it and talks to usplash as a progress bar-style thing.
[05:36] <mdz> Kamion: did you already add the integrity check option to the isolinux menu?  I didn't look
[05:36] <Keybuk> mdz: ie. not 127/7
[05:36] <mdz> Mithrandir: eh?  we've had md5sum.txt on the live CD for ages
[05:36] <Keybuk> uh /8
[05:36] <Kamion> mdz: no, was waiting for instructions from Mithrandir on that
[05:36] <sabdfl> Kamion: does ubuntu-ppc support resizing existing partitions? i have a new apple desktop...
[05:36] <mdz> sabdfl: welcome back
[05:36] <hyperactivecrond> for gfxboot for dapper... does one have to get a patched grub?
[05:37] <hyperactivecrond> never mind.
[05:37] <Mithrandir> Kamion: I was planning on giving you those tomorrow.
[05:37] <Keybuk> sabdfl: Apple stores are damned hard to walk past, aren't they; I swear they use extra-special glass that makes things look shinier
[05:37] <fabbione> sabdfl: depends from the partition
[05:37] <Kamion> sabdfl: there's nothing special about powerpc in that regard, although if it has new Mac OS X on it you'll need to use dapper in order for parted to understand the new partition types
[05:37] <Kamion> hyperactivecrond: we won't be doing graphical grub in dapper
[05:37] <sabdfl> Kamion: ok, cool, dapper it is then :-)
[05:37] <Kamion> syslinux only
[05:37] <mdz> Keybuk: it fails, but the interface ends up configured correctly
[05:38] <Keybuk> mdz: yeah, I'm not really sure what that's all about
[05:38] <Kamion> Mithrandir: cool, thanks
[05:38] <sabdfl> will try a dapper live-cd to see if it will even boot on a dual dual-core
[05:38] <mdz> Keybuk: maybe it's the ipv6 bit, which uses add/del
[05:38] <Mithrandir> mdz: we'd probably want to use something which did have a more useful progress bar than what you'd get with the naive approach when you have a single file which is 5/7 of all you're checking.
[05:38] <Keybuk> SIOCSIFADDR is "set interface address"
[05:38] <Kamion> sabdfl: you'll have to use 'live-powerpc64', unless you have a time machine to get tomorrow's live CD which'll make that unnecessary. :)
[05:38] <mdz> Mithrandir: usplash wolud be reasonable
[05:38] <pitti> BenC: the eject issue should be fixed in -10, right?
[05:38] <fabbione> sabdfl: if it doesn't work. please let us know. I know benh is working on his SMP dual core to fix some stuff and not all of it is in the kenrel yet
[05:39] <sabdfl> Kamion: i have patience, no need for time machine :-)
[05:39] <Keybuk> pitti: fixed which way?
[05:39] <BenC> pitti: should be
[05:39] <pitti> Keybuk: properly :)
[05:39] <Keybuk> devices still get ejected properly?
[05:39] <pitti> Keybuk: the kernel should now be able to eject all usb devices
[05:39] <Keybuk> and vanish off the system?
[05:39] <pitti> Keybuk: 'still'?
[05:39] <pitti> Keybuk: they didn't for many devices, see #5049
[05:39] <Keybuk> well, my laptop when I unmount a usb device ejects it properly
[05:39] <Keybuk> ah, as long as you didn't turn off ejecting :p
[05:39] <BenC> it's very random
[05:40] <Keybuk> some people have been moaning about the fact it ejects at all, and doesn't just unmount
[05:40] <BenC> Keybuk: oh, I found out something: eject -t will get a USB device back in most cases, after an eject
[05:41] <Keybuk> that's kinda cute
[05:41] <mdz> Keybuk: anyway, I don't think we need to try to support ridiculous configurations like that; the defaults will work and users will be able to add additional addresses to loopback, which is about the only reasonable thing to do to it
[05:41] <Keybuk> in a totally sick, "get your own ioctls and stop abusing them" kind of way
[05:41] <mdz> Keybuk: so for 1), "add an ifconfig/ip call to initramfs" seems like the right way to go
[05:41] <BenC> keybuk: fact is that some USB devices do the right thing, and "eject" the memory card (or whatever) virtually
[05:42] <Keybuk> mdz: right; we still need a way of bringing up interfaces for hardware though
[05:42] <BenC> it's a hardware thing more than a driver issue
[05:42] <mdz> Keybuk: right, on to 2)
[05:42] <Keybuk> mdz: allow-hotplug seems ... wrong; it brings us back yet again to having two options which mean the same thing in different ways
[05:42] <Keybuk> mdz: just having "auto eth0" seems best, that way if udev doesn't bring it up, S40networking will
[05:43] <Keybuk> BenC: my video camera switches back to ordinary camera mode, instead of dumb USB mode; I'm reasonably glad it doesn't eject its own hard disk <g>
[05:43] <mdz> lemme read the bug about allow-hotplug
[05:44] <mdz> hmm
[05:44] <mdz> so it sounds like allow-hotplug is meant to mean "run ifup when the hardware is detected"
[05:45] <mdz> while "auto" continues to mean "run ifup from S40networking"
[05:45] <Keybuk> right
[05:45] <Keybuk> which is crackful
[05:45] <Keybuk> imo
[05:45] <mdz> I agree that the distinction doesn't seem particularly useful
[05:45] <mdz> either it should be brought up at boot, or not
[05:45] <Keybuk> if the hardware hasn't be detected, trying to bring it up in S40networking isn't going to work
[05:46] <mdz> your --auto-only approach sounds sane
[05:46] <mdz> except for the fact that it involves touching ifupdown code, but c'est la vie
[05:46] <Nafallo> Keybuk: isn't that network-manager's job to cover 2)? :-)
[05:47] <mdz> the magic 8-ball says network-manager's future is unclear
[05:47] <mdz> in the dapper timeframe
[05:47] <Keybuk> network-manager would be great, if it did what it says on its tin
[05:47] <Nafallo> well, wfm ;-)
[05:47] <jsgotangco> :(
[05:48] <mdz> whether or not we're able to integrate it for desktop purposes, I think we need a more trustworthy fallback for server configurations anyway
[05:48] <Keybuk> mdz: do I have to resist the urge to write incredibly sarcastic documentation alongside the change? :p
[05:48] <mdz> Keybuk: I think the literate programming syntax requires that you do so
[05:49] <Kamion> this is starting to remind me of Intercal's politeness rules
[05:51] <mjg59> Keybuk: What are the dates for the distro sprint? It's the week after LCA?
[05:51] <Keybuk> mjg59: yes
[05:51] <mjg59> Keybuk: Ok. I may be able to be there for some of it, if you're planning on covering network stuff
[05:52] <elmo> whine, X segfaults on this stupid laptop
[05:52] <Keybuk> cool; I'm not doing LCA this year after all, I just can't spare the time homewise, let alone anything else
[05:52] <mjg59> elmo: Nngh. You're sure it's Intel?
[05:53] <mdz> mjg59: it seems pretty likely that we'll have things to talk about in that area
[05:53] <fabbione> elmo: dapper/ppc ?
