[12:05] <mdz> what is "e-baking"?
[12:05] <mdz> oh, maybe he meant banking
[12:17] <jdub> now i have battery info, i have to figure out why this machine chews so much battery
[01:14] <thom> mjg59: no ooops, but no firmware load either
[02:02] <ddaa> Thanks everyone for the help in setting up the T42p. I have updated wiki.ubuntulinux.org. The page contains older more complicated instructions for T42 and a different model of T42p, if you could get the original authors to check and aggregate the instructions, that would be great.
[02:39] <drbyte> jdub: and what machine is that
[02:39] <drbyte> (so i can avoid buying one :P)
[02:42] <pl0vs> ddaa, you should no longer use wiki.ubuntulinux.org bur use www.ubuntulinux.org/wiki
[03:13] <jdub> ARGH
[03:13] <jdub> hahahaha
[03:13] <jdub> man
[03:14] <jdub> now that acpi is a-ok
[03:14] <jdub> i can't close the lid of my laptop when i'm using my external monitor/mouse/keyboard
[03:14] <jdub> because it locks the screen and chvts
[03:17] <amu> ;) 
[03:43] <jdub> woo
[03:43] <mjg59> thom: Works fine here...
[03:44] <jdub> yo mjg59 
[03:44] <jdub> mjg59: did you look into paul's kernel patches much?
[03:44] <jdub> my laptopt totally chugs the battery life
[03:44] <mjg59> Paul?
[03:44] <jdub> paul drain
[03:44] <jdub> i linked them on -devel
[03:44] <mjg59> Never heard of him, I'm afraid
[03:45] <jdub> he's the garnome maintainer
[03:45] <jdub> but also
[03:45] <mjg59> (Sorry, just back from robtaylor's. He was in the process of having the council ask him to turn the music off)
[03:45] <jdub> has recommended patches linked by me from -devel
[03:45] <jdub> haha
[03:45] <mjg59> Cool
[03:45] <mjg59> I'll take a look
[03:46] <mjg59> (V, v drunk)
[03:47] <jdub> what's the best way to replicate the kernel builds the buildds do? apt-get source linux?
[03:47] <jdub> linux-source-... was poo
[03:47] <jdub> roughly speaking
[03:53] <mjg59> linux-source-foo, dpkg-buildpackage
[03:53] <amu> 2 error's hoary compared with warty 
[03:53] <mjg59> It'll build lots of them
[03:54] <amu> E: Couldn't find package x3270
[03:54] <amu> E: Couldn't find package pr3287
[03:54] <jdub> mjg59: but not how the buildds do it...
[03:55] <mjg59> Uh. What do the buildds do, then?
[03:55] <jdub> dunno
[03:56] <mjg59> How is it different to what the buildds do?
[04:01] <jdub> linux-source... is not the same as apt-get -b source <blah>
[04:02] <mjg59> Oh, I see what you mean
[04:02] <mjg59> I /believe/ that the buildds do it from linux-source, but produce individual packages that have different source dependencies
[05:08] <Mitario> hi everyone
[05:28] <Mitario> anyone knows who's maintaining apt in ubuntu?
[05:29] <bob2> in what sense?
[05:29] <bob2> it's still mdz's baby
[05:29] <Mitario> package maintainer
[05:29] <Mitario> ah, ok
[05:29] <Mitario> ty
[05:30] <bob2> ubuntu doesn't have the same sort of package maintainership as debian
[05:30] <Mitario> well, call it 'standard uploader' then ;)
[05:31] <Mitario> i mean, i guess someone is a bit more responsible for one package then the rest of the ubuntu maintainers
[07:10] <mdz> ddaa: thanks for writing up your laptop findings for the wiki
[07:12] <fabbione> morning guys
[07:12] <mdz> good morning
[07:12] <fabbione> hey mdz
[07:13] <fabbione> mdz: probably today we will do another internal drop
[07:13] <fabbione> and it should be the last one
[07:13] <fabbione> tuesday -> archive
[07:13] <mdz> ready to test when you are
[07:13] <fabbione> buildd has completed
[07:13] <fabbione> no failure
[07:14] <fabbione> mdz: yup.. we are going to split some libraries
[07:14] <fabbione> and we need to test a new autoconfiguration feature of X.org
[07:14] <fabbione> if the latter works we can get rid of 90% of the external tools we are using
[07:14] <mdz> great news on the build test
[07:15] <mdz> what kind of autoconfiguration does the feature provide?
[07:15] <mdz> if it could only choose the driver, that would make me so happy
[07:15] <mdz> because we could ditch discover1 for hoary
[07:15] <fabbione> mdz: it does everything
[07:16] <fabbione> just remove /etc/X11/xorg.conf
[07:16] <fabbione> and try to startx
[07:16] <fabbione> it detect all my hardware perfectly
[07:16] <fabbione> including video card
[07:16] <fabbione> but the keyboard layout
[07:17] <fabbione> or you can also try to remove one single section only
[07:17] <fabbione> like the Monitor one
[07:17] <fabbione> and it should create an autodetected one
[07:17] <fabbione> problem is that it doesn't save it in the config
[07:17] <fabbione> there is a tool in X.org to do so, but we need to test it 
[07:18] <mdz> why save it, if it can do it automatically?
[07:18] <fabbione> because you might want to customize after?
[07:20] <mdz> nice to have
[07:22] <bob2> networkmanager looks awesome
[07:24] <mdz> E: Couldn't find package networkmanager
[07:24] <bob2> thombot's packaging it
[07:24] <mdz> yeah
[07:24] <bob2> which reminds me, his Sources file is broken
[07:26] <mdz> wow, if you search for 'blackdown deb', the Ubuntu wiki resources are in the top 5 :-)
[07:26] <mdz> (google)
[08:22] <fabbione> goody
[08:22] <fabbione> wanna build and quinn-diff are ready
[08:22] <fabbione> time to configure sbuild
[08:24] <mdz> fabbione: why so much infrastructure for just a build test?
