[12:05] <nakeee> I want to translate things to hebrew using rosetta, I thought I would start with xchat
[12:05] <nakeee> I know xchat is partly translated
[12:05] <nakeee> how can I see what version of the hebrew po file is being used?
[12:05] <nakeee> so I won't retranslate things which were translated by the gnome team already
[12:13] <mdz> nakeee: /join #rosetta
[12:14] <nakeee> mdz: where there are even less people?:)
[12:15] <mdz> nakeee: there may be fewer, but they are the ones who can answer your question (I can't)
[12:15] <mdz> you can also mail the rosetta-users mailing list
[12:16] <Nafallo> there is no way in current hoary to empty an CD-RW?
[12:16] <mdz> Nafallo: only from the command line, afaik
[12:17] <mdz> I usually don't bother blanking them; that's one less write cycle that could be used for real data
[12:18] <Nafallo> mdz: hmm, need to write an ISO to it. I HAVE to blank it before?
[12:18] <Nafallo> can't get cdrecord to show me what dev= to use :-P
[12:19] <Nafallo> /dev/hdc should be ATA:something?
[12:22] <nakeee> 2
[12:23] <nakeee> no?
[12:24] <crimsun> Nafallo: -dev=/dev/hdc
[12:25] <Nafallo> nakeee: 0,2,0 worked. crimsun: ah, THAT easy :-)
[12:26] <Nafallo> -dev=/dev/hdc worked best :-)
[12:30] <mdz> Nafallo: right-click on the ISO in nautilus, click "write to CD"
[12:30] <mdz> Nafallo: it will automatically blank it if necessary
[12:31] <Nafallo> mdz: DOOH! to easy ;-)
[12:31] <mdz> there is no way to tell it "just blank the CD and don't write anything to it", though, which is what I thought you were asking
[12:32] <Nafallo> mdz: well, I was. but with the goal to write an ISO-image ;-).
[12:34] <jdub> daniels: ping
[12:35] <elmo> how do I see what's making outgoing DNS calls?
[12:35] <elmo> as, in what process
[12:36] <Nafallo> while I see alive ubs, what does debzilla@ubuntu.com as assigned to do in bugzilla?
[12:39] <mdz> elmo: hmm...you can match on pid in netfilter, but I don't think you can convince it to log it afaik
[12:39] <elmo> damn.. something cached my old resolv.conf from when I was in the data centre
[12:39] <mdz> elmo: filter outbound DNS so that it has to timeout, then use netstat to see what has the socket open?
[12:40] <elmo> AHA
[12:40] <elmo> it's spamassassin
[12:40] <elmo> POS
[12:40] <mdz> you knew it was a POS already, no sympathy
[12:41] <elmo> mdz: meh, I'll subscribe you to ftpmaster, admin and some of the inanely-widely-publicized-since-the-dawn-of-time debian addresses and then see who needs the sympathy
[12:41] <Keybuk> SA3 is more POS than SA2
[12:43] <elmo> is there anything else that's as good and most importantly doesn't require training?
[12:45] <Treenaks> elmo: echo "elmo:|cat>/dev/null" >> /etc/aliases ; newaliases
[12:47] <elmo> Treenaks: thanks dude, you're a star
[12:48] <Treenaks> elmo: it gets rid of all spam and doesn't require training
[12:52] <mdz> elmo: I already get a huge amount of spam
[12:52] <mdz> elmo: my addresses are, er, widely publicized :-P
[12:53] <mdz> bogofilter does very well for me
[12:54] <elmo> isn't bogofilter the one that uses libdb and constantly needs upgrades every second week?
[12:54] <elmo> and does it work without training, a la.. err, crm114?
[12:55] <Clint> yes, and no.
[01:44] <bob2> so, yeah, python2.4 doesn't like this whole installing thing
[01:45] <bob2> Setting up python2.4 (2.4-2ubuntu5) ...
[01:45] <bob2> Compiling python modules in /usr/lib/python2.4 ...
[01:45] <bob2> /usr/bin/python2.4: can't open file '/usr/lib/python2.4/compileall.py': [Errno 2]  No such file or directory
[01:49] <jdub> elmo: bogofilter is the one that changes command line parameter semantics -> to their polar opposites ;-)
[01:53] <Clint> that only happened once
[02:01] <jdub> i only pulled the trigger of the nailgun once
[02:02] <Clint> next fun breakage might be to store database words in UTF-8
[02:02] <Clint> is everyone excited?
[02:13] <Keybuk> . o O { if only zsh supported utf-8 }
[02:14] <Clint> . o O {if only people wouldn't discuss zsh and utf-8 every day without submitting patches}
[02:14] <Keybuk> I looked once, I got scared
[02:14] <Clint> it's scary
[02:14] <Keybuk> there did seem to be a Tom Lord "reimplement the world" going on inside zsh
[02:14] <Clint> well, a lot of that predates the world
[02:15] <Clint> like all the funky memory management stuff predates mmap
[02:16] <Keybuk> then again, I can't talk about packages not supporting utf-8
[02:16] <Keybuk> ho-hum
[02:17] <doko> bob2: from which python2.4 version do you upgrade?
[02:17] <Clint> at some point, someone will get frustrated enough to write some really hacky patches to make it look like it does utf-8, a la readline
[02:19] <bob2> doko: 2.4-2ubuntu1 -> ubuntu5
[02:22] <sladen> Keybuk:   bash  Provides: zsh
[02:23] <doko> bob2: and python2.4-minimal is already installed?
[02:23] <Keybuk> sladen: except it doesn't
[02:24] <bob2> doko: hrm, yes, at ubuntu1
[02:24] <Keybuk> bash doesn't do anywhere near as much of the weird (but infinitely useful) shit zsh dose
[02:24] <bob2> doko: maybe aptitude is being stupid and put it on hold
[02:24] <Keybuk> mv ^Sources(@) ..
[02:24] <Keybuk> "move all symlinks, except Sources to the parent directory"
[02:24] <Clint> bash is only about 5 years behind zsh in terms of usability
[02:25] <doko> bob2: ok, I see, please update -minimal as well. the dependency is still unversioned. will fix it.
[02:25] <bob2> doko: ahh
[02:25] <bob2> doko: cool, thanks!
[02:26] <sladen> find -type l | xargs mv --target-directory=..     # nothing could be simpler
[02:26] <Keybuk> sladen: you forgot -not -name Sources
[02:26] <sladen> yeah, still might fit on one line with a 132 column display
[02:27] <sladen> and to cope with spaces, you'd need  -print0 | xargs -0  too
[02:28] <Keybuk> sorry, remind me, are you defending bash or not? :)
[02:29] <bob2> doko: ah, perfect, thanks
[02:34] <Keybuk> foreach FILE (*) mv $FILE ${FILE/^\([^.] \)\./\1\/}
[02:34] <Keybuk> and friends too :p
[02:35] <davyd> is the gnome-panel package maintainer around?
[02:35] <bob2> davyd: it's probably a gtk bug
[02:35] <davyd> (who appears to be seb)
[02:35] <davyd> bob2: hmm? I don't want to chat about bugs
[02:37] <Keybuk> wrong timezone for seb really
[02:37] <davyd> yeah
[02:37] <davyd> I really want a CVS gnome-panel package
[02:37] <sivang> jdub: you have that streaming video you took of me and some others over the crystal room? :)
[02:37] <davyd> I was going to roll myself one
[02:37] <davyd> but you gents have much patchage crack in there
[02:38] <Keybuk> davyd: hmm?  you should just be able to copy the debian/ directory in
[02:38] <mjg59> I so desperately need more RAM
[02:38] <sivang> mjg59: how much do you have?
[02:38] <davyd> Keybuk: that's what I would have thought
[02:38] <Keybuk> you might get bitten by debian/patches/01_layout
[02:38] <davyd> Keybuk: indeed... what is that?
[02:39] <Keybuk> changes Application/System to Application/Places/System
[02:39] <Keybuk> just nuke the patches that don't work
[02:39] <davyd> Keybuk: aah
[02:42] <mjg59> sivang: 256MB
[02:42] <sivang> davyd: well, for me it's 3:42, so for seb around 2:42...:)
[02:43] <davyd> sivang: yeah, they're on +1 at the moment?
[02:43] <davyd> WEST?
[02:44] <sivang> davyd: I think so, due to non daylight savings, when they are using daylight savings it's the same time as here..
[02:44] <sivang> mjg59: oh, I bet your system chokes when running OOo together with moz ..:-/
[02:45] <sivang> mjg59: I have 0.5G ram, and still machine gets slow when using OOo/Web together with evo|moz|thunderbird
[03:02] <zenrox> sivang,  have you tried prelinking
[03:03] <zenrox> sivang,  try this http://ubuntuforums.org/showpost.php?p=9864&postcount=80
[03:03] <sivang> zenrox: no, can't say I have. You have a link for how to make it work?
[03:04] <zenrox> does help reduce load time on firefox and OOo
[03:04] <zenrox> i have bonic running and my dvds dont slow down
[03:05] <sivang> zenrox: this is documented somehwere other then the uf.o ?
[03:05] <zenrox> works for me
[03:05] <zenrox> and suse uses it
[03:05] <zenrox> at install of every program
[03:10] <sivang> zenrox: nice
[03:11] <zenrox> does that help
[03:12] <sivang> zenrox: I will install beofre I go to sleep, currently working on something can't afford to let the system become less responsive
[03:13] <zenrox> wont slow it down too much
[03:14] <davyd> hmm, I am going to have to build every gorram module under the Sun I think
[03:14] <davyd> this is annoying
[03:28] <jdub> pants off davyd 
[03:29] <davyd> PANTS!
[03:29] <jdub> having a fun pre-freeze hack weekend?
[03:29] <davyd> also hi
[03:29] <davyd> jdub: yeah, except I realized the one problem in my new, no work dev environment
[03:29] <davyd> I need CVS gnome-panel, and I can't pull the same hack with gnome-panel as I can with applets
[03:30] <davyd> this makes me sad :(
[03:31] <jdub> :-)
[03:31] <jdub> hrm, you can use the funny bonobo env vars to do the same thing
[03:31] <davyd> jdub: panel is not started by bonobo though
[03:31] <davyd> so my bonobo magic is no good
[03:32] <jdub> you can start panel on its own
[03:32] <jdub> hold on, on phone
[03:32] <davyd> jdub: yeah, that's worth a try
[03:33] <davyd> I was going to just roll a package, but it's proving to be a bitch
[04:08] <mdz> Keybuk: how do you think we should handle it if we were to switch to udevsend as default?  have init.d/udev set kernel.hotplug, and then init.d/hotplug set it iff it's not set to udevsend, or something like that?
[04:08] <Keybuk> yeah, that's how I was handling it
[04:08] <Keybuk> udev's first init script sets it
[04:08] <Keybuk> and hotplug won't spam over the top if it sees udevsend in there
[04:14] <mdz> Keybuk: what's this ipw/firmware thing about?  ipw2200 seems happy enough with udevsend
[04:15] <Keybuk> which version of udev?
[04:15] <mdz> ii  udev           0.050-3ubuntu1 /dev/ management daemon
[04:16] <Keybuk> the problem is that the kernel issues a firmware hotplug event to get the firmware loaded
[04:16] <Keybuk> and the driver only waits 10s
[04:16] <mdz> I don't have an ipw2100 to test, though
[04:16] <Keybuk> so if udev doesn't (for whatever reason) handle the event straight away, you get no firmware
[04:16] <Keybuk> and the driver sulks
[04:21] <mdz> daniels: X autoconfiguration for the live CD
[04:23] <daniels> mdz: hoary chroot building now
[04:23] <daniels> mdz: i've literally walked in the door 15 minutes ago
[04:23] <bob2> mdz: I can test it
[04:24] <mdz> bob2: would you?
[04:24] <mdz> bob2: brief instructions are above
[04:25] <daniels> GNAR
[04:28] <davyd> hey daniels !
[04:29] <bob2> hrm, is there a 'test' flag to check the contents of a file?
[04:30] <mdz> daniels: hoary chroot wha?
[04:32] <daniels> mdz: when I do sudo dpkg-reconfigure -plow xserver-xorg, predictably it works :P
[04:32] <daniels> mdz: working on the whole 'testing' thing
[04:32] <daniels> davyd: hey dude :)
[04:32] <davyd> daniels: you never mailed me your address ;)
[04:32] <daniels> davyd: heh
[04:32] <davyd> and my psychic powers are lacking
[04:32] <daniels> davyd@maddley.id.au?
[04:32] <davyd> madeley
[04:34] <daniels> sent
[04:35] <davyd> daniels: my spidey sense tells me it was your birthday on Thursday, is that vaguely true, or am I smoking expensive crack?
[04:37] <daniels> davyd: it's true, in a vague and more concrete sense
[04:37] <mdz> Keybuk: do you happen to have tracked down why udev is quite so abhorrently slow under d-i?
[04:37] <davyd> daniels: happy birthday ;)
[04:37] <davyd> 19 or 20?
[04:37] <davyd> "By the definitions I use,  and  seem strictly equivalent to me. Would be it fair to say that   " haha
[04:37] <Keybuk> mdz: I didn't know it was
[04:38] <mdz> Keybuk: it takes about 5 seconds before /dev/fb/0 appears after the module is loaded, most of the time
[04:38] <mdz> and sometimes it never appears at all, but I expect that'll be fixed by udevsend
[04:39] <Keybuk> odd, can't say I've looked at it
[04:40] <mdz> daniels: I was under the distinct impression that 'Driver "radeon"' was nonsense and there was no reason ever to do that
[04:40] <mdz> daniels: but someone just stated in the wiki that it got things working for them
[04:41] <Keybuk> mdz: could certainly be an issue with hotplug trying to do the event too early, and then backing off too long
[04:42] <daniels> mdz: it is not nonsense, and sometimes the ati driver is so staggeringly shit that it fails to hand over to the right submodule
[04:42] <daniels> mdz: my personal opinion is that we should use atimisc/r128/radeon in preference to tai
[04:43] <mdz> Keybuk: does udev have some magic to prevent it from having itself called again via its own hotplug hook?