[05:53] <elmo> mjg59: oh, sorry, I'm confusing you, I'm working on two different laptops here
[05:53] <elmo> mjg59: the i915 thing was on a sony vaio, X is fine there
[05:53] <mdz> mjg59: the basic agenda is going to be outstanding dapper issues
[05:53] <Keybuk> the depressing thing about literate programming is that the only way I've ever been able to understand the code is to generate it, read the resulting code, change it, then use grep to work out how to change the literate version
[05:53] <elmo> mjg59: X segfaulting, and the BCM4318 is on a random amd64 turion laptop I'm trying to setup for someone else
[05:53] <elmo> fabbione: breezy/amd64
[05:53] <fabbione> elmo: ok.
[05:54] <mjg59> elmo: Oh. ATI chipset?
[05:54] <Keybuk> of course, I COULD WRITE THE UBUNTU CODE IN A DIFFERENT COLOUR! :D
[05:54] <elmo> mjg59: yeah
[05:54] <mjg59> elmo: Option "NoAccel" "tru"
[05:54] <elmo> oh, right, that molarky
[05:54] <elmo> will it work in x86 mode?
[05:54] <mjg59> Except "true", not "tru"
[05:54] <mjg59> No
[05:54] <elmo> okay, so I'll not needlessly reinstall then
[05:54] <mjg59> Our X is broken on a small number of ATIs
[05:54] <elmo> mjg59: thanks :)
[05:55] <janimo> mjg59, there's a patch in bugzilla for that brokage though
[06:03] <mdz> Keybuk: 3) seems a bit tricky
[06:04] <mdz> Keybuk: since those abstract interfaces sometimes  have implicit dependencies on hardware-oriented interfaces
[06:04] <OpsVentus> Good day, I have started to develop a GUI for wireless networking for Ubuntu because I felt this was missing from Ubuntu, an first glimp of what it should look like and what it can do can be found at http://www.opsventus.com/wifigui
[06:04] <Robot101> Keybuk: I have weirdness with udev 077-0ubuntu5 and dvb devices, it successfully made the /dev/dvb/adapterX/fooY stuff once, but hasn't since
[06:05] <Robot101> Keybuk: how do I debug this?
[06:05] <OpsVentus> can some people please give me an opinion on whether I should invest more time to make it work properly or just give up?
[06:05] <mdz> OpsVentus: what is it meant to do specifically?
[06:05] <OpsVentus> give an easy graphic solution to connect to wireless networks
[06:06] <Keybuk> Robot101: does it make the wrong names for them now?
[06:06] <HiddenWolf> OpsVentus, networkmanager seems to do what you want... 
[06:06] <Treenaks> network-manager
[06:06] <OpsVentus> not for wireless networks
[06:07] <Keybuk> mdz: true, but if we use ifup + auto for 2), and keep the S40networking script, then the ifup -a call there will "sort out" all of those interfaces
[06:07] <Robot101> Keybuk: it makes /dev/dvb0.dvr0 etc
[06:07] <Keybuk> and it'll work the same as it does in breezy
[06:07] <OpsVentus> with my program you can select the wireless network you want
[06:07] <Keybuk> Robot101: can you run "udevmonitor -e" and plug in the device?
[06:07] <seb128> is /dev/fd0 supposed to be 660 or 640?
[06:07] <Keybuk> seb128: yes :)
[06:07] <mdz> Keybuk: assuming all the important bits have finished executing by the time S40networking runs, no?
[06:08] <Keybuk> seb128: theoretically 660
[06:08] <seb128> Keybuk: it was a "or" but I will take that as a "660" :)
[06:08] <mdz> OpsVentus: network-manager does in fact do that for wireless networks
[06:08] <seb128> Keybuk: udev bog? Should I bugzilla it?
[06:08] <Mithrandir> seb128: 640 | 660 is 660, so that sounds sane. :-)
[06:08] <Keybuk> seb128: not that I'm aware of, why, what do you think the permissions should be; and why?
[06:09] <seb128> I think it should be 660 because with the current 640 gfloppy complains that you don't have any right on the device when you are a member of the floppy group
[06:09] <Keybuk> seb128: if it's 640, it means that the kernel is now advertising floppy block devices as "removable"
[06:09] <Keybuk> and that sounds like a bug in gfloppy, no?
[06:09] <Keybuk> what does write permission on the device give you?
[06:09] <seb128> do you need to write to format it?
[06:10] <Keybuk> we need pitti for this, he explicitly asked for removable devices to NOT have write permissions
[06:10] <Robot101> Keybuk: http://rafb.net/paste/results/TnX4TL77.html
[06:10] <Keybuk> the only exception is CD drives, which need write for eject
[06:10] <elmo> somone hit me with build-depends heavy packages I can prime porter chroots with, things like evolution, firefox and openoffice
[06:10] <Robot101> elmo: gnucash :)
[06:10] <Kamion> elmo: debian-installer
[06:10] <mdz> Keybuk: weird, I wonder why he would want that
[06:11] <mdz> writing directly to a floppy or USB key is a reasonable use case
[06:11] <jordi> elmo: nautilus
[06:11] <Keybuk> we'll have to ask him when he comes back
[06:12] <seb128> k
[06:12] <jordi> elmo: some kde package... kdemultimedia I guess
[06:12] <Mithrandir> elmo: mplayer
[06:12] <mdz> Diziet: around?
[06:13] <Nafallo> gaah. I hilight network-manager :-P
[06:13] <Nafallo> anyway, see ya :-)
[06:15] <elmo> Robot101, Kamion, jordi, Mithrandir: thanks
[06:15] <Mithrandir> elmo: ant, possibly, to get some java stuff in
[06:15] <mdz> seb128: is there any new information about the disappearing applications menu?  is it fixed in newer kernels?
[06:15] <mdz> the latest kernels hang during boot on my machine, so I can't quite test
[06:16] <mjg59> mdz: Hm. Fun.
[06:16] <seb128> mdz: it's a gam_server bug, and it triggers by /var/lib/menu-xdg/menus/debian-menu.menu beeing missing
[06:16] <Keybuk> Robot101: I guess somebody in kernelland changed things, or I mucked up the device name script
[06:16] <seb128> s/it triggers/it's triggered/
[06:16] <fabbione> mdz: does that machine have a PCI I/O controller?
[06:16] <mdz> seb128: ah, above the kernel then
[06:16] <Robot101> Keybuk: I tested the device name script and it seemed to dtrt, which is why I'm so confused
[06:16] <seb128> mdz: /etc/xdg/menus/debian-menu.menu points to /var/lib/menu-xdg/menus/debian-menu.menu, when the second is not here gam_server has some issue
[06:16] <mdz> fabbione: I'm not sure what that is.  do SCSI and IDE controllers count?
[06:16] <seb128> mdz: will be fixed soon, there is some patching pending review upstream
[06:17] <elmo> Mithrandir: good plan
[06:17] <fabbione> mdz: yes they do count
[06:17] <Keybuk> Robot101: you have nothing there that would cause it to be run
[06:17] <mdz> fabbione: then yes
[06:17] <Robot101> $ ./dvb_device_name -e dvb0.dvr0
[06:17] <Robot101> 0.0
[06:17] <fabbione> mdz: ok
[06:17] <seb128> mdz: workaround: run update-menus
[06:17] <Keybuk> you just have SUBSYSTEM==dvb things, not SUBSYSTEM==dvb_device
[06:17] <Robot101> Keybuk: aha
[06:18] <mdz> seb128: s/run/install menu and &/ ;-)
[06:18] <mdz> hmm, update-menus fails here with no output
[06:18] <Keybuk> that's wrong anyway
[06:18] <Keybuk> it should output 0 net0 or something
[06:18] <dholbach> menu has update-menus in its postinst
[06:18] <Robot101> Keybuk: ah right
[06:18] <dholbach> installing it should suffice
[06:19] <Keybuk> cool, thanks for that; I can fix that now
[06:19] <Robot101> Keybuk: np
[06:19] <seb128> dholbach: seems that it's bugged, some people get no /var/lib/menu-xdg/menus/debian-menu.menu after running it
[06:19] <elmo> doko: grr, why doesn't apt-get build-dep  work for gcc-4.0 ?
[06:19] <mdz> that does seem to fix it, thanks
[06:19] <dholbach> seb128: strange :(
[06:19] <Robot101> Keybuk: can I get a pointer to how I might make some persistent names for usb serial devices, regardless of what order they probe in?
[06:19] <seb128> mdz: np
[06:20] <Keybuk> yes, easy
[06:20] <Keybuk> use udevmonitor -e, and plug the device in
[06:20] <Keybuk> the [UEVENT]  will tell you the useful information for the device
[06:20] <OpsVentus> mdz: I have to take a deeper look into network-manager then, I'm still on Warty, maybe in Breeze this has been updated
[06:21] <Keybuk> do udevinfo -a -p $DEVPATH  (from the [UEVENT]  bit) to see everything in sysfs you can use
[06:21] <Keybuk> and build up a description of it, probably something like:
[06:22] <Keybuk> SUBSYSTEM=="usb", SYSFS{product}=="iAUDIO M3 Digital Audio Player", ...
[06:22] <Keybuk> and then add a persistent name using SYMLINK+="iaudio"
[06:24] <doko> elmo: which architecture?
[06:24] <elmo> doko: amd64
[06:24] <elmo> something about libc6-686-dev not available
[06:25] <doko> yes, but the alternative is available (ia32-libs-dev)
[06:25] <doko> anyway, jbailey's next upload should fix that and introduce this package
[06:25] <elmo> yeah, it kind of sucks that apt-get build-dep bombs tho
[06:25] <elmo> ah, ok, cool
[06:29] <Keybuk> Robot101: try to void BUS== or anything referring to physical devices, et. al. otherwise your head will explode
[06:29] <Robot101> Keybuk: mm, ok, cool. thanks.
[06:31] <Keybuk> SUBSYSTEM guarantees SYSFS attributes, so use that for custom matches
[06:35] <poningru> what kernel version is planned for dapper?
[06:35] <Keybuk> 2.6.15
[06:36] <poningru> oh
[06:36] <poningru> wait isnt that a in development version?
[06:36] <Keybuk> so is dapper
[06:36] <HiddenWolf> poningru, it's final since yesterday
[06:37] <poningru> ok nm thought that wasnt going to be ready till next year
[06:37] <poningru> err this year
[06:37] <Burgwork> poningru, got released today
[06:37] <poningru> later this year*
[06:37] <poningru> thanks guys
[06:37] <Keybuk> kernels are on roughly a 2-3 month release cycle, I believe
[06:37] <Keybuk> so 2.6.16 will be out before dapper, but too close to dapper's release for us to be confident in it
[06:38] <poningru> hmm ic
[06:38] <Keybuk> we've been packaging 2.6.15 since the first rc
[06:38] <Keybuk> thus already have a lot of testing on it
[06:39] <poningru> cool
[06:40] <poningru> geez the changelog is down it seems
[06:40] <jordi> Keybuk: is the ubuntu kernel team working with Debian's?
[06:40] <poningru> nm
[06:40] <jordi> (I ask because debian has also been packaging the rcs)
[06:41] <Keybuk> jordi: I suspect the Debian kernel team take from the Ubuntu kernel team's git repository as they see fit; it's published on kernel.org
[06:42] <jordi> nod
[06:43] <jordi> Keybuk: ie, no real cooperation
[06:43] <Keybuk> would you really want to go near the Debian kernel team? :p
[06:43] <jordi> well. Not ALL of them no :)
[06:43] <jordi> I know where you're going :p
[06:43] <Keybuk> where am I going? O:-)
[06:44] <jordi> let's be friendly and say no more :)
[06:48] <poningru> how come the debian kernel team doesnt real
[06:48] <poningru> ly like ubuntu?
[06:49] <Robot101> spoons vs forks...? :)
[06:51] <HiddenWolf> poningru, part pride, part real issues, part imagined issues
[06:51] <poningru> ic
[06:51] <Keybuk> various bits of Debian still don't like Ubuntu for particular reasons
[06:52] <HiddenWolf> poningru, both debian and ubuntu have trolls in their ranks, and those like bashing eachother.
[06:52] <jordi> I haven't said the Debian kernel team doesn't like Ubuntu's
[06:52] <Keybuk> it's rather deeper than that
[06:52] <jordi> I just owndered what the situation is
[06:53] <Keybuk> a few Debian maintainers have fairly large problems with, for a lack of a better description, "being patched"
[06:53] <Keybuk> there's not much inter-developer trolling
[06:53] <Keybuk> that kind of thing (as with KDE/GNOME) tends to just take place in the user ranks
[06:53] <poningru> I think its due to the coc
[06:54] <ogra> thats what i meant
[06:54] <HiddenWolf> Keybuk, and having their system of work disrupted by ubuntu developers, by not having a designated maintainer to talk to in ubuntu, and loads of other reasons. :)
[06:54] <poningru> well even within dev for gnome and kde they have other subject trolls
[06:54] <poningru> like the recent israel bout
[06:54] <Keybuk> jordi: I suspect the situation is really just that we don't have a kernel team, we have a BenC; who sends patches directly upstream
[06:54] <jordi> Keybuk: ah I see
[06:55] <Riddell> poningru: where was that?
[06:55] <Keybuk> HiddenWolf: curiously, I've noticed that the Debian developers who complain the most about Ubuntu tend to have the worst relationships with their upstream too; it's not a hard/fast rule, but it's a pattern
[06:55] <poningru> gnome irc dev channel
[06:55] <Keybuk> make of that what you would
[06:55] <poningru> channels*
[06:55] <Riddell> poningru: comedy, I was blogging only yesterday about the exact same thing happening in #kde-devel
[06:56] <Keybuk> jordi: and it's better for us for BenC to do that, and spend his time fixing bugs, etc. than spoon-feeding patches to Debian, or anyone else, for that matter
[06:56] <poningru> rofl
[06:56] <janimo> elmo, xubuntu-docs/NEW/GPL ping :)
[06:56] <HiddenWolf> Keybuk, it takes a certain attitude to work openly. If you like to take pride in and out of _your_ work, the debian system tends to give you more credit and possibly "power" and "protection"  than ubuntu and upstream
[06:56] <Riddell> janimo: did your key get added to the ring?
[06:56] <janimo> elmo, also jani@ubuntu.com GPG key in main
[06:56] <Keybuk> HiddenWolf: you're aware that I'm a Debian developer too, right?
[06:56] <elmo> janimo: right, sorry. you said you were going to contact the {k,}ubuntu doc authors, what happened about that?
[06:57] <janimo> heh, trying too ;)
[06:57] <HiddenWolf> Keybuk, I know that many people here are. :)
[06:57] <HiddenWolf> Keybuk, and that wasn't ment personally, but as a possible explanation
[06:57] <janimo> elmo, I did not say I would contact them, just said that the docs in xubuntu are based on their work
[06:57] <HiddenWolf> Keybuk, s/you/one
[06:57] <janimo> but here's Riddell so we can talk it out no
[06:57] <janimo> now
[06:57] <Keybuk> HiddenWolf: I actually find that hypocritical; because Debian is supposed to be all about open-ness and sharing, not about power or protection
[06:58] <janimo> Riddell, elmo raised the issue that the xubuntu-docs license is GPL
[06:58] <Keybuk> when you get a maintainer complaining that somebody has patched their package, I start to wonder why they so freely patch their upstream without taking the same consideration
[06:58] <janimo> I based it on ubuntu/kubuntu docs so followed their license
[06:58] <HiddenWolf> Keybuk, but the maintainer has the final word about a package. thus the maintainer has "power"
[06:58] <ogra> janimo, that should be cc-sa 2.5 
[06:58] <janimo> ogra, since when?
[06:58] <dilinger> Keybuk: some people just like to complain.  that's what i have a blog for, for example.  ;p
[06:59] <janimo> not two weeks ago for sure
[06:59] <Keybuk> dilinger: ah yes, the "b" in "blog" stands for "bitch" :p
[06:59] <ogra> janimo, see edubuntu-artwork (whih contains the docs in edubuntu dapper)
[06:59] <Keybuk> (old jwz-related joke)
[06:59] <janimo> ogra, I am talking about ubuntu-docs :)
[06:59] <janimo> I know artwork is CC
[07:00] <Keybuk> HiddenWolf: it's not final though, because the package is freely licenced, anyone (Ubuntu, DCC, users, etc.) can change it again
[07:00] <Keybuk> and that seems to upset some people
[07:00] <Keybuk> *shrug*
[07:00] <dholbach> janimo: i test-installed xubuntu on an old 350mhz box - it runs nicely :)
[07:00] <Keybuk> personally I think those people have somewhat misfired priorities
[07:00] <janimo> ogra, which btw if you based on ubuntu-docs you're a GPL violator ;) 
[07:00] <ogra> janimo, i never touched ubuntu-docs, but generally all packages that are originated in ubuntu and can contain artwork or screenshots should be under this license
[07:00] <janimo> dholbach, cool! the one in the corner of your room ?
[07:00] <dholbach> janimo: yeah
[07:01] <HiddenWolf> Keybuk, like I said, it takes a certain attitude to be truly open. :)
[07:01] <janimo> ogra, should vs are not :)
[07:01] <Keybuk> but then I guess I'm politically at the other end of the scale ;)  I don't even maintain my own software, and let other people do it :p
[07:01] <dilinger> jordi: the debian and ubuntu kernel team have slightly different goals, as well
[07:01] <ogra> janimo, i based on ubuntu-artwork ... since my only doc in dapper is the firefox homepage
[07:01] <janimo> ogra, xubuntu-docs only contains the text of the ff start page
[07:01] <dilinger> jordi: which make enough of a difference that sharing the codebase probably won't happen anytime soon
[07:01] <janimo> that is in ubntu and kubuntu-docs under GPL
[07:01] <Kamion> Keybuk: speaking of, bubulle's been trying to hunt you down to do a dpkg release
[07:02] <Keybuk> Kamion: yeah, most people are hunting me down for that :p
[07:02] <ogra> janimo, i can only talk aboutthe rules ;) i didnt break them, ask the ubuntu-doc package maintainer why his is having a sourcecode license :)
[07:02] <Burgwork> janimo, ubuntu and kubuntu docs are GFDL and CC-by-sa
[07:02] <janimo> ogra, yeah I am just doing that :)
[07:02] <janimo> Burgwork, since when? I looked in their changelogs and saw only GPL
[07:03] <Burgwork> janimo, since Dec 2004
[07:03] <Burgwork> janimo, that is a mistake in packaging. can you file a bug on it?
[07:03] <janimo> Burgwork, oh nice.So the debian/copyright file is wrong
[07:04] <janimo> Riddell, ^^ do you take care of kubuntu-docs or should I file a bug on that too?
[07:04] <jbailey> doko: Errm should be libc6-i386-dev, I think, shouldn't it?
[07:05] <Pygi> pitti: ping
[07:05] <janimo> elmo, it seems it's a packaging oversight, do I upload another version to NEW?
[07:06] <janimo> alternately you allow it in and I upload the fix to debian/copyright
[07:06] <elmo> janimo: sure
[07:06] <elmo> no, it's been rejected, a new upload is needed
[07:06] <janimo> ok
[07:06] <janimo> elmo, also the main upload priv if you please ;)
[07:06] <elmo> gstreamer0.8 is scheduled for kitten-pulping, right?
[07:07] <Riddell> janimo: I'll take care of it
[07:07] <sabdfl> seb128: is there any easy way to test the new gstreamer 0.10 capabilities?
[07:07] <sabdfl> mdz: ^^^ re xine vs gstreamer
[07:07] <Riddell> elmo: gstreamer0.8 is still used by kaffeine and amarok
[07:07] <Keybuk> "kitten-pulping" ?!
[07:07] <janimo> Burgwork, so I file a bug on ubuntu-docs and one of the uploaders fixes it then I copy the same license to xu-docs. Sounds ok?
[07:08] <elmo> Riddell: is that going to be fixed before dapper?
[07:08] <elmo> Keybuk: "removal" is such a cliche
[07:08] <Riddell> elmo: the gstreamer thing?  I certainly hope so but can't 100% guarantee
[07:08] <elmo> Riddell: if it's not, I'll set infinity on you
[07:09] <Burgwork> janimo, yes
[07:09] <elmo> Riddell: (seriously, carrying two versions of such a big package would suck, it'd be Really Nice if we could get rid of 0.8)
[07:09] <seb128> sabdfl: what do you mean by "capabilities"?
[07:09] <seb128> elmo: the plan was to get 0
[07:09] <seb128> 0.8 moved to universe if possible
[07:09] <jordi> dilinger: nod
[07:10] <seb128> but seems that KDE guys are slow :p
[07:10] <jordi> seb128: don't expect any other human to be like you mkay
[07:10] <dholbach> seb128: there are quite some other rdepends still, no? :)
[07:11] <seb128> dholbach: from main/GNOME? gnome-media/sound-juicer, we don't really care about the first and the second has a patch upstream I've just not updated yet
[07:12] <Riddell> elmo: totally agree, we don't want to support an obsolete gstreamer for 3 years, I'm doing what I can to make sure we have gstreamer0.10 versions in time
[07:12] <dholbach> seb128: i just had a quick look at the rdepends of libgstreamer0.8-0
[07:13] <seb128> dholbach: that includes universe :)
[07:13] <dholbach> yeah
[07:13] <janimo> dholbach, does the gdm package still patch gdm.conf?is it expected it will used gdm-custom or is that left for admins?
[07:14] <dholbach> janimo: it patches config/gdm.conf.in
[07:14] <janimo> I am thinking about how to integrate xubu gdm settings as non-intrusively as possible
[07:14] <dholbach> janimo: yeah, i installed wdm for the fun of it :)
[07:14] <janimo> and?
[07:15] <janimo> I did too befor ebreezy but it's ugly and needs config love too
[07:15] <janimo> gdm is more suitable
[07:16] <dholbach> yeah, i think so
[07:16] <janimo> btw did you notice faster startup with 2.13.04 by any chance?
[07:16] <janimo> did the degnomification have any visible effect?
[07:17] <dholbach> maybe a bit
[07:18] <dholbach> the bootchart might know :)
[07:25] <mdz> sabdfl: I'd start with playing a video and seeing if the sound is in sync
[07:26] <mdz> sabdfl: totem-gstreamer seems to be using 0.10 now
[07:26] <Burgwork> sabdfl, https://wiki.ubuntu.com/Media/Testing
[07:26] <Keybuk> . o O { the plinky music is synchronised with the video? }
[07:32] <janimo> would it be a bad idea to include the common doc licenses too in /usr/share/common-licenses? 
[07:32] <janimo> pasting many pages in debian/copyright is not much fun :)
[07:32] <Burgwork> janimo, seems sensible
[07:33] <janimo> should I file a bug in ubuntu bugzilla or debian BTS?
[07:35] <Burgwork> janimo, ubuntu
[07:35] <janimo> ok
[07:35] <janimo> much nicer
[07:43] <janimo> pitti, can you estimate when you have time to review xfce related packages for main inclusion? thanks
[07:44] <Diziet> Joy.  `firefox invalid to show in thai language'.  Kudos to the reporter for managing to be vaguely coherent in the report but what a nice one to try to fix.
[07:45] <mdz> Diziet: are you on top of the firefox certificate breakage?
[07:45] <mdz> Diziet: thai language -> sounds like one for upstream
[07:47] <Diziet> mdz: cert breakage> I'm going to fix it tomorrow.  I hope that's soon enough.
[07:48] <Diziet> I knew that it was too good to be true when no-one reported any problems with the nspr rearrangements for a whole day after I uploaded.
[07:49] <jbailey> Diziet: Right, reminds me that I have to file a bug on FF not working without LANG=C for me.  I had totally forgotten about that. =)
[07:49] <Diziet> jbailey: Yes, please, all I need is more bugs :-).
[07:50] <jbailey> Diziet: I take the simpler solution and I generally use epiphany-browser instead. =)
[07:50] <jbailey> Diziet: That way I don't feel compelled.
[07:50] <jbailey> *seb128* gets my bugs instead. ;)
[07:55] <pvanhoof> for when is a new version of firefox expected in dapper? The current version is extremely unstable :-\
[07:56] <Diziet> pvanh: Tomorrow, and there'll probably be another one shortly after that.
[07:56] <Diziet> In what way(s) is it unstable ?
[07:56] <Diziet> (There are some known bugs, but I would like to be sure yours are those ...)
[07:57] <pvanhoof> The downloads don't showup in the download manager
[07:57] <pvanhoof> and it very often crashes
[07:57] <pvanhoof> I've already tried installing a different theme, for the download manager problem. Didn't solve the issue
[07:58] <Diziet> The download manager is a known problem.  I investigated it for a day or so before Christmas but I need to work on it some more.
[07:58] <Diziet> `Very often crashes'.  Hrm.
[07:58] <pvanhoof> Yeah :), "very often"
[07:58] <pvanhoof> Like: most pages
[07:59] <Diziet> After you do anything in particular ?  What architecture ?  Do you have extensions/plugings/etc. ?
[07:59] <pvanhoof> well, not most pages. Just a lot websites and webpages
[07:59] <pvanhoof> i386 with 686 kernel, no additional plugins installed
[07:59] <Diziet> One example where it always crashes with a fresh profile would be ideal.
[07:59] <pvanhoof> fresh hoary (with not a single change) to dapper upgrade
[08:00] <Diziet> As in, file a bug with `steps to reproduce: mv .mozilla .mozilla-aside; firefox&; visit http://....'.
[08:01] <Diziet> Or is it not that reproduceable ?
[08:01] <pvanhoof> The problem is that the problems look like races. Often I can after firefox restart load the crashing page perfectly
[08:01] <Diziet> Oh.
[08:01] <pvanhoof> I know that sucks
[08:01] <pvanhoof> :)
[08:02] <pvanhoof> Oh wait, I do have Java(TM) Plug-in 1.5.0_06-b05 and Flash Movie player Version 0.4.12 as plugins
[08:02] <Diziet> Can you find a recipe that makes it nearly-always crash for you ?  As in, `start it, visit this url and then this one and hit reload a lot' or some such.
[08:02] <pvanhoof> I'll try
[08:02] <Diziet> Oh.  Can you take those out and see if it still does it ?  Easiest is to move your .mozilla aside.
[08:02] <Diziet> If you  mv .mozilla .mozilla-aside  it will come up fresh.
[08:03] <Diziet> When you're done testing you can  rm -rf .mozilla  and  mv .mozilla-aside .mozilla  to put everything back.
[08:03] <Diziet> Um, assuming you installed them via ffox rather than synaptic.
[08:04] <pvanhoof> this one is reproducable
[08:04] <pvanhoof> www.vw.be, choose "Dealer Locator"
[08:04] <pvanhoof> Check bock checkboxes in the new window, fill in .. for example "Sint Nicklaas"
[08:04] <Diziet> Excellent.  Can you file a bug ?  My testbed isn't booted atm; I'm just wading through hundreds of emails.
[08:04] <pvanhoof> go back to the parent window: it doesn't respond to keys on your keyboard
[08:05] <pvanhoof> bugzilla on ubuntu or mozilla?
[08:05] <Diziet> Ubuntu.
[08:05] <Diziet> Oh dear.  `New: downloading by firefox display exceptionally'
[08:06] <Diziet> Ah, screenshots.
[08:06] <sivang> rehi all
[08:07] <salsan> hi
[08:07] <pvanhoof> http://bugzilla.ubuntu.com/show_bug.cgi?id=21836
[08:09] <Diziet> pvanhoof: Thanks.  I'll try to look into that soon.
[08:09] <pvanhoof> ok
[08:09] <pvanhoof>  update-notifier: Depends: notify-daemon but it is not installable
[08:09] <pvanhoof> is somebody going to fix that one? :)
[08:10] <pvanhoof> E: Couldn't find package update-daemon
[08:12] <sivang> pvanhoof: that was fixed for me this afternoon
[08:12] <sivang> weird
[08:12] <pvanhoof> perhaps are the mirrors out of sync?
[08:12] <pvanhoof> which one are you using?
[08:13] <sivang> pvanhoof: UK one, a.u.c
[08:13] <sivang> mvo told me to use those when I had problems with update-* as well
[08:14] <pvanhoof> that one does give me a new gnome-session ;)
[08:14] <pvanhoof> but update-notifier stays broken
[08:15] <pvanhoof> Package notify-daemon is not available, but is referred to by another package.
[08:15] <pvanhoof> This may mean that the package is missing, has been obsoleted, or
[08:15] <pvanhoof> is only available from another source
[08:15] <pvanhoof> that's another error that the first one
[08:15] <Diziet> Right, off for dinner now.  TTFN
[08:15] <pvanhoof> s/that/than
[08:17] <Kamion> pvanhoof: just needs to be promoted to main, that's all; doing now
[08:17] <pvanhoof> http://uk.archive.ubuntu.com/ubuntu/pool/main/u/ doesn't contain a update-daemon
[08:17] <pvanhoof> ok, thanks
[08:18] <Kamion> oh, hmm, I'm not sure that this is a continuation of notification-daemon
[08:19] <Kamion> somebody will need to do up a main inclusion report for it, then
[08:19] <Kamion> sivang: it's fixed for you because you have universe in your sources.list; pvanhoof doesn't
[08:19] <pvanhoof> *adding universe*
[08:19] <Kamion> in the meantime, just ignore it
[08:19] <pvanhoof> wieerd, I have this one: deb http://be.archive.ubuntu.com/ubuntu dapper universe
[08:20] <pvanhoof> aha, be's universe isn't synced
[08:20] <Kamion> be.archive.ubuntu.com isn't the master archive
[08:21] <pvanhoof> getting lots of new updates ... the be mirror guys should fix their stuff :)
[08:21] <Kamion> and from the looks of it it hasn't updated much at all since November
[08:21] <Kamion> Znarl: ^-- be.archive hopelessly out of date
[08:21] <Kamion> though I suppose it's still OK for breezy
[08:22] <pvanhoof> mplayer-586: Depends: libjack0.80.0-0 (>= 0.99.0) but it is not going to be installed
[08:22] <pvanhoof> new problem :)
[08:22] <Kamion> you know where the bug tracking system is, and it's not here ...
[08:22] <pvanhoof> :p, sorry
[08:23] <pvanhoof> Using unstable distro's is fun. Just don't break development stuff AND the stuff VMWare needs .. doing a new project at newtec for which I need at least vim, gmake and g++ :p
[08:24] <sivang> Kamion: ah , right. thanks for the note
[08:25] <sivang> pvanhoof: you better have a breezy installation next to you, trust me
[08:26] <pvanhoof> sivang, yeah .. that's why you guys shouldn't break VMWare :)
[08:26] <pitti> pvanhoof: we need to do that to count the dapper user base :)
[08:26] <sivang> ROTFL
[08:26] <pvanhoof> :p
[08:27] <pvanhoof> by the way, Xen would be nice
[08:27] <pvanhoof> I tried it on a dapper .. it does work a little bit :)
[08:27] <pvanhoof> after building an initrd of course
[08:28] <pvanhoof> Mainly issues with the mouse. But for example X11 works
[08:28] <pvanhoof> of course not fglrx and stuff like that
[08:29] <pvanhoof> and I would like to use the video card on a guest, rather than the monitor (for that I assume I'd have to assign the pci device of my ati card to the guest)
[08:29] <pvanhoof> Using ubuntu mainly for development .. I can assure virtual machines are used a lot for a lot projects and testing
[08:30] <pvanhoof> Until that day, I'll pay vmware :)
[08:33] <elmo> Kamion: what's that feature in ssh called where you can have one machine explicitly trust another machine, and let bob ssh from foo to bar as bob without creds?
[08:34] <pvanhoof> private and public keys?
[08:34] <Kamion> elmo: HostbasedAuthentication in protocol 2
[08:34] <pvanhoof> ah
[08:35] <elmo> Kamion: right, thanks
[08:35] <elmo> Kamion: how insane is that, do you think?
[08:36] <Kamion> elmo: it's fine if you trust DNS
[08:37] <Kamion> hostbased auth still checks the host keys (unlike old .rhosts)
[08:37] <elmo> ok
[08:41] <Mez> elmo: can you sync python-feedparser from debina please-  no ubuntu package :D
[08:41] <elmo> feedparser | 3.3+cvs20051220-1 | dapper/universe | source
[08:41] <elmo> mez: nothing to sync
[08:41] <Mez> oh :D lol - It wasnt there yesterday :D
[08:42] <elmo> yes it was
[08:42] <elmo> it's been there since 30 Dec
[08:42] <Mez> well whenever i tried to install ipodder :P
[08:42] <Mez> lol
[08:42] <Mez> musta not update'd
[08:42] <Mez> :D
[08:42] <Mez> lol
[08:42] <Mez> sorry :D
[08:42] <Mez> elmo - any ideas why I'm not getting katie emails to my mez@ubuntu.com email ? It's starting to annoy me now
[08:44] <Mez> oh, and cheers for the backports :D
[08:47] <elmo> mez: no idea, no, I'll look a bit later
[08:47] <Mez> elmo: cheers :D
[08:57] <jdahlin> what's the state of dapper, is it somewhat usable/stable yet?
[08:58] <seth_k|lappy> I've been using it since day 1, jdahlin 
[08:58] <seth_k|lappy> I've never had showstopper problems (similar to the ones encountered during Breezy cycle)
[08:58] <zyga> did anyone notice that restart seriously hangs and requires hard reboot / remote reboot 
[08:58] <zyga> even doing ps aux hangs at that point
[09:04] <zyga> how to force dpkg to extract configuration files again
[09:07] <zyga> nn
[09:07] <BenC> --force-confmiss
[09:08] <zyga> dpkg could use a smaller, more helpfull --help message 
[09:09] <zyga> I don't really understand what's the purpose of --license in the default help screen ....
[09:09] <zyga> BenC: thanks :-)
[09:09] <zyga> (as if gnu/linux systems had insufficient amount of copies of GPL around)
[09:16] <Amaranth> zyga: there should only be one copy
[09:16] <zyga> Amaranth: yeah right common-licenses 
[09:20] <Kamion> zyga: dpkg --license almost certainly predates /usr/share/common-licenses/
[09:21] <zyga> Kamion: If I ever hack dpkg I'll sure to patch that and the help screen :)
[09:21] <Kamion> dpkg's --help is helpful, when you need it
[09:22] <Kamion> if you have problems with it, you probably shouldn't be using dpkg directly anyway ...
[09:22] <zyga> Kamion: it's not organized right IMHO; I could compare it to tla --help and bzr help
[09:23] <zyga> s/tla --help/tla help/
[09:23] <zyga> the former spits out gobs of text, the latter spits out just enough
[09:24] <Kamion> when I'm reaching for dpkg --help, it's because I need the obscure stuff *shrug*
[09:25] <zyga> Kamion: I wouldn't touch it if the tool of choice, apt, had 'fix this package for me, I'm dumb' command (aka full-reinstall)
[09:25] <zyga> :-)
[09:25] <Kamion> so add it to apt?
[09:26] <Kamion> you can do it in apt with ultra-weird options, but it would be nice to have an easier way to get at --force-conf{miss,def} at least
[09:26] <zyga> Kamion: I always wanted apt-get reinstall instead of install --reinstall
[09:26] <zyga> I might look at it some day
[09:26] <Kamion> I'm sure a wrapper script would be trivial anyway
[09:26] <Kamion> there's wajig if you want that sort of wrapper ...
[09:27] <zyga> wajig?
[09:27] <zyga> oh :)
[09:27] <Kamion> (don't particularly care for it myself, but hey)
[09:44] <stratus> mvo, new gdebi revision (the most important error handling needed atm, i think - console only)
[09:45] <mvo> stratus: merging now
[09:46] <stratus> mvo, is there gdebi somewhere in launchpad? i was trying to start translating it there, just to see how it works
[09:48] <mvo> stratus: https://launchpad.net/distros/ubuntu/dapper/+sources/gdebi/+translate
[09:48] <mvo> but it seems to be not yet ready for translation in rosetta, may carlos knows what to do
[09:48] <carlos> mvo, dapper translations are not yet imported
[09:49] <mvo> aha, thanks carlos 
[09:49] <carlos> some of them appear because a script was run before it should
[09:49] <jordi> carlos: oh
[09:51] <hyperactivecrond> does the current ubuntu grub support gfxboot?
[09:51] <stratus> mvo, no problem i'll follow the old way
[09:51] <Burgwork> hyperactivecrond, in dapper yes
[09:51] <hyperactivecrond> Burgwork: so no in breezy then...
[09:51] <Burgwork> hyperactivecrond, oh wait, I am tired. I have no idea
[09:52] <Kamion> hyperactivecrond: no, and it won't
[09:52] <Kamion> as I said to you earlier
[09:52] <Kamion> not in dapper, anyway
[09:52] <hyperactivecrond> i'm asking this b/c i planned on making a customization of ubuntu, to run on a dvd, that has ubuntu, kubuntu, and xubuntu-desktop's instlaled. ok thx Kamion
[09:52] <janimo> jordi, which is the largest project that uses rosetta for translations
[09:53] <Kamion> if you need it, you'll have to fish out the patches from SuSE
[09:53] <shaya> seb128: evolution bug: evolution: symbol lookup error: /usr/lib/evolution/2.6/components/libevolution-mail.so: undefined symbol: mail_tool_get_local_movemail_path
[09:53] <hyperactivecrond> Kamion: so the ubuntu cd uses syslinux for booting,then?
[09:53] <Kamion> hyperactivecrond: yes, we always have
[09:53] <hyperactivecrond> ah. thanks.
[09:53] <mvo> stratus: thanks for the diff :) merged 
[09:53] <Kamion> I'd wait until gfxboot-theme-ubuntu matures a bit more, though, if I were you
[09:53] <seb128> shaya: does it crash?
[09:53] <shaya> yes
[09:53] <shaya> 2.5.4-0ubuntu3
[09:53] <seb128> shaya: evolution --force-shutdown && evolution?
[09:53] <hyperactivecrond> Kamion: ok... 
[09:53] <shaya> tried
[09:53] <shaya> try again I will
[09:54] <stratus> mvo, i think it cover the most common mistakes that the user will do after install the package
[09:54] <jordi> janimo: well, some language teams are successfully using Rosetta to translate OpenOffice (Esperanto and Kurdish)
[09:54] <mvo> definitely!
[09:54] <shaya> same crash
[09:54] <shaya> actually, less of a crash, then an exit
[09:54] <stratus> mvo, the only one left out is when user try to run gdebi manually against a non-deb package, it stills spits out a backtrace on him
[09:54] <shaya> yes, no crash, exit
[09:54] <Kamion> hyperactivecrond: (i.e. I'd suggest just using the old syslinux in breezy for now and documenting the boot parameters)
[09:54] <shaya> gdb says Program exited with code 0177
[09:54] <janimo> jordi, I am wondering if rosetta is mature enough to propose it being used exclusively by a project
[09:55] <Kamion> but obviously it's up to you
[09:55] <hyperactivecrond> Kamion: planned on
[09:55] <janimo> are gnome/kde not playing with it yet?
[09:55] <jordi> janimo: many smaller projects are using it already.
[09:56] <janimo> jordi, is there a current status/feature roadmap for rosetta somewhere?
[09:56] <jordi> janimo: KDE and GNOME have their own translation infrastructure. We're working on a few things that will make it possible to transalte GNOME/KDE via Rsoetta without conflicting with their infrastructure though.
[09:56] <Burgwork> janimo, you thinking of moving all of xfce onto it?
[09:56] <janimo> what are the main hurdles you see right now before it could suit most projects needs
[09:56] <jordi> janimo: for future plans, just look at the specs registered for http://launchpad.net/products/rosetta
[09:56] <jordi> th
[09:56] <jordi> that's our roadmap
[09:57] <janimo> Burgwork, exactly, but before proposing this to xfce I need to see if it;s feasible :)
[09:57] <stratus> mvo, i'll translate it to pt_BR and see if there's something more that i can bring into the code (probably not, i'll be short in time) in two or three hours. It's up to you when a new version could be released then. =)
[09:57] <janimo> jordi, thanks. As Burgwork noted I am thinking what it would take to try convincing xfce to try using it
[09:57] <jordi> janimo: xfce, as a project with multiple products under the same umbrella, would currently find some problems with the permission handling of product groups
[09:58] <shaya> seb: nothing seems to have "mail_tool_get_local_movemail_path"
[09:58] <mvo> stratus: your changes justify a release already, feel free to upload it :)
[09:58] <seb128> shaya: move on #ubuntu-desktop please, it's way too noisy here
[09:58] <janimo> jordi, ah so maybe a bit further on?
[09:58] <shaya> downloading source package, to see where its used
[09:58] <stratus> mvo, great i'll do.
[09:58] <jordi> it basically depends on your needs. Do you need that only some designated people by xfce admins are able to work on those translations, or would you allow anyone to contribute to, say, the French translation?
[09:58] <Burgwork> jordi, janimo shall we move to #launchpad?
[09:59] <mvo> stratus: thanks
[09:59] <janimo> ok
[09:59] <janimo> moved over
[09:59] <jordi> Burgwork: sure
[09:59] <hyperactivecrond> how did we get syslinux to support gfxboot? is it in a deb?
[10:00] <hyperactivecrond> never mind.
[10:14] <wftl> Is there a list of major packages to be included in the final Dapper? I'm not looking for individual debs, but specific applications. For example, under the Sound & Video menu, will the basic CD player app still be there (it's not in Flight 2)
[10:14] <tseng> some desktop items are hidden
[10:14] <Burgwork> wftl, dapper is not final yet. And there has been some programs hidden. See https://wiki.ubuntu.com/MenusRevisited
[10:15] <tseng> the menus were too crowded
[10:17] <mdke> wftl, put an audio cd in, it should open automatically
[10:18] <sivang> night all
[10:58] (mdz/#ubuntu-devel) especially given that it's a regression from breezy
[10:59] <Mithrandir> you think the new livecd is a regression from breezy?
[11:00] <mdz> no, the usplash support has regressed though
[11:00] <Mithrandir> but that'll be fixed with the initramfs-usplash patches and we'll all in all end up with a lot smoother live than before
[11:01] <mdz> yes, that's the plan, but I'm saying that there's a little more to it than waiting for usplash-initramfs to land
[11:01] <Mithrandir> sure
[11:02] <mdz> usplash support was a part of the summary for the simplified-livecd spec; the text of the spec delegates that concern to usplash-initramfs, but that spec doesn't cover the livecd-specific aspect that I see
[11:02] <Mithrandir> I'll make sure that we have usplash working, even if it falls between the crack between the specs.
[11:02] <mdz> I appreciate it
[11:03] <Mithrandir> I'm more worried about how it looks as a whole than whether it's my responsibility or not to fix it from fine-reading the spec. :-)
[11:05] <mdz> functionally it's looking good
[11:05] <mdz> have you talked with lamont or infinity about getting livefs images in squashfs format in addition to the existing ones?
[11:06] <mdz> it should be pretty easy for them to arrange
[11:06] <Mithrandir> yes, or rather, infinity said "we'll talk about it on friday"
[11:06] <mdz> good
[11:07] <Mithrandir> it's not a hard thing at all.  I want to play with squashfs-lzma as well, since that saves us even more.
[11:08] <lamont> mdz: you're making me cry
[11:08] <mdz> Mithrandir: you mentioned a figure of ~100M?
[11:08] <mdz> that's fairly astonishing
[11:08] <mdz> lamont: am I?
[11:08] <lamont> du -s public_html/LiveCD/
[11:08] <lamont> 20131140        public_html/LiveCD/
[11:08] <Amaranth> squashfs-lzma saves ~100M?
[11:08] <lamont> but hey, disk is cheap
[11:09] <mdz> lamont: how much per build?
[11:09] <Mithrandir> 15:34 < Mithrandir> -r--r--r--  1 root root 520M 2005-12-19 12:34 /home0/filesystem.cloop
[11:09] <Mithrandir> 15:34 < Mithrandir> -rwx------  1 root root 495M 2005-12-19 13:52 /home0/filesystem.squashfs
[11:09] <Mithrandir> 15:34 < Mithrandir> -rwx------  1 root root 436M 2005-12-19 15:33 /home0/filesystem.squashfs-lzma
[11:09] <lamont> that's on the order of 4 builds
[11:09] <Amaranth> Mithrandir: holy shit
[11:09] <lamont> and has both the cloop and fsimage for each
[11:09] <mdz> lamont: we can scrap the uncompressed version when the build is done
[11:09] <mdz> lamont: oh, you need it for the next one
[11:09] <lamont> mdz: not if you want rsync-love
[11:09] <lamont> yeah
[11:09] <Amaranth> hey, you could fit more languages and/or mono on there with the extra room ;)
[11:10] <lamont> well, we could uncompress it at the start, but sigh.
[11:10] <mdz> only the N-1th build though
[11:10] <Mithrandir> mdz: yeah, and squashfs (both lzma and regular) are rsyncable.
[11:10] <mdz> but unless the machine is out of disk space...;-)
[11:10] <mdz> Mithrandir: is decompression performance significantly different?
[11:10] <lamont> mdz: that 20 is 2 each * 4 at ~2.3GB/build
[11:11] <mdz> lamont: 2 each?
[11:11] <lamont> yeah, the other 1 or 2 failed...
[11:11] <lamont> the prune script just nukes things > 3 days old
[11:11] <Amaranth> i remember lzma being slightly slower than gzip and a lot faster than bz2
[11:11] <Amaranth> dunno what cloop is/uses though
[11:11] <lamont> unless it's the latest or current (last successful), in which case, it's timeless
[11:11] <mdz> cloop uses gzip mostly
[11:11] <lamont> Amaranth: gzip
[11:12] <Amaranth> ok then
[11:12] <lamont> our images are almost entirely gzip -9, according to the log
[11:12] <Amaranth> this is probably one of those things where the small size makes it faster to get off disc so it ends up being a tie or win
[11:12] <lamont> hrm... istr maybe we just forced that, instead of letting it try -1 through -9 before deciding that -9 was best.
[11:13] <mdz> it supports gzip and 7zip
[11:13] <lamont> mdz: it wouldn't be to much of a tweak to have the prune script nuke everything _except_ latest&current
[11:13] <Amaranth> 7zip's "native" format is lzma
[11:14] <mdz> hmm, we could try cloop+7zip
[11:14] <mdz> I doubt it'll be as good as squashfs-lzma though
[11:14] <Mithrandir> mdz: devmapper+cloop: ~367s till disk activity is settled, unionfs+cloop: ~326s, unionfs+squashfs: ~231s
[11:14] <mdz> Mithrandir: !
[11:14] <Mithrandir> mdz: basically, since you're pulling stuff off the CD, it's the cd which is the bottleneck
[11:14] <daniels> Kamion: ping
[11:14] <Amaranth> it's faster than cloop _and_ 100M smaller?
[11:14] <daniels> Kamion: have you seen the xkb-on-console-keymap work?
[11:15] <Mithrandir> yeah, and squashfs gives us the ability to sort the filesystem however we want
[11:15] <Amaranth> Mithrandir: yeah, that's what i was guessing
[11:15] <mdz> Amaranth: not entirely surprising, given that cloop knows nothing about the filesystem
[11:15] <Amaranth> Mithrandir: you can use the same argument for using lzma for packages too :)
[11:15] <mdz> every metadata block has to be decompressed too
[11:15] <Mithrandir> so we can do some cool readahead-based magic and probably just grab the first 100M or so off the CD in one long read.
[11:16] <Amaranth> it'd be nice to have a livecd that didn't take 3 minutes to start
[11:16] <Mithrandir> that'd be fairly sweet.
[11:16] <Amaranth> or longer
[11:16] <Amaranth> i stopped counting at 3 :P
[11:16] <Mithrandir> Amaranth: http://people.ubuntu.com/~tfheen/live-bootcharts/unionfs-squashfs-dapper-20051216-1.png is from my machine (which is fast and has plenty of ram, but still.)
[11:16] <mdz> Mithrandir: when we do data collection to create the readahead list, we can use that same data to sort the filesystem
[11:17] <Mithrandir> mdz: exactly what I was thinking.
[11:17] <mdz> is squashfs going upstream?
[11:18] <lucas> do members of ubuntu-dev have the right to upload multiverse packages ? or does it have to be an ubuntu-core-dev ?
[11:18] <Mithrandir> mdz: hmm, we could do something like bootchart=upload.  readahead=upload, so we could collect usage stats and see what apps people actually started in the first couple of minutes after boot and put those early (and readahead them too, if possible)..
[11:18] <mdz> lucas: ubuntu-dev can upload to multiverse
[11:18] <lucas> ok, thx
[11:18] <Mithrandir> mdz: I don't know; it seems to be nice and stable, so it could, at least.
[11:18] <mdz> more likely than cloop, certainly
[11:19] <Amaranth> Mithrandir: that dang "init" thing seems to be running the whole time, you should try to kill it ;)
[11:19] <Mithrandir> Amaranth: ;-P
[11:19] <mdz> Mithrandir: someone suggested using inotify to track file access during boot; that sounded like the right idea
[11:20] <Mithrandir> mdz: I have suggested it multiple times, I just haven't had time to actually write the code. :-)
[11:20] <mdz> I didn't realize inotify could detect read access
[11:20] <dholbach> have a nice evening
[11:20] <Mithrandir> it can do everything, including summoning dragons and making pigs fly.
[11:20] <Mithrandir> but, I'm off to bed now.  See you around.
[11:21] <Amaranth> night
[11:21] <mdz> good night
[11:21] <psusi> wait a second
[11:21] <psusi> lzma for the squashfs saved 100M?  using the same block size?
[11:22] <psusi> 7zip can save a LOT of space but it seems to do so mostly through use of extreamly large dictionaries... which only apply to data blocks that are larger
[11:23] <psusi> cramfs/squashfs/cloop all use small block sizes to allow random access
[11:24] <Mithrandir> psusi: it works, at least, so yes.
[11:24] <Mithrandir> psusi: that is, I haven't actually booted squashfs-lzma, but it should work.
[11:24] <Mithrandir> it works for the openwrt people.
[11:24] <Mithrandir> but, night, for real.
[11:55] <doko> Kamion: do you have a schedule for the next live CD?
[11:57] <doko> mdz: would it make sense to name packages/package sets in UpstreamVersionFreeze, for which we want make an exception or propose an exception?