[08:24] <fabbione> mdz: ????
[08:24] <fabbione> that's for the sparc port of hoary
[08:24] <mdz> oh, I thought it was for x.org
[08:24] <fabbione> x.org has been done already
[08:24] <fabbione> as i wrote before main compiled perfectly
[08:24] <mdz> so it was doubly surprising :-)
[08:25] <fabbione> we are not going to kick in universe
[08:25] <fabbione> but yes.. we did use wanna-build & co also for that
[08:25] <fabbione> but we had the setup ready before the Xsprint
[08:31] <bob2> hm
[08:31] <bob2> cpufreqd's default config uses acpi, which is (afaik) useless on ppc
[08:31] <bob2> is it a bug that it doesn't modify that based on the arch?
[08:37] <mdz> bob2: doesn't particularly matter, since cpufreqd is unsupported :-)
[08:44] <sjoerd> bob2: cpufreqd sets the right type in postinst
[08:45] <sjoerd> (in debian thatis)
[08:50] <mdz> hoary and sid have the same cpufreqd
[08:50] <bob2> yeah, thought I was in a different -devel channel :)
[09:41] <fabbione> Automatic build of choose-mirror_1.06ubuntu1 on vultus5 by sbuild/sparc 1.170.5
[09:41] <fabbione> Built successfully
[09:41] <fabbione> Purging chroot-hoary/build/sparcbuildd/choose-mirror-1.06ubuntu1
[09:41] <fabbione> YEPPA
[09:41] <fabbione> time to kick the buildd
[09:42] <daniels> cool :)
[09:45] <fabbione> daniels: yeah.. i had to fight with sbuild..
[09:45] <fabbione> i couldn't understand why it was not using the chroot :-)
[09:45] <fabbione> daniels: are you on the way here?
[09:46] <daniels> fabbione: i have to do my laundry at some stage, was thinking of doing that this morning instead of tonight
[09:47] <fabbione> daniels: ok, take your stuff here
[09:47] <fabbione> i will drive you to a laundry
[09:47] <daniels> fabbione: oh cool, thanks
[09:47] <fabbione> daniels: so we can take a break in the middle of the day
[09:47] <fabbione> i might have to do some laundry too
[09:49] <fabbione> daniels: do you think you can do me a little favour coming here?
[09:49] <fabbione> i am really too lazy to go out right now ;)
[09:49] <fabbione> nah well
[09:49] <fabbione> i will go out
[09:49] <fabbione> getting some fresh air is good :-)
[09:50] <daniels> fabbione: yeah sure :) that's fine
[09:50] <daniels> heh
[09:50] <daniels> so what's the plan?  i was just about to jump in the shower and catch the train
[09:50] <fabbione> just come here...
[09:50] <fabbione> it's fine
[09:51] <SuperL4g> is there a way to pass ESSID and WEP info before you start the install, so I can get connectivity before the first reboot? as it stands right now, I have to do the install without the network, then reboot, then update.  If I could pass the options pre-install it would update everything up front and save about 20 minutes. :)
[09:52] <daniels> ok, cool
[10:48] <fabbione> State changes at 2004 Nov 07 10:43:44 for distribution hoary:
[10:48] <fabbione> --merge-packages(hoary): db3_3.2.9-20 changed from Building to Installed by sparcbuildd as sparcbuildd
[12:24] <thom> mdz: will probably upload NM on monday or tuesday
[01:12] <cenerentola> any lovable canonical ppl in here?
[01:12] <daniels> i don't know about 'lovable' as such ...
[01:14] <thom> Mithrandir: fixed fakeroot yet? ;-)
[01:14] <thom> (i saw the mips bug, which looked more or less identical)
[01:14] <Mithrandir> thom: possibly, yes.
[01:14] <thom> oh? rock.
[01:14] <Mithrandir> thom: it's probably a libc compile bug, I haven't tried upgrading to the new-compiled libc yet, but that's a simple thing to do
[01:15] <thom> if you need me to test just shout
[01:15] <thom> i saw your conversation with drow as well. so was it bad nptl/tls options in the build?
[01:15] <cenerentola> daniels: id like to know all you can tell about the italian ml...
[01:16] <cenerentola> since the btlug.it [italian ubuntu linux community]  site is now up
[01:16] <Mithrandir> thom: I think so, but I haven't tracked it yet.  I need to fiddle around a little
[01:17] <Mithrandir> thom: I miscompiled glibc again, it seems.
[01:42] <daniels> Mithrandir: yo dude
[01:51] <zopi> Hi 
[01:51] <zopi> sorry to bother you
[01:51] <zopi> for my problem of yesterday
[01:51] <zopi> slow installation
[01:52] <zopi> the problem has been solved I have made my Hard Disk in Master
[01:52] <zopi> in Slave mode the installation is very slow !
[01:52] <mjg59> zopi: Do you mean the disk was a slave without a master?
[01:57] <ddaa> mjg59: thanks for pointing out the old wiki issue. Though I put the comments on new wiki and just got the name wrong when writing my message here.
[01:58] <zopi> mjg59 : yes
[01:58] <ddaa> Duh... zwiki is terminally confusing...
[01:59] <ddaa> The tabs change (or do not change) in a completely unpredictable manner :-(
[02:04] <tuo2>  Umm.. random question.. Is there a good reason why we are using a self-signed certificate on https://www.ubuntulinux.org ?
[02:04] <bob2> hahaha
[02:04] <bob2> yes
[02:04] <tuo2> bob2: care to enlighten me?
[02:05] <bob2> not sure if I can :-)
[02:05] <bob2> elmo or thom would be the people to ask
[02:05] <mjg59> ddaa: Hmm. You shouldn't have /dev/eth*
[02:05] <azeem> nobody at canonical has expertise with internet certificates I guess ->;)<-
[02:06] <mjg59> But do you mean that there's no network interfaces?