[04:43] <mdz> Keybuk: or does it just deal with that when it comes?
[04:45] <davyd> daniels: I opted not to go to Melbourne next month
[04:45] <Keybuk> I thought there was, but I can't see the magic now
[04:45] <davyd> since I'm going to be doing a lot of travelling otherwise
[04:45] <Keybuk> I think udevd sets an environment variable, and udevsend bails out if it finds that
[04:45] <davyd> so it'll obviously be April then
[04:47] <daniels> davyd: ahr, bugger
[04:48] <davyd> daniels: also, you didn't answer, did you turn 19 or 20?
[04:48] <daniels> davyd: oh, sorry, missed that one -- 19
[04:48] <davyd> cool
[04:59] <sivang> davyd: g_warning is great in debugging. thanks :)
[05:00] <sivang> davyd: it appears there nothing wrong iwht the model , as expected :) I'll have to check the xml creation functions then....:-/
[05:01] <daniels> 'The V3200 is a 16Mb Banshee board, which means it can support some rather impressive 2D resolutions. It can manage a 75Hz refresh rate at 1792 by 1344 with 32 bit colour. This is probably enough to stretch a startlingly expensive 30 inch monitor to its limits, and much higher than you should bother using on smaller screens like the 17 and 19 inch monitors many people use.'
[05:05] <bob2> mdz: doesn't work
[05:06] <bob2> mdz: afaict from dmesg it is that it's not loading the firmware in time
[05:06] <mdz> bob2: thanks
[05:06] <mdz> daniels: you have an ipw2100, right?
[05:06] <bob2> he has atheros
[05:06] <mdz> oh
[05:06] <daniels> mdz: ath
[05:06] <bob2> thombot has ipw2100
[05:06] <bob2> lifeless has ipw2200
[05:06] <mdz> I have 2200 as well
[05:07] <mdz> I guess thom can fix it when it breaks
[05:07] <mdz> Keybuk: unless you already know exactly what needs to be done
[05:07] <mdz> Keybuk: I'm thinking about doing the udevsend switch tonight
[05:07] <Keybuk> no, I don't -- just kill it anyway :p
[05:07] <davyd> daniels: ooh ooh, where do I buy one?
[05:08] <daniels> davyd: i have one on my desk; if it wasn't the only tdfx I had, I'd give it to you
[05:08] <davyd> I remember back in the day
[05:08] <bob2> er, last
[05:08] <davyd> when I got a 2MB Cirrus Logic
[05:08] <davyd> I could play Wolfenstein 3D
[05:08] <davyd> and it was quick too
[05:09] <bob2> daniels: mjg59 says you have a fix for the gl-is-fucked-after-resume i855 thing?
[05:09] <daniels> i miss my rendition verite (3d blaster pci)
[05:09] <daniels> that was totally ahead of its time
[05:09] <daniels> bob2: yep
[05:09] <jdub> so
[05:09] <daniels> bob2: but if you value your job, don't ask me to do it before xorg autoconfiguration for livecd is done :P
[05:09] <bob2> haha, alrighty :)
[05:10] <mdz> Keybuk: is there even any point in having hotplug check if udev is kernel.hotplug, and if not, putting itself there?
[05:10] <mdz> Keybuk: it is the default and all
[05:11] <bob2> I was just having udev overwrite sys.kern.hotplug always
[05:11] <Keybuk> mdz: not really :p
[05:12] <jdub> ok
[05:12] <jdub> so
[05:12] <mjg59> daniels: 1792x1344? 30 inch?
[05:12] <bob2> jdub: put your pants on and finish a sentence!
[05:12] <jdub> i'm getting "can't write to /dev/null" (and /dev/console) errors during initrd when booting software raid1
[05:13] <mdz> I wonder if it still makes sense for init.d/hotplug stop to set the helper to /bin/true
[05:13] <daniels> mjg59: FEAR THE AWESOME VOODOO POWAH
[05:13] <jdub> it occurs to me that /dev on the target is entirely empty
[05:13] <jdub> because there is no udev
[05:13] <jdub> so i'm trying a reboot with /dev/null and /dev/console on there
[05:13] <jdub> (i'm assuming it's target disk stuff here)
[05:14] <jdub> this is a bit of a problem
[05:14] <jdub> if it works with those nodes created, then i'll be worried
[05:14] <mdz> jdub: how did you end up in this mess?
[05:14] <mdz> jdub: initrds generated by mkinitrd have /dev/console and /dev/null on them (statically) for exactly this reason
[05:14] <jdub> yes, they are there
[05:15] <jdub> i've done mkinitrd -r /dev/md0 -o pants
[05:15] <jdub> pia says the machine successfully booted
[05:15] <jdub> so it needed null and console on the *target*
[05:16] <jdub> which means that it would be very hard for someone to convert a fresh ubuntu install to software raid1 easily
[05:16] <jdub> (also, i really dislike the hardcoded mdadm call after redoing the mkinitrd)
[05:17] <jdub> i'm sure we can come up with a better way of making robust initrds
[05:17] <jdub> that can handle configuration changes like this
[05:18] <jdub> hrm
[05:19] <jdub> unless our /dev already has those
[05:19] <jdub> i did a cp -vax
[05:19] <jdub> though that should've copied them if they were there
[05:20] <mdz> Keybuk: hmm...should init.d/hotplug run udevstart rather than run_rcs if udevsend is the helper?
[05:20] <mdz> I don't suppose it matters
[05:20] <mdz> probably best to leave it alone
[05:20] <Keybuk> mdz: that's what most of my work in Mataro was trying to do -- then you end up racing X to load your mouse driver
[05:21] <bob2> you should start gdm out of a udev event script
[05:21] <Keybuk> bob2: which one?
[05:22] <bob2> hah, touche'
[05:22] <bob2> silly english keyboard
[05:25] <daniels> le sigh
[05:25] <daniels> mdz: we don't use radeonfb, do we?
[05:25] <mdz> daniels: we deliberately pretend that it doesn't exist
[05:33] <mdz> daniels: https://bugzilla.ubuntu.com/show_bug.cgi?id=3993
[05:35] <mdz> jdub: the best way to do it is to use EVMS
[05:35] <mdz> jdub: and forget about mdadm and all that other cruft
[05:35] <mdz> jdub: if you set up fstab so that it uses a /dev/evms/foo volume as root, it will magically set up the initrd for you
[05:38] <daniels> mdz	sweet
[05:47] <daniels> mdz: thoughts on this for lrm? http://www.heby.de/ltmodem
[05:50] <Keybuk> You know what tla's killer feature is?  How it doesn't get affected by a change of umask
[05:50] <Keybuk> OH WAIT
[05:50] <Keybuk> descent dpkg-1.13% tla changes | wc -l
[05:50] <Keybuk> 2369
[05:51] <bob2> Keybuk: "don't do that"
[05:51] <bob2> (and/or file a bug)
[05:51] <Keybuk> bob2: I have different umasks on different machines
[05:51] <Keybuk> tla cannot cope
[05:53] <bob2> changing it causes tla to do a massive permission change changeset?
[05:53] <Keybuk> apparently
[05:54] <Keybuk> all the patch logs have different permissions on this machine
[05:54] <bob2> erm
[05:54] <Keybuk> 644 vs. 664
[05:57] <daniels> bbiaw
[06:11] <jdub> mdz: mmm, i wasn't bold enough to go evms this time around
[06:12] <jdub> mdz: should we regard evms as our one true raid/volume-management system; no md?
[06:12] <lamont> mdz: we could just fetch the architecture list from the archive...
[06:12] <lamont> that is, if there's a ubuntu/dists/hoary/main/binary-foo, then foo needs to be done.
[07:43] <mdz> jdub: no mdadm, right
[07:46] <lamont> mdz: what do you think of fetching the list of main/binary-* and using that?
[07:46] <lifeless> bob2: theres a bug open on keybuk...'sproblem
[07:46] <lamont> mdz: for ubuntu-meta
[07:51] <bob2> be nice.
[07:51] <lifeless> how ?
[07:51] <lifeless> :}
[08:14] <daniels> keybuk touched it last
[08:18] <crimsun> by default ~/.fonts.conf overrides everything
[08:52] <mdz> lamont: I don't think that we add new architectures frequently enough to justify the complexity, to be honest
[08:52] <jdub> mdz: but we wouldn't recommend md.ko usage either?
[08:53] <mdz> jdub: why wouldn't we?
[08:54] <mdz> EVMS uses md for raid5
[08:54] <mdz> until such time as device-mapper gets equivalent functionality, anyway
[08:54] <jdub> oh right
[08:54] <jdub> sorry, thought dm already did that
[08:55] <mdz> it does mirroring only, afaik
[08:55] <jdub> lamont: we should kill all our local.conf changes
[08:55] <mdz> but the nice thing about EVMS is that you don't have to give a damn
[08:55] <mdz> you tell EVMS you want a raid-1 volume and a raid-5 volume, and it does what needs to be done
[08:55] <mdz> including figuring out how to configure the kernel for it this week
[08:55] <jdub> so is there a simple path from md raid1 to evms, or will i need to fail one drive and re set stuff up?
[08:56] <mdz> I believe EVMS will read md metadata, so it will recognize and activate your volume
[08:56] <mdz> I'm not sure whether it uses raid1 or dm-mirror in that case
[08:57] <mdz> jdub: fire up evmsgui and see what it shows you
[08:58] <jdub> hmm
[08:58] <jdub> ok :)
[08:58] <sivang> mdz: can I install evms on an already running lvm system?
[08:59] <mdz> of course
[08:59] <sivang> mdz: coool :) apt-get instal evms? ;-)
[08:59] <jdub> mdz: is there a non-gtk evms interface?
[08:59] <jdub> oof
[08:59] <mdz> sivang: it's in base
[08:59] <mdz> jdub: evmsn is an ncurses interface
[08:59] <jdub> particularly a non-gtk1 ui
[09:00] <mdz> evms-cli has a CLI, but I don't recommend it
[09:00] <mdz> I use evmsn and evmsgui
[09:00] <jdub> whoa
[09:00] <jdub> /dev/evms/md/md0
[09:00] <mdz> sivang: evms understands both lvm1 and lvm2 metadata
[09:00] <jdub> so i can set my / to that and go?
[09:00] <sivang> mdz: cooooooooooool
[09:00] <sivang> :)
[09:01] <mdz> jdub: yeah, or you can create an EVMS volume on top of it and use that
[09:01] <mdz> EVMS sees md volumes as a lower level storage object out of which it can build things
[09:01] <mdz> but you can use them directly as well
[09:03] <jdub> makes sense
[09:03] <jdub> i might just stick with that for now
[09:03] <sivang> mdz: is colin going to integrate it into d-i ?
[09:04] <mdz> not likely
[09:04] <mdz> but it could become a bounty
[09:04] <mdz> integrating EVMS in the installer would be a much bigger project than the existing LVM support though
[09:04] <mdz> there's just a lot more you can do
[09:05] <mdz> there is an evms udeb which should in theory let you use evmsn from the installer
[09:05] <jdub> on this system i seem to have to modprobe loop
[09:05] <mdz> jdub: haven't upgraded in a while?
[09:05] <mdz> jdub: grep loop /etc/udev/links.conf
[09:05] <jdub> it's a warty server
[09:05] <jdub> not listed
[09:06] <jdub> hrm
[09:06] <jdub> script in the initrd still has mdadm
[09:07] <jdub> hrm, it doesn't have any evms related binaries in it
[09:08] <jdub> # update-grub
[09:08] <jdub> Searching for GRUB installation directory ... found: /boot/grub .
[09:08] <jdub> /dev/evms/md/md0 does not have any corresponding BIOS drive.
[09:08] <jdub> 
[09:08] <jdub> d'oh ;)
[09:08] <sivang> jdub: grub doesn't play well with it?
[09:09] <jdub> i'm pissing aroudn in the dark atm dude :)
[09:09] <sivang> jdub: hehe 
[09:09] <sivang> jdub: does it boot at all? :)
[09:09] <jdub> haven't tried
[09:13] <jdub> erm
[09:13] <jdub> hrm
[09:15] <jdub> "This means your /boot volume must be based on a single partition."
[09:15] <jdub> bong!
[09:17] <sivang> jdub: /boot is adviced to not reside on a non evms/lvm volume, if to sum up to not so much I've read about it on the web :)
[09:17] <sivang> jdub: correction, not on _a_ evms/lvm
[09:17] <mdz> jdub: ah, that was fixed post-warty
[09:18] <jdub> mdz: which bit?
[09:18] <mdz> jdub: both, actually
[09:18] <mdz> jdub: the loop device thing, and the initrd thing
[09:18] <mdz> the initrd fix should be a one-liner
[09:18] <jdub> oh right
[09:18] <mdz> the logic just doesn't get triggered on /dev/evms/md/foo, only on /dev/evms/foo
[09:18] <mdz> /usr/share/initrd-tools/probe.d/evms
[09:19] <fabbione> morning
[09:19] <mdz> fabbione: morinng
[09:19] <jdub> mdz: change to /dev/evms/*/*, or /dev/evms/* /dev/evms/*/* ?
[09:19] <mdz> jdub:     for vol in /dev/evms/* /dev/evms/md/* /dev/evms/lvm/*; do
[09:19] <mdz> is what I have in current hoary
[09:19] <jdub> oh, ok
[09:20] <fabbione> AHAHAH
[09:20] <fabbione> oh god
[09:21] <mdz> you'll be way happier if you update to the hoary version of evms if you're doing this kind of stuff
[09:21] <fabbione> "All hail fabbione,
[09:21] <fabbione> Grandmaster of the all omnipotent and omniprescient ubuntu linux kernel.
[09:21] <fabbione> I offer up my sacrifice of some signed and encrypted gpg keys in the
[09:21] <fabbione> hope that they satisfy.
[09:21] <fabbione> I only pray that my acpi may work in the next version.
[09:21] <fabbione> "
[09:21] <jdub> ok, initrd fixed - thanks :)
[09:21] <mdz> jdub: one thing to watch out for
[09:22] <mdz> jdub: the evms initrd magic in warty doesn't know how to figure out which disks make up your raid array
[09:22] <jdub> only now update-grub is completely spewing
[09:22] <jdub> mdz: ... ugh
[09:22] <mdz> jdub: so you need to ensure that the modules it needs are loaded in the initrd, yourself
[09:22] <mdz> jdub: it can detect the array fine, you just need to make sure that it can see your disks before it gets to that point
[09:22] <jdub> so, 'raid1'
[09:22] <mdz> no, stuff like your SCSI driver
[09:22] <mdz> will need to get added to /etc/mkinitrd/modules
[09:23] <mdz> the hoary version has more magic
[09:23] <sivang> fabbione: LOL
[09:23] <jdub> oh, i see, yeah
[09:23] <sivang> fabbione: morning
[09:23] <jdub> it missed sata_sil and friends
[09:23] <fabbione> eheh hi sivang
[09:23] <mdz> making evms-udeb a reality would be a pretty cool bounty project, actually
[09:24] <jdub> ok,
[09:24] <jdub> sata_sil
[09:24] <jdub> sd_mod
[09:25] <jdub> mdz: recommendations wrt grub, lilo, etc.? give up and have a separate /boot?
[09:25] <mdz> jdub: I would use grub
[09:26] <sivang> hehe, someone promoted my news flash about the arstechnica awards to a (!) mark on main page :)
[09:27] <jdub> grr, bong, using the same config (fstab, menu.lst) as before, update-grub isn't working
[09:29] <sivang> ok, enough g-s-t hacking for now, sleep 4 hours, then fininsh the backends hopefully :)
[09:31] <sivang> night all!
[09:31] <crimsun> night.
[09:33] <sivang> crimsun: night Zzzz
[09:37] <bob2> so, after suspend-to-disk, ipw2100 says "ipw2100: eth1: Failed to start the card." every second until I reboot
[09:42] <mdz> er
[09:42] <mdz> when did xfsprogs get added to base?
[09:42] <fabbione> bob2: buy serious hardware...
[09:42] <fabbione> if it works for me.. everything else is buggy hardware
[10:32] <jdub> http://alan.aspuru.com/archives/2005/01/08/ubuntu-linux-on-a-middle-school-computer-lab-a-linux-enthusiasts-adventure/
[10:32] <jdub> mako: ^
[10:33] <^_^> i am a devel
[10:33] <^_^> ?
[10:33] <^_^> ok why can't ubuntu put a bounty for a gui installer?
[10:33] <^_^> i am willing to do it for $750USD
[10:34] <^_^> i might even do it for $500 but
[10:35] <^_^> if you paying for a hack in firefox to show some html pages $500 then your standards are diffrent to what i thought they might be
[10:36] <cartman> ^_^: start with getting a real nick
[10:36] <^_^> why is ubuntu paying $500 for a $10 job
[10:37] <^_^> cartman: i change it at a rate of 1 per 4hrs
[10:37] <^_^> brb
[11:32] <^_^> back
[11:43] <^_^> BOOOOOOOOOOOOOOOOO
[11:43] <^_^> ooops
[11:43] <^_^> wrong tab
[12:58] <Kamion> lamont: http://terranova.warthogs.hbd.com/~buildd/livecd/livecd-current.cloop gives a 403; would appreciate a permissions fix
[01:05] <^_^> Kamion: lag?
[01:13] <pitti> elmo: ping
[01:13] <^_^> pitti: pong
[01:14] <pitti> ^_^: you are certain that you are elmo? :-)
[01:14] <^_^> pitti: pong
[01:15] <pitti> elmo: I'm releasing the Warty kernel update now (16.8)
[01:18] <^_^> don't
[01:18] <^_^> you will brake some systems
[01:18] <^_^> just black mail ati and nvidia
[01:20] <daniels> ^_^: please take it somewhere else, ok?  if this channel gets cluttered with crap, we cannot do development.
[01:21] <Kamion> ^_^: huh?
[01:21] <^_^> daniels: umm all i see is parts and joins and ping requests
[01:21] <^_^> OMG
[01:21] <pitti> hey, it's Sunday after all
[01:21] <^_^> Kamion: you have way too much lag
[01:21] <Kamion> ^_^: there's a lot of interesting stuff from before you joined
[01:21] <cartman> sigh
[01:21] <Kamion> ^_^: dude, some of us do not pay attention to IRC 24/7
[01:22] <^_^> Kamion: you had to stop the intresting stuff :(
[01:22] <daniels> ^_^: there's every chance Kamion was actually doing interesting work instead of listening to this
[01:22] <^_^> ok here is my wishlist
[01:22] <daniels> ^_^: you are not here for development, and you are actively hindering it.  please be quiet, or leave.
[01:22] <^_^> 1) ignore me
[01:23] <^_^> 2) tell me to leave
[01:23] <^_^> thanks
[01:23] <azeem> hmm, and I was half-convinced ^_^ was daniels in disguise
[01:23] <azeem> 2004-08-#gnome-hackers.log:11:25 -!- daniels is now known as ^_^
[01:23] <azeem> there, evidence
[01:24] <cartman> hmmm
[01:24] <daniels> azeem: heh
[01:24] <^_^> LOL
[01:24] <^_^> wtf
[01:24] <azeem> 2004-08-#gnome-hackers.log:18:28 -!- ^_^ was kicked from #gnome-hackers by ^-^ [get your own damned name] 
[01:24] <azeem> ^_^: beware
[01:24] <^_^> oh ow
[01:24] <Kamion> hm, diffing loop-mounted ISOs takes a while
[01:25] <^_^> ok here is my wish list
[01:25] <^_^> ubuntu should start selling computer with ubuntu pre installed
[01:27] <Kamion> I doubt Canonical will be in that business in the near future, but we'd be happy to support other people doing that, and indeed there's some installer stuff in planning stages to help with OEM installations
[01:27] <^_^> yes and please GET THOSE NAKED PICTURES OFF UBUNTU
[01:28] <^_^> not all of us like the female body
[01:28] <^_^> or let alone the male
[01:28] <^_^> that one thing m$ did right no pictures of people on the desktop
[01:28] <Kamion> they're not the default, plus the artwork argument is *so* September
[01:29] <^_^> yes they should even be part of ubuntu
[01:30] <^_^> that is a reason why some of my friends didn't install the os because they saw they cd cover
[01:31] <daniels> we've already had that argument, please read the mailing list archives.
[01:35] <^_^> link
[01:36] <^_^> btw a lot of people like how not all apps look the same grey...
[01:36] <Kamion> look in the ubuntu-users archives for September and October and I find it hard to imagine that you could miss it
[01:36] <tuo2> gah
[01:36] <Kamion> the current situation was decided following a community meeting
[01:36] <^_^> look most media players are umm whats the word
[01:38] <martink> "ugly"? "in-consistent with your desktop"? "fruit salad"?
[01:39] <daniels> ^_^: this is not appropriate for #ubuntu-devel.  if you insist on continuing, please take it to #ubuntu.
[01:39] <^_^> daniels: ahh the people who would do anything would be in here 
[01:40] <^_^> and as you said you are not at irc 24x7
[01:40] <^_^> you'll miss of it
[01:40] <Kamion> no, that just means we have to read it in scrollback
[01:41] <^_^> you are telling me you are going to read 10k lines of stuff just to find 1 line that you need to read
[01:41] <Treenaks> ^_^: /help lastlog
[01:41] <daniels> i read scrollback of #ubuntu-devel, because it's not filled with morons trying to hijack it
[01:41] <daniels> right now, i'm trying to watch discussion of the live cd development, which is now very, very difficult.
[01:42] <^_^> ohh one question
[01:42] <Kamion> I read scrollback of #ubuntu-devel because I do this stuff for a living and I feel slightly obliged to pay attention to scrollbacck
[01:42] <pitti> nice Sunday everybody
[01:42] <^_^> why not let gnoppix be the live cd for ubuntu
[01:43] <^_^> it uses the ubuntu packages....
[01:43] <^_^> it is far better than the current cd
[01:43] <Kamion> Gnoppix *is* the Warty live CD
[01:43] <^_^> and it has an installer on the cd
[01:43] <Kamion> they're bit-for-bit identical
[01:43] <^_^> Kamion: not the one i got
[01:43] <^_^> the one i got doesn't have a installer on the cd
[01:43] <Kamion> perhaps you should look at http://www.gnoppix.org/ and read the stuff on the front page
[01:44] <daniels> ^_^: look, there is only one cd, and it's gnoppix.  and i still fail to see the relevance to ubuntu development.
[01:45] <^_^> well the screeny on that site show an insatller
[01:46] <Kamion> this is something we're working on for the next release, by means of making the live CD bootstrap process use the installer code
[01:46] <^_^> hey the gnoppix cd is 0.8warty
[01:46] <^_^> and in a screeny of 0.7
[01:46] <^_^> there is already a installer
[01:46] <^_^> as shown in this screeny
[01:47] <^_^> http://www.gnoppix.org/pages/screenshots/gnoppix07beta/gnoppix_installer.png
[01:47] <daniels> take it to #gnoppix
[01:47] <^_^> what happened to it
[01:47] <Kamion> no, we aren't going to use the gnoppix installer, we already have an installer and we don't want to have to support two installation systems
[01:48] <^_^> Kamion: ah ha so the gnoppix project is now your live cd project and you guy desided to dump the installer then
[01:48] <^_^> now it makes sence
[01:49] <^_^> btw #gnoppix has one person in there
[01:49] <Kamion> just because other places don't have enough targets for rants doesn't mean this is the right place
[01:50] <^_^> ok ok
[01:52] <^_^> one thing where do i rant about gnoppix's site saying its based on debian woody and not ubuntu? there is no one in #gnoppix to rant to, and as ubuntu took over the project i'll guess this would be the place
[01:53] <Kamion> perhaps you should try finding e-mail addresses or bug tracking systems rather than irc channels; bug reporting on irc is generally not so useful
[01:53] <^_^> ok
[01:54] <daniels> don't know what he's on, the gnoppix site has ubuntu all over it
[01:55] <Kamion> wow, my security-updated warty CD actually seems to be vaguely correct
[01:55] <daniels> heh, nice :)
[01:57] <Kamion> and I can basically build it with DEBUG=1 DIST=warty ARCHES=i386 cron.daily
[01:57] <daniels> rad
[02:12] <daniels> Kamion: what's the default debconf priority in casper; high?
[02:12] <Kamion> should be, yeah, everything defaults to high now
[02:14] <daniels> phat
[02:14] <daniels> we should be able to do live cd autoconfiguration when I fix this stupid stupid stupid bug
[02:15] <daniels> note to self: dpkg --print-architecture works without dpkg-dev; dpkg-architecture doesn't
[02:15] <Kamion> heh, yeah, dpkg-architecture's only intended for stuff in package builds IIRC
[02:16] <daniels> well, I thought, 'what gets the value of the architecture?', and I was in a debian/ dir at the time :P
[02:28] <daniels> mdz: ok, listen up, yo
[02:28] <daniels> export XORGFORCEPROBE=yes
[02:28] <daniels> md5sum /etc/X11/xorg.conf > /var/lib/xfree86/xorg.conf.md5sum
[02:28] <daniels> dpkg-reconfigure -phigh xserver-xorg
[02:28] <daniels> mdz: for the meantime, you'll need to have dpkg-dev installed, but I have the fix for that locally
[02:29] <daniels> mdz: that *should* do autodetection fine; the only real problem is prompting for the keyboard layout atm, and also lack of a fallback (probably 640x480?) if we can't pick the default resolution.  i'll do 640, but how do you want to handle the keyboard layout?
[02:31] <daniels> mdz: (by '*should*', I mean 'does on my freshly-installed hoary chroot, with dpkg-dev, laptop-detect and xresprobe installed')
[02:35] <daniels> daniels@nanasawa:/etc/X11% export XORGFORCEPROBE="yes"
[02:36] <daniels> daniels@nanasawa:/etc/X11% sudo sh -c 'echo > xorg.conf'
[02:36] <daniels> daniels@nanasawa:/etc/X11% sudo sh -c 'md5sum /etc/X11/xorg.conf > /var/lib/xfree86/xorg.conf.md5sum'
[02:36] <daniels> daniels@nanasawa:/etc/X11% sudo dpkg-reconfigure -phigh xserver-xorg
[02:36] <daniels> daniels@nanasawa:/etc/X11% wc -l xorg.conf
[02:36] <daniels> 126 xorg.conf
[02:36] <daniels> Kamion: what state does casper come in?  i assume unpacked and configured?
[02:37] <daniels> Kamion: or does it do this on the fly?
[02:39] <mjg59> daniels: Still around?
[02:40] <azeem> he's like flooding this chan currently
[02:40] <daniels> mjg59: represent
[02:41] <mjg59> daniels: VIA again
[02:41] <daniels> mjg59: ...
[02:41] <mjg59> On the craptop
[02:41] <mjg59> It doesn't switch the LCD off now, but it doesn't actually seem to start X
[02:41] <daniels> still arse with Xorg?
[02:41] <daniels> oh, hooray
[02:42] <mjg59> Hang on, I'll just try with vesafb disabled
[02:42] <daniels> want to bounce me Xorg.1.log from sudo Xorg :1 -logverbose 9999 -notvtswitch?
[02:42] <daniels> yeah, fb is generally badness
[02:42] <daniels> the via driver is probably flakey enough to not deal with coming from !text
[02:45] <mjg59> daniels: It just stops with Using VT number 8
[02:46] <mjg59> Ok, it's better without -novtswitch
[02:47] <mjg59> www.codon.org.uk/~mjg59/tmp/Xorg.1.log
[02:47] <mjg59> (II) VIA(0): Not using mode 1024x768 (no mode of this name)
[02:47] <mjg59> Hmm.
[02:49] <daniels> looks like it's failing in VIAGetBIOSTable
[02:50] <daniels> VGA BIOS image is wrong!! CheckSum = 5e403a
[02:50] <daniels> you'd probably need to poke libv about that one
[02:50] <mjg59> libv?
[02:51] <daniels> oh, arse.
[02:51] <daniels> luc varhaegen, one of the via guys
[02:51] <daniels> it looks like we can't get any sensible output without rebuilding the via driver
[02:53] <mjg59> Hahaha
[02:53] <daniels> http://amnesiac.heapspace.net/~daniel/via-drv.o
[02:53] <daniels> er, via_drv
[02:53] <daniels> that should have way more debugging funk
[02:53] <daniels> well, in a sec
[02:54] <mjg59> The one I just grabbed?
[02:54] <daniels> the incomplete one, yes :P
[02:54] <daniels> my link is flogged atm
[02:55] <daniels> ok, grab it now
[02:56] <daniels> if you run with -logverbose 9999, you should get some sweet debug love
[02:56] <mjg59> Ah, yes, that's bigger
[02:56] <daniels> a bit, yeah
[02:59] <mjg59> daniels: Ok, www.codon.org.uk/~mjg59/tmp/Xorg.1.log again
[03:00] <daniels> and it just hangs there?
[03:00] <daniels> or is the mode stuff broken?
[03:00] <mjg59> It just stops there
[03:01] <daniels> oh, haha
[03:01] <daniels> (II) VIA(0): Generic Monitor: Using default hsync range of 28.00-33.00 kHz
[03:01] <mjg59> Nothing on screen
[03:01] <daniels> (II) VIA(0): Generic Monitor: Using default vrefresh range of 43.00-72.00 Hz
[03:01] <mjg59> Ah. Yes, it's not going to get 1024x768 out of that, is it?
[03:01] <mjg59> 33kHz?
[03:01] <mjg59> Christ
[03:02] <daniels> try setting HorizSync 28-49, VertRefresh 43-72
[03:03] <daniels> via seems to be one of the drivers which will happily do everything for the modes except set up sensible sync ranges so the only modes it sets up and will accept can validate
[03:04] <mjg59>         Option          HorizSync     28-49
[03:04] <mjg59>         Option          VertRefresh   43-72
[03:05] <mjg59> In the monitor section?
[03:05] <daniels> (II) VIA(0): VIABIOS_GetVideoMemSize
[03:05] <daniels> (II) VIA(0): Return Value Is: -1073544704
[03:05] <daniels> very encouraging, that is
[03:05] <daniels> nope, just HorizSync\t28-49\nVertRefresh\t43-72
[03:05] <mjg59> Hm. It's still setting the default
[03:05] <daniels> in the Monitor section
[03:06] <mjg59> Oh, right
[03:06] <daniels> (i made the same mistake)
[03:06] <mjg59> Some day I'll understand the semantics of this
[03:07] <mjg59> Hrm. Still nothing. Hang on.
[03:07] <mjg59> www.codon.org.uk/~mjg59/tmp/Xorg.1.log
[03:11] <mjg59> Interesting. I can't find a single machine here that doesn't have a bad video BIOS checksum.
[03:11] <Treenaks> mjg59: try using Option          "NoDDCValue"    "on"
[03:12] <Treenaks> mjg59: the via driver kept choosing the lowest res. when I didn't have that.. somehow the (incorrectly read/parsed) values override the X config
[03:13] <mjg59> Treenaks: Is that in the monitor section?
[03:13] <Treenaks> mjg59: no, Device
[03:13] <daniels> hold on, what mode *is* it using?
[03:14] <daniels> i'm beginning to suspect I'll never have a hope of correctly configuring the via driver
[03:14] <daniels> via and apple need their hardware licences revoked
[03:15] <mjg59> Treenaks: Nope, stops in the same place
[03:16] <mjg59> I'm almost at the point of wondering whether there's a newer BIOS available for this thing
[03:16] <mjg59> The one in it is a Dixons special
[03:18] <ericf> why is it that a symlink like '/dev/dvd -> /dev/hdd' is gone after reboot? Maybe it's offtopic here, but in #ubuntu no-one could tell me...
[03:18] <Treenaks> ericf: because udev manages /dev in ubunti
[03:18] <Treenaks> ericf: and if you want that symlink, add it to /etc/udev/links.conf
[03:19] <ericf> Treenaks: Great! Thanks :)
[03:19] <daniels> mjg59: with a capital SPECIAL
[03:20] <kagou> hi
[03:20] <kagou> i'v found a bug on hoary 64 with openoffice 1.1.3
[03:21] <kagou> first launch -> crash ... i had to manually edit ~/.sversionrc and change 1.1.2 by 1.1.3. After openoffice is ok
[03:21] <kagou> hi seb128 
[03:21] <Nafallo> kagou: could you find it on bugzilla.ubuntulinux.org?
[03:22] <kagou> mmmm let'"s go
[03:24] <kagou> no Nafallo 
[03:25] <Nafallo> kagou: then you should file a bug against the package.
[03:25] <kagou> okay
[03:27] <mjg59> Damnit. I need to get DOS on here.
[03:28] <mjg59> ...and, of course, it has no floppy drive
[03:30] <Nafallo> mjg59: you might want one of those; ubcd.sf.net
[03:30] <Nafallo> :-)
[03:30] <mjg59>  14:30:24 up 376 days, 21:21,  4 users,  load average: 0.29, 0.08, 0.02
[03:30] <mjg59> Nafallo: Nah, I'll just netboot it
[03:31] <mjg59> Pff. That'd be 450 days if my housemate hadn't switched it off last christmas.
[03:31] <daniels> haggai: see kagou's comment above
[03:31] <mjg59> Best. Laptop. Ever.
[03:31] <daniels> mjg59: the craptop, or another laptop?
[03:31] <Nafallo> mjg59: hehe, nice one :-).
[03:32] <mjg59> Haha. No. That's an ancient P75 Toshiba
[03:32] <daniels> nice
[03:32] <daniels> mjg59: do you know anything about GTK's craptasmic input method, BTW?
[03:33] <mjg59> In what way?
[03:33] <daniels> well, I've been trying to hack it to generate a unicode heart on compose-<-3
[03:33] <daniels> took me 3 seconds with the X mappings, but adding it to gtk_compose_map in GTK didn't do any good
[03:34] <daniels> so it works with the X input method, but not with the default one
[03:34] <mjg59> Ah. No idea, I'm afraid.
[03:34] <daniels> ditto compose-i-j for the ij ligature
[03:34] <daniels> ah well
[03:36] <daniels> christ, the original imac advertises 640x480@112Hz
[03:36] <daniels> that's psychopathic
[03:43] <mjg59> I love the feeling of flashing hardware with (theoretically) unsupported BIOSes
[03:43] <daniels> even if it doesn't work, it's an improvement
[03:44] <mjg59> Haha
[03:44] <mjg59> memdisk is the best thing ever
[03:44] <daniels> then again, setting it on fire, or using it to beat an old woman to death on the street, would be an improvement
[03:44] <elmo> seb128: ?
[03:44] <elmo> or any mono guys?
[03:44] <seb128> elmo: I'm not a mono guy
[03:45] <mjg59>                 UUID: 22222222-2222-2222-2222-222222222222
[03:45] <daniels> elmo: is this about mcs/m-a-b?
[03:45] <mjg59> Best. UUID. Ever.
[03:45] <daniels> mjg59: word
[03:45] <seb128> elmo: but since you are here, sync for python-gtk2-tutorial python-gtk2-doc please :)
[03:45] <mjg59> Oh, WOW
[03:45] <mjg59> The new BIOS makes the fan EVEN LOUDER
[03:46] <mjg59> When it's idling
[03:46] <elmo> seb128: no, but you might know if there's any reason gtk-sharp would be using an old copy of gtkhtml, gal and soup?
[03:47] <mjg59> daniels: Ok, X is still just as fucked
[03:47] <elmo> why the heck is python-gtk2-doc in multiverse?
[03:48] <daniels> mjg59: awesome
[03:48] <daniels> mjg59: i think this might be a courier-it-to-the-via-hackers thing
[03:48] <daniels> mjg59: (he's in belgium, iirc, so maybe you could just ride your bike over)
[03:48] <mjg59> daniels: It's weird, though. At least one of the via hackers had dealt with the G320.
[03:48] <daniels> mjg59: i'm not entirely sure libv had dealt with one, as much as heard a report from someone using one once
[03:48] <mjg59> And it used to work better than this (as in, it started X fully and then turned the LCD off...)
[03:48] <daniels> heh
[03:49] <mjg59> Haha
[03:49] <daniels> i dunno, i assume the horizsync/vertrefresh lines are the same as from xfree86
[03:49] <daniels> if not, you might want to steal the xfree86 ones
[03:49] <mjg59> Oh, it never worked at all with XFree
[03:49] <daniels> oh, sweet
[03:49] <mjg59> Remember? It hung the machine
[03:49] <mjg59> This was an earlier Xorg package
[03:49] <daniels> oh yeah
[03:50] <daniels> looks like the unichrome merge broke it then, maybe
[03:50] <daniels> you'll have to poke libv, I'm afraid (he hangs out in #freedesktop)
[03:50] <daniels> my downstream's hammered, so I'm going downstairs to get some more food now
[03:51] <seb128> elmo: I've no idea, python-gtk2-doc should be in main
[03:51] <seb128> elmo: and dunno for gtk-sharp
[03:52] <elmo> seb128: well it's not seeded - could you get it seeded?  in the meantime, I'll move it to universe
[03:52] <seb128> elmo: ok, thanks
[03:52] <elmo> and both syncs' done
[03:53] <mjg59> Bah. Upgrading the BIOS hasn't fixed ACPI.
[03:53] <seb128> thanks :)
[04:08] <Kamion> daniels: casper's just the piece that finds the live filesystem, munges it a bit, and pivots into it; you mean the live filesystem?
[04:09] <Kamion> daniels: AFAIK it comes configured
[04:14] <lamont> Kamion: grumble the non-symlinks are fine...
[04:14] <lamont> fixing
[04:16] <fabbione> huffff
[04:16] <fabbione> another room has been sand papered :-)
[04:23] <daniels> Kamion: ok, that's alright.
[04:24] <daniels> mdz: so if you ship an /etc/X11/xorg.conf with its md5sum already in /var/lib/xfree86/xorg.conf.md5sum, then all you need to do is call dpkg-reconfigure -phigh xserver-xorg, with XORGFORCEPROBE=yes
[04:25] <daniels> mdz: (with dpkg-deb installed)
[04:25] <daniels> er, dpkg-dev
[04:34] <Kamion> lamont: thanks, rebuilding now
[04:35] <Nafallo> is there anyway to but a cd from a grub-shell?
[04:35] <Nafallo> (i.e. ubuntu livecd)
[04:35] <lamont> Kamion: still dealing with it - not very apache2 literate.
[04:36] <lamont> thom around?
[04:36] <lamont> or anyone else wanna give me the snippet?
[04:36] <daniels> lamont: it's probably on AllowSymlinksIfOwnerMatch, which does what it says on the box
[04:36] <fabbione> lamont: what are you searching for?
[04:36] <daniels> lamont: if you want to unconditionally walk symlinks, add Option FollowSymlinks to the Directory or Location stanza
[04:36] <lamont> ah, that could be it...
[04:36] <daniels> (if you don't have a Location stanza, make one)
[04:37] <daniels> (i think it's SymlinksIfOwnerMatch, rather than Allow..., but meh)
[04:37] <Kamion> lamont: oh, ok, I'm off for a bit now so let me know when it's done
[04:37] <gsuveg> thom: are you here ?
[04:37] <lamont> yeah - is fixed on terranova now.
[04:37] <daniels> gsuveg: it's Sunday afternoon in the UK
[04:38] <gsuveg> daniels: like here. thanks
[04:38] <gsuveg> daniels: np. email sent
[04:38] <gsuveg> i've problem with NM on warty
[04:38] <daniels> networkmanager is known to be rather broken
[04:38] <daniels> the suggested thing to do is to use netapplet, which is far more likely to actually work
[04:39] <gsuveg> i see
[04:42] <lamont> Kamion: is better now.
[04:42] <lamont> back in 4-5 hours or more
[05:00] <daniels> woah, they've done the i830 lvds stuff *right*.
[05:02] <Treenaks> daniels: who have?
[05:03] <fabbione> God, Jesus and the Holy spirit
[05:03] <fabbione> nobody else could have done it otherwise
[05:03] <daniels> Treenaks: tungsten graphics
[05:03] <daniels> they're setting sensible sync ranges per default
[05:04] <daniels>    /* With panels, we're always assuming a refresh of 60Hz */
[05:04] <daniels>    pScrn->monitor->nHsync = 1;
[05:04] <daniels>    pScrn->monitor->nVrefresh = 1;
[05:04] <daniels>    /* With panels, we're always assuming a refresh of 60Hz */
[05:04] <daniels>    pScrn->monitor->nHsync = 1;
[05:04] <daniels>    pScrn->monitor->nVrefresh = 1;
[05:04] <daniels> erk, stupid mouse -- sorry
[05:04] <fabbione> hey
[05:04] <fabbione> they have copied our setup!
[05:04] <fabbione> 60Hz was OUR idea!
[05:05] <Treenaks> fabbione: sue them!
[05:05] <daniels> heh :)
[05:05] <daniels> i'd rather buy them a beer than sue them
[05:05] <Treenaks> "60Hz should be enough for everyone"
[05:05] <daniels> this gets rid fo the need for 855resolution/865patch and all that crap, as far as I can tell
[05:05] <fabbione> we can buy many beers :-)
[05:33] <crimsun> gah, missed Martin. I need to ask him if he's going to provide linux-headers-2.6.10-hardened*
[05:35] <daniels> Kamion: ping
[05:39] <fabbione> crimsun: i am pretty sure he will once the packages will ake shape
[05:39] <fabbione> take
[05:39] <mako> jdub: that's awesome
[05:40] <mjg59> daniels: So, dude, are we getting i830 love today?
[05:40] <daniels> mjg59: it's 3:40am
[05:40] <crimsun> fabbione: yeah, I'm hoping so. I'm attempting to test a newer ALSA [alsa-driver 1.0.8rc2, with fixes for a nasty OOPS in snd-pcm-oss] 
[05:40] <daniels> mjg59: by the time I do test builds around all three architectures (i386 is frigging *slow), I'll want to be asleep
[05:40] <mjg59> You suck
[05:40] <mjg59> GIVE ME THE LOVE
[05:41] <Keybuk> I'll give you more loving than you can handle, bitch! :p
[05:41] <fabbione> crimsun: you can do that in the normal kernel :-)
[05:42] <crimsun> fabbione: I'm trying to stay with -hardened :-)
[05:49] <daniels> mjg59: 
[05:49] <daniels> daniels@catsby:~/x/xorg/bleh/xc% wc -l yay-tungsten.diff
[05:49] <daniels> 6219 yay-tungsten.diff
[05:50] <mjg59> Yay tungsten
[05:50] <Keybuk> +# Copyright (C) 2002, 2003, 2005  Free Software Foundation, Inc.
[05:50] <Keybuk> +#
[05:50] <Keybuk> +# This file is free software; the Free Software Foundation
[05:50] <Keybuk> +# gives unlimited permission to copy and/or distribute it,
[05:50] <Keybuk> +# with or without modifications, as long as this notice is preserved.
[05:51] <Keybuk> fear.  RMS approved too
[05:51] <mjt> Keybuk: pls excuse me for a silly question... what's the reason for grepmap? ;)  I mean, why another piece of software is "better" than 3-line patch for modprobe (in m.i.t.)?
[05:52] <Keybuk> mjt: modprobe doesn't do the same thing
[05:52] <fabbione> fabbione@gordian:~$ glxgears 
[05:52] <fabbione> 37840 frames in 5.0 seconds = 7568.000 FPS
[05:52] <fabbione> 46718 frames in 5.0 seconds = 9343.600 FPS
[05:52] <fabbione> THIS IS SO RAAAAD!!!!
[05:52] <fabbione> daniels: ubuntu9 + 2.6.10 + l-r-m
[05:52] <mjt> Keybuk: think /lib/modules/$version/modules.alias 
[05:52] <fabbione> see above :-)
[05:53] <crimsun> gordian, eh? an allusion to gordian knot, praps?
[05:53] <Keybuk> mjt: as far as I remember, those don't work with device drivers that just say "anything with this set of flags"
[05:53] <fabbione> crimsun: all my boxes are named after jap's robot
[05:53] <Keybuk> grepmap is basically just an implementation of the hotplug shell code in C
[05:53] <Keybuk> so it's drop-in compatible
[05:53] <fabbione> like gundam, trider, vultus5
[05:54] <crimsun> fabbione: ah, ok. I mistook that for the prog metal-jazz band, Gordian Knot
[05:54] <mjt> Keybuk: "anything with this set of flags" -- such as?
[05:54] <fabbione> crimsun: nope.. you are not the first one :-)
[05:55] <daniels> fabbione: GLXGEARS IS NOT A BENCHMARK
[05:55] <fabbione> daniels: AHAHHAHA
[05:55] <daniels> key	nice
[05:55] <daniels> Keybuk: ^ that's you
[05:55] <fabbione> daniels: anyway.. it GLX works only on one of the 2 heads.. but that's nothing new
[05:55] <Keybuk> mjt: things like the mousedev driver need to be loaded for anything with a given USB class, etc.
[05:55] <daniels> fabbione: yeah
[05:55] <daniels> fabbione: get a Radeon, they have MergedFB support :)
[05:55] <daniels> fabbione: overlays on both heads
[05:56] <fabbione> daniels: that's why i get 9K FPS on it... and 300 FPS on the other ;)
[05:56] <fabbione> daniels: nah.. i don't care until xv works on both
[05:56] <mjt> Keybuk: that mousedev example works just fine with modules.alias
[05:56] <daniels> fabbione: doesn't TwinView do that?  IIRC that was the main revision for TV over Xinerama
[05:56] <fabbione> i can watch 8 differnt pron movies full screen...
[05:56] <mjg59> daniels: Make the Tungsten people do mergedfb support
[05:56] <fabbione> that's enogh
[05:56] <daniels> fabbione: Xv works on both with Radeon -- all overlays do
[05:56] <daniels> mjg59: for i8xx?
[05:56] <mjg59> Yes
[05:56] <daniels> mmm, that would be cool
[05:56] <fabbione> daniels: dunno... i think it does.. but hounetly i don't really care
[05:57] <fabbione> daniels: i only need headS
[05:57] <Keybuk> mjt: hotplug upstream didn't seem to think it was sufficient (otherwise hotplug would use it itself)
[05:57] <fabbione> time to prepare some dinner.....
[05:57] <fabbione> later folks
[05:57] <Keybuk> in theory it's just a temporary solution, and will go away eventually
[05:58] <Keybuk> descent 2.6.8.1-4-686% grep mousedev modules.alias
[05:58] <Keybuk> zsh: exit 1     grep mousedev modules.alias
[05:58] <mjt> hmm... hotplug code evolved since 2.4 at least
[05:58] <daniels> fabbione: seeya
[05:59] <mjt> Keybuk: do you really have that module? ;) Isn't it hiddev?
[05:59] <Keybuk> mjt: yes, everyone really has that module
[05:59] <Keybuk> descent 2.6.8.1-4-686% grep tsdev modules.alias
[05:59] <Keybuk> zsh: exit 1     grep tsdev modules.alias
[05:59] <Keybuk> descent 2.6.8.1-4-686% grep joydev modules.alias
[05:59] <Keybuk> zsh: exit 1     grep joydev modules.alias
[05:59] <Keybuk> modules.alias is missing quite a few things
[05:59] <mjt> it's from inputmap, right? 
[06:00] <Keybuk> yes
[06:00] <mjt> i mean, all those modules are listed in modules.inputmap
[06:00] <mjt> oh
[06:00] <mjt> that's the only map which isn't in modules.alias ;)
[06:01] <Keybuk> it was easier to clone the hotplug code than futz around with stuff that's not yet finished :)
[06:02] <mjt> but grepmap is so large tarball!.. ;)
[06:02] <Keybuk> it's not that big ?
[06:02] <mjt> 280Kb for 0.1.0 ;)
[06:03] <Keybuk> heh, that's not big :p
[06:03] <mjt> the whole module-init-tools is 122Kb ;)
[06:03] <mjg59> mjt: Most of that will be build infrastructure...
[06:04] <Keybuk> 104     po
[06:04] <Keybuk> 108     tests
[06:04] <Keybuk> 176     m4
[06:04] <Keybuk> 368     configure
[06:04] <Keybuk> 480     intl
[06:06] <mjt> well.. just kidding, really
[06:07] <Keybuk> mjt: (digs out his e-mail) and modules.alias doesn't yet correctly do usb versions
[06:07] <Keybuk> given a version in usbmap, you have to match that or later within certain bounds
[06:07] <Keybuk> mit only does an exact match at the moment
[06:08] <Keybuk> it also only does precise pci class matches, rather than free ones (but I don't think there's a driver that relies on that yet)
[06:08] <mjt> i'm trying to "get in touch" with someone who "can get the end to all this nonsense" ;) for quite some time -- there's hotplug and udev, m.i.t, kernel module stuff (file2alias etc, the same is repeated in mit), everyone complains hotplug is too slow, and every distro solves the problem it's own way
[06:12] <Keybuk> the end theory is that you'd have udev handling kobject events from the kernel, preparing the device node if necessary and running a set of handlers when they're ready to be dispatched.  One of these handlers would be to check the output of m-i-t and load any appropriate module (current hotplug), others would include hal and so on.
[06:12] <mjt> i just wanted a small tool to do module loading in initrd, somehing akin a tiny scan_pci tool used with ltsp (i think) - which had its own pci->module file.  And after digging into that stuff, i see there's only very small stuff needs to be done to solve it all (it seems -- your comments above makes me think again).. before it's too late and everyone does it their own incompatible way... ;)
[06:12] <Keybuk> on boot, you still need to coldplug the modules (scan the bus and load them)
[06:13] <mjt> yes sure
[06:13] <Keybuk> udevstart seems a reasonably logical place to do that
[06:14] <mjt> i wrote a small shell script to do that -- based on /sys/bus/pci/devices and modules.alias (aliases works faster in shell than *.*map files)
[06:14] <mjt> but that's yet another way to do it ;)
[06:14] <Keybuk> that's basically /etc/hotplug/pci.rc ? :)
[06:15] <mjt> yes but that works in initrd (small etc)
[06:15] <Keybuk> it depends on the discretion of the problem you want to solve
[06:15] <Keybuk> 1) dynamic /dev, or static /dev
[06:15] <Keybuk> 2) built-in modules, or dynamic modules
[06:15] <Keybuk> 3) no hardware reporting, hardware reporting
[06:16] <mjt> and i just changed my pci.rc to call that same my script too (it completes in 3 sec instead of almost 20 sec in sid)
[06:16] <Keybuk> your script, for example, wouldn't be able to inform hal of the new hardware -- so hal would have to do /yet another/ cold plugging scan when it starts
[06:16] <mjt> yes sure
[06:16] <Keybuk> an ideal goal is to reduce the number of hardware scans on boot to just one
[06:16] <mjt> isn't it impossible when using initrd?
[06:18] <Keybuk> not necessarily, you just start udevd and hald earlier
[06:18] <Keybuk> before init :p
[06:18] <mjt> so initrd becomes persistent?
[06:18] <Keybuk> (which actually makes some sense, having hardware managing code running before any userland)
[06:18] <Keybuk> well, you're already mounting /proc and /sys in initrd at that point
[06:19] <Keybuk> it's a bit blue-sky at the moment, when kobject ide drivers come along, it'll be a very sexy thing to have
[06:19] <Keybuk> so rather than the 100-odd lines of modprobe that's statically generated, you'd just start the hardware in the system] 
[06:20] <mjt> "it'll be .." what's "it" ?'
[06:20] <Keybuk> udev/hotplug-in-initrd
[06:20] <Keybuk> right now there's no reason to do it
[06:21] <Keybuk> but if things normally loaded in initrd are kobjected, then there becomes a very compelling reason to do it
[06:22] <mjt> hmm
[06:22] <Keybuk> if you can "hotplug" your 
[06:23] <Keybuk> ide/scsi drivers and filesystem drivers, then it makes sense to do that before you try and mount the root partition
[06:23] <Keybuk> at that point, you may as well hotplug everything else as well
[06:23] <mjt> udevD isn't really needed (not even in real root), i think -- provided one does not insert/remove hardware 1000 times a sec (as /dev is in tmpfs and not in udevd memory)
[06:23] <Keybuk> udevd is very needed
[06:24] <Keybuk> e.g. usb storage remove events take about a tenth of the time of an insert event
[06:24] <Keybuk> so you can plug in and out a device reasonably quickly (within a second or two) and have the remove event finish before the start event has even finished
[06:24] <Keybuk> it's job is to correctly serialise events
[06:25] <mjt> well yes
[06:25] <mjt> just too much stuff for initrd... ;)
[06:25] <Keybuk> yup
[06:26] <mjt> i liked the days when a single floppy with kernel+initrd image was sufficient to boot any of 500+ our different machines... ;)
[06:26] <Keybuk> if you hotplug in initrd, you're going to be mounting every storage device on the system at the same time as the root
[06:27] <Keybuk> I've often wondered how MS deal with scsi drives
[06:28] <Keybuk> one idea is that you do away with initrd almost entirely; and in userland do hotplugging and detect things like scsi drives and filesystem types
[06:28] <Keybuk> this writes a configuration onto the minimal initrd to just load those modules
[06:28] <Keybuk> though this doesn't work if you swap hardware
[06:29] <mjt> heh.  grub comes to mind, with its ext2fs etc implementation... ;)
[06:30] <Keybuk> indeed, if grub could actually prepare an initrd for the kernel ...
[06:30] <mjt> in solaris there's 2nd stage bootloader which loads 16-bit drivers for hardware
[06:30] <Keybuk> so the kernel would load the right single ide/scsi and filesystem module it needed
[06:30] <Keybuk> and then the first thing you do in userspace is hotplug the rest of the hardware
[06:33] <Keybuk> (ignoring, for now, strange people who want separate or even nfs-mounted /usr)
[06:33] <mjt> oh well.
[06:36] <Keybuk> that's a fun one
[06:36] <Keybuk> if you want hotplug to use stuff in /usr, you need to do it after remote filesystems are mounted
[06:36] <Keybuk> but you can't access remote filesystems until you've hotplugged the network device
[06:36] <Keybuk> and the network device might depend on firmware in /usr
[06:36] <Keybuk> *hello!*
[06:45] <sivang> Keybuk: hehe
[06:45] <Keybuk> as time goes on, I'm more and more heading towards a "well, don't configure your machine like that" stance on some things :)
[06:45] <mjt> in that case the initrd (pxe, whatever) should contain the firmware
[06:46] <mjt> mkinitrd job -- yet another piece of software that um.. does "offline hot^Wcouldplug"
[06:47] <mjt> i already got some probs with tg3 and network (pxe) booting. but for now it works, ie the firmware, while good to have, is optional
[06:49] <Keybuk> my tg3 seems ok, I've not yet dropped the firmware into place
[06:50] <sivang> does anybody know what happens privilege wise, if I have one prog that is executed using sudo, which forks/executes another program ? Does the "child" program  also get the sudo'd privileges?
[06:50] <Keybuk> sivang: yes
[06:50] <Keybuk> (this is not generally true of setuid programs though)
[06:51] <sivang> Keybuk: so sudo makes an excepction and assigns explicitly the child's as root privs?
[06:51] <Keybuk> no
[06:52] <sivang> Keybuk: k, thanks
[06:52] <Keybuk> do you know the difference between real, effective and saved user/group id?
[06:52] <sivang> Keybuk: hmmm, /me goes on googeling :)
[06:52] <mjt> heh. that's quite a story...
[06:52] <Keybuk> ok, well, simply put
[06:53] <Keybuk> processes have both a real and effective user id
[06:53] <Keybuk> generally they are the same
[06:53] <mjt> (all that different uids, that is -- historic mess)
[06:53] <Keybuk> processes are entitled to swap them as well
[06:53] <Keybuk> or make them both the same
[06:53] <mjt> fsvp 
[06:53] <mjt> er. fsvo "both" ;)
[06:53] <Keybuk> mjt: shush, I'm simplifying :p
[06:54] <sivang> mjt: what's fsvp if I may ask? ;)
[06:54] <Keybuk> now, when you fork a child process, that child process is created under your real user id
[06:54] <Keybuk> so if your real user id is 1000, your child will be also 1000 for both real and effective
[06:54] <sivang> Keybuk: but I can reser it?
[06:54] <mjt> sivang: a typo: fsvo it should be, or "for some value of"
[06:54] <Keybuk> your effective user id is used whenever you do anything on disk
[06:54] <Keybuk> so if your effective user id is 0, you can roughly behave as root
[06:55] <Keybuk> a setuid program (o+s) is run with its real user id unchanged, but it's effective user id changed to 0
[06:55] <mjt> s/o+s/u+s/
[06:55] <Keybuk> so if you run a setuid program as you (1000), it's real user id is still 1000 and it's effective user id is 0
[06:55] <Keybuk> mjt: gah, typo, thanks
[06:56] <Keybuk> if that setuid program was to fork a child, the child process would run as you
[06:56] <Keybuk> so setuid programs that don't know any better can't fork root children
[06:56] <sivang> Keybuk: ok, that explains alot! Thank you!
[06:56] <Keybuk> sudo knows better, it knows it's setuid so one of the things it does after it's checked your credentials is sets its real user id to its effective user id (so both become 0)
[06:57] <sivang> Keybuk: uh ha. got it now.
[06:57] <Keybuk> setuid gives you a helping hand to root, you still need to pull yourself up
[06:57] <sivang> Keybuk: as in having the child process set effetive uid to 0
[06:57] <Keybuk> it also means (for example) that a setuid program can do something quickly as root, and then drop its root privileges and run as just the user again
[06:58] <sivang> Keybuk: ok, but it has to be setuid prior to runnig right? it can't ask to change it i runtime?
[06:58] <Keybuk> yup
[06:58] <Keybuk> you can only set your real, effective or saved user id to the value of one of the others
[06:58] <Keybuk> you can't stick anything arbitrary in there
[06:58] <mjt> unles you're God^Wroot ;)
[06:59] <Keybuk> and once you've dropped root, (so real, effective and saved are 1000) you can no longer switch back to it
[06:59] <Keybuk> or if you never had root in the first place, you can't stick 0 in there
[06:59] <daniels> Keybuk: so how come you can only drop your uids to 1000? that sounds like a poor limitation to me
[06:59] <daniels> unix sucks
[06:59] <Keybuk> daniels: well, if you're root, you can change your uid to whatever you like <g>
[07:00] <Keybuk> but you could have a "setuid daniels" process, which when I run I can flip between being daniels or scott until I make a decision
[07:00] <sivang> Keybuk: so a nont setuid proc could never gain root privs, unless executed by root, right?
[07:00] <daniels> Keybuk: yhbt yhl hth hand kthxbye
[07:00] <Keybuk> children would be mine by default, unless I became you in which case they'd be yours
[07:01] <crimsun> sounds like a handy avoid-blame resolution
[07:01] <Keybuk> sivang: kernel bugs not withstanding, yes
[07:01] <sivang> crimsun: hehe
[07:01] <mjt> nah, it's like "don't let everyone do evil things with the system"
[07:03] <Nafallo> then sudo -u other_user will gain root to become other_user, right?
[07:04] <mjt> sudo is setuid *root*
[07:04] <Keybuk> Nafallo: yup.  sudo is setuid root, checks creds, becomes root, becomes over user, runs process
[07:05] <Nafallo> good to have that one confirmed :-).
[07:13] <crimsun> mdz: looks like more soname fun with liboggflac* ? [from Flac-dev] 
[07:21] <fabbione> elmo: i think that apt-ftparchive is confused about update-notifier
[07:21] <fabbione> elmo: there are sources in both main and universe
[07:21] <fabbione> elmo: and binaries only in universe
[07:22] <fabbione> while the packages file for main reports the binary there
[07:31] <doko> elmo: please sync graphviz and drdsl from non-free to multiverse
[07:32] <seb128> elmo: have you synced glib2.0 2.6.1-1 and easytag 1.99.2-2 ?
[07:34] <mdz> gah, my mail is bouncing
[07:44] <mdz> daniels: we need a better interface to tell it "just clobber the config" than writing the md5sum to that file
[07:44] <Kamion> daniels: pong?
[07:44] <daniels> mdz: what, such as CLOBBERXORGCONFIG=yes?
[07:45] <mdz> fabbione: main should have update-notifier, universe upgrade-notifier
[07:45] <mdz> fabbione: upgrade-notifier can be removed
[07:45] <daniels> Kamion: could you please clag me your Option"Xkb(...)" settings?
[07:45] <daniels> Kamion: (from the powerbook)
[07:45] <mdz> daniels: something like that, perhaps
[07:45] <mdz> daniels: or
[07:45] <mdz> daniels: if the config is missing entirely, always create a new one
[07:45] <fabbione> mdz: well something is wrong
[07:46] <Kamion>         Option          "XkbRules"      "xfree86"
[07:46] <Kamion>         Option          "XkbModel"      "macintosh"
[07:46] <Kamion>         Option          "XkbLayout"     "gb"
[07:46] <fabbione> update-notifier is in both universe and main
[07:46] <mdz> I don't see it in universe, only main
[07:46] <Kamion> mdz: new daily-live build up now
[07:46] <Kamion> publishing actually worked automatically \o/
[07:46] <mdz> Kamion: how many architectures?
[07:47] <Kamion> just i386 for now, amd64 and ia64 didn't have fs builds last time I checked and I haven't fixed the powerpc yaboot.conf yet
[07:47] <fabbione> mdz: it's in both...
[07:47] <fabbione> http://archive.ubuntu.com/ubuntu/pool/universe/u/
[07:47] <daniels> Kamion: thanks
[07:47] <fabbione> either my mirror was mirroring during the sync
[07:48] <daniels> mdz: mmm.  presumably branden had a reason for his paranoia and I'd like not to differ too much from Debian in that regard if possible, since it's already obtuse and a complete shit to deal with as it stands.
[07:48] <Kamion> (powerpc is relatively straightforward to fix, but I'm concentrating on this updated warty CD for Linux+ at the moment)
[07:48] <mdz> daniels: that is not necessarily a safe assumption
[07:48] <daniels> mdz: point
[07:48] <daniels> Keybuk: what does dpkg do in this case?
[07:48] <daniels> it respects removed conffiles, no?
[07:49] <mdz> yes
[07:49] <daniels> right ...
[07:49] <mdz> which is insane and nothing deals with it properly and users get horribly confused
[07:50] <daniels> let me have a think about it later (i'd rather just do grunt work right now); itmt, that recipe should do it
[07:50] <mdz> Kamion: what do you think about udevsend
[07:50] <mdz> Kamion: if you mailed me, please hang onto it for a bit until my mail stops permanently bouncing
[07:51] <Keybuk> mdz: for the Debian definition of config file, it makes sense
[07:51] <mdz> daniels: just tried your method; it gives me the video modes question
[07:51] <mdz> daniels: standard Ubuntu install on the same machine asks no questions
[07:52] <Kamion> mdz: sounds sane but I have to reserve judgement until I've tried it out :-)
[07:52] <daniels> mdz: please run with DEBUG_XFREE86_PACKAGE=yes DEBUG_XRESPROBE=yes
[07:52] <mdz> daniels: believe it or not, I tried this already
[07:52] <Kamion> which I'll do
[07:52] <daniels> mdz: (are you sure you have xresprobe and laptop-detect installed?)
[07:52] <daniels> mdz: i got the same until I installed those two
[07:52] <mdz> lamont: ?
[07:54] <mdz> fabbione: http://archive.ubuntu.com/ubuntu/pool/universe/u/update-notifier/
[07:54] <mdz> fabbione: it is an empty directory
[07:54] <mdz> Kamion: does it also include the increased ramdisk_size?
[07:54] <fabbione> mdz: i think my mirror was mirroring during a sync...
[07:55] <fabbione> and got everything mixed
[07:58] <fabbione> Kamion: i started mirroring the weekly-dvd locally...
[07:58] <fabbione> i will let you know how it goes :-)
[07:58] <mdz> daniels: laptop-detect is present, but xresprobe is missing
[07:58] <Kamion> mdz: oh, no, not yet, sorry, this has been a very busy weekend for me and I've only just been keeping up with you lot :-)
[07:58] <mdz> Kamion: no worries
[07:58] <Kamion> mdz: remind me what you wanted it bumped up to?
[07:59] <Kamion> 65536 or something?
[07:59] <mdz> Kamion: 65536 would be good for starters
[07:59] <daniels> mdz: righto
[07:59] <mdz> something which runs during dpkg-reconfigure xserver-xorg is calling dpkg-architecture
[07:59] <mdz> when it wants to be calling dpkg --print-architecture
[07:59] <Kamion> mdz: done for the next build
[07:59] <Kamion> hm, let's kick off another build now
[07:59] <daniels> mdz: yeah, already fixed
[07:59] <mdz> ok
[07:59] <daniels> mdz: hence the note about needing dpkg-dev installed
[08:00] <mdz> Kamion: are lamont's blobs showing up daily now? or will this be the one from friday?
[08:00] <Kamion> mdz: they seem to be daily
[08:00] <Kamion> [   ]  livecd-current.cloop    09-Jan-2005 06:38  501M
[08:00] <Kamion> new image up
[08:00] <mdz> cool
[08:01] <Kamion> Purging daily-live images older than 4 days ...
[08:01] <Kamion> this is just LOVE
[08:01] <mdz> :-)
[08:02] <fabbione> mdz: would it be possible to ship a non compressed .iso for the live? together with a tool that will compress?
[08:02] <fabbione> that might increase of a few N the rsyncability of it
[08:02] <mdz> daniels: dpkg-dev is installed, in fact (though I'm not sure why), but of course gcc isn't
[08:02] <mdz> ah, for alien, via lsb
[08:03] <mdz> fabbione: we could distribute an uncompressed filesystem image, I suppose
[08:03] <mdz> fabbione: I was thinking we could make the compressed image rsyncable if we really wanted to
[08:04] <fabbione> mdz: something that we can just toolfoo /path/to/uncompressed | cdrecord or something
[08:04] <fabbione> mdz: how?
[08:05] <mdz> fabbione: gzip --rsyncable
[08:05] <fabbione> is there any negative impact on that?
[08:06] <mdz> slightly less compression
[08:06] <mdz> a few %
[08:06] <mdz> we would turn it off for the final, presumably
[08:06] <fabbione> that can be quite a lot...
[08:07] <mdz> fabbione: 1% ~= 5M
[08:08] <daniels> great, more lib64 braindamage
[08:08] <daniels> that means bedtime
[08:08] <mdz> fabbione: in exchange for rsyncability, it is a good tradeoff
[08:08] <mdz> daniels: No devices detected
[08:08] <daniels> mdz: xorg.conf and Xorg.0.log please
[08:08] <mdz> RADEON: No matching Device section for instance (BusID PCI:1:0:0) found
[08:09] <mdz> sent
[08:09] <daniels> thanks
[08:09] <mdz> don't bother replying to the mail; my mail is fucked
[08:09] <daniels> if you could clag me lspci in /msg, that would be handy also
[08:11] <mdz> daniels: it looks like the PCI busID of the old config remained
[08:12] <mdz> I desperately need for it to _ignore_ everything that came before
[08:12] <mdz> and create a completely new config based only on probes
[08:12] <mdz> users need this too
[08:12] <daniels> yeah
[08:14] <Kamion> guh, please stop changing evms soname, kthxbye
[08:15] <mdz> Kamion: it changed once in September, and once in January
[08:16] <Kamion>   * [hoary]  libevms-2.4 -> libevms-2.5.
[08:16] <Kamion>  -- Colin Watson <cjwatson@canonical.com>  Sun,  9 Jan 2005 19:10:08 +0000
[08:16] <Kamion>   * [hoary]  Add libelfg0; libevms-2.3 -> libevms-2.4; libgcrypt7 ->
[08:16] <Kamion>  -- Colin Watson <cjwatson@canonical.com>  Tue,  2 Nov 2004 23:47:39 +0000
[08:16] <mdz> evms (2.4.0-1) unstable; urgency=low
[08:16] <mdz>   * New upstream release
[08:16] <mdz>  -- Matt Zimmerman <mdz@debian.org>  Fri, 17 Sep 2004 17:55:21 -0700
[08:17] <mdz> oh, that was the warty->hoary dealy
[08:17] <mdz> delay
[08:17] <Kamion> although admittedly the latter was the first relevant debootstrap change after the great hoary merge
[08:17] <daniels> fwiw, debootstrapping hoary seems broken, unless my laptop is out of date
[08:17] <mdz> daniels: see above
[08:17] <daniels> doesn't have ubuntu-keyring and ethtool
[08:18] <Kamion> your laptop's out of date
[08:18] <daniels> phat
[08:18] <daniels> alright, important appointment with a pillow
[08:18] <Kamion> you must have debootstrap << 0.2.45ubuntu16; current is 0.2.45ubuntu19
[08:18] <Kamion> (well, counting the upload of two minutes ago)
[08:19] <mdz> Kamion: is there any straightforward way to convince a CD-booted installer to act like a netboot one, and download udebs from the archive?
[08:19] <mdz> Kamion: that'd be incredibly handy for live CD testing
[08:19] <Kamion> not that I can think of
[08:20] <Kamion> well, not since we mutilated the installer in the name of fewer questions :-P
[08:20] <Kamion> hm, you need net-retriever on the CD for that
[08:21] <Kamion> but you also need to include download-installer in the initrd rather than load-cdrom, and I suspect that that would break things
[08:22] <Kamion> mdz: just use the netboot initrd, which also has casper-check in it
[08:22] <Kamion> there's a netboot/mini.iso somewhere if actual netbooting is inconvenient for you
[08:23] <Kamion> http://archive.ubuntu.com/ubuntu/dists/hoary/main/installer-i386/current/images/netboot/mini.iso
[08:23] <Kamion> try booting that with casper/enable=true and see what happens
[08:24] <mdz> actual netbooting could be made to be convenient for me
[08:24] <mdz> I have PXE stuff set up already
[08:25] <mdz> hm, no I don't
[08:25] <mdz> but I have tftpd/bootp stuff set up
[08:25] <Kamion> since you depend on cdrom-detect, /cdrom should be available despite netbooting by the time casper-udeb.postinst runs
[08:26] <mdz> yay for depends
[08:38] <Kamion> ... and starts downloading
[08:39] <Kamion> ETA three hours or so, so off to relax
[08:39] <fabbione> ehhe
[08:39] <fabbione> this new live methods are nice :-)
[08:40] <Kamion> (bootstrapping off other people's porting work)++
[08:45] <mdz> ETA 38 minutes
[08:47] <mdz> Kamion: we need to find a reasonable way to enable DMA on the CD-ROM
[08:47] <mdz> it's fairly essential if you're using an IDE CD-ROM drive
[08:48] <mdz> and the installer would benefit as well
[08:48] <mdz> we could add a simple "nodma" option for the command line, and start enabling it by default again
[08:48] <mdz> nocdromdma or so
[08:50] <mdz> Kamion: current i386 hoary-live seems to work great
[08:50] <mdz> Kamion: and thanks to the hotplug/udev changes I made yesterday, no more problem with /dev/input/mice missing
[08:56] <fabbione> mdz: i am not too happy to enable DMA by default
[08:56] <sivang> mdz: we should start a "derivation howto" on the wiki :)
[08:57] <sivang> mdz: for the livecd at least, if your infra is already mature.
[08:57] <fabbione> it tends to do much more mess if either the cdrom or the ide controller or a combinantion of all the chain is faulty
[08:58] <mdz> sivang: brian sutherland was interested in this as well
[08:58] <mdz> fabbione: that may be, but it is worse to have a live CD without DMA enabled
[08:58] <fabbione> mdz: i will check something on the kernel cofig
[08:58] <fabbione> config
[08:58] <fabbione> afaik we enable DMA by default only on harddisk
[08:59] <fabbione> we could remove that option
[08:59] <sivang> mdz: you have his email?
[08:59] <fabbione> and iirc at the first sign of crappiness the kernel disable the DMA on cdroms
[08:59] <sivang> mdz: hmmm , come to think, I can look for him on the mailing list. don't bother.
[09:00] <doko> hmm, is /proc mounted in the buildd chroots by default? gcc-3.4 builds are failing :-(
[09:01] <fabbione> doko: yes...
[09:01] <robtaylor_> mdz: sivang: i'll
[09:01] <robtaylor_> bah
[09:01] <fabbione> also /dev/pts
[09:01] <robtaylor_> bi intereseted in helping out
[09:01] <sivang> Kamion: Is there anything ready for derivation of ubuntu itself? that is, the installer cd. (not pushing or anything, curious)
[09:01] <robtaylor_> also. take care enabling dma on cdroms
[09:02] <robtaylor_> a lot of older drives just fail to work#
[09:02] <fabbione> doko: ppc: make[6] : stage2/xgcc: Command not found
[09:03] <mdz> robtaylor: however, it seems to work in a majority of cases, and the tradeoff when it's disabled is _huge_ for the live CD
[09:03] <mdz> for an install, it just takes longer during a noninteractive period
[09:03] <mdz> on the live CD, it's constantly slowing user interaction
[09:03] <doko> fabbione: IIRC this only happens without /proc being mounted.
[09:04] <fabbione> doko: it failed much much later on ia64
[09:04] <robtaylor_> mdz: i dont suppose it's dynamically settable?
[09:05] <mdz> it is
[09:05] <robtaylor_> mdz: ideal would be to have dma if it works..
[09:05] <mdz> the trick is figuring out whether it works
[09:05] <mdz> it's a combination of chipset, drive, etc.
[09:05] <mdz> afaik
[09:05] <robtaylor_> cant switch it on, attemt to read a file, if failiure occurs, run without it?
[09:07] <doko> fabbione: the build succeeded on my machine. almost the same build succeeded on the debian buildd
[09:07] <robtaylor_> it'd be a shame if, after these changes that give us a live cd that should run on every system that ubuntu will install cleanly on, we then lose that with this :)P
[09:09] <robtaylor_> wish i still had access to asystem where i new dma on cdrom failed :/
[09:09] <Nafallo> my girl's machine fails
[09:10] <Nafallo> faulty hardware
[09:15] <robtaylor_> Nafallo:so it works fina normally and fails if you hdparm -d1 /dev/hdc?
[09:16] <Nafallo> robtaylor_: she installed warty today. I had to guide her to ALT+F2, hdparm -d0 /dev/cdroms/cdrom0 cause it hanged.
[09:17] <robtaylor_> Nafallo: righty, cool stuff
[09:17] <robtaylor_> Nafallo: dont uppose you have access to the machine atm? ;)
[09:18] <Nafallo> robtaylor_: it's on my left side :-).
[09:18] <robtaylor_> (didn't know warty set DMA by stanbdard on install =) .. it doesnt after install, it seems =))
[09:19] <lamont> E: Couldn't download libevms-2.4
[09:20] <lamont> mdz/kamion: non-i386 is not currently debootstrapable...
[09:20] <robtaylor_> Nafallo: ok, could you try a script that say, hdparm -d1 /dev/cdroms/cdrom0;  ls /media/cdrom/ ; echo $?; hdparm -d0 /dev/cdroms/cdrom0 ...
[09:21] <robtaylor_> i cant remember quite what goes wrong =)
[09:21] <Nafallo> robtaylor_: from wartyinstall?
[09:22] <robtaylor_> Nafallo: just on the box as is will do fine
[09:22] <Nafallo> oki
[09:22] <robtaylor_> (with a cdrom in, of course ;) )
[09:22] <Nafallo> *booting*
[09:22] <robtaylor_> (and mounted)
[09:25] <Nafallo> damn cdrom won't open with the ejectbutton :-P
[09:26] <robtaylor_> Nafallo: /usr/bin/eject? ;)
[09:26] <Nafallo> on my way, just have to create myself an account ;-)
[09:26] <robtaylor_> Nafallo: heh
[09:29] <robtaylor_> mdz: i'm thinking it may well be possible to sanely detect the problem with an ls {some file on cdrom} && dmesg |grep 'DriveReady' or somesuch might suffice for detection
[09:31] <Nafallo> hehe, doesn't like mount /media/cdrom ;-)
[09:31] <Nafallo> hdc: Timeout waiting for DMA
[09:31] <Nafallo> hdc: drive not ready for command
[09:32] <Nafallo> Buffer I/O error on device hdc, logical block xx
[09:32] <robtaylor_> Nafallo: how long did it take to fail?
[09:32] <Nafallo> like, *enter, BOOM!* ;-)
[09:32] <robtaylor_> ah thats ok then :)
[09:33] <robtaylor_> can you paste me the output of dmesg?
[09:35] <robtaylor_> Nafallo: does something like d if=/dev/cdrom count=1 of=/dev/null give you the same result?
[09:36] <robtaylor_> s/d/dd
[09:36] <Nafallo> http://www.magicalforest.se/~nafallo/dmesg
[09:37] <Nafallo> last command succedded
[09:40] <Nafallo> LOL, forgot how dmesg worked :-P
[09:41] <Nafallo> fixed
[09:41] <Nafallo> same url :-)
[09:41] <robtaylor_> Nafallo: hmm, what about with bigger values for count?
[09:42] <Nafallo> robtaylor_: 9,99,999,9999 succedded
[09:42] <Nafallo> trying 99999 now ;-)
[09:43] <Nafallo> worked to
[09:44] <Nafallo> how large counts should I try?
[09:44] <Nafallo> s/count/block/
[09:44] <robtaylor_> Nafallo: mmm
[09:44] <robtaylor_> maybe bigger block sizes..
[09:44] <Nafallo> trying six 9's now ;-)
[09:45] <robtaylor_> try bs=655536, count=1
[09:45] <robtaylor_> or 2
[09:46] <robtaylor_> oops one less 5's ther ;)
[09:46] <jdub> lifeless: http://www.kerneltraffic.org/kernel-traffic/kt20050109_293.html#3
[09:46] <robtaylor_> Nafallo: or bs=512K, count=1 something like that..
[09:48] <Nafallo> bs=65536 count=[1,2]  succeded
[09:50] <Nafallo> testing bs=65536 count=99999 now
[09:50] <robtaylor_> no, try larger values of bs
[09:50] <robtaylor_> Nafallo: and small values of count
[09:50] <Nafallo> oki
[09:51] <robtaylor_> Nafallo: you can use K and M in specifying bs..
[09:51] <Nafallo> dooh! bs=99999999 ;-)
[09:52] <Nafallo> 100M would have been better :-P
[09:52] <Nafallo> transferred. and I believe I know why it isn't failing ;-)
[09:53] <robtaylor_> Nafallo: oh?
[09:53] <Nafallo> hdparm /dev/cdrom shows DMA as off ;-)
[09:53] <Nafallo> the kernel does that automagically
[09:53] <robtaylor_> doh!
[09:53] <robtaylor_> double doh!
[09:54] <Nafallo> I turned it on again ;-)
[09:54] <Nafallo> and runs bs=100M count=1 :-)
[09:55] <robtaylor_> Nafallo: well starting small is probably the sane option ;)
[09:56] <Nafallo> same errors and then kernel dropped DMA. test succedded :-P
[09:57] <Nafallo> 10M gives error
[09:57] <robtaylor_> ok dma back on, start with bs=512 count=1, keep doubling till failure
[09:58] <Nafallo> oki
[09:59] <robtaylor_> Nafallo: thanks :)
[09:59] <Nafallo> 8M worked, so I tried 10M again. worked this time. 16M fails :-P.
[10:00] <Nafallo> using warty pressed livecd for info.
[10:01] <Nafallo> turned dma on and tried 10M again, and it fails :-P
[10:01] <Kamion> lamont: I beat you to spotting that one; upgrade debootstrap :)
[10:01] <robtaylor_> Nafallo: hmm
[10:01] <Kamion> sivang: not my department, sorry, dunno
[10:02] <sivang> Kamion: k, np
[10:03] <Nafallo> robtaylor_: dd starts from the beginning of the CD every time?
[10:05] <robtaylor_> Nafallo: yep, unless you use skip=bla
[10:06] <Nafallo> robtaylor_: even more odd then ;-)
[10:06] <Nafallo> 10M fail, succedd, fail :-P
[10:06] <lamont> Kamion: yeah - done.
[10:06] <lamont> I need to update ubuntu-meta too.
[10:06] <lamont> ia64 is pretty booooring otherwise
[10:08] <robtaylor_> Nafallo: ah well, guess we can conlcude that dding is a bad way to cause the failure :)
[10:08] <Nafallo> robtaylor_: agreed ;-)
[10:08] <jdub> mdz, Kamion: http://www.advogato.org/person/mbrubeck/diary.html?start=90
[10:09] <Nafallo> robtaylor_: mounting is much more effective :-P
[10:13] <lamont> *** Error -5 compressing block 0! (compressed=0x7fbffff950, len=548682070304, uncompressed=0x526010, blocksize=65536)
[10:13] <lamont> /usr/sbin/livecd.sh: line 170:  9566 Segmentation fault      create_compressed_fs $IMG 65536 >livecd.cloop
[10:13] <lamont> ew!
[10:16] <Kamion> lamont: didn't think ubuntu-base directly depended on libevms-2.5?
[10:16] <Kamion> it's not in the seed list itself, it's just a dependency ...
[10:18] <robtaylor_> mdz, Kamion: where can i find casper?
[10:20] <mdz> robtaylor: apt-get source casper
[10:22] <robtaylor_> hmm
[10:22] <robtaylor_> mdz: thx :)
[10:23] <Nafallo> robtaylor_: you need silverfairy anymore? :-)
[10:28] <lamont> Kamion: another round of builds with latest debootstrap and ubuntu-meta will run at 22:06 london time.  should take about 25-30 minutes to complete.
[10:30] <Kamion> lamont: let me know when amd64/ia64 are available; it'll be easier to finish automating the builds if they have the same set of architectures
[10:30] <Kamion> otherwise, sure, I can kick off another build then
[10:31] <mdz> Kamion: powerpc live CD didn't work
[10:31] <mdz> it's missing some modules we need
[10:31] <Kamion> really? I haven't trimmed udebs yet
[10:32] <Kamion> what was missing?
[10:32] <mdz> dm-snapshot
[10:32] <mdz> looks like i386 uses common/md-modules
[10:32] <Kamion> huh
[10:32] <mdz> while powerpc has a separate one for some reason
[10:32] <lamont> btw, livecd-current.manifest
[10:32] <mdz> fabbione: ping
[10:33] <mdz> Kamion: the udeb stuff in linux-source could probably use some review, to clean up inconsistencies like this
[10:33] <Kamion> yes
[10:33] <mdz> likewise for the configs, really, but that's harder
[10:33] <Kamion> it was probably because the same set of modules weren't built for powerpc at the time I created that file
[10:34] <Kamion> the udeb review has to go hand in hand with the config review to some extent
[10:34] <jdub> mdz: how's casper going?
[10:34] <mdz> jdub: rad
[10:34] <lamont> Kamion: question for you... if the build fails, should livecd-current.* go away, or remain so that at least you get something?
[10:34] <lamont> currently they do the latter
[10:35] <Kamion> lamont: hm, think I'd prefer it to remain
[10:37] <jdub> mdz: is this dc-build related discussion?
[10:37] <mdz> jdub: that, and "damn, powerpc doesn't work because of something silly"
[10:37] <jdub> sucks being a second class citizen
[10:37] <lamont> Kamion: was thinking you would./
[10:37] <Kamion> dm-snapshot was only added to common/md-modules in September
[10:37] <jdub> house got robbed this morning :|
[10:37] <lamont> tomorrow I'll go leach bandwidth and download the then-current CD
[10:37] <Kamion> jdub: erk
[10:37] <lamont> jdub:???
[10:37] <mdz> Kamion: just confirmed that i386 and powerpc build identical modules, so it should be safe to fix that
[10:37] <jdub> couple of laptops stolen, wish they'd taken one or all of the three tvs we don't want
[10:37] <mdz> amd64 looks to be OK
[10:37] <Kamion> mdz: yes, I was checking it in Debian as well
[10:38] <mdz> ia64 and sparc are also OK, seems like only powerpc was weird in this respect
[10:38] <Kamion> shall I file a bug on our kernel?
[10:38] <mdz> I'd prefer to just fix it, but fabbione has a kernel upload pending I think
[10:38] <mdz> so yeah, I guess a bug
[10:39] <lamont> mdz: does ubuntu-meta_0.16 have ia64 as well?
[10:39] <mdz> lamont: no
[10:40] <lamont> grumble.
[10:40] <mdz> lamont: I thought you were going to do it
[10:40] <lamont> wanna roll 0.17, or shall I?
[10:40] <lamont> yeah - I tried, we stepped on each other.
[10:40] <mdz> Jan 08 13:50:00 <mdz>   lamont: and you can do that whenever you're ready
[10:40] <lamont> I'll roll it now.
[10:40] <mdz> go for it
[10:40] <lamont> again. :-(
[10:41] <Kamion> #5372
[10:42] <Kamion> I wonder why dm-snapshot was added to kernel-wedge in the first place? Nothing in mainline d-i uses it.
[10:42] <mdz> Kamion: maybe it went in with the rest of dm-*
[10:43] <Kamion> no, it was added explicitly and individually
[10:43] <Kamion>   * Martin Michlmayr
[10:43] <Kamion>     - Add dm-snapshot.o to md-modules if it's available.
[10:44] <gsuveg> can i use gimp22 under warty ?
[10:46] <lamont> 0.17 uploaded
[10:49] <kent> gsuveg, ask those kind of questions in #ubuntu
[10:52] <mdz> Kamion: I wonder what tbm was up to
[10:52] <mdz> it would not be at all common to configure a snapshot during installation
[10:53] <Kamion> I think he was working on RAID stuff at the time, but I don't remember the details
[10:59] <lamont> sigh. local mirror is 852 files out-of-date. :-(
[11:13] <lamont> Kamion: ia64 requeued to run at 22:36, this time with dependencies.. :-)
[11:18] <gsuveg> kent: ok. sorry
[11:20] <Kamion> lamont: is amd64 long-term screwed, or will it start working with the current build?
[11:23] <mdz> lamont: is the issue corrected which was preventing amd64 live fs builds from being possible?
[11:24] <Nafallo> it's only the livecd that's broken on amd64?
[11:29] <lamont> 8458 Segmentation fault      create_compressed_fs $IMG 65536 >livecd.cloop
[11:29] <lamont> you mean that bug?
[11:29] <lamont> *** Error -5 compressing block 0! (compressed=0x7fbffff950, len=548682070304, uncompressed=0x526010, blocksize=65536)
[11:29] <lamont> still there, cause unknown
[11:29] <Mithrandir> that's amd64?
[11:30] <lamont> yep
[11:30] <Mithrandir> ugh :/
[11:30] <Mithrandir> it used to work?
[11:30] <lamont> Mithrandir: unknown
[11:30] <lamont> it used to die because there was no loopback support in the kernel.
[11:30] <Mithrandir> ok
[11:30] <lamont> OTOH, amu built home-edition OK, iirc
[11:30] <Mithrandir> he has an amd64
[11:30] <Mithrandir> ?
[11:32] <mdz> lamont: no, I was talking about the lack of loop modules and what not on the build system
[11:32] <mdz> lamont: this is the first I think I've heard about it segfaulting
[11:33] <mdz> lamont: that 'len' value looks quite wrong
[11:33] <mdz> I have this bad feeling that nobody has really used cloop on anything but i386
[11:33] <mdz> powerpc should be fun
[11:34] <lamont> mdz: kernel was fixed yesterday.
[11:35] <lamont> and then it was failing because of evms-2.4
[11:35] <lamont> we're finally up to the segv failure.
[11:35] <lamont> note that if you want to test a live-dvd, the uncompressed fsimg can be made available now... :(
[11:36] <mdz> lamont: try advfs?
[11:36] <mdz> never mind; they're the same now
[11:37] <mdz>  * * Sun Okt 26 01:05:29 CEST 2003 Klaus Knopper
[11:37] <mdz>  * - Changed format of index pointers to network byte order
[11:37] <mdz> thanks, Klaus
[11:37] <mdz> so there's hope
[11:38] <mdz> jdub: I need some mbox love from mailman
[11:38] <Kamion> combo install/live DVDs need a bit more work at the cdimage end
[11:38] <mdz> jdub: my mail was fucked this morning and I missed a bunch of list mail
[11:38] <mdz> lamont: reproduced the amd64 crash
[11:38] <Kamion> at the moment I just have a LIVE flag; I need to invent another flag called INSTALL or something to allow for combinations
[11:39] <Kamion> shouldn't take too long, because I can find all the places where it's needed by grepping for LIVE :)
[11:39] <mdz> hmm
[11:39] <mdz> that weird 'len' value comes from it printing the value of a pointer as an integer
[11:40] <lamont> that would do it.
[11:40] <mdz> doesn't explain the crash, though
[11:40] <lamont> did mono move into main?
[11:41] <mdz> #define Z_BUF_ERROR    (-5)
[11:42] <mdz>      compress2 returns Z_OK if success, Z_MEM_ERROR if there was not enough
[11:42] <mdz>    memory, Z_BUF_ERROR if there was not enough room in the output buffer,
[11:42] <mdz>    Z_STREAM_ERROR if the level parameter is invalid.
[11:42] <mdz> lamont: should have
[11:42] <mdz> so it thought there was not enough room in the output buffer
[11:44] <lamont> muine (in main) Build-depends a bunch of mono stuff...
[11:45] <robtaylor_> Nafallo: umm, hope you took my lack of reply as a 'no' ;)
[11:46] <Nafallo> robtaylor_: indeed ;-)
[11:48] <lamont> Kamion: i386 and ppc are there.
[11:48] <lamont> ia64 is, um, borked
[11:48] <lamont> amd64 has the segv issue
[11:48] <lamont> mdz: or is shipping openoffice.org on ia64 a requirement?
[11:49] <lamont> and what's the best way to strip the OO.o stuff from the desktop list for just one arch?
[11:49] <jdub> mdz: ok
[11:49] <Mithrandir> lamont: I don't think that's advisable if we want elmo to retain his sanity.
[11:49] <mdz> lamont: ia64 seems pretty fucked overall as far as hoary is concerned
[11:50] <lamont> mdz: if we get oo.o gone from the u-d deps, then we should have a livecd.
[11:50] <lamont> then again, anna doesn't seem to like it very much
[11:50] <mdz> I'm not inclined to bother
[11:50] <lamont> to bother with?  ia64/hoary
[11:50] <lamont> ?
[11:50] <mdz> ia64 live cd
[11:50] <lamont> ah, yes.
[11:50] <mdz> and, if things don't change, ia64 hoary
[11:51] <lamont> it has a glibc now...
[11:51] <magnon> ia64 does seem pretty screwed overall
[11:51] <mdz> the community people who said they would look after it have disappeared, as far as I can tell
[11:52] <lamont> I've heard from t-bone, but only that he was hip deep until january started...
[11:53] <lifeless> jdub: thanks!
[11:54] <lifeless> fabbione: http://www.kerneltraffic.org/kernel-traffic/kt20050109_293.html#3 <--- can we get this, please please please
[11:54] <jdub> heh
[11:54] <mxpxpod> lamont: when I go to upgrade libmono0 on powerpc, every mono app wants to uninstall... is that being fixed?
[11:54] <lamont> so back to the mono question....
[11:54] <lamont> mdpxpod: it's ftbfs, and known broken upstream.
[11:55] <mxpxpod> ftbfs?
[11:55] <lamont> the last I've heard, they planned to upload what was needed once debian had built all the architectures they cared about.
[11:55] <lamont> fails to build from source.
[11:55] <mxpxpod> ah, ok
[11:55] <mxpxpod> thanks
[11:55] <lamont> see people.ubuntu.com/~lamont/buidlLogs/m/mcs/1.0.4-1
[11:56] <lamont> mdz/jdub: so does muine belong in universe, or did mono enter main while it was ftbfs
[11:56] <lamont> ?
[11:56] <Mithrandir> I should set up a local mirror, I think.
[11:56] <jdub> lamont: mono shouldn't be in main...
[11:56] <lamont> Mithrandir: of at least main/restricted for at least amd64...
[11:57] <jdub> (it's not here atm)
[11:57] <mdz> lamont: is t-bone back in the game and not going off on a ship somewhere?
[11:57] <lamont> jdub: the muine needs some love.
[11:57] <lamont> mdz: nfc
[11:58] <jdub> lamont: what's up with it?
[11:58] <Mithrandir> what's the current recommended mirror script?  Debmirror?
[11:58] <jdub> (turns out muine 0.8.0 just came out, so my quick rebuild upload was pretty pointless)
[11:58] <lamont> Checking for already installed source dependencies...
[11:58] <lamont> debhelper: missing
[11:58] <lamont> mono-jit: missing
[11:58] <lamont> mono-mcs: missing
[11:58] <lamont> c-sharp-compiler: missing
[11:58] <lamont> that's from the build of muine_0.6.3-4ubuntu1
[11:59] <jdub> i thought all the mono foo was fixed up now?
[12:00] <lamont> jdub: see 5052
[12:00] <lamont> it's understood... it's ftbfs.
[12:00] <lamont> built upstream, but not buildable in our archive.