[00:50] BoS_ (n=BoS@dslb-088-072-035-084.pools.arcor-ip.net) left irc: Remote closed the connection
[00:50] BoS_ (n=BoS@dslb-088-072-034-227.pools.arcor-ip.net) joined #rocklinux.
[00:52] blindcod1r (i=id@ joined #rocklinux.
[00:52] blindcoder (n=blindcod@tor/session/x-741058c940cfd555) left irc: Nick collision from services.
[00:53] Nick change: blindcod1r -> blindcoder
[00:57] <fake> if i don't pass any root= option to the kernel, it says:
[00:58] <fake> VFS: Mounted root (cramfs filesystem) readonly.
[00:58] <fake> VFS: Cannot open root device "<NULL>" ...
[00:58] <th> sounds nice
[00:58] <th> uh
[00:58] <fake> jep.
[00:58] <th> that's not initramfs, is it?
[00:59] <th> well i only talked about skipping the init= root parameter
[00:59] <fake> it's a cramfs image created using mkcramfs
[00:59] <fake> i know
[00:59] <fake> i was just curious
[00:59] <th> so how do you pass the image?
[00:59] <th> via initrd?
[00:59] <fake> you were right...
[01:00] <fake> if i call the /linuxrc /sbin/init, i can exec the real /sbin/init
[01:00] <fake> nice! feels better already.
[01:00] <th> who was calling your /linuxrc before?
[01:00] <fake> th: i use the mkcramfs-created image as 'initrd (hd0,0)/boot/initrd.img'
[01:00] <fake> th: init=/linuxrc
[01:00] <th> ah ok
[01:00] <th> yea!
[01:01] <th> exactly i remember
[01:01] <fake> th: if it's called /linuxrc, it seems to get called from somewhere else
[01:01] <fake> th: if init= is suppressed
[01:01] Action: fake grepping kernel sources..
[01:01] <th> i now remember that it really should be /sbin/init
[01:01] <th> else you get lots of work with later init
[01:02] <fake> though Documentation/initrd.txt says it should be /linuxrc...
[01:03] <th> Documentation/initrd.txt is obsolete
[01:03] <th> our next step should be to move from initrd to initramfs
[01:03] <fake> yes
[01:04] <fake> Documentation/early-userspace/*
[01:04] <th> ah cool
[01:04] <th> damn subversion bindings
[01:05] <th> i dont even know which software i should hate
[01:09] kasc_ (n=kasc@dslb-084-060-102-179.pools.arcor-ip.net) joined #rocklinux.
[01:15] kasc (n=kasc@dslb-084-060-120-238.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out)
[01:15] Nick change: kasc_ -> kasc
[01:16] <fake> perl
[01:22] <th> actually i'm talking of ruby bindings here
[01:22] <fake> lucky for perl ;)
[01:22] <th> so perl is one of the few which i can't reasonable blame
[01:22] <th> fake: will your journal be committed tomorrow morning?
[01:23] <th> (with tomorrow being the 14th)
[01:23] <fake> th: yes
[01:23] <th> rocks
[01:23] <th> i feel like we are making good progress
[01:23] <fake> th: i'm just testing the 2 patches i sent to sm
[01:24] <th> 334 builds total, 7 completed fine, 0 with errors.
[01:24] <th> i hope that crystal will finish over night
[01:40] <fake> so, let's see...
[01:41] <th> trying to scare us?
[01:46] <th> so i'm off
[02:32] <fake> so.
[02:33] <fake> now i'll check wether i broke something ;-)
[03:39] <fake> looking good - time for some sleep.
[05:08] netrunner (n=andreas@anvame.net) left irc: Remote closed the connection
[05:14] netrunner (n=andreas@anvame.net) joined #rocklinux.
[07:50] [raphael] (n=raphael@raphael.netpark.at) left irc: "using sirc version 2.211+KSIRC/1.3.12"
[08:04] mnemoc_ (n=amery@user4-2.tutopia-dialup.ifxnw.cl) joined #rocklinux.
[08:06] Nick change: mnemoc_ -> mnemoc
[08:46] <blindcoder> moin
[08:49] <blindcoder> fake: well,when we switch to initramfs then size shouldn't be an issue anymore anyway
[09:37] th (n=th@montana.hbsn.de) left irc: "leaving"
[09:38] th (n=th@montana.hbsn.de) joined #rocklinux.
[09:40] <th> damn
[09:40] <th> 334 builds total, 7 completed fine, 0 with errors.
[09:40] <th> still
[09:40] <th> must be some endless loop
[09:40] <th> == 00:01:25 =[1]=> Building base/glibc=glibc23 [2.3.6 0].
[09:40] <blindcoder> yes
[09:41] <blindcoder> check for ld.so processes
[09:41] <th> that's with this patch applied...
[09:41] <th> yea there is a ld.so
[09:41] <th> so my crystal testbuild with this journal is screwed
[09:41] <th> do we have a fix for this?
[09:41] <blindcoder> no
[09:42] <blindcoder> it also only happened here with a pmmx optimisation
[09:42] <blindcoder> while killall ld.so ; do sleep 1 ; done
[09:42] <blindcoder> the package build then finishes successfully
[09:43] <th> pentium2 here i think
[09:44] <blindcoder> worked fine for me
[09:44] <th> ok - so if we have no fix i have to drop that journal
[09:44] <blindcoder> it's not the journal
[09:44] <blindcoder> I get it with an month-old tree
[09:44] <th> well crystal built before i applied it
[09:44] <blindcoder> cut-down LVP tree
[09:46] <blindcoder> the same tree built with generic, p2, p3, pm optimisation
[09:46] <blindcoder> but not with pmmx
[09:46] <th> export ROCKCFG_X86_OPT='pentium2'
[09:46] <blindcoder> *shrug*
[09:47] <blindcoder> still, I can say for certain that it's not your journal since the month-old lvp 0.5.2 tree also has it
[09:47] <blindcoder> and I sync the lvp tree only to vanilla rock trunk
[09:49] ija_ (n=ija@ joined #rocklinux.
[09:52] <th> must be some sort of deadlock
[09:52] <th> it's doing no library calls nor syscalls
[09:54] <th> #2  0x000063b0 in _dl_rtld_di_serinfo ()
[09:54] <th> hmm
[09:54] <th> gdb does not help a lot
[10:05] ija (n=ija@ left irc: Read error: 113 (No route to host)
[10:06] Nick change: ija_ -> ija
[10:38] <th> blindcoder: i think the reason for these locks is that you did not set a locale in your build environment
[10:41] <blindcoder> o_O
[10:41] <th> fake: [rock-sm] 2006012100164101355 fake status discarded
[10:41] <th> ugh
[10:41] <th> ah 
[10:41] <th> no matter that
[10:50] icelbox (i=simon@roedelkiste.net) left irc: Remote closed the connection
[10:55] <th> blindcoder: no it's not the local
[10:55] <th> blindcoder: i think that's a flwrapper problem
[10:57] <th> hmm perhaps i'm still in the blind
[10:58] <daja77_> hah! that's the solution http://strips.packetloss.biz/2006.03.14/Dilbert-2006.03.14.gif
[10:58] <th> hehe
[11:05] <th> the interesting thing is that i have this problem only in crystal
[11:06] <blindcoder> th: I had it in LVP
[11:06] <th> no problem in bbs
[11:06] <blindcoder> so it's not directly related to crystal
[11:06] <blindcoder> and in lvp only once
[11:06] <blindcoder> seems to be a sporadic thing
[11:06] <th> pretty reproducable here
[11:07] <th> ldconfig: Can't open configuration file /etc/ld.so.conf: No such file or directory
[11:07] <th> owell
[11:08] <th> well this is provided by glibc
[11:28] clifford (n=clifford@213-229-1-138.sdsl-line.inode.at) joined #rocklinux.
[11:48] <th> hey cliff
[11:49] <th> is that you? ;-)
[11:49] <clifford> no. ;-)
[11:49] <th> even better ;-)
[11:58] <clifford> Downloading download/mirror/d/dropbear-0.48.1.tar.gz ...
[11:58] <clifford> ######################################################################## 100.0%
[11:58] <clifford> Finished downloading 1473114 bytes in 16.457 seconds (89511.000 bytes/sec).
[11:58] <clifford> bzip'ing + cksum-test: download/mirror/d/dropbear-0.48.1.tar.gz
[11:58] <clifford> Segmentation fault (core dumped)
[11:58] <clifford> root@murphy:/data/SubMaster/rock.test# file core
[11:58] <clifford> core: ELF 32-bit LSB core file Intel 80386, version 1 (SYSV), SVR4-style, SVR4-style, from 'Download'
[11:58] <clifford> ... what trhe fuck ist that?
[11:59] <clifford> looks like there is a race in bash for handling SIGCHILDs!
[12:02] <th> perhaps broken hardware?
[12:06] <clifford> (gdb) bt
[12:06] <clifford> #0  0x0806da71 in kill_current_pipeline ()
[12:06] <clifford> #1  0x0806eb6a in kill_pid ()
[12:06] <clifford> #2  0x0806efd1 in kill_pid ()
[12:06] <clifford> #3  <signal handler called>
[12:06] <clifford> #4  0x4007d6f6 in free () from /lib/libc.so.6
[12:06] <clifford> #5  0x0806d7cc in save_pipeline ()
[12:06] <clifford> #6  0x0806d816 in cleanup_the_pipeline ()
[12:06] <clifford> .. no. that looks very much like a race for me.
[12:54] <daja77_> bash3?
[13:28] <owl> moin
[13:44] <stf^rocklinux> moin 2
[13:44] <owl> hi stf^rocklinux 
[13:44] <stf^rocklinux> hi owl :)
[14:25] <stf^rocklinux> th: how can I reproduce the bug?
[14:26] <th> stf^rocklinux: well
[14:26] <th> stf^rocklinux: i dont really know
[14:26] <th> stf^rocklinux: i can do anything you want on the current build where it is reproducable
[14:26] <stf^rocklinux> Crystal target, optimized for P2, any patches, special Config options?
[14:27] <stf^rocklinux> I think I'll try a default Crystal target without patches, optimized for P2 first...
[14:28] <stf^rocklinux> (I'll use tmpfs, but that doesn't count ;))
[14:32] <th> yes my journal
[14:32] <th> i can give you a full patch
[14:33] <th> blindy says he knows this issue
[14:33] <th> and its sporadic
[14:33] <th> Revision: 7183
[14:34] <th> http://thzn.de/journal.diff
[14:38] <clifford> daja77_: nope. bash2.
[14:40] <daja77_> hm
[14:41] <daja77_> btw what is the reason for not updating?
[14:52] <stf^rocklinux> daja: bash 3.1 as default bash causes some builds to fail (haven't investigated further). bash 3.0 otoh replaces 2.95 without problems.
[14:53] <clifford> daja77_: pure kainolophobia ;-)
[14:55] <stf^rocklinux> th: thanks
[14:55] <stf^rocklinux> th: I'll build with the journal then. btw. it applies to rev. 7210 as well.
[14:55] <th> stf^rocklinux: and my build host is a bbs package selection
[14:56] <daja77_> :D
[14:56] <stf^rocklinux> from ROCK 2.0.x?
[14:57] <th> NO
[14:59] <stf^rocklinux> whe have a bbs package selection in trunk?
[14:59] <stf^rocklinux> oh, really. wow ^^
[15:00] <th> yea
[15:00] <th> that should contain everything need for building and comitting
[15:04] <stf^rocklinux> nice. I've always envied 2.0.x for that...
[16:43] daja77 (n=daja77@dslb-088-072-035-235.pools.arcor-ip.net) joined #rocklinux.
[16:54] daja77_ (n=daja77@dslb-088-072-047-063.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out)
[17:15] Nick change: BoS_ -> BoS|afk
[20:33] [raphael] (n=raphael@raphael.netpark.at) joined #rocklinux.
[20:33] <[raphael]> hi!
[20:33] <[raphael]> you got some stuff done yesterday?
[20:34] <th> well 
[20:34] <th> yea sort of ;)
[20:37] <[raphael]> good, sort of is probably bettern than nothing
[20:37] <[raphael]> do you think you get a usable trunk sometime soon?
[20:38] <[raphael]> that is, I guess trunk is usable anyway, but has some udev quirks
[20:38] Action: daja77 subscribes [raphael] to the mailinglists ;)
[20:40] <[raphael]> daja77: oh, right, well, you have no obligation to answer any of my stupid questions that result from me being unsubscribed from the lists
[20:41] <[raphael]> daja77: and, please, don't subscribe me, not at the moment at least
[20:41] Action: [raphael] is trying to reduce mail traffic again, which is always a tough job
[20:42] <daja77> well i couldn't do that anyway
[20:42] Action: [raphael] thanks god
[20:42] <daja77> :p
[20:44] <blindcoder> heh, I know the feeling
[20:44] <blindcoder> just unsubscribed from a good dozen lists I've only been reading in for months anyway
[20:44] <blindcoder> if I read them at all
[20:45] <[raphael]> yes :)
[21:01] <th> of course [2006021511203329434] helped with my glibc[1] lockup...
[21:04] <blindcoder> being the locale patch of mine?
[21:05] <th> ack
[21:35] <fake> mornin'
[21:36] <owl> hi fake ^^
[21:36] <fake> clifford: uh-ow
[21:36] <fake> clifford: and that with my first updated package -_-
[21:38] <th> that's a good sign
[21:38] <th> i guess
[21:38] <th> ...
[21:38] <th> somehow
[21:39] Action: fake checking regresseion test base + bootdisk build...
[21:42] <fake> arghl
[21:42] <fake> udev failed
[21:42] <fake> because i didn't honor my own content... damn me.
[21:44] <fake> *comment
[21:44] <fake> hotfix...
[21:46] <fake> th: i take it your shadow/cracklib/pam + fixes journal got delayed?
[21:46] <[raphael]> night all!
[21:46] <fake> night [raphael] 
[21:47] <th> fake: due to that damn lockup in glibc[1] while creating locales
[21:47] <th> (localedef blocked)
[21:47] <fake> th: ic... interesting. never had that...
[21:47] <th> i know applied [2006021511203329434] which disables locale creation in stages < 2
[21:47] <th> now it's building
[21:47] <fake> th: okay
[21:47] <th> 334 builds total, 89 completed fine, 0 with errors.
[21:48] Action: fake saw Battlestar Galactica S2 E20 (season finale) yesterday
[21:49] <fake> boy, that series rocks
[21:50] <fake> damn, i need another hotfix... the whole bootdisk 2nd stage building patches were a mess!
[21:50] <fake> 1st stage i mean
[21:51] <th> details?
[21:51] <fake> th: especially your patch was way off
[21:51] <th> details?
[21:52] <fake> th: either i missed a patch (which i doubt), or something else weird happened, but i.e. 'udevstart' was still in build_stage1.sh at the time i applied your patch
[21:53] <th> yes it was
[21:53] <th> thats what i gathered from some patches
[21:53] <th> so it was not exactly my patch
[21:53] <th> lemme see
[21:53] <fake> th: nevermind
[21:53] <fake> th: i merged by hand (and made a small mistake)
[21:54] <fake> applying hotfix now...
[21:55] <th> are you talking of `udevstart` in liunuxrc.sh?
[21:55] <th> udevstart
[21:55] <th> cd /dev
[21:55] <th> rm -rf fd
[21:55] <fake> no
[21:55] <th> ...
[21:55] <th> ?
[21:55] <fake> see http://www.rocklinux.net/submaster/smadm.cgi?i=2006031421535608676
[21:56] <fake> there was 'udevstart' instead of 'udevd' in there and stuff
[21:56] <th> hehehe
[21:56] <fake> and then you invented the line break ;)
[21:56] <th> i did exactly THAT mistake 2 weeks ago ;)
[21:56] <th> line breaks rule!
[21:56] <th> we need so much more of them
[21:57] <fake> we need a policy about 80x25 terminals, you mean
[21:57] <th> something like that
[21:57] <fake> *shivers*
[21:58] <th> perhaps more a policy that we should not try to save lines of code
[21:58] <th> by putting everything in one line
[21:59] <fake> th: that's a given...
[21:59] <fake> testinstall running...
[22:00] <fake> looking good so far
[22:04] <fake> YES
[22:04] <fake> install works like a champ, machine comes up just fine
[22:05] <fake> now i'm waiting for your journal to finish, with the shadow, pam & cracklib fixes, trunk is usable again.
[22:05] <fake> @th
[22:06] <daja77> can you test if the permissions are set correctly when a usb device is plugged in?
[22:06] <daja77> later of course
[22:06] <fake> daja77: sure...
[22:06] <daja77> that bothers me in the latest install ..
[22:06] <fake> uh
[22:06] <fake> wait
[22:07] <th> 334 builds total, 98 completed fine, 0 with errors.
[22:07] <fake> a bluetooth dongle might not be the finest choice ;)
[22:07] <daja77> a digicam or scanner perhaps
[22:08] <fake> hm... usb storage stick?
[22:08] <fake> i posess neither ;)
[22:08] <fake> at least not at work..
[22:08] <daja77> storage is interesting too
[22:08] <fake> okay
[22:08] <daja77> can a user mount it
[22:09] <daja77> but i guess it is different than devices that are accessed by libusb
[22:09] <th> ok - we need a proposal for udev naming scheme
[22:09] <th> we could have different schemes at the distribution builders choice
[22:10] <fake> daja77: no usb stick around, anyways :(
[22:10] <fake> th: first thing now is to change /etc/udev/scripts back to /lib/udev
[22:10] <fake> th: 'like the author intended'-policy ;)
[22:11] <fake> and the udev initscript has to die
[22:14] <daja77> it just sucks to do chmod on /proc nodes ;)
[22:19] <stf^rocklinux> fake: don't forget to add the /lib/udev/dev directory :)
[22:20] <th> i think it's called "devices"
[22:20] <th> iirc
[22:21] <th> i'd really like this. that makes it easier to bootstrap a build-environment
[22:21] <stf^rocklinux> there's one open issue with the new shadow config files: only members of the 'wheel' group can use su...
[22:22] <stf^rocklinux> th: yes, it's called .../devices
[22:22] <th> well that's not an issue in it's nature. some people like it that way.
[22:22] <stf^rocklinux> The problem is we have no wheel group registered so far
[22:23] <fake> stf^rocklinux: i agree. we should patch that out...
[22:23] <fake> stf^rocklinux: i take it the shadow people use it more like an example of what you can do, than stating it is a default
[22:23] <stf^rocklinux> My idea was to create a 'wheel' group for shadow...
[22:23] <th> we should patch this feature out, cause we dislike it?
[22:23] <daja77> .oO (we should not put the rootpasswd into the installer's log ..)
[22:23] <th> or cause it would mean work to handle it the right way?
[22:23] <daja77> do we have an installer's log?
[22:24] <th> daja77: oh yea ;) that's fun
[22:24] <fake> daja77: no
[22:24] <daja77> :D
[22:24] <fake> am i missing a funny reference here?
[22:24] <th> yea
[22:25] <th> http://it.slashdot.org/it/06/03/13/0525254.shtml
[22:25] <daja77> fake: http://www.heise.de/newsticker/meldung/70754
[22:25] <th> "What's the problem? Open source passwords make it more secure. "
[22:25] <fake> LOL
[22:26] <daja77> we should create an installer's log just to show that ours is safe
[22:26] <fake> daja77: scriptable installs or on my agenda
[22:27] <fake> daja77: for that we'll need a logfile
[22:27] <daja77> kickstart ;)
[22:27] <stf^rocklinux> daja: we have postinstall/postremove logs 
[22:28] <stf^rocklinux> daja: and they don't contain passwords ;)
[22:28] <daja77> so they are safe :)
[22:42] <fake> ARGH!!!
[22:42] <fake> from the 'udev' build log:
[22:42] <fake> killall udevd
[22:42] <fake> udevd: no process killed
[22:42] <fake> make: [install-bin] Error 1 (ignored)
[22:42] <fake> rm -rf /dev/.udev
[22:42] <fake>  /sbin/udevd --daemon
[22:42] <fake> i can't believe what i'm seeing
[22:55] <stf^rocklinux> fake: udev-084/Makefile, line 298...
[22:56] <fake> stf^rocklinux: i guess i should patch that..
[22:56] <stf^rocklinux> fake: yes
[22:56] <fake> their assumption is right though.
[22:57] <fake> if you're not cross-compiling, and installing to /, a change in the rules file could wrack your system havoc
[22:57] <th> 84???
[22:57] <fake> 42 * 2
[22:57] <th> perhaps better do 43*2
[22:57] <fake> ?
[22:58] <th> why are you playing with old udev?
[22:58] <blindcoder> fake: they're starting udevd from the MAKEFILE?
[22:58] <th> (86 has the same killall + rm -f ...
[22:58] <fake> blindcoder: re-starting, yes
[22:58] <blindcoder> <comment />
[22:58] <fake> th: 'old' is a funny term wrt. udev
[22:58] <fake> th: i applied the update 84 just yesterday ;)
[22:59] <fake> th: i will send an update patch once my current /lib/udev / own rules file / own scripts patch is through
[23:00] <blindcoder> sleep happens
[23:00] <blindcoder> bye
[23:00] <fake> bye blindcoder !
[23:00] <fake> i hope you didn't take my rockinitrd move personal ;)
[23:00] <blindcoder> fake: the move to base?
[23:00] <fake> jep
[23:00] <blindcoder> since rockinitrd should be CORE that's fine with me
[23:01] <blindcoder> hell, I don't care where it is as long as it works :P
[23:01] <fake> ;)
[23:01] <blindcoder> anyway, I'm off
[23:02] <fake> daja77: /dev/bus/usb/1:
[23:02] <stf^rocklinux> cu blindcoder
[23:02] <fake> all nodes are owned by root:root, crw-rw----
[23:04] <th> there are some changes regarding this from 84->86
[23:04] <th> i just checked that diff
[23:14] <fake> mknod-calls are not caught by flistwrapper??
[23:39] <fake> now i'm curious wether the lib/udev/devices thing really works
[23:39] <fake> nope
[23:45] <fake> we need to copy those devices ourselves
[00:00] --- Wed Mar 15 2006