[02:06] <bob2> hehehehehe
[02:06] <mjg59> Well, based on thom...
[02:06] <fabbione> hahah
[02:07] <ddaa> mjg59: well, okay... there are not here with warty kernel either...
[02:07] <bob2> do any unices have /dev/eth?
[02:07] <ddaa> Neverthell, networking works with warty kernel, but not with yours.
[02:07] <daniels> bob2: if you asked thom about the ul.o certificate, he'd give it to you :P
[02:08] <mjg59> ddaa: Weird. I'm running it fine on here.
[02:08] <bob2> daniels: hahahaha
[02:08] <mjg59> ddaa: Does suspend to RAM work with it?
[02:08] <ddaa> mjg59: yup
[02:08] <mjg59> Cool. Did you need any kernel arguments?
[02:09] <mjg59> acpi_sleep=foo ?
[02:09] <tuo2> azeem: funny. ;)
[02:09] <ddaa> I used the same acpi_sleep=s3_bios as needed with the current kernel
[02:09] <ddaa> Did not test w/o it.
[02:12] <mjg59> ddaa: It probably does need it, but there's something else you could test
[02:12] <mjg59> ddaa: If you could install the video-post package from http://www.srcf.ucam.org/~mjg59/laptops and then add /usr/sbin/video_post to the start of your resume script, that would be good
[02:12] <mjg59> With a bit of luck, that'll avoid the need for the kernel argument
[02:13] <ddaa> mjg59: let me test things properly...
[02:18] <ddaa> about eth1, the error I get is "no such device".
[02:19] <ddaa> However the device appears in the Gnome Device Manager (I do not know if that means anything)
[02:21] <mjg59> ddaa: Does lsmod show that the modules are loaded? Does dmesg say anything?
[02:22] <daniels> mjg59: what does video_post do?
[02:22] <mjg59> daniels: soft-boots the video card
[02:23] <daniels> vbe?
[02:24] <mjg59> It's pre-vesa, I think
[02:24] <ddaa> ipw2200 shows in lsmod output
[02:24] <daniels> mjg59: oh
[02:25] <mjg59> daniels: But yeah, it does int 10 stuff
[02:25] <mjg59> ddaa: Hmm. Anything in dmesg?
[02:25] <daniels> oh my god, it's almost straight ripped from the xfree86 soruce
[02:25] <ddaa> Yes. Detected, but "Unable to load firmware: 0xFFFFFFFE"
[02:26] <mjg59> daniels: It is indeed
[02:26] <mjg59> ddaa: Grr. Weird.
[02:26] <mjg59> I thought I had that fixed.
[02:27] <daniels> Keybuk: yo
[02:28] <Keybuk> yo, hey what's happenin' dude?
[02:28] <mjg59> ddaa: Oh, damn. Sorry, I managed to miss half the patch.
[02:29] <mjg59> Ugh. I'll fix that.
[02:29] <ddaa> mjg59: the network problem is the big show-stopper. The other issue is the cpu freq applet. I have not checked whether freq scaling actually works though.
[02:30] <mjg59> ddaa: Right, I'll sort out the networking stuff
[02:32] <daniels> Keybuk: not much, just archifying xorg
[02:33] <ddaa> Mh... interesting... without acpi_sleep, the machine did not actually hang... the video and the sound were out, but I was able to reboot it by blind-typing.
[02:33] <Keybuk> fun
[02:33] <daniels> seems plausible
[02:33] <ddaa> Did not actually went as far as trying that with the warty kernel...
[02:34] <mjg59> ddaa: Ok - it'd be great if you could see if video_post brings it back
[02:34] <ddaa> mjg59: next step
[02:34] <ddaa> however, after that there is another time consuming test: checking how much batter power it actually consumes over 8 hours.
[02:35] <ddaa> maybe I'm having false impressions, but I thought the acpi_sleep parameter might be needed to get the proper sleep state.
[02:36] <mjg59> No, acpi_sleep just changes how the video is reinitialised on resume
[02:37] <daniels> no, acpi_sleep=s3_bios just runs through the video bios to reinitialise the video card when it comes back
[02:41] <ddaa> mjg59: kudos! video_post at the start of the resume script makes it happy
[02:42] <mjg59> Rock
[02:43] <daniels> noice
[02:44] <thom> mjg59: "miss half the patch" is that ipw2100 firmware loader related?
[02:47] <ddaa> mjg59: do you want me to check something for the cpu scaling support, or is the issue well known?
[02:47] <mjg59> thom: All firmware related
[02:47] <mjg59> ddaa: If you could, though I think that there's something broken in the realm of ACPI processor management
[02:48] <mjg59> If you have the cpufreq-acpi module (can't remember what it's actually called) loaded, changing to cpufreq-centrino may be better
[02:48] <ddaa> mjg59: not that I can have a lot of initiative, but I'd happily run any test you want.
[02:48] <ddaa> (as long as it does not involve throwing the the lappy on the floor)
[02:49] <mjg59> ddaa: If you could stop powernowd, rmmod acpi, modprobe cpufreq-centrino and then restart powernowd, that would be good
[02:49] <thom> mjg59: ah
[02:49] <ddaa> mjg59: ack
[02:50] <mjg59> Oh, lord. More of last night is coming back to me.
[02:50] <ddaa> hu... no acpi module at all...
[02:50] <ChrisH> mjg59: Did you update anything laptoppy? I could check it out tomorrow at work. (Toshiba Tecra 8100)
[02:50] <mjg59> ChrisH: Yeah, though the kernel is a bit broken at the moment :)
[02:50] <mjg59> ddaa: Hmm. Does modprobe speedstep-centrino work?
[02:50] <mjg59> (sorry, not cpufreq-centrino)
[02:50] <ChrisH> mjg59: Just update to newest hoary?
[02:50] <mjg59> ChrisH: No, this is outside hoary
[02:52] <ddaa> Mh... modprobe acpi => no such device. No helpful output in dmesg or syslog.
[02:52] <daniels> mjg59: oh, oh.  is he in trouble?
[02:53] <mjg59> ddaa: modprobe speedstep-centrino ?
[02:53] <daniels> mjg59: i have an ogg for that song to hand
[02:53] <mjg59> daniels: That's... unnecessary
[02:53] <mjg59> Unless you want to upset anyone in your vicinity
[02:53] <daniels> mjg59: just wondering if you wanted a copy
[02:54] <ddaa> modprobe speedstep-centrino: no such device. dmesg says "no table support for CPU model "Intel(R) Pentium(R) M processor 1.70GHz" and "try compiling with CONFIG_X86_SPEEDSTEP_CENTRINO_ACPI enabled"
[02:55] <mjg59> Hm. Right.
[02:55] <mjg59> So the problem is in the ACPI processor module.
[02:55] <mjg59> Thanks, that narrows it down
[02:56] <ddaa> You're welcome. Thanks for working on making my lappy work :-)
[02:57] <ddaa> The nice thing with working in Canonical, is that you can whine at the people who actually have the ability to fix the issues :-)
[02:57] <daniels> mjg59: last night, had you been out all night, and hadn't been home?
[02:58] <bob2> ddaa: haha
[02:58] <fabbione> ddaa: remember that some Canonical emploies are dangerous...
[02:58] <fabbione> ;)
[02:59] <ddaa> bob2: getting hardware specs from Apple would involve a commando raid...
[02:59] <bob2> ddaa: I just want someone to break into broadcom's offices 
[02:59] <thom> broadcom would be much more productive
[02:59] <thom> has no-one tried using darwin drivers?
[02:59] <ddaa> fabbione: bah, you are just posturing as a Dangerous Evil X Maintainer. We all know you are a nice dude :-P
[03:00] <mjg59> daniels: Strictly, no
[03:00] <bob2> thom: I get errores during modprobe!
[03:00] <mjg59> thom: Secret plans are in motion
[03:00] <mjg59> There are no darwin drivers, AFAIK
[03:01] <mjg59> There are binary-only Linux drivers for MIPS
[03:01] <bob2> yeah
[03:01] <bob2> which I begged kinnison to reverse engineer
[03:01] <thom> or s/darwin/OS X ; which should be more or less the same no?
[03:01] <bob2> benh has some tools to sniff what mac os x drivers do
[03:01] <mjg59> thom: Sure, there are OS X drivers
[03:02] <mjg59> So I guess you could do something ndiswrapper like...
[03:02] <thom> mjg59: that was kinda what i was thinking, yeah
[03:02] <mjg59> But we've got someone working on a better solution
[03:02] <daniels> mjg59's secret laptop cabal
[03:02] <bob2> oh, mac ndiswrapper, eeevil
[03:02] <daniels> (which is later revealed to be a room full of monkeys with albooks)
[03:02] <mjg59> New packages soon
[03:03] <bob2> there's a standing offer of as much beer as you can drink for whoever hacks up a working broadcom driver
[03:04] <daniels> bob2: you forgot the most important stipulation
[03:04] <daniels> bob2: *cold* beer
[03:04] <bob2> yeah
[03:04] <bob2> if they claim it in london, I'll bring an esky to cool it
[03:04] <ddaa> btw, I was quite surprised at the sobriety of the Canonical dudes in Oxford.
[03:04] <ddaa> I would have expect much more liberal beer-gulping.
[03:05] <azeem> they drank whiskey?
[03:05] <thom> in the hotel? NFW
[03:05] <ddaa> azeem: stevea drank expensive (and excellent) whisky
[03:05] <thom> it was more expensive than norwegian beer
[03:05] <bob2> spain had better be cheaper
[03:05] <mjg59> We had poor quality wine and vodka from the petrol station
[03:05] <daniels> the wine wasn't abysmally bad
[03:05] <daniels> it was drinkable
[03:05] <ddaa> but really, you'd have to be a bit out of your mind to spend that much money on perfumed ethanol...
[03:05] <bob2> and poor quality movies to go with it
[03:06] <daniels> bob2: that you didn't even watch, I might add
[03:06] <bob2> I tried to watch it
[03:06] <mjg59> In the real world, when you kill people, they die
[03:06] <thom> bob2: first time i've seen quality used in any sense about antitrust
[03:06] <bob2> but it was in the distro room
[03:06] <bob2> and I was sleepy
[03:06] <bob2> distro room = bob2's nap room
[03:06] <ddaa> daniels: It might even been watchable, if it had some kind of sound besides the built-in speakers of a powerbook...
[03:07] <thom> as demonstrated during the pygtk talk
[03:07] <daniels> mjg59: it's not just ones and zeroes!
[03:07] <bob2> thom: hah
[03:07] <mjg59> 6 minutes to upload this stuff
[03:07] <mjg59> I want more bandwidth
[03:07] <daniels> bob2: so how many times did you apologise to jamesh?
[03:07] <thom> hrm, time to get to sainsburys and back
[03:07] <bob2> daniels: a dozen on the bus back to london
[03:08] <daniels> bob2: heh
[03:08] <bob2> so, I hate you all, i368 owners
[03:08] <daniels> i368? wassat?
[03:09] <bob2> bah
[03:09] <thom> bob2: join the x40 cabal
[03:09] <mjg59> If this works, then I just need to fix the acpi processor stuff
[03:10] <daniels> bob2: one day you'll wake up with an iBook LCD in your bed
[03:11] <bob2> thom: don't tempt me
[03:11] <ddaa> mjg59: video_post seems to do the right thing with the warty kernel (with acpi_sleep).
[03:12] <daniels> ddaa: and without acpi_sleep?
[03:12] <thom> bob2: you know you will, it's just a matter of time
[03:12] <ddaa> daniels: Hu. I meant _without_
[03:12] <bob2> daniels: yeah, with a broadcome wireless chip through it's head
[03:12] <bob2> thom: hah, it is, sadly
[03:12] <ddaa> daniels: feels like including it in you thinkpad-acpi stuff?
