[00:02] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) left #rocklinux.
[00:14] blindcoder (i=id@tor/session/x-526e9ed2634caad2) left irc: Remote closed the connection
[00:20] <owl> gn8
[00:50] daja77 (n=daja77@dslb-088-072-033-076.pools.arcor-ip.net) joined #rocklinux.
[00:50] BoS_ (n=BoS@dslb-088-072-035-201.pools.arcor-ip.net) left irc: Remote closed the connection
[00:50] BoS_ (n=BoS@dslb-088-072-035-084.pools.arcor-ip.net) joined #rocklinux.
[01:00] daja77_ (n=daja77@dslb-088-072-033-073.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out)
[01:08] kasc_ (n=kasc@dslb-084-060-120-238.pools.arcor-ip.net) joined #rocklinux.
[01:17] kasc (n=kasc@dslb-084-060-096-083.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out)
[01:17] Nick change: kasc_ -> kasc
[02:13] annesh (n=hannes@cl-1568.ham-01.de.sixxs.net) left #rocklinux.
[04:53] netrunner (n=andreas@anvame.net) left irc: Remote closed the connection
[04:54] netrunner (n=andreas@anvame.net) joined #rocklinux.
[08:29] blindcoder (n=blindcod@onion.silicon-verl.de) joined #rocklinux.
[08:29] <blindcoder> moin
[09:48] ija_ (n=ija@212.80.242.103) joined #rocklinux.
[09:57] ija (n=ija@84.19.222.144) left irc: Nick collision from services.
[09:58] Nick change: ija_ -> ija
[10:35] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) joined #rocklinux.
[11:25] <th> ok - let's do some commits.
[11:53] <blindcoder> hehe
[12:09] <th> i think that's it
[12:12] <blindcoder> :)
[12:12] <blindcoder> gotta check sync with rock-community within the next few days then
[12:12] <th> well
[12:12] <th> i'm working to get it sync as well
[12:13] <blindcoder> did you change anything in the patches to get them to apply?
[12:13] <th> perhaps we should really drop that tree and see if we can get FAST commits for public things
[12:13] <th> blindcoder: few things. yes.
[12:15] <blindcoder> th: when we can get fast commits for public things I'll be more than happy to drop rock-community
[12:16] <th> blindcoder: i can apply public patches at once if it complies to what we decided at clt
[12:16] <th> blindcoder: just give me sm patch id's
[12:17] <th> they must apply to trunk
[12:17] <th> and they must have the necessary votes
[12:17] <th> (e.g. of the repository owner)
[12:17] <th> or any vote for public non-core packages
[12:17] <th> or target owner
[12:19] <th> e.g. i could apply 2006021415292611718 now - if you like.
[12:30] <blindcoder> well, in theory all patches should apply to trunk :)
[12:30] <blindcoder> let me compile a list
[12:31] <th> would be cool if your list contains the description of each patch
[12:31] <blindcoder> 2006021415292611718 2006021415390913035 2006021415563114033 2006022010210200943 ...
[12:32] <blindcoder> okay, wait
[12:32] <blindcoder> lvp doesn't need to, new version is only hours away
[12:32] <blindcoder> 2006021415390913035 (updated wesnoth)
[12:32] <blindcoder> 2006021415563114033 (updated hddtemp)
[12:32] <blindcoder> 2006021516085401581 (pkgmapper.in for ksimus)
[12:33] <blindcoder> 2006022010210200943 (update tor)
[12:33] <blindcoder> 2006022719001611926 (tor should shed its root privileges)
[12:33] <blindcoder> 2006030815553131500 (licq should use pkgsplit instead of pkgfork)
[12:33] <blindcoder> okay, that's it
[12:34] <blindcoder> all that only affect my packages
[12:36] th (n=th@montana.hbