[12:02] <mjg59> Ah
[12:02] <mjg59> 2.6.9 has better USB suspend/resume
[12:02] <mjg59> Should probably still unload them, though
[12:02] <bob2> ah
[12:02] <bob2> will that block if any usb stuff is in use?
[12:03] <chrisa> There's usually a problem on wakeup if usb was loaded and a device was removed between sleep and resume
[12:03] <mjg59> Shouldn't
[12:03] <bob2> hm, right, so it's a good idea to detach stuff anyway
[12:04] <bob2> I'll try fabio's new 2.6.9 in the morn
[12:05] <Kamion> call me bored
[12:06] <bob2> I figured that when I saw you looking for typoes in the wiki :)
[12:07] <Kamion> well, we were kinda collectively going through the quotes
[12:07] <Mithrandir> Kamion: you're not considering just going to bed? :)
[12:07] <bob2> I was trying to capture the accent
[12:07] <Kamion> bob2: mind you, you're subscribed to Quotes, a fine one you are to talk :)
[12:07] <Kamion> Mithrandir: soon ...
[12:07] <bob2> poorly, of course
[12:07] <Kamion> bob2: yeah, Scott told me and changed it back
[12:08] <mjg59> I'm really not sure that I understand this
[12:09] <Keybuk> so, you don't want me to run this then? :)
[12:09] <mjg59> Hrm. If I cat something into stdin and then read from it, read should stop on the EOF, right?
[12:10] <Mithrandir> like cat foo | while read thing; do .. ?
[12:10] <mjg59> No, into a C program
[12:11] <mjg59> And it works now.
[12:11] <mjg59> Hrmph.
[12:11] <Keybuk> depends whether the shell actually closes stdin or not
[12:12] <Mithrandir> mjg59: cat foo | c-program-which-does-read?
[12:12] <mjg59> Yeah
[12:12] <Mithrandir> read in said c program will return zero when EOF is detected.
[12:21] <bob2> good idea
[12:26] <mjg59> ARGH.
[12:27] <mjg59> There's a maximum length on environment variables?
[12:28] <Keybuk> mjg59, UNIX; UNIX, mjg59.  I'm sure you'll be friends.
[12:29] <mjg59> Keybuk: Right.
[12:29] <mjg59> Keybuk: Can you grab http://www.codon.org.uk/
[12:29] <mjg59> Keybuk: Can you grab http://www.codon.org.uk/~mjg59/tmp/vbestate
[12:30] <mjg59> And then try VBESTATE=`tempfile`; sudo vbestate save >$VBESTATE; sudo video_post; sudo vbestate restore <$VBESTATE; rm $VBESTATE
[12:30] <Keybuk> want me to grab the former or the latter?
[12:30] <Keybuk> I have enough bandwidth <g>
[12:30] <mjg59> The latter
[12:30] <Keybuk> in a console
[12:30] <mjg59> I don't have enough bandwidth :p
[12:30] <mjg59> Yaeh, console
[12:31] <Keybuk> back after the system crash and reboot then
[12:31] <mjg59> ?
[12:31] <mjg59> Heh
[12:31] <Keybuk> ....where do I get "video_post" from ?
[12:31] <mjg59> Oh, right. use vm86_video_post
[12:34] <thom> mjg59: scott's singing that song again
[12:34] <mjg59> Haha
[12:34] <mjg59> Is he using a framebuffer?
[12:34] <thom> no
[12:34] <mjg59> Hrmph
 do I need one?
[12:36] <mjg59> Nope
[12:36] <mjg59> Works better without
[12:36] <thom> FSVO works :-)
[12:36] <mjg59> His is the only machine I've found where it breaks this badly
[12:37] <thom> he's closed his laptop and is sulking now ;-)
[12:37] <mjg59> Scott's laptop is broken in a wide variety of amusing ways
[12:38] <thom> including the STUPIDLY PLACED LID SWITCH
[12:38] <mjg59> Anyway, we can do this from userspace without using X
[12:38] <mjg59> So that's good enough
[12:39] <thom> anyway
[12:39] <thom> gnight
[12:46] <shaya> anyone know if there's a linux-restricted-modules to go w/ the 2.6.9 kernel floating around somewhere?
[12:47] <Kamion> shaya: not done yet AFAIK
[12:48] <Kamion> it's on daniels' to-do list
[12:48] <Kamion> http://people.ubuntu.com/~daniels/l-r-m/ has preliminary bits it seems
[12:48] <Kamion> dunno what state they're in though; they're from six days ago
[12:50] <shaya> hmm
[12:51] <shaya> splitting the packages up?
[12:55] <Kamion> shaya: don't think so?
[12:55] <Kamion> doesn't look split up to me
[01:09] <mirak> there is a bug with gnome-menus
[01:10] <shaya> is there a way to make esd start wrapped by aoss?
[01:13] <mjg59> I've managed to get my X session entertainingly broken
[01:21] <mjg59> Ok, that seems to approximately work
[01:21] <mjg59> I've added a new, heavily crack-laden acpi-support package to www.srcf.ucam.org/~mjg59/laptpos/
[01:21] <mjg59> Make that www.srcf.ucam.org/~mjg59/laptops/
[01:26] <shaya> mjg59: where's the changelog?
[01:27] <mjg59> shaya: In the package
[01:28] <shaya> mjg59: but what about us who dont want to install it to see it
[01:29] <mjg59> shaya: You can extract the control archive
[01:29] <shaya> for just look at the .tar.gz like I just did
[01:29] <shaya> s/for/or/
[01:29] <mjg59> Yes, since it's a Debian-native package
[04:02] <jdodson> knock, knock.
[07:55] <lifeless> daniels: tarball time methinks.. please
[09:04] <daniels> lifeless: ok
[09:05] <thom> marning
[09:06] <daniels> thom: last night I just got back and basically collapsed, sorry
[09:06] <thom> no worries
[09:06] <thom> figured that was the case
[09:06] <thom> i have new GROOVY graphs
[09:08] <Mithrandir> url?
[09:08] <daniels> thom: are they also sexy?
[09:09] <thom> www.planetarytramp.net/bootchart/bootchart-20041208-2219.png
[09:09] <bob2> mjg59: massive amounts of sleep love now I disabled ehci_hcd
[09:09] <bob2> it also reminds me that daniel had an oops when suspending while copying data to a usb2 disk
[09:11] <daniels> thom: .11/~daniels/
[09:11] <carlos> fabbione: https://bugzilla.ubuntu.com/show_bug.cgi?id=2197
[09:11] <daniels> thom: (Xorg and i810_drv.o)
[09:11] <fabbione> bob2: that's a problem in the scsi layer
[09:11] <fabbione> we already have the fix
[09:13] <thom> daniels: nod
[09:14] <thom> daniels: what i have is all ready to upload,pretty much :-)
[09:15] <bob2> fabbione: in -3?
[09:15] <daniels> thom: nice!
[09:16] <bob2> pimpalicious
[09:16] <daniels> thom: what're the changes thus far?
[09:16] <thom> daniels: well, depmod might need some cleaning ;-)
[09:16] <daniels> oooo, I see readahead
[09:16] <bob2> someone needs to unfuck gnome's hoary so I can dist-upgrade ;)
[09:16] <thom> daniels: check out the vicious readahead in parallel with hotplug :-)
[09:16] <daniels> yeah, maybe the whole 'sh -x' thing might need to change ;)
[09:16] <daniels> thom: yeah, that's awesome :)
[09:16] <daniels> wtf is hwclock spinning so much for, tho?
[09:17] <Keybuk> that's how it works
[09:17] <Mithrandir> thom: nice
[09:17] <daniels> Keybuk: oh?
[09:17] <thom> hwclock spins on the cpu until the end of the second
[09:18] <daniels> thom: can we defer cupsys to 21?
[09:18] <Keybuk> gnome-cups-icon in user's session
[09:18] <daniels> thom: i reckon if we move it to 21 and you drop in the new X stuff, we can get it 40, or sub-409
[09:18] <daniels> Keybuk: so that will absolutely refuse to turn up if cupsys isn't up?
[09:18] <thom> i think so
[09:18] <thom> per user cups love! ;-)
[09:19] <Keybuk> exactly, so if cups turns up a second later, you get no cups icon
[09:19] <daniels> mmm, that would be ill
[09:22] <daniels> thom: losing postfix and a2 will obviously be a win also
[09:28] <daniels> thom: also, is it worth reading Xorg ahead?
[09:29] <daniels> thom: -rw-r--r--  1 daniels daniels 14M 2004-12-09 09:29 /home/daniels/public_html/xorg-trace
[09:29] <thom> daniels: yeah, intend to
[09:29] <thom> oh, sweet
[09:29] <thom> can you put that somewhere?
[09:29] <thom> i really need to trace gdmgreeter too, or will that be in the trace?
[09:30] <daniels> thom: ... /home/daniels/public_html ...
[09:30] <daniels> thom: gdmgreeter isn't in there, no, I'll trace it later
[09:30] <thom> oh, duh
[09:30] <thom> thanks
[09:30] <daniels> np
[09:30] <daniels> i'm doing one with -e trace=file now tho
[09:31] <daniels> thom: xorg-trace-file
[09:31] <thom> i did trace=open which was not quite right i guess
[09:31] <thom> daniels: that has gdmgreeter? sweet
[09:31] <daniels> argh, WHY ARE WE STATTING ALL THE FILES
[09:31] <daniels> not with gdmgreeter
[09:31] <thom> oh, right
[09:31] <daniels> wow, it's psychopathic about /proc/bus/pci/devices
[09:32] <daniels> and stats all the drivers again ...
[09:32] <daniels> what a hog
[09:32] <daniels> it's statting all the files in a loop.
[09:32] <fabbione> ahahaha
[09:32] <daniels> daniels@catsby:~/canonical/gdm/gdm-2.6.0.4% grep radeon_drv.o ~/public_html/xorg-trace-file | wc -l
[09:32] <daniels> 41
[09:32] <daniels> that's forty-one accesses to a driver I'm not even loading.
[09:33] <thom> yeah
[09:36] <fabbione> Kamion: here?
[09:36] <haggai> wooerwewrrewrewrew~.
[09:37] <haggai> uh
[09:37] <Kamion> fabbione: yo
[09:37] <haggai> sorry, my network got confused during suspend/resume
[09:37] <fabbione> Kamion 
[09:37] <fabbione> do we officially support pegagos2?
[09:37] <haggai> ..and the suspend took 3 minutes BTW
[09:38] <daniels> why does X use regexs?
[09:38] <Kamion> fabbione: not yet, no
[09:38] <fabbione> Kamion: i am going to kill that driver temporary. removing the name from d-i is enough fo make d-i happy again?
[09:38] <Kamion> fabbione: I should think we'll only do it if it's low-effort or contributed; I do plan to see if I can make it work in my spare time though
[09:38] <Kamion> fabbione: put a "-" in front of it in the d-i file
[09:38] <Kamion> that makes the driver optional
[09:39] <fabbione> Kamion: ok thanks
[09:39] <fabbione> i will give it another hour love before reverting
[09:39] <Kamion> ok
[09:39] <fabbione> but it doesn't compile yet.
[09:41] <daniels> four loops that all read drivers in!  wa-hey!
[09:43] <thom> daniels: that Xorg took way longer but it still has all the printing
[09:43] <fabbione> daniels: taht could be the new autodetection thingy
[09:44] <thom> added 8 seconds
[09:44] <Mithrandir> thom: what do you use for generating those graphs?
[09:44] <thom> http://www.klika.si/ziga/bootchart/
[09:44] <thom> run from the top of inittab
[09:46] <Mithrandir> hm, have to install java, then
[09:46] <daniels> thom: er, yeah, the printing shouldn't have any performance impact
[09:46] <daniels> thom: really??
[09:47] <thom> Mithrandir: i've just been uploading the logs ;-)
[09:49] <Mithrandir> heh, that should work as well, yes.
[09:51] <thom> i'm building it now before the poor guy kills me
[09:51] <thom> or trying to 
[09:51] <thom> gcj 3.3 appears not to have java.util.logging.Logger
[09:54] <thom> did i mention i hate java?
[09:54] <Keybuk> many times :)
[09:54] <Keybuk> cupsys_1.1.20final+rc1-10ubuntu5_source.changes ACCEPTED
[09:54] <Keybuk> wow, a mail from katie that's actually *intended* for me ... :o)
[09:56] <thom> heh
[10:00] <bob2> jdub: thom considered using the windows firefox icon instead of the plain planet one?
[10:01] <jdub> bob2: that's a trademark issue more than a platform issue :)
[10:04] <bob2> oh, suck
[10:04] <bob2> that's highly obnoxious
[10:07] <Keybuk> s/postfix/nullmailer/ -- yay, five whole seconds
[10:09] <bob2> jdub: what does the HIG say about "escape" in a dialog?  it should exit?
[10:10] <daniels> daniels@catsby:~/canonical/xorg/arch/pristine/xorg-6.8.1/build-tree/xc/programs/Xserver% grep radeon_drv xorg.trace | wc -l
[10:10] <daniels> 0
[10:14] <Keybuk> bob2: escape tends to be bound to cancel, but I don't think the HIG is specific about it
[10:14] <Keybuk> it doesn't make a huge amount of sense in instant-apply dialogs
[10:15] <bob2> just talking to a windows user who's pointing out possibly confusing things
[10:16] <jdub> bob2: there's an argument about it going on. :)
[10:16] <Keybuk> I think the general theory is that if you dialog *has* Escape/Cancel functionality, you bind the key to the Cancel button ... e.g. the Open/Save dialogs
[10:17] <Keybuk> but Instant-Apply dialogs which only have Close use Enter, as you can't escape from them
[10:19] <Keybuk> ...though a argh-put-it-all-back-how-it-was escape functionality for i-a dialogs would be cute
[10:19] <Keybuk> but Jeff might beat me up for saying that
[10:19] <Keybuk> :p
[10:24] <pitti> jdub, thom: we need a name for this hardware detection thingy
[10:25] <pitti> jdub, thom: what about hwfu?
[10:25] <bob2> hardwarepants!
[10:25] <pitti> NOOOOOOO
[10:25] <fabbione> Kamion: i got the driver to compile
[10:25] <Mithrandir> chainsaw
[10:25] <Kamion> fabbione: bonus
[10:26] <Mithrandir> on the basis of it ripping through your hardware.
[10:26] <fabbione> Kamion: :-)
[10:27] <pitti> "Ubuntu Unified Python Hardware Detection and XMLification Extensible Framework"
[10:27] <pitti> (SCNR)
[10:29] <lamont_r> pitti: UUPHDaXEF?
[10:29] <pitti> lamont_r: easy to pronounce, right?
[10:30] <lamont_r> maybe for you Germans...
[10:30] <azeem> at least for those playing in humpa-bands
[10:30] <Mithrandir> that could almost have been one of my hostnames.
[10:30] <Keybuk> Python Hardware Autodetection Tool
[10:31] <lamont_r> Keybuk: PyHAT?
[10:31] <Keybuk> PHAT!
[10:31] <pitti> Keybuk: this sounds good
[10:31] <lamont_r> Pedantic HW detection tool?
[10:31] <lamont_r> shouldn't really have the language name in the name.
[10:32] <pitti> Keybuk: btw, it does not really detect hw, it only calls several backends which do and unifies them into an XML output
[10:32] <Keybuk> so basically it converts /sys into XML?
[10:32] <pitti> Keybuk: not only /sys, but also /proc/foo, dmidecode, hal, whatever
[10:33] <pitti> Keybuk: I wrote a plugin architecture to make it expandable easily
[10:33] <ogra> see here, it uses the ouptunt of the named tools: https://ubuntulinux.org/wiki/HardwareDatabase 
[10:33] <Mithrandir> why XML?
[10:33] <Mithrandir> xml is so icky
[10:33] <Mithrandir> gives me scratches
[10:33] <pitti> Keybuk: this is supposed to unify d-i's, X's and base's HW detection
[10:33] <pitti> Mithrandir: easy to throw into dbs
[10:33] <pitti> Mithrandir: easy to process
[10:34] <pitti> Mithrandir: standard format, works with many tools
[10:34] <pitti> Mithrandir: why _not_ XML? What to take instead?
[10:34] <Mithrandir> pitti: have you worked much with XML?
[10:34] <Keybuk> why base's?
[10:34] <pitti> Mithrandir: enough to like it :-)
[10:34] <Mithrandir> but not enough to really, really dislike it, I understand.
[10:34] <Keybuk> the kernel does hardware detection in base
[10:34] <Mithrandir> (:
[10:34] <pitti> Keybuk: it's not really kernel detection
[10:35] <pitti> Keybuk: it's supposed to become a large hardware database
[10:35] <pitti> Keybuk: devices which are supported/not supported by Ubuntu
[10:35] <Keybuk> right
[10:35] <pitti> Keybuk: it could say "this laptop works fine", "S3 does not work with this BIOS", etc.
[10:35] <pitti> Keybuk: also it could help X configuration, like "widescreen is not detected on this laptop"
[10:35] <Keybuk> *nods*
[10:36] <thom> pitti: ZDHW
[10:36] <pitti> -> yesterday's bof
[10:36] <pitti> thom: zero day hard ware?
[10:36] <fabbione> lamont_r, elmo: 
[10:36] <fabbione> sparcbuildd@vultus5:~$ wanna-build --list=needs-build
[10:36] <fabbione> Total 0 package(s)
[10:36] <ogra> zero data
[10:36] <pitti> thom: what does "zero day" mean?
[10:36] <fabbione> this is golden :-)
[10:36] <pitti> fabbione: congrats
[10:36] <thom> zerodayhardwarez
[10:36] <fabbione> we only need the last X from danicrack to give back a few packages
[10:36] <lamont_r> pitti: in theory, the answers you want are there the day your hardware is purchasable
[10:36] <thom> pitti: like zero day warez or sploits
[10:37] <sivang> fabbione : recalling you asked for my key, aren't we going to have a key signing party?
[10:37] <pitti> sivang: tomorrow evening
[10:37] <lamont_r> fabbione: is that phase2, and just main, or did it finally finish universe?
[10:37] <ogra> sivang: dont forget to print your keylist !
[10:38] <sivang> ogra : I won't! Thanks for reminding me :-))
[10:38] <fabbione> lamont_r: phase2 main
[10:39] <lamont_r> fabbione: nice
[10:39] <Treenaks> lamont_r: just shout his name very, very loudly
[10:39] <fabbione> sivang: yes.. i understood you were going to leave before that
[10:39] <fabbione> sivang: but @ the keysign is ok with me
[10:39] <Treenaks> daniels: uh.. how many?
[10:39] <fabbione> daniels: i need X sometime soon
[10:40] <seb128> lamont_r, just take a ticket and wait for your turn, a lot of people want to track me down first apparently :p
[10:40] <daniels> Treenaks: three thousand five hundred
[10:40] <daniels> fabbione: will upload before lunch
[10:41] <fabbione> daniels: thanks
[10:41] <daniels> also just ditched 150 open()s.
[10:41] <Treenaks> daniels: ho-ly cow on crack
[10:41] <daniels>   5702 /home/daniels/public_html/xorg-trace-file
[10:41] <daniels>   1904 xorg.trace-1040
[10:41] <daniels> (before and after, looking at file accesses)
[10:41] <fabbione> hmmmm
[10:46] <lamont_r> daniels: and 150 closes? or were the opens all failures?
[10:47] <daniels> lamont_r: 150 open/close cycles
[10:47] <daniels> NOT ANY MORE
[10:47] <bob2> wooohoo
[10:53] <root_> fabbione: aren't there any 686 linux-restricted-kernel for 2.69?
[10:53] <fabbione> root_: daniels is working on l-r-m
[10:54] <fabbione> if they are not there, they are not there yet
[10:54] <fabbione> patience... it's hoary
[10:54] <ogra> wow, 2.69 ? thats nice :)
[10:55] <root_> yes
[10:55] <root_> I love to wait
[10:55] <root_> thx for ur info
[10:57] <jdub> Keybuk: i have no opinion on the dialogue escape stuff
[10:58] <Keybuk> I think the current mix is about right ... I don't think I've ever hit Escape and have nothing happen, yet I'm sure I do hit Escape from time to time
[10:58] <Keybuk> so it must work where I expect it to
[11:00] <sivang> fabbione : I'll be here tommorow eveing  , I guess I missunderstood.
[11:01] <jdub> lamont: ooh, what were the postfix startup fixes?
[11:04] <Keybuk> :0,$s/^/#/
[11:05] <sivang> ogra : how do I install the naked lady artwork? :)
[11:06] <Mithrandir> sivang: install ubuntu-calendar
[11:06] <Mithrandir> and then change to "ubuntu monthly calendar" background
[11:07] <Keybuk> thom: \o/ udev 04[78]  includes the udevd-as-hotplug patches
[11:08] <sivang> Mithrandir : thaks alot!
[11:08] <Keybuk> 048 being a brown-paper-bag "oops, that didn't build" of 047
[11:08] <Keybuk>  - wait_for_sysfs is now gone.
[11:08] <Keybuk>  Oh, happy day...
[11:09] <Kamion> Keybuk: rock!
[11:09] <daniels> wooooo
[11:12] <thom> GAR
[11:12] <thom> fabbione: !!!!!!
[11:12] <thom> 11:11 /tmp% sudo invoke-rc.d dbus-1 restart
[11:12] <thom>  * Restarting system message bus...
[11:12] <thom>  * Stopping Hardware abstraction layer:                                  [ ok ] 
[11:12] <thom>  * Starting Hardware abstraction layer:
[11:13] <thom> 11:12:05.020 [W]  hald.c:302: Your kernel does not support capabilities; some features will not be available.                               
[11:13] <Kamion> fun
[11:13] <thom> ii  linux-image-2.6.9-1-686        2.6.9-3
[11:13] <pitti> thom: you know how to beat up :-)
[11:13] <Kamion> hm, I wonder if waldi was mistaken when I was talking to him about multi-component d-i Packages files
[11:13] <Kamion> I don't see where the problem would be ...
[11:14] <daniels> yet more proof that ed is crap
[11:14] <fabbione> thom: doh!
[11:14] <pitti> thom: works fine on linux-image-2.6.9-1-powerpc
[11:15] <Treenaks> ah so it's not just me
[11:15] <fabbione> thom: let me check here
[11:15] <fabbione> thom: no errors here
[11:15] <Keybuk> and then realises, that as he can see Mithrandir doing that, the problem is more subtle than the line just dropping
[11:16] <Treenaks> dns crap
[11:16] <Keybuk> no, isn't dns
[11:16] <Treenaks> no?
[11:16] <thom> no
[11:16] <Treenaks> what is it then?
[11:16] <Kamion> oh, but I can't test because there's no l-r-m 2.6.9 yet, gah
[11:16] <Keybuk> new connections in general aren't happening
[11:16] <daniels> seems to be drop SYN on the floor
[11:16] <thom> routing new connections go boom
[11:16] <fabbione> Kamion: ask daniles :-)
[11:17] <daniels> Kamion: bleh
[11:17] <Treenaks> even connections to the proxy?
[11:17] <Keybuk> isn't just SYN
[11:17] <Keybuk> ICMP isn't happening either
[11:17] <thom> ISP breakage
[11:17] <Keybuk> ARP-fuckage?
[11:17] <Mithrandir> thom: NAT table overflow?
[11:17] <elmo> it did this yesterday too
[11:17] <Keybuk> back now, anyway
[11:17] <elmo> it'll resolve itself in a bit probably
[11:18] <seb128> lamont_r, #4495 ?
[11:18] <seb128> lamont_r, on which arch ?
[11:18] <bob2> "the ISP died for our sins"
[11:18] <seb128> lamont_r, according to the build log that's ok (out of ia64)
[11:20] <lamont_r> seb128: yeah - ia64.. but the error sure looked generic...
[11:22] <lamont_r> jdub: postfix goes kinda paranoid at startup... I commented out the checks. :-)
[11:22] <lamont_r> seb128: it's because xorg hasn't built on ia64 yet.
[11:23] <seb128> lamont_r, so, I'm supposed to do something with the build-dep ? or we just wait ?
[11:23] <jdub> lamont_r: hrm. would wietse kill you?
[11:23] <lamont_r> jdub: it's ubuntu-only. :-(
[11:23] <lamont_r> seb128: we should beat on doko
[11:23] <seb128> ok :)
[11:24] <seb128> thanks
[11:24] <jdub> lamont_r: is it sane on servers?
[11:24] <lamont_r> jdub: probably more so than on desktops.
[11:25] <jdub> oh?
[11:25] <lamont_r> servers tend to shutdown nicely.
[11:26] <jdub> hrrrmmmmm
[11:26] <jdub> but when they haven't, you want them to start up paranoid :)
[11:26] <lamont_r> but, yeah. it should really schedule a postfix check run for some point in the future...
[11:27] <lamont_r> hrm... cron.daily here we come.. :-)
[11:28] <lamont_r> jdub: it's the part that runs through the entirety of /var/spool/postfix and does sanity checks on it.
[11:28] <jdub> yeah
[11:28] <lamont_r> so yeah, it should really be scheduled to happen sometime..
[11:28] <elmo> kamion/fabbione/mdz/anyonewhocares: anastacia wants to demote *2.6.8.1* to universe - ok?
[11:28] <lamont_r> woot!
[11:28] <jdub> lamont_r: you can do that at any time, right?
[11:29] <lamont_r> jdub: exactly
[11:29] <Kamion> elmo: might be nice to wait for 2.6.9 l-r-m
[11:29] <Kamion> otherwise *shrug* as far as I'm concerned
[11:29] <doko> lamont_r: I'll do the merge today. found the bug, testing a fix
[11:29] <lamont_r> I suppose it's possible that email that was in the system _might_ wind up being magically deferred until it is run.
[11:30] <jdub> lamont_r: maybe start postfix and add another init at the end for the check?
[11:30] <lamont_r> jdub: I like that...
[11:30] <lamont_r> end of rc2?
[11:31] <jdub> i guess
[11:31] <daniels> lamont_r: um, dude, xorg should've built on ia64 long ago
[11:31] <lamont_r> daniels: should is the operative word there.
[11:31] <daniels> thom: /~daniels/gdm-greeter-strace
[11:31] <daniels> lamont_r: anything I need to do?
[11:31] <daniels> thom: (.11)
[11:32] <thom> nice
[11:32] <lamont_r> daniels: help doko? :-)
[11:32] <lamont_r> glibc is ftbfs, doko's going to upload a fix soon
[11:32] <thom> i just fixored my hdparm settings (all this was with default hdparm so far)
[11:32] <lamont_r> then all will proceed according to the master plan.
[11:32] <jdub> thom: eeek!
[11:32] <fabbione> Mithrandir: swig1.3-1.3.22 is a FTBFS on 4/5 arches
[11:32] <Mithrandir> fabbione: ew.
[11:32] <daniels> lamont_r: ah, right
[11:32] <fabbione> Mithrandir: mind to take a look to it since you did the last uplaod?
[11:32] <Mithrandir> sure
[11:33] <fabbione> thanks
[11:33] <Mithrandir> don't we have gcj on all arches?
[11:33] <fabbione> i think it's related to the version
[11:34] <lamont_r> Mithrandir: gcc-3.4 is still waiting for a working binutils on i386, which is tcl8.4 bug
[11:34] <doko> Mithrandir: shoud be there for all arhcs.
[11:34] <Mithrandir> fabbione: it's really depwait, then
[11:35] <Mithrandir> lamont_r: it ftbfs on !i386
[11:35] <lamont_r> that's because i386 builds arch-all, and therefore life sucks on !i386
[11:38] <kanto_so> cough, cough too freezing here, a gift for all developers in X-mas, work and listen to musics from LimeWire, I love u guys all
[11:38] <kanto_so> http://sales.limewire.com/040903/gc59uxYlbhPoxaBKtyVc3k2dKuillFxf/download/LimeWireLinux.bin
[11:38] <Mithrandir> lamont: ok, so it should magically fix itself in a little while, or you need to tickle it?
[11:38] <lamont_r> I think I have to kick it.
[11:38] <mjg59> bob2: Ok, my latest acpi-support stuff unloads USB modules
[11:39] <Mithrandir> is haggai around still?
[11:39] <kanto_so> mjg59, are u Herbert Xu?
[11:39] <Mithrandir> kanto_so: no, he's Matthew Garret
[11:39] <Mithrandir> Xu doesn't irc.
[11:40] <lamont_r> Keybuk: got a second?
[11:40] <Keybuk> sure
[11:40] <Keybuk> you don't have five though :p
[11:40] <fabbione> mjg59: does that require more kernel love?
[11:40] <kanto_so> sorry
[11:41] <haggai> Mithrandir: here
[11:41] <Kamion> kanto_so: /whois is your friend
[11:41] <Mithrandir> haggai: did you or rene get anywhere with the ooo-vfs dependency we discussed?
[11:42] <bob2> mjg59: woo, cool
[11:42] <haggai> Mithrandir: uuh, what was I supposed to do?
[11:42] <Mithrandir> haggai: make ooo-bin not depend on gnomevfs. :)
[11:42] <haggai> Mithrandir: ah, _rene_'s dirty hack :)  I think he looked at it but I don't remember seeing any checkin
[11:42] <Mithrandir> I think you possibly punted it to _rene_ 
[11:43] <Mithrandir> ok
[11:43] <Mithrandir> _rene_: around?
[11:43] <haggai> no, he'll be at college
[11:43] <Mithrandir> ok.
[11:43] <Mithrandir> I'll take a look, then.. *grumble*  ooo takes _forever_ to build.
[11:44] <mjg59> fabbione: Nope
[11:44] <haggai> Mithrandir: if you want to look then lets discuss it together
[11:44] <mjg59> fabbione: One thing it /might/ be worth looking at in future is the experimental USB suspend/resume support
[11:44] <haggai> Mithrandir: where are you?
[11:45] <haggai> mjg59: the network module unload stuff didn't work for my card because there is no driver entry in sys for the card
[11:45] <lamont_r> jdub: on further thought and discussion...  I'm going to redo that a bit...
[11:45] <Mithrandir> haggai: hack room.
[11:45] <Mithrandir> haggai: you're downstairs?
[11:45] <haggai> Mithrandir: yes
[11:46] <Mithrandir> I'll come down
[11:46] <haggai> ok
[11:46] <Keybuk> thom: it looks like simply issuing udevstart in the udev init script is sufficient to cause hotplug to happen \o/
[11:46] <mjg59> haggai: Hm. How so?
[11:46] <thom> Keybuk: score!
[11:47] <elmo> GOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOAAAAAAAAAAAAAAAAAAAAAAAAAAAL
[11:47] <elmo> doh.
[11:47] <elmo> AL
[11:47] <mjg59> haggai: Built in or PCMCIA?
[11:50] <jdub> lamont_r: :-)
[11:52] <robtaylor> hmm
[11:55] <robtaylor> so. what are the plans for tonight? i'll be getting into mataro some point this evening :)
[11:59] <fabbione> mjg59: yes.. i saw the option but the help text scared me a bit to enable it right now
[12:00] <fabbione> is mdz close to anybody?
[12:02] <Treenaks> Keybuk: it's better to remember it now than at some other time at which we'd have to hit you...
[12:03] <Keybuk> *whimper* don't hit me ... I'm fragile
[12:07] <Keybuk> ok...
[12:13] <silbs> fabbione: I'm sitting next to mdz - need him?
[12:14] <fabbione> silbs: can you ask him to join irc when he has time please?
[12:16] <fabbione> hey mdz
[12:16] <fabbione> mdz: the scsi fix appears not to be the only one needed
[12:17] <fabbione> willy has been locking at it too
[12:17] <fabbione> i might as well ask him later today (when he is awake) what's the best one around
[12:19] <fabbione> anyway the patch we found is applied
[12:19] <mdz> ok
[12:19] <mdz> I am in the middle of a BOF at the moment
[12:20] <fabbione> ok
[12:21] <daniels> Keybuk: how'd it go?
[12:21] <Keybuk> ok ... so it helps if I actually start udevd in the initscript
[12:23] <Keybuk> though I don't see why udevstart isn't starting it
[12:25] <rburton> daniels: so the fd.o bugzilla isn't mailing me when there are changes to bugs i'm CCd on
[12:25] <daniels> rburton: interesting, it should be
[12:25] <daniels> @burtonini?
[12:25] <rburton> yeah
[12:26] <daniels> daniels@gabe:~% sudo grep burtonini /var/log/mail.log
[12:26] <daniels> Dec  8 12:35:48 gabe postfix/smtp[5691] : 6B9659E8CE: to=<ross@burtonini.com>, relay=mail.burtonini.com[195.54.228.42] , delay=4, status=sent (250 Ok: queued as C4AF7581A)
[12:26] <seb128> daniels, just use "ross" you will get a list of the names matching it
[12:26] <daniels> i count 4 mails that have gone to you in the last 24h
[12:26] <seb128> ouips
[12:26] <seb128> nm
[12:26] <daniels> seb128: yeah, and all the spam that uses generic names
[12:26] <daniels> Dec  9 00:42:42 gabe postfix/smtp[1383] : B793B9E7C7: to=<ross@keithp.com>, relay=mail.keithp.com[63.227.221.253] , delay=14, status=bounced (host mail.keithp.com[63.227.221.253]  said: 550 Unknown local part ross in <ross@keithp.com> (in reply to RCPT TO command))
[12:27] <rburton> daniels: argh
[12:27] <rburton> thom: ping?
[12:29] <daniels> rburton: he's next to me
[12:30] <rburton> daniels: poke him and tell him to check amnesiac
[12:30] <haggai> ok folks, quick survey please
[12:31] <haggai> run: dmesg | grep "Simple Boot"
[12:31] <haggai> if you do not see a message like
[12:31] <haggai> Simple Boot Flag at 0x<foo> set to 0x1
[12:31] <haggai> you are not benefiting from the fastboot spec and we'd like to know why
[12:32] <Mithrandir> : tfheen@thosu ~ > dmesg | grep "Simple Boot"
[12:32] <Mithrandir> Simple Boot Flag at 0x35 set to 0x1
[12:32] <Mithrandir> X40 love!
[12:32] <haggai> the 0x1 means you have a PNP Bios and everything is fine and fastboot will happen next time you boot
[12:32] <daniels> Mithrandir: simple boot?
[12:32] <lamont_r> mdz?
[12:32] <daniels> Simple Boot Flag at 0x35 set to 0x1
[12:33] <daniels> x40 4 lyf
[12:33] <Mithrandir> daniels: see what haggai wrote.
[12:33] <fabbione> x40 is the gayest lappy i have ever seen
[12:33] <lamont_r> daniels: you know that you're going to fall in love with a new model long before you die...
[12:33] <fabbione> even daniels has one...
[12:33] <fabbione> i mean
[12:33] <mdz> lamont_r: ?
[12:33] <Treenaks> hmm... mozilla people are _weird_
[12:33] <Treenaks> http://forums.mozillazine.org/viewtopic.php?t=5306
[12:33] <lamont_r> elilo Depends: efibootmgr.
[12:33] <lamont_r> so I think we seed elilo, and it all works out from there...
[12:34] <ogra> haggai: if your system runs already a little loger it should be: cat  /var/log/dmesg | grep "Simple Boot"
[12:34] <rburton> what's fastboot?
[12:34] <haggai> ogra: heh, good point
[12:34] <lamont_r> rburton: that's where things boot fast.
[12:34] <rburton> lamont_r: aah i see ;)
[12:34] <haggai> rburton: tells the BIOS that everything's good and not to bother checking/initing the hardware -> faster boot
[12:34] <lamont_r> rburton: effort is being made to speed up the users boot experience
[12:35] <mdz> lamont_r: ok, sounds good.  go ahead and add it to the seed
[12:35] <lamont_r> will do.
[12:35] <rburton> haggai: oh nice. is this only in hoary?
[12:36] <lamont_r> rburton: as opposed to?
[12:36] <rburton> warty and sid
[12:37] <lamont_r> dunno about sid, but warty is _DONE_.
[12:37] <rburton> yeah, i was wondering if fastboot was in warty, or can i turn it on, or what
[12:37] <mjg59> ARGH.
[12:38] <lamont_r> the hardware/bios fastboot is there regardless of the OS...  setting it is the only question
[12:38] <fabbione> mdz: what is the bug number you opened on bugzilla for the scsi stuff?
[12:38] <rburton> lamont_r: ah i see.
[12:40] <haggai> rburton: that's what I want to know.. how widespread is the support and where it is not already working
[12:40] <haggai> rburton: if you get the message then you are ok
[12:41] <rburton> i seeee
[12:41] <rburton> haggai: my X22 sets 0x35 to 0x1
[12:42] <Mithrandir> haggai: wrapping one-line code in #ifdef USE_GNOMEVFS feels so useful. ;)
[12:44] <thom> rburton: bugzilla on freedesktop sending bad headers
[12:44] <rburton> thom: thanks
[12:44] <haggai> Mithrandir: hehe
[12:45] <rburton> thom: hadess mailed me last night too, can you grep for hadess? :)
[12:45] <haggai> rburton: as I said, that's fine
[12:45] <thom> rburton: daniels is gonna look when he stops playing with fonts, i think
[12:51] <daniels> RESOLVED/WORKSFORME
[12:53] <lamont_r> thom: ECHAN?
[12:53] <lamont_r> in the other one, that is.
[12:53] <daniels> thom: In-Reply-To: <bug-2039@bugs.freedesktop.org>
[12:53] <daniels> thom: hth hand kthxbye
[12:54] <mjg59> daniels: vm86-video-post now actually gets the PCI ID calculation right
[12:54] <mjg59> I've still no idea why it fucks up on so many machines, though
[12:55] <mjg59> daniels: But you should be able to disable VBERestore in X, I hope
[12:55] <daniels> thom: should be a test comment flowing through now
[12:55] <daniels> mjg59: how were you doing the pci id calculation?
[12:55] <mjg59> daniels: Shifting one bit too far to the left
[12:55] <Mithrandir> haggai: I should be able to rip out both the section below "hack" and most of the one after the next comment as well, right?
[12:56] <mjg59> Which gave nice looking answers that were wrong
[12:56] <mjg59> The answers now look ugly, but are right
[01:02] <thom> lamont_r: ah, yeah
[01:05] <daniels> rburton: thom and I are looking at it
[01:06] <daniels>   5702 /home/daniels/public_html/xorg-trace-file
[01:06] <daniels>   1093 xorg.trace-1303
[01:07] <rburton> that's just silly
[01:07] <lamont_r> rburton: that it was that wasteful before, certainly. :-)
[01:08] <lamont_r> although I have to wonder exactly what daniels is ripping out.......
[01:09] <rburton> lamont_r: i think before he started he put some loops in which opened files repeatedly so it looks good when he removes them
[01:10] <lamont_r> heh
[01:12] <seb128> daniels, should I open a bug (as a reminder) against X.org about that m"t"'$tm resolution on my laptop or that's ok and you keep it in mind ?
[01:13] <Treenaks> seb128: m"t"'$tm resolution?
[01:13] <seb128> Treenaks, don't bother, that's for daniels 
[01:15] <Keybuk> hmm... nothing I'm poking into udevsend seems to be coming out in udev
[01:15] <daniels> seb128: i'll close it as a duplicate of 'i855: catch-all for terrible VBE implementations'
[01:15] <daniels> seb128: it's really not something we can fix, sadly
[01:16] <daniels> not without NDAs from 40 different BIOS vendors
[01:16] <daniels> lamont_r: silly sleeps from i810, and kicking the current loader in the testicles
[01:16] <seb128> is there any way to workaround it on my box with this 855resolution stuff ? 
[01:16] <daniels> seb128: apparently, but if it doesn't work ... i don't know
[01:16] <seb128> ok, thanks
[01:17] <seb128> I'll google a bit about it, perhaps there is some useful pointer hidden somewhere :)
[01:17] <daniels> sorry I'm not more useful
[01:17] <seb128> np, not your fault
[01:17] <Keybuk> (yay at disgusting sandwiches ... FOOD FIGHT!)
[01:19] <bob2> lunch was one thing oxford did well
[01:20] <Keybuk> ENOSUMMERPUDDING ?
[01:20] <elmo> that and the Crowded House in the lobby
[01:21] <bob2> don't dreaaaaaaaaaaaaaaam it's over
[01:28] <plovs> mdz, did you report the bug to zwiki, if not i will
[01:29] <daniels> goddamnit
[01:29] <daniels> do not talk about summer pudding unless you are buying it for me
[01:33] <haggai> Mithrandir: yup
[01:34] <haggai> Mithrandir: it's called DEB_BUILD_OPTIONS=ccache
[01:35] <haggai> Mithrandir: can you punt me your mods over and I'll check them?  I could even test them against my prebuilt tree
[01:36] <Mithrandir> just a sec
[01:42] <Mithrandir> haggai: http://err.no/tmp/rules and http://err.no/tmp/vfs-uno-uri.diff
[01:43] <mjg59> daniels: Does dropping VBERestore significantly help things, then?
[01:44] <daniels> mjg59: ah, right, heh.  i'd imagine we could ditch vberestore in that case, then.
[01:46] <mjg59> I'll need a test machine to be sure of this, but I think it can probably be dropped if my stuff actually works
[01:46] <mjg59> (which it seems to)
[01:52] <Mithrandir> haggai: *prod*
[01:52] <Keybuk> syndicate udev: run_program: running /etc/hotplug.d/default/default.hotplug
[01:52] <Keybuk> \o/
[01:54] <thom> nice
[01:54] <Mithrandir> thom/keybuk: could either of you prod haggai to look at his IRC screen?
[01:55] <Keybuk> nowhere near him
[01:58] <Mithrandir> k
[01:58] <thom> Mithrandir: he's away from his laptop
[01:59] <Mithrandir> how useless
[02:01] <seb128> thom, where is everybody ?
[02:01] <thom> "everybody"?
[02:01] <seb128> 2 people in the BOF room atm
[02:01] <seb128> where are you guys :)
[02:01] <thom> daniels, mvo, me are in the downstairs room
[02:01] <thom> quiet room in guess?
[02:02] <seb128> probably yep
[02:16] <Kamion> only three in the quiet room
[02:17] <sjoerd> 6 in the hacking room...
[02:20] <Keybuk> thom: did you know that hotplug doesn't invoke the hal or udev-fu when it starts?
[02:24] <thom> Keybuk: how do you mean?
[02:24] <Treenaks> should "dbus-monitor --system" work?
[02:25] <Mithrandir> Treenaks: dbus-monitor is broken, iirc
[02:25] <Keybuk> well, because it execs the agent directly it never runs anything else in /etc/hotplug.d
[02:25] <Treenaks> Mithrandir: that explains it
[02:27] <daniels> Dear Bong, 
[02:27] <daniels> Your business with IBM Developer Kit for Linux, Java 2
[02:27] <daniels>             Technology Edition is very important to us.  We are sorry that your 
[02:27] <daniels> previous transaction with us did not go through.  We have fixed this 
[02:27] <daniels> error in your account.  Please return to the IBM Developer Kit for Linux, Java 2
[02:27] <daniels>             Technology Edition web site to resubmit 
[02:27] <daniels> your transaction.  Thank you and sorry for the inconvenience.
[02:27] <thom> rofl
[02:27] <thom> Keybuk: hrm, interesting
[02:28] <Keybuk> was the the Aieezheikawopostan one?
[02:28] <daniels> Treenaks: in warty, maybe (probably not); in hoary, yes
[02:28] <daniels> Mithrandir: works in hoary
[02:28] <daniels> Keybuk: azerbaijan?
[02:28] <daniels> yeah, that was azerbaijan
[02:28] <daniels> they never mailed my so-called australian one, though
[02:29] <daniels> (or was it US?)
[02:31] <Treenaks> daniels: in hoary it's b0rken
[02:31] <Treenaks> daniels: it tells me:
[02:31] <Treenaks> 8094: arguments to dbus_connection_send_with_reply_and_block() were incorrect, assertion "(error) == NULL || !dbus_error_is_set ((error))" failed in file dbus-connection.c line 1999.
[02:31] <Treenaks> This is normally a bug in some application using the D-BUS library.
[02:31] <Treenaks> Failed to set up match "dbus-monitor": 
[02:37] <lamont_r> Kamion: now we're up to not finding d-i modules:-)
[02:40] <Mithrandir> rock, "Welkom bij Mozilla Thunderbird"
[02:42] <trulux> doko, ping
[02:43] <Mithrandir> Treenaks: I'd appreciate if you could test m-t-l-nl when it hits the archive (approx 30min to 1hour, I think.)
[02:44] <Treenaks> Mithrandir: ok, I think
[02:53] <thom> mjg59: dude, you gonna package netapplet 1.0?
[02:53] <mjg59> thom: Oh, is it out?
[02:53] <mjg59> It is. Hot damn.
[02:54] <mjg59> Still without a changelog
[02:54] <thom> i love answers like that
[02:54] <thom> heh
[02:58] <mjg59> thom: Only difference to 0.99.4 seems to be gettext
[02:58] <mjg59> Oh, and netline has been removed
[02:59] <mjg59> Yeah, basically it's just translations between 0.99.4 and 1.0
[03:00] <trulux> lamont, ping
[03:05] <Keybuk> man... 14.8s -> 0.4s ... I RULE!
[03:06] <Treenaks> Keybuk: 0.4s to generate an error message? :)
[03:06] <Keybuk> nope, is actually loading all the PCI modules and stuff
[03:06] <thom> DUDE
[03:07] <ogra> Keybuk: you rule !
[03:07] <rburton> Keybuk: hotplug?
[03:08] <Keybuk> yeah
[03:08] <rburton> rock
[03:08] <rburton> i get totally bored when hotplug starts on my x22
[03:09] <ogra> rburton: so get a faster laptop with less hw :-P
[03:09] <rburton> ha
[03:09] <ogra> :)
[03:12] <lamont_r> trulux: yo
[03:12] <trulux> hey lamont_r !
[03:12] <trulux> lamont_r, i've got an ubuntu-running box
[03:13] <trulux> P4, a lot of mem, 120G hd, only for Ubuntu and sarge
[03:14] <trulux> lamont_r, i have a few questions about Ubuntu, mainly why fonts look weird in the crt screen
[03:14] <trulux> i want to get a lcd panel for that box but still using a hp mx70
[03:14] <lamont_r> those would be #ubuntu questions
[03:16] <trulux> kay, then next questions about dev
[03:16] <trulux> btw, my cd was broken and i had a real pain using it
[03:16] <trulux> lamont_r, the selinux packages create conflicts with gnome
[03:17] <trulux> tseng, ping
[03:17] <lamont_r> bummer.  selinux isn't in main
[03:18] <trulux> also, i think i'm going to send a bug report about adding some patches to the linux-sources
[03:18] <trulux> and maybe about bootsplash, which i've made some optimized pkgs
[03:19] <Kamion> we're not going to be using bootsplash.
[03:19] <trulux> in the other hand, usb support seems broken in my box, my slim disk (ide-to-usb) doesn't work (it's just a simple usb-storage compliant dev)
[03:20] <trulux> Kamion, why? it's eye-candy for new users
[03:20] <Kamion> bootsplash broke our installer.
[03:20] <Kamion> last time we tried this
[03:20] <Kamion> we're doing a pure user-space solution instead
[03:20] <Mithrandir> trulux: we'll use other, better crack
[03:20] <fabbione> trulux: what kernel are you running?
[03:20] <trulux> i mean not in the installer, i mean in the linux-sources pkgs
[03:20] <Kamion> the installer uses the stock kernel; we are not going to change that, sorry
[03:20] <trulux> fabbione, 2.6.8.1-p4-1 (my own)
[03:20] <trulux> Kamion, ok
[03:21] <fabbione> trulux: if it's broken why would you open a bug on our linux-source?
[03:21] <trulux> bootsplash?
[03:21] <trulux> no, it's about it
[03:21] <fabbione> WONTFIX
[03:21] <fabbione> bootsplash will stay outside the kernel
[03:21] <trulux> i mean about other things, mainly security-related patches
[03:21] <fabbione> trulux: such as which one?
[03:22] <trulux> fabbione, kay, i'm not talking about it now
[03:22] <fabbione> trulux: i am the temporary kernel maintainer
[03:22] <trulux> fabbione, openswan ipsec stack
[03:22] <fabbione> so either you tell me what is about
[03:22] <fabbione> that's not security
[03:22] <fabbione> these are enanchment
[03:22] <trulux> grsecurity (when it comes available)
[03:22] <trulux> fabbione, yep
[03:22] <fabbione> but it's not security
[03:22] <trulux> fabbione, let me find you the patches
[03:23] <fabbione> ipsec is already in the kernel dude
[03:23] <Mithrandir> ah, new X crack.
[03:23] <fabbione> Mithrandir: yup
[03:23] <fabbione> finally it will unlock a bunch of packages on sparc
[03:24] <fabbione> Kamion: want to start adding sparc too?
[03:24] <Mithrandir> lamont_r: what's the problem with mono stuff?
[03:24] <trulux> fabbione, not the openswan stack
[03:24] <trulux> kernel ipsec stack is different
[03:24] <lamont_r> Mithrandir: source package?
[03:24] <Mithrandir> lamont_r: mono
[03:25] <lamont_r> 2.6 kernel openswan was using native kernel stuff (no patch), I thought
[03:25] <Kamion> fabbione: will do later, want to check with elmo what'll happen if I seed sparc stuff though
[03:25] <Mithrandir> lamont_r: unupgradeable since a few days, you mumbled "universe" when somebody talked about it two days ago
[03:25] <lamont_r> Mithrandir: right.  let me check
[03:25] <fabbione> lamont_r: so did I
[03:25] <fabbione> Kamion: sure :-)
[03:25] <fabbione> probably the implementation is different, but for christ sick the RFC's are the same
[03:27] <fabbione> actually i should ask Iari about it
[03:30] <lamont_r> Mithrandir: mcs is version 1.0.2, needs to be 1.0.4, it would appear.
[03:31] <Mithrandir> lamont_r: ok, so Debian's broken as well?
[03:31] <lamont_r> guess.
[03:34] <lamont_r> sure looks that way
[03:38] <lup|gone> daniels you there?
[03:39] <daniels> lup|gone: sup
[03:40] <lup|gone> I was looking at your bootcharts
[03:40] <daniels> right
[03:40] <lup|gone> and I wondered why like windows you can't start network stuff after X is started
[03:47] <daniels> a large part of it is hostname changing, but we can hack around it
[03:47] <daniels> we've been seriously considering it
[03:48] <lup|gone> yeah
[03:48] <lup|gone> seems like almost 10 sec
[03:48] <lup|gone> the pc is doing almost nothing :s
[03:50] <thom> with udevsend that becomes async
[03:51] <sjoerd> thom: going to do ifup from dev.d ?
[03:52] <thom> sjoerd: i think that's gonna be how it winds up, yeah
[03:52] <sjoerd> nice
[03:52] <thom> scott's boot currently is "interesting"
[03:52] <Mithrandir> thom: how so?
[03:52] <Treenaks> ja
[03:52] <Treenaks> uh
[03:52] <Treenaks> oops
[03:53] <thom> Mithrandir: a stack of stuff now runs before their modules are loaded ;-)
[03:54] <Mithrandir> heh
[03:54] <lup|gone> does hotplug start the scripts parallel ?
[03:55] <sjoerd> udevd doing hotplug stuff serializes the stuff that depends on eachother.. the rest runs parallel
[03:56] <lup|gone> nice
[04:01] <Mithrandir> elmo_away: please sync autofs
[04:02] <Mithrandir> anybody know where elmo is?
[04:04] <Kamion> Mithrandir: he's not feeling well; went to his room for a nap
[04:05] <Mithrandir> ok
[04:05] <Mithrandir> I guess saying "please sync" and assigning the bug to him ought to work, then.
[04:14] <seb128> elmo_away, gnome2-user-docs sync please
[04:17] <Treenaks> sjoerd: do it on the wiki
[04:25] <daniels>   1093 xorg.trace-1303
[04:25] <daniels>    877 xorg.trace-1625
[04:25] <bob2> hah
[04:25] <daniels> that's a micro-optimisation at this point, but ber
[04:26] <bob2> 20%!
[04:26] <daniels> thom: .11/~daniels/Xorg, as usual
[04:26] <thom> noice
[04:26] <thom> gonna finish bootchart packaging
[04:27] <thom> is there a "real" way of modifiying inittab?
[04:27] <daniels> in postinst/whatever?
[04:27] <thom> yeah
[04:28] <daniels> 'echo'
[04:28] <thom> i don't really want to just apply a patch in the postinst
[04:28] <thom> but i guess i will
[04:29] <daniels> and we ditch another 80 open()s (not stat()s) by sanitising font configuration
[04:29] <lup|gone> so when gdmgreeter ends the user sees the gdmlogin window?
[04:29] <jdub_> I CAN SMELL ROCKING :)
[04:29] <thom> lup|gone: basically yes
[04:30] <daniels> (and another 100 by sanitising extension configuration ...)
[04:30] <daniels> lup|gone: yeah
[04:30] <thom> we're not gonna get much faster than 35-40 secs on the x40 i think
[04:30] <daniels> thom: .11/~daniels/xorg.conf for superstreamlined configuration
[04:30] <daniels> (with no functionality loss)
[04:30] <mjg59> thom: 35-40 seconds until what?
[04:30] <daniels> mjg59: boot -> gdm login screen
[04:30] <mjg59> Ok
[04:30] <thom> mjg59: start of rcS -> end of rc2
[04:31] <mjg59> But gdm could be moved further up rc2?
[04:31] <thom> mjg59: it's at 13 already
[04:31] <mjg59> thom: Uh. So is 40 seconds end of rc2 or appearance of login window?
[04:32] <thom> mjg59: there's maybe a second's difference between those here
[04:32] <mjg59> thom: Really? Wow.
[04:32] <mjg59> But that's now in parallel, rather than in series, I guess
[04:33] <mjg59> Are you doing any readahead yet?
[04:34] <Keybuk> udevd is cute, it parallelises events on different DEVPATH, but serialises events on the same
[04:34] <thom> mjg59: http://www.planetarytramp.net/bootchart/bootchart-20041209-1318.png
[04:35] <daniels> hmm, down to 571 now
[04:35] <mjg59> (Incidentally, how does this compare to FC?)
[04:35] <thom> mjg59: we're smacking the crap out of the numbers we can find
[04:35] <daniels> that's a 90% saving
[04:35] <mjg59> thom: That's some pretty fucking solid disk i/o
[04:35] <thom> mjg59: yeah
[04:35] <mjg59> Any idea how much of it's thrash?
[04:36] <thom> mjg59: i imagine much of the readahead is seek, but besides that i can't really tell
[04:36] <mjg59> And this is on an X40?
[04:36] <mjg59> With a 4200RPM disk?
[04:36] <thom> yep
[04:36] <mjg59> Man
[04:36] <mjg59> How's the readahead configured?
[04:36] <Keybuk> thom: fuck yeah, cupsd's init is totally nuked off that :p
[04:37] <thom> pretty manually right now - we've straced X, gdm, gdmgreeter, cupsys and worked out what they're opening
[04:37] <lup|gone> daniels you said in your blogs about some issues with gdm speed what do you exactly mean by that?
[04:38] <mjg59> Hrm. Should the readahead be carrying on past rc2?
[04:38] <thom> mjg59: bit under a thousand files in the readahead
[04:38] <thom> ideally not
[04:38] <mjg59> That looks like it'd thrash heavily for 7 seconds or so
[04:38] <mjg59> Eurgh. Yeah, you're getting readahead during X startup.
[04:38] <mjg59> That can't be good.
[04:39] <thom> you'll note that there's no network startup in that chart - with that readahead and network finish about the same time
[04:39] <mjg59> Maybe rc2 should block until readahead is done?
[04:39] <thom> yeah, i think something like that
[04:39] <mjg59> Just drop something into rc2.d at the start
[04:39] <mjg59> I love how ipw2100 is a good 2 seconds of D state
[04:40] <mjg59> Must be firmware loading
[04:40] <thom> you should see scott's soundcard load
[04:40] <thom> there's a bootchart-keybuk-1.png in that dir
[04:40] <mjg59> keybuk's bootchart is teh suckitude
[04:41] <thom> yeah
[04:41] <mjg59> Oh, the 10 seconds of nothing are waiting for DHCP...
[04:41] <mjg59> That should so be backgrounded, and individual scripts block if necessary
[04:41] <daniels> lup|gone: it seems to take a *lot* of CPU
[04:41] <daniels> mjg59: right
[04:41] <Keybuk> oh, dhcp is gone from my boot now :p
[04:41] <thom> we're talking about dependency init
[04:41] <thom> it's where this is going, totally
[04:41] <mjg59> HOLY SHIT
[04:42] <mjg59> How does it manage 10 seconds of module loading?
[04:42] <Keybuk> *shrug* that alsa driver sucks
[04:42] <thom> that's the soundcard of death
[04:42] <mjg59> What's it trying to do? Copy the entire OS over a single wire bus?
[04:43] <Keybuk> there's probably sleep(10) in there or something
[04:43] <lup|gone> I wonder if hotplug can't start parallel to X
[04:43] <daniels> no, it's spinning the entire time, innit?
[04:43] <mjg59> I'm amused at how you're micro-optimising it on machines that are probably going to be rebooted something like once a month in normal usage
[04:43] <azeem> it plays the free software song muted to check the card, I assume
[04:43] <daniels> lup|gone: think: sound
[04:43] <mjg59> daniels: Kernel-space sleep spins
[04:43] <lup|gone> why sound?
[04:43] <mjg59> On the other hand, you can't sleep for more than 0.1 seconds in the kernel now
[04:43] <mjg59> There's a magic macro that breaks it
[04:43] <lup|gone> gdmgreeter needs sound
[04:43] <lup|gone> X doesn't I think?
[04:44] <daniels> down to 538.  if I attempt to optimise it further, please smack me in the head.
[04:44] <mjg59> daniels: 538 opens?
[04:44] <daniels> lup|gone: nope, X doesn't, but if you start it parallel, then you run the risk of having your GNOME session start before your sound card comes up, which is massive suck.
[04:44] <daniels> mjg59: 538 opens/stats/whatever
[04:44] <daniels> mjg59: as opposed to 5702 originally
[04:45] <daniels> mjg59: x server startup is a *lot* faster
[04:45] <mjg59> thom: Well, David Zuthen is claiming 40 seconds wallclock from grub to desktop
[04:45] <daniels> thom: usual place
[04:45] <daniels> mjg59: yes, but that was bullshit
[04:45] <mjg59> Ah
[04:45] <mjg59> Heh
[04:46] <daniels> mjg59: he replaced init with a small shell script that loaded the modules for his machine
[04:46] <mjg59> Yeah
[04:46] <daniels> didn't bother with trivial services such as network
[04:46] <daniels> and slammed 193MB into the page cache
[04:46] <mjg59> And it starts little else
[04:46] <mjg59> But it gives possibilities
[04:46] <daniels> also, not starting the gdm greeter is a pretty big win
[04:46] <thom> yeah
[04:47] <thom> gdmgreeter is pretty hellish - it has a stack of stuff that it needs to load
[04:47] <lup|gone> daniels that is correct but is there no way to start X and then start gnome-session if sound is found or all devices have been detected by hotplug?
[04:47] <daniels> (the other problem with gdmgreeter is that right now, it needs to scale down a 1600x1200 PNG and then overlay another PNG with an alpha channel on top of it)
[04:48] <daniels> (this is not terribly cheap.)
[04:48] <mjg59> daniels: You've dropped VBERestore now?
[04:48] <Keybuk> can't we start the sound-server as part of the boot process?
[04:49] <thom> mjg59: my graphs are all with vberestore on
[04:49] <daniels> mjg59: i'm getting quite blinding speed without it
[04:49] <daniels> Keybuk: isn't esd per-user?
[04:49] <mjg59> daniels: Dude, drop VBErestore
[04:49] <daniels> mjg59: talk to thom
[04:49] <mjg59> We can do it in the ACPI path
[04:49] <mjg59> thom: DROP VBERESTORE
[04:49] <daniels> thom: (YOU MUPPET)
[04:50] <daniels> thom: (NOT REALLY, BUT IT WAS ENTERTAINING)
[04:50] <Keybuk> daniels: sure, but it should be per-machine ... different users don't have different soundcards normally :)
[04:50] <daniels> Keybuk: polyp?
[04:50] <daniels> Keybuk: anyway, is there IBM power over there I can steal?
[04:50] <Keybuk> thom's on battery
[04:50] <thom> bob2?
[04:50] <daniels> me too ... all 3% of it
[04:50] <Keybuk> if you can take mdz's plug, you could steal his I guess :p
[04:50] <daniels> his stuff appears to be elsewhere
[04:51] <daniels> Keybuk: who's is that?
[04:51] <Keybuk> mdz
[04:51] <daniels> ah
[04:51] <daniels> well, mdz volunteers, I guess :)
[04:51] <bob2> thom: ?
[04:51] <mjg59> thom: The stuff I hacked up yesterday lets us do the VBE save/restore on our way through suspend/resume, rather than doing it in general
[04:52] <mjg59> If it turns out that we still need it in X space, the right thing to do is hack it into the X suspend/resume code rather than do it on every VT switch
[04:52] <daniels> right
[04:52] <mjg59> (running my stuff while X is running kills X)
[04:53] <thom> oh, sweet
[04:53] <mjg59> vbestate save >foo
[04:54] <mjg59> vbestate restore <foo
[04:55] <pitti> o
[04:56] <thom> sans vberestore
[04:56] <bob2> ffoh, I'm on battery too
[04:56] <daniels> ffoh?
[04:57] <Keybuk> hmm... I wonder why things like ntpdate and mountnfs aren't done in if-up.d
[04:57] <bob2> ff = hitting keys to see if the network is fucked
[04:58] <mjg59> Keybuk: Because, uh, erm...
[04:59] <mjg59> (because that's not the way things are done, damn it)
[04:59] <Keybuk> heh
[04:59] <Keybuk> they are now :p
[05:01] <daniels> bob2: ah
[05:03] <thom> we take 12 seconds to mount and uncompress the initrd
[05:04] <thom> we so should try with an uncompressed one
[05:04] <daniels> yeah
[05:04] <daniels> i was going to say, maybe it's worth shooting for no compression if it's viable
[05:04] <fabbione> thom: just check cramfs options
[05:04] <fabbione> that's easy to hack
[05:05] <daniels> hm, so gdmgreeter uses a *lot* of cpu
[05:06] <thom> i think for us we're probably safe not to bother
[05:06] <rburton> daniels: considered trying a pure SVG theme to avoid the PNG rescale? you'll be trading one complex algorithm for another, but people claim SVG icons are faster than PNG in the scheme of things in gtk
[05:06] <thom> ber, that run got totally bitten by laptop mode
[05:07] <daniels> rburton: i'm trying something now
[05:07] <daniels> thom: hold up
[05:08] <daniels> thom: humour me -- replace /usr/share/gdm/themes/Human/background.png with .11/~daniels/background.png
[05:09] <daniels> thom: it's not perfect (positioning is off), but I want to see how much we win by not rescaling and overlaying
[05:09] <daniels> i.e. how much of it is needing to optimise gdmgreeter to hell
[05:12] <daniels> thom: btw, remove xfonts-base-transcoded if you have it
[05:13] <thom> daniels: background.png not on .11/~daniels/
[05:14] <daniels> daniels@catsby:~/public_html% sudo mv background.png /usr/share/gdm/themes/Human
[05:14] <daniels> daniels@catsby:~/public_html% cp /usr/share/gdm/themes/Human/background.png .
[05:14] <daniels> lesson learnt.
[05:14] <thom> heh :-)
[05:14] <thom> ok, lets try this again with no laptop-mode
[05:15] <daniels> thom: DOUBLE THE FIST
[05:15] <lup|gone> using imlib is out of the question? :) (I heard it was very fast for image handling)
[05:15] <daniels> to be fair, raster's tests were written to show up render and make imlib look awesome
[05:17] <rburton> gdkpixbuf shouldn't be way slower than imlib
[05:18] <daniels> thom: .11/~daniels/Human.xml
[05:18] <daniels> thom: and dude, ditch ntpdate
[05:21] <maswan> daniels: dude, if you don't ntpdate, your clock might be way off.
[05:21] <daniels> maswan: ntpd
[05:22] <daniels> having ntpdate block sucks
[05:22] <tseng> openntpd++
[05:22] <tseng> i should package that
[05:22] <daniels> (it's blocking for a good 10-15sec)
[05:22] <maswan> daniels: Is that one less chicken of adjusting the clock multiple hours at startup?
[05:22] <maswan> .. these days
[05:22] <daniels> maswan: huh?
[05:22] <daniels> oh, right
[05:22] <thom> bootchart-20041209-1721.png
[05:23] <daniels> it won't do multiple hours per default, but I think it could be configured to do so
[05:23] <haggai> does anyone speak danish?
[05:23] <Clint> subverting ntp isn't such a good idea
[05:23] <maswan> that's the classical reason for running ntpdate at boot, because if the hwclock says 1970-01-01 00:00...
[05:23] <daniels> haggai: very, very badly, but I have a phrasebook on my room
[05:23] <daniels> s/on/in/
[05:23] <maswan> but if ntpd can be fixed to do large adjustments at boot/startup, that's good.
[05:24] <maswan> haggai: it is close enough to swedish for me to usually be able to read it.
[05:24] <haggai> daniels: I'm looking at 2 translations and wondering which one is the most up to date
[05:24] <haggai> "En logind-hndtering er et program, der giver mulighed for grafisk logind "
[05:24] <haggai> is it "En" or "Den" ?
[05:24] <haggai> (at the beginning)
[05:24] <daniels> haggai: i don't have the grammar to tell you, sorry
[05:24] <maswan> "En", most likely.
[05:25] <haggai> hmm, ok I was expecting the other way
[05:25] <maswan> hmm.. yeah, "_A_ login-thingie is _a_ program"
[05:25] <haggai> "A display manager is a program that provides graphical login capabilities "
[05:25] <haggai> was the original
[05:25] <maswan> yeah
[05:26] <haggai> Please select which display manager should "
[05:26] <haggai> "run by default.""run by default."
[05:26] <haggai> 1. "som udgangspunkt skal kres."
[05:26] <haggai> 2. "skal kres som standard."
[05:27] <haggai> 1 or 2 ?
[05:27] <thom> i'm gonna look at caching /dev
[05:27] <daniels> ok, so we're looking at around 40 seconds so far
[05:27] <daniels> (that's subtracting ntpdate)
[05:27] <maswan> haggai: Hmm.. That's a hard one, it seems to say about the same thing.
[05:28] <daniels> /dev seems to be a major hit (in the order of 5-7 seconds)
[05:28] <haggai> maswan: ok, thanks
[05:28] <thom> don't wanna rebuild initrd on battery
[05:28] <maswan> haggai: so the change is probably because one is better danish than the other, and I'm not native enough to tell which.
[05:28] <haggai> maswan: heh :) ok, I'll take the one from KDE CVS not the deb package
[05:28] <Keybuk> thom: the main problem is the udev device table
[05:29] <thom> Keybuk: expand?
[05:29] <haggai> maswan: that's the one with the later date, and the En
[05:29] <maswan> haggai: ok, glad I could help
[05:29] <lamont_r> ntpdate is almost entirely I/O bound, could just be backgrounded without hurting things much at all.
[05:29] <Keybuk> you know udev has a table in /dev/.udevdb right? :)
[05:29] <Keybuk> though that should be attarrable
[05:29] <thom> Keybuk: oh, right. yes
[05:30] <lamont_r> Keybuk: which font is that from??
[05:31] <Keybuk> ntpdate is going into if-up.d
[05:31] <Keybuk> no point doing it until the first device pops its head up
[05:31] <lamont_r> even better
[05:33] <Kamion> anyone at the conference got a Thinkpad T42?
[05:33] <lamont_r> doko: gcc-3.3 and 3.4 built for i386. yippee!
[05:33] <Keybuk> Kamion: mdz.
[05:33] <Kamion> bonus, I may hijack him later for testing then
[05:33] <Keybuk> syndicate tmp% sudo time tar xzf ~/dev.tar.gz
[05:33] <Keybuk> 0.05user 0.39system 0:01.24elapsed 36%CPU (0avgtext+0avgdata 0maxresident)k
[05:33] <Kamion> thanks
[05:33] <Keybuk> 0inputs+0outputs (0major+460minor)pagefaults 0swaps
[05:33] <Keybuk> syndicate tmp% sudo time udevstart
[05:33] <Keybuk> 7.38user 2.68system 0:11.79elapsed 85%CPU (0avgtext+0avgdata 0maxresident)k
[05:33] <Keybuk> 0inputs+0outputs (0major+588855minor)pagefaults 0swaps
[05:33] <Keybuk> ^ good 10s there
[05:34] <daniels> Kamion: yeah, mdz has one in front of me
[05:34] <thom> nice
[05:34] <daniels> thom	/boot is calling you
[05:35] <thom> heh
[05:36] <doko> lamont_r: yes, after tcl8.4 built
[05:38] <daniels> thom: paging in /usr/X11R6/bin/Xorg, /usr/X11R6/lib/modules/linux//libint10.a,  /usr/X11R6/lib/modules/libvbe.a, /usr/X11R6/lib/modules/libscanpci.a, /usr/X11R6/lib/modules/libpcidata.a, /usr/X11R6/lib/modules/fonts/libfreetype.a, /usr/X11R6/lib/modules/extensions/libGLcore.a, and /usr/X11R6/lib/modules/extensions/libglx.a highly recommended
[05:38] <lamont_r> doko: yes.  thank you.
[05:38] <thom> daniels: already done
[05:38] <daniels> thom: rockin'
[05:38] <daniels> thom: those are the big ones
[05:38] <thom> grabbed from the straces
[05:39] <daniels> ah, one step ahead
[05:39] <daniels> is it just me, or does anyone else in the NM bof think mdz is posing like a life model atm?
[05:39] <thom> rofl
[05:40] <thom> we need a photo to go with the line on quotes
[05:40] <Kamion> daniels: haha
[05:40] <Kamion> must be the turtle-neck
[05:40] <thom> lamont_r++
[05:43] <lamont_r> Mithrandir: swig1.3 is still ftbfs on amd64 (pike7.4 issues)
[05:44] <lamont_r> 4146
[05:44] <Mithrandir> phun
[05:44] <Mithrandir> ew
[05:46] <Mithrandir> I'll whack at it
[05:54] <Treenaks> Mithrandir: the NL thunderbird is not on crack today
[05:55] <fabbione> daniels: http://debian.linux-systeme.com/dists/unstable/main/source/xorg_6.8.1-0.4_i386.changes
[05:56] <fabbione> daniels: i think i found that guy was mumbling about problems building xorg with us
[05:56] <fabbione> but he seems to have a bunch on patches
[05:56] <fabbione> that we can pull in
[06:00] <Mithrandir> Treenaks: goodie
[06:03] <seb128> elmo_away, libgnomeui sync please
[06:04] <fabbione> seb128: he is sleep in his room.
[06:04] <fabbione> he was not feeling good
[06:05] <seb128> fabbione, ok. Do you think he'll read the IRC messages when he comes back or I should rather note the packages to sync somewhere ?
[06:06] <fabbione> seb128: if he will survive, i would suggest to remind him
[06:07] <seb128> ok
[06:07] <seb128> fabbione, and if he doesn't who should I ping ? :p
[06:09] <fabbione> seb128: god
[06:09] <fabbione> he is the only one one step up elmo
[06:09] <fabbione> ;)
[06:10] <seb128> ah ah
[06:19] <Treenaks> New Global Directory of OpenPGP Keys --> http://yro.slashdot.org/yro/04/12/09/1446203.shtml?tid=93&tid=158
[06:20] <Treenaks> http://www.pgp.com/downloads/beta/globaldirectory/index.html
[06:22] <mxpxpod> lamont_r: ping
[06:22] <seb128> seb128, carlos says ping
[06:22] <seb128> oups
[06:22] <lamont_r> mxpxpod: ack
[06:24] <mxpxpod> lamont_r: ok, I know you got the mount segfault fixed, but my partitions still don't unmount on shutdown because /.dev gets unmounted before everything else
[06:24] <lamont_r> with 2.21j-2?
[06:24] <lamont_r> 2.12j-2 that is
[06:24] <lamont_r> er, -2ubuntu1 or so
[06:25] <mxpxpod> lamont_r: yes
[06:25] <lamont_r> hrmpf
[06:26] <mxpxpod> lamont_r: hold on, let me check my logs
[06:27] <mxpxpod> lamont_r: it's not in my logs, but I know that when it goes to unmount local filesystems, it says it can't find '/'
[06:28] <mxpxpod> lamont_r: I'm also getting a warning on reboot that says I still have a .udevdb in /dev
[06:29] <lamont_r> yeah - I see that one too, but I don't think that's mine...
[06:29] <mxpxpod> ok, just making sure
[06:35] <Mithrandir> seb128: do you mind others nicking packages and merging them underneath you, like eel, bug-buddy and so on?
[06:39] <seb128> Mithrandir, I'm doing merging right now, I'm just closed 4 bugs about that
[06:40] <Mithrandir> ok
[06:45] <lamont_r> mdz??
[06:53] <doko> mvo: http://lists.debian.org/debian-legal/2004/12/msg00017.html
[06:54] <fabbione> mdz: ping
[06:55] <mdz> fabbione: pong
[06:56] <fabbione> mdz: amu was asking about including cloop as part of the kernel since it is required for the LiveCD
[06:56] <fabbione> mdz: do you have objections to it?
[06:56] <fabbione> i think it can live as module and be loaded in the initrd.gz
[06:56] <fabbione> and i also remember that we need to some kind of fs module
[06:57] <fabbione> to simulate writing
[06:57] <mdz> fabbione: yes, we need it
[06:57] <mdz> but no, we do not need the filesystem module
[06:57] <mdz> only cloop
[06:57] <mdz> it's packaged in Debian and should be quite simple, one file
[06:57] <fabbione> mdz: yes i am already looking at the code
[06:58] <fabbione> but since one requirement for hoary was to have teh same kernel for live and normal...
[06:58] <fabbione> ok.. it will be in in -5 :-)
[06:59] <mdz> fabbione: yes, that is the idea
[06:59] <mdz> if we can get the cloop module in, we can use the same kernel
[06:59] <mdz> we are using a different mechanism to make the filesystem writable
[07:00] <mdz> using device-mapper, which is already there
[07:02] <fabbione> mdz: sure i can get it in
[07:02] <fabbione> mdz: -4 with the scsi fix is up
[07:03] <fabbione> at least the one liner we found
[07:03] <fabbione> all the google search didn't report anything else
[07:03] <fabbione> the other crap i found this morning was rc2 -> rc3 related and it shouldn't be an issue for us
[07:05] <thom> bunch new bootcharts up
[07:14] <jdub_> mdz: what was the dm thing you're using for the livecd?
[07:15] <mdz> jdub_: dm-snapshot
[07:15] <jdub_> thanks
[07:16] <jdub_> that keeps an overlay of stuff in ram?
[07:16] <mdz> it does copy-on-write from one device-mapper target to another
[07:16] <jdub_> aha
[07:16] <jdub_> very very nice
[07:16] <jdub_> *very* nice
[07:16] <mdz> I don't think anyone has used it for a live CD yet
[07:16] <mdz> so shhhhh until I get the prototype out :-)
[07:16] <jdub_> :-)
[07:17] <mirak> what must I do to use the ati driver with kernel 2.6.9 ?
[07:18] <jdub_> mdz: it came up in discussion of readahead block optimisation
[07:18] <mirak> is just building the source package of the driver enough ?
[07:18] <jdub_> mdz: craaaaaack :)
[07:24] <lamont_r> mirak: if the modules package for 2.6.9 is there, it's trivial.  if not, then it hasn't been uploaded yet, but is in plan for "real soon now".
[07:29] <mirak> ok
[07:29] <thom> mirak: and this is a question better asked in #ubuntu, by the way :-)
[07:31] <mirak> thom: and not better answered
[07:31] <mirak> if at all
[07:31] <thom> mirak: that's as maybe, but this really is not a support channel
[07:36] <fabbione> mdz: cloop-utils -> main?
[07:38] <fabbione> amu:   Compressed Loop Support (BLK_DEV_CLOOP) [N/m/?]  (NEW) m
[07:39] <fabbione> and now let see if it compiles :-)
[07:39] <jdub> Keybuk: can you come upstairs for a minute?
[07:39] <ogra> hehe
[07:39] <jdub> Keybuk: need to run a dpkg issue by you
[07:39] <lamont_r> jdub: you should point mdz at fabbione's question
[07:39] <jdub> (if Keybuk isn't watching, someone nudge him)
[07:39] <ogra> fabbione: can we see you jumping around again ?
[07:40] <jdub> fabbione: mdz says yes, i'll put it in supported
[07:43] <jdub> thom: ping
[07:45] <daniels> SHIT HOT.
[07:45] <daniels> mjg59: more optimising, this time on the stupidly CPU-heavy parts
[07:45] <daniels> mjg59: X is now a blip on the bootchart -- we're now spending more time in VBE than initialising X and handing over to gdmgreeter by a factor of, oh, 7
[07:46] <Kamion> daniels: how goes l-r-m?
[07:47] <daniels> Kamion: just looking at that now, actually :) i'm pretty confident I can upload this run
[07:48] <Mithrandir> Treenaks?
[07:49] <Treenaks> Mithrandir?
[07:49] <Mithrandir> where are you?
[07:49] <Treenaks> Mithrandir: BOF room
[07:49] <Mithrandir> oh, ok
[07:49] <Mithrandir> I just left.
[07:49] <Treenaks> Mithrandir: where are YOU
[07:49] <Mithrandir> you wanted to sign keys; we can do it afterwards.
[07:49] <Mithrandir> cristal
[07:49] <Treenaks> ok, I can get there
[07:50] <thom> jdub: ack
[07:51] <Mithrandir> *sigh*, keyslips left in my room, give me two minutes.
[07:52] <fabbione> ogra: ehehe
[07:52] <fabbione> jdub: fine for me.. if it is ok with amu
[07:53] <jdub> thom: n/m
[07:55] <fabbione> amu: looks good: compressed_loop.o
[08:00] <Keybuk> keybuk: 1 - hotplug: 0
[08:09] <fabbione> dinner bof
[08:09] <fabbione> :-)
[08:10] <sivang> fabbione : exactly
[08:10] <sivang> fabbione :  :-)
[08:13] <moquist_> i see there's a bounty for Python scripting interfaces for the GIMP.  Has GIMP-Python (http://www.jamesh.id.au/software/pygimp/) been considered?
[08:15] <jamesh> I'm probably not eligible
[08:16] <jamesh> (of course, yosh has been doing most of the pygimp maintaintenance recently
[08:18] <mjg59> daniels: That is so much coolness
[08:18] <mjg59> thom: Keybuk: Is there any reason ntpdate can't be backgrounded?
[08:18] <mjg59> Is there anything in the init scripts that depends on time being monotonic?
[08:19] <mjg59> daniels: ntpd won't step time, it'll just alter the clock speed to get you there gradually. That way, time never goes backwards or jumps too far
[08:19] <mjg59> (see what happens to xscreensaver if you ntpdate from inside X)
[08:19] <daniels> heh :)
[08:19] <daniels> Keybuk is working on heavy crack for that
[08:20] <daniels> so we can effectively parallelise it
[08:21] <mjg59> Rocking
[08:22] <moquist_> jamesh: i wasn't thinking so much about the payment of the bounty; i was wondering if the need leading to the bounty offer is already being met.  apparently it's not - so what else is needed?
[08:25] <jamesh> moquist_: well, there is the talk of "a common document object model", which is a large project (and not an Ubuntu only one).
[08:25] <jamesh> moquist_: I'm not sure what the exact available bounties would be.
[08:27] <moquist_> jamesh: re: common DOM - right.  it isn't clear from the description whether work is wanted on OO, Abiword, GIMP, etc. individually, or only within the context of developing a common DOM.
[08:28] <jamesh> moquist_: have you ever programmed in VB or similar on Windows?
[08:28] <moquist_> jamesh: yes, though that was in school.  i've been doing C in UNIX for three years since then...
[08:29] <jamesh> moquist_: okay.  VB is a pretty crappy language, but its power comes from the ability to script pretty much any COM interface available
[08:30] <jamesh> moquist_: there were a lot of components provided with the OS, and pretty much every decent sized app provided COM scripting support
[08:30] <moquist_> if it isn't clear, I'm looking for an entry-point into Ubuntu dev.  i'm not interested so much in a bounty as I am in a defined problem space...
[08:31] <moquist_> jamesh: right.  so the goal is a common set of interfaces to GIMP, OO, Abi, etc. functionality.
[08:32] <jamesh> the idea is to have something similar on Linux desktops, so that people find it worthwhile writing language bindings for the object model, and applications find it worthwhile supporting it.
[08:33] <jamesh> (this is a pretty big project, as you can guess)
[08:33] <moquist_> yah.
[08:34] <jamesh> is there anything in particular that you are interested in?
[08:35] <moquist_> jamesh: no, and that's part of my problem.  :)
[08:35] <moquist_> i have UNIX kernel experience, and general application experience (though not professionally there)
[08:37] <moquist_> so far I've been looking over the bounties, because they're [relatively]  well-defined areas that I *know* need work.  can you help direct me more precisely?
[08:38] <moquist_> i looked a bit at the kernel module autoloading problem, but I didn't think that would necessarily be a good jumping-in point.
[08:39] <jamesh> I'm not exactly sure what work is covered there.
[08:39] <moquist_> so I'm interested in diverse things, and I'm having trouble narrowing my focus.  knowing what the strongest needs are would help...
[08:40] <moquist_> (sorry - i'm laboring under a *very* slow connection ATM, so I can't respond quickly)  And now you're gone, but I'll post this anyway.  :-p
[08:41] <eruin> wish I had skills in C
[08:42] <eruin> but alas, back to my communist-russia-inspired designing ;D
[08:43] <moquist_> heh
[09:56] <mjg59> Working ACPI on a Toshiba with nvidia!
[09:58] <azeem> mjg59: do you have a patch relative to the ubuntu kernel source for the kernel image you got on srcf.ucam.org?
[09:59] <mjg59> azeem: It's all in the Hoary 2.6.9
[09:59] <azeem> ah, okie
[09:59] <mjg59> But the patches are at http://www.codon.org.uk/~mjg59/patches/
[10:01] <azeem> thanks
[10:54] <mjg59> daniels: Ping?