[03:13] <daniels> bob2: eveeerrrryboooodddyyyyy comes to x40
[03:13] <daniels> alright, time for me to go do laundry and head back to the hotel
[03:13] <daniels> in no particular order
[03:13] <mjg59> Need to check the license situation on video_post first
[03:13] <mjg59> The guy from Intel hasn't actually included any copyright information
[03:13] <bob2> I shoulda waited until we had a distro team before buying, so I could copy them
[03:13] <daniels> mjg59: most of it is ripped *verbatim* from various places in xfree86
[03:13] <daniels> i'd be stunned if his additions were even copyrightable
[03:13] <mjg59> Oh, yeah
[03:14] <mjg59> It's all X11 license based stuff
[03:14] <daniels> so it'll all be either mit or bsd to egbert eich
[03:14] <daniels> anyway, I'm out
[03:17] <mjg59> ddaa / thom: New kernel on http://www.srcf.ucam.org/~mjg59/laptops should fix the firmware stuff
[03:28] <thom> mjg59: does the suspend-to-disk button work for you, ooi?
[03:29] <thom> mjg59: and still the firmware load problem
[03:29] <mjg59> thom: With ibm-acpi, yes
[03:29] <mjg59> thom: Gah. Fuck.
[03:30] <mjg59> thom: Remember that it won't load the firmware unless /sys is mounted
[03:30] <mjg59> You're doing a full boot rather than just init=foo?
[03:31] <thom> yeah, full boot
[03:32] <mjg59> And it's the same error as before?
[03:32] <thom> yup
[03:32] <mjg59> Arse
[03:32] <thom> yup :/
[03:32] <thom> ok, i really have to go get milk
[03:32] <thom> back in about 15
[04:02] <Mitario> hello everyone
[04:03] <sivang> Hey Mitario, what's new?
[04:03] <Mitario> nothing much really :)
[04:07] <thom> Mitario: hey, so when i logged out and logged in again, i ended up with about 7 upgrade notification icons
[04:08] <Mitario> you probably started them all manually? :)
[04:08] <thom> pretty sure i didn't, but maybe
[04:09] <Mitario> upgrade-notifier puts every instance of itself in gnome-session
[04:10] <thom> NetworkManager has some code to ensure it's only in the session once, iirc. might be worth stealing?
[04:13] <Mitario> ah, good idea
[04:14] <sivang> Mitario : are you working on the auto update applet?
[04:15] <Mitario> with michael yeah
[04:16] <sivang> must be cool thing
[04:16] <sivang> :)
[04:19] <Mitario> hmm, no hoary updates the last day?
[04:24] <mjg59> This all gives me increased respect for the kernel maintainers
[04:45] <play> hi everyone
[04:51] <play> i have some troubles with K3B and Warty... K3B doesn't detect my DVD burner :( It seems it's a problem with the 2.6.8 kernel. Does somebody know a patch ?
[04:51] <play> i enjoy very well Ubuntu ! good work ! ;)
[04:52] <thom> play: #ubuntu is the support channel
[04:53] <thom> thanks :-)
[04:54] <play> ok Thom :) Sorry. bye
[05:25] <ddaa> mjg59: I confirm persistent of the firmware loading problem with your newer kernel.
[05:25] <ddaa> *persistence
[05:27] <fabbione> wow
[05:27] <fabbione> i found a warty chroot from debconf4 :-)
[05:27] <fabbione> lost in one my machines 
[05:56] <mjg59> Mm.
[05:57] <mjg59> Lovely kernel crack.
[06:03] <mjg59> Argh. Except swsusp has now broken.
[06:47] <mjg59> Rock
[06:48] <Keybuk> mjg59 is KERNEL BOY
[06:55] <mjg59> ddaa: http://www.srcf.ucam.org/~mjg59/laptops/http://www.srcf.ucam.org/~mjg59/laptops/
[06:55] <mjg59> Working firmware, working processor stuff
[06:56] <ddaa> mjg59: well, actually I am going to way for unison to finish munching my disks :-)
[06:56] <ddaa> Unless you are going to bed soon, then I'll postpone unison.
[06:58] <mjg59> ddaa: Heh. No, I'll be awake for a while yet
[06:59] <ddaa> Mithrandir: not thinking time, but working on the Arch team teached me that people can go to bed at any hour, irrelevent of where they actually live.
[06:59] <ddaa> * not thinking that
[07:00] <Mithrandir> even in the early afternoon?  I can go to bed anything between 22-ish and 12-ish, but not between 12 and 22, that's just wrong.
[07:00] <mjg59> Can C do switch statements on strings?
[07:00] <Mithrandir> mjg59: no
[07:00] <Mithrandir> how would you switch on a pointer?
[07:00] <mjg59> Yeah, thought not
[07:00] <mjg59> It'd make life massively easier, though
[07:00] <Keybuk> well, you can, provided you're switching on literal equality
[07:01] <Keybuk> not just strcmp() returning a particular value
[07:02] <ddaa> mjg59: if you are concerned by performance, you could probably hack something nifty as a "internable string", with compile-time interned values.
[07:02] <ddaa> But it would probably not fit any definition of "simple".
[07:05] <Keybuk> yeah, there's a nice quark layer in glib though :)
[07:14] <herzi> hmm, if i switch to vt01 by pressing ctrl-alt-f1 i cannot get back or get to another vt
[07:19] <mjg59> If people could test the dpms package at http://www.srcf.ucam.org/~mjg59/laptops/ that would be great
[07:20] <mjg59> Note: it'll turn the screen off, but won't turn it back on unless you run it with the on argument. It doesn't track mouse or keyboard events.
[07:20] <daniels> heh, funny you shoudl mention that
[07:21] <daniels> I was tooling with some Radeon DPMS code in the laundromat
[07:21] <daniels> just as junkcode
[07:21] <mjg59> This is really, really simple
[07:21] <daniels> it compiled; probably didn't even work, but it was fun in any case
[07:21] <mjg59> And should be entirely hardware independent
[07:22] <daniels> using VBE?
[07:23] <mjg59> Yeah
[07:23] <mjg59> Just call 0x4f10 with an argument
[07:31] <mjg59> Arse. The craptop only produces one lid event per open/close
[07:31] <ChrisH> mjg59: well... "dpms standby" worked. But when I try to get back I have a changed resolution. Far wider than high.
[07:31] <mjg59> ChrisH: Hmm. Was this under X?
[07:31] <mjg59> Or on the console?
[07:31] <ChrisH> mjg59: yes, sir
[07:31] <ChrisH> mjg59: X
[07:32] <mjg59> ChrisH: How did you try to get back? Just dpms on?
[07:32] <ChrisH> mjg59: yes
[07:32] <ChrisH> mjg59: switching to the console (Alt-Ctrl-F1) and back (Alt-F7) restored the resolution.
[07:32] <mjg59> Weird.
[07:33] <mjg59> ChrisH: What graphics hardware is this?
[07:33] <ChrisH> mjg59: Not a notebook actually. ;)
[07:34] <ChrisH> mjg59: GeForce 5900 XT
[07:34] <mjg59> Hm. nv or nvidia drivers?
[07:34] <ChrisH> mjg59: The evil non-free nvidia module (nvidia-kernel-2.6.6_1.0.6106-3).
[07:35] <mjg59> Heh
[07:35] <mjg59> Tricky for me to track down, then :)
[07:35] <ChrisH> mjg59: I could try with "nv" in a few minutes. Just need to finish a Wiki article.
[07:36] <mjg59> Sure, no problem
[07:42] <daniels> i blame nvidia
[07:43] <cenerentola> what about mataro... 
[07:44] <daniels> cenerentola: yeah, that's nvidia's fault, too
[07:44] <mjg59> daniels: I'm thinking of this for the stuff that's in /usr/share/acpi-support/screenblank at the moment - downside is that in the absence of a lid event on open, the screen isn't going to come back
[07:45] <daniels> mjg59: do we know of anything other than the craptop that don't issue a separate lid event?
[07:46] <ddaa> the only reason left i have to use my 5 years old destkop is the ability to have _two_ screens... well, and the fact that external video does not quite work yet, too :-)
[07:46] <mjg59> daniels: I don't know of anything else, but it's a possibility
[07:47] <mjg59> Unfortunately, I doubt there's a programatic way of us working that out...
[07:48] <daniels> mjg59: gah
[07:48] <daniels> ddaa: what sort of video chipset?
[07:49] <mjg59> T42 ought to be ATI
[07:49] <daniels> should work OK with atitvout
[07:50] <ChrisH> mjg59: Do you happen to know where the 'nv' module can be found in the 2.6.8 kernel config? I seem to be blind.
[07:50] <ChrisH> mjg59: I'm so stupid. Forget it. ;)
[07:50] <daniels> ddaa: you might want to try xorg and the 'Option "UseFWPLL"' option
[07:50] <mjg59> ChrisH: There isn't one, as far as I know
[07:50] <mjg59> It's only a 2D driver, so it doesn't need kernel support
[07:51] <daniels> ddaa: that should do the trick (possibly in combination with atitvout) when the monitor's plugged in before X is started
[07:52] <mjg59> daniels: Have you had a chance to test the i830 dual head stuff under xorg?
[07:53] <daniels> mjg59: if you could ship a big-arse CRT to room 531, First Hotel Vesterbro, Vesterbro 23-29, Kbenhavn, I promise to do my best :)
[07:54] <daniels> mjg59: it looks pretty promising, though -- supports both pipes separately
[07:54] <daniels> (iirc you can also pick which pipe overlays go to)
[07:56] <mjg59> daniels: xorg-common is missing a depends on lsb
[07:56] <mjg59> /etc/init.d/xorg-common: line 10: /lib/lsb/init-functions: No such file or directory
[07:56] <daniels> GNAR
[07:56] <daniels> craaaack.
[07:56] <daniels> you can just sub in the xfree86-common init script anyway, does the same stuff
[07:56] <daniels> thanks for the heads-up
[07:58] <mjg59> /etc/init.d/xorg-common: line 22: log_begin_msg: command not found
[07:58] <mjg59> Heh
[07:58] <mjg59> That's just a deb/ubuntu thing, right?
[07:59] <daniels> yeah, we have all kinds of LSB initscript crack
[08:08] <ChrisH> mjg59: same effect... nvidia and nv. I have taken photos and will upload them. Give me a sec.
[08:08] <ddaa> daniels: I can only assume xorg is hoary stuff...
[08:10] <mjg59> ChrisH: Weird
[08:10] <daniels> ddaa: hoary + external archive
[08:10] <daniels> ddaa: read your list mail
[08:10] <mjg59> daniels: Massive breakage shifting from xfree->xorg if I was using the i830 drm previously
[08:10] <ddaa> daniels: sure... just have a few hundreds of them
[08:10] <daniels> mjg59: the bios probably needs poking in a non-vesa way for non-vesa modes
[08:10] <mjg59> Oopses all over the place
[08:10] <daniels> mjg59: does it go away if you restart?
[08:10] <mjg59> daniels: For the DPMS stuff, it shouldn't really matter
[08:10] <mjg59> But yeah...
[08:11] <daniels> mjg59: you'd think not ...
[08:11] <mjg59> Doing it on the console is likely to be safe
[08:11] <daniels> yah
[08:11] <daniels> hence the chvts in thinkpad-x40-support
[08:11] <daniels> which I need to resurrect now I've lost my sources for the new stuff
[08:12] <mjg59> Christ. Now the firmware loader has broken again.
[08:12] <ChrisH> mjg59: http://workaround.org/dpms/
[08:13] <daniels> mjg59: not my fault!
[08:13] <mjg59> ChrisH: Haha
[08:13] <mjg59> That is *so* cool
[08:13] <ChrisH> mjg59: Very funny. ;)
[08:13] <mjg59> I've no idea why it's happening, though
[08:13] <ChrisH> mjg59: Looks like C=64 retro style. Perhaps you should offer a "--no-commodore-64" option.
[08:15] <daniels> whoa, that's bong
[08:15] <mjg59> daniels: Is composite on by default?
[08:15] <mjg59> Ah, yes
[08:15] <daniels> mjg59: no
[08:15] <mjg59> Oh, maybe that was elft over from me palying with stuff at some point, then
[08:15] <daniels> mjg59: Section "Extensions"\n\tOption\t"Composite"\t"Enabled"\nEndSection
[08:15] <daniels> mjg59: noting that composite and GLX are currently mutually exclusive
[08:16] <mjg59> Ha, no - I wanted it off
[08:16] <daniels> what, you don't enjoy X crashing repeatedly? :P
[08:17] <mjg59> daniels: Strange as it may seem...
[08:18] <daniels> mjg59: but it's fun ... you can get backtraces and send them to us
[08:18] <thom> daniels: he's not allowed to play with X till he fixes the firmware loader ;-)
[08:18] <daniels> mjg59: btw, xorg fixes the only-one-X-instance-at-once problem
[08:18] <daniels> thom: that's what you think.  but which is more shiny? :)
[08:18] <thom> daniels: swsusp
[08:19] <daniels> thom: fwiw, it's 2C right now, and I just went running before in a singlet and shorts.  does London get much colder?
[08:19] <daniels> thom: yeah, fair point
[08:19] <thom> is the ultimate shiny
[08:19] <thom> ok, you're a freak
[08:19] <daniels> thom: although, in winter, not having any suspend at all and starting X compiles is the ultimate shiny
[08:19] <thom> london probably won't be any colder than that, no
[08:19] <daniels> thom: you put it in your backpack and it warms your back a bit :)
[08:19] <ddaa> mjg59: kernel news: cpufreq works however the firmware for ipw2200 still fails to load.
[08:19] <thom> heh
[08:19] <daniels> thom: to be fair, as I said, I was running
[08:19] <mjg59> ddaa: Hmm.
[08:20] <ddaa> mjg59: suspend to ram... suspended... and that's about all.
[08:20] <ddaa> It won't wake up.
[08:20] <mjg59> ddaa: Weird.
[08:21] <ddaa> Want me to run some additional diagnostic?
[08:22] <mjg59> ddaa: What do you mean by won't wake up?
[08:22] <ddaa> The moon led stays lit, no cpu or hd activity, no reaction to keyboard or lid actions.
[08:23] <ddaa> It's sleeping for all I can tell.
[08:23] <mjg59> Right
[08:23] <nasdaq4088> interesting article: http://www.eweek.com/article2/0,1759,1714680,00.asp#talkback
[08:23] <mjg59> Have you tried the power button?
[08:23] <ddaa> Short presses have no effect.
[08:23] <ddaa> A long press will probably shut it down though.
[08:23] <mjg59> Try for a couple of seconds
[08:23] <mjg59> Less than 4, but more than 1
[08:24] <ddaa> Done. No effect.
[08:24] <mjg59> Hrm.
[08:24] <mjg59> Can you reboot it and take a look at the contents of /proc/acpi/wakeup ?
[08:24] <mjg59> Meanwhile...
[08:24] <mjg59> Any hotplug gurus around?
[08:25] <ddaa> mjg59: btw, it's still w/o any acpi_sleep option.
[08:26] <mjg59> ddaa: Shouldn't make any difference
[08:26] <ddaa> Not that it should make any difference from what you said.
[08:26] <mjg59> (Well, in theory)
[08:27] <ddaa> mjg59: LID
[08:28] <ddaa> mjg59: LID and SLPB have sleep state = 3 and status = *enabled.
[08:29] <ddaa> all the rest is sleep state 3 and status disabled except for PCI1 and AC9M which have sleep state 4.
[08:29] <ddaa> not the * in "*enabled"
[08:29] <ddaa> *note the *
[08:32] <mdz> mjg59: what about hotplug?
[08:32] <mjg59> mdz: How many arguments should the kernel be passing to hotplug when it's loading firmware?
[08:33] <mdz> mjg59: I'm not sure it passes any arguments; it should send a few environment variables
[08:33] <mdz> #       ACTION=%s [add or remove] 
[08:33] <mdz> #       DEVPATH=%s [in 2.5 kernels, /sys/$DEVPATH] 
[08:33] <mdz> #       FIRMWARE=%s
[08:33] <mjg59> Ah
[08:34] <zopi> does acpi=off is writting on the wiki for Via Chipset ?
[08:34] <zopi> because before enable it Ubuntu was very slow to install and booting !
[08:34] <mdz> zopi: I have a VIA chipset and it works fine without acpi=off
[08:34] <zopi> mdz : arf 
[08:38] <mjg59> ARGH, YOU BLOODY USELESS PIECE OF SOFTWARE.
[08:41] <mjg59> ddaa: Can you try umount /initrd and then rmmod ipw2100; modprobe ipw2100
[08:42] <mjg59> Hrm. No, that /shouldn't/ be it
[08:45] <ddaa> mjg59: umount: /initrd: device is busy
[08:47] <ddaa> device is still busy
[08:49] <mjg59> ddaa: Ah - umount /initrd/sys
[08:49] <mjg59> Then umount /initdr
[08:49] <mjg59> I see what the problem is...
[08:51] <ddaa> seems to work better
[08:54] <ddaa> okay, that I had to fiddle a bit to unfuck resolvconf, but I can now ping the outside world.
[08:54] <mjg59> ddaa: There's a new initrd-tools at http://www.srcf.ucam.org/
[08:54] <mjg59> ~mjg59/laptops
[08:55] <ddaa> no additional action requested besides installing it?
[08:55] <mjg59> If you could install that and then either reinstall the same kernel image or do mkinitrd -o /boot/initrd.whatever
[09:01] <ChrisH> Some of you guys are probably still maintaining Debian packages. Are you using Ubuntu on your workstations and doing packages from there? Or do you have a seperate installation for Sid. I'd like to reinstall to Ubuntu but am not sure if I will break anything.
[09:02] <ddaa> mjg59: dude, you rock
[09:02] <mjg59> ddaa: Excellent
[09:02] <thom> ChrisH: use a debian chroot, but i'd strongly recommend you keep a sid partition around for testing
[09:03] <ddaa> but suspend to ram is stiff fubared
[09:03] <mjg59> ddaa: That's a bit weird. I've just tested it here.
[09:03] <ddaa> mjg59: actually, on wake-up, it emits a small bip and the battery led turns on.
[09:03] <ddaa> Then that's all...
[09:04] <mjg59> Hrm.
[09:04] <mjg59> ddaa: Can you try with init=/bin/bash ?
[09:04] <ChrisH> thom: I'd probably use pbuilder or debootstrap anyway. So I guess it could be possible.
[09:04] <ddaa> mjg59: not right now, I'm called for dinner.
[09:04] <ddaa> mjg59: but I'll do when I get back.
[09:05] <mjg59> ddaa: Heh, no problem :)
[09:07] <__randy___> Where to do I go to submit a package request for Hoary?
[09:10] <thom> mjg59: ok, that initrd-tools has fixed wireless for me
[09:11] <mjg59> ddaa: Hmm, I can reproduce that now
[09:21] <mjg59> ddaa: Ah, got it
[09:21] <mjg59> ddaa: It needs the PCI hotplug drivers blacklisted
[09:24] <Mithrandir> YAY!
[09:24] <Mithrandir> fakeroot has stopped segfaulting
[09:24] <thom> cool
[09:24] <Mithrandir> seems I got completely rid of TLS in the process, though :/
[09:24] <thom> ah
[09:24] <thom> possibly not ideal
[09:25] <Mithrandir> possibly very non-ideal. :/
[09:27] <lupus_> the 2 icons in the panel
[09:27] <lupus_> firefox, evolution and help
[09:28] <lupus_> by which package are they put there?
[09:28] <seb128_> gnome-panel
[09:28] <lupus_> because the evolution icon points to evolution-2.0
[09:28] <lupus_> instead to evolution
[09:29] <seb128> will be fixed in the next upload
[09:29] <lupus_> k
[09:44] <mjg59> Hmm. I'm now very, very confused about the suspend/resume issue.
[09:58] <ddaa> mjg59: still want me to try with init=/bin/bash?
[09:58] <mjg59> ddaa: If you could, that'd be great
[09:58] <ddaa> I s'pose that I should remove ro option too?
[09:58] <mjg59> init=/bin/bash seems to work fine for me. If I run hotplug, things don't work. If I load the same modules by hand, things do work.
[09:58] <mjg59> There's something subtle going on
[10:01] <lupus_> k I have music on a vfat partition
[10:01] <lupus_> but if you mount a filesystem
[10:01] <lupus_> you can not point to the link on the desktop
[10:02] <lupus_> to the filesystem
[10:02] <lupus_> created by gnome-vfs (g-v-m,hal)
[10:04] <lupus_> nm
[10:05] <lupus_> the problem seems to be that the old app does not support this kind :S
[10:05] <lupus_> in it's fileselector
[10:06] <ddaa> what's the command to remount root as rw?
[10:07] <thom> mount -o remount,rw /
[10:14] <ddaa> I'm definitely not managing to get something out of that init=/bin/bash situation
[10:27] <ddaa> mjg59: okay, if I hack off /etc/init.d/hotplug, then s3 works okay in single user mode.
[10:30] <mdz> elmo: around?
[10:32] <Mithrandir> doko: around?
[10:42] <ddaa> mjg59: okay, with hotplug hacked off, I can suspend-to-ram and it wakes up okay.
[10:43] <ddaa> even if full multi-user mode.
[10:44] <ddaa> of course, suspend to disk, cpufreq, and the ipw2200 (once loaded manually) still work okay
[10:46] <ddaa> however, the hibernate key is inoperative... but I suppose that's because I am missing some module.
[11:20] <ddaa> mjg59: wow...
[11:20] <ddaa> did /etc/init.d/hotplug stop, then pressed the sleep button.
[11:20] <ddaa> It worked....
[11:21] <ddaa> Now, try to figure that one out :-)
[11:22] <ddaa> nevermind, was using warty kernel.... (me blushes)
[11:25] <mjg59> ddaa: I think  there's some strange timing issue or something
[11:25] <ddaa> Mh... anything to back that up, or it's just because you just have no idea?
[11:27] <mjg59> I'm finding that I get some strange messages from acpi under certain circumstances with this kernel
[11:27] <mjg59> Odd stuff about functions returning AE_TIME
[11:28] <Mithrandir> thom: it seems libc is at fault, but we build glibc the same as debian, where the problem doesn't show.
[11:28] <Mithrandir> thom: so I'm thinking of blaming doko.
[11:28] <Mithrandir> (or gcc)
[11:29] <mjg59> Which is odd, because there's almost no difference between 2.6.10-rc1-bk17 and 2.6.10-rc1
[11:29] <mjg59> In terms of acpi, anyway
[11:29] <ddaa> Well, maybe you can make me a 2.6.10-rc1 kernel so I can test what is going on?
[11:30] <ddaa> That would narrow the problem somewhat.
[11:30] <mjg59> Ah - the previous kernel I gave you had more recent acpi
[11:30] <mjg59> Hang on, let me try something...
[11:33] <mjg59> Hrm. Damnit, this may require rebuilding all the modules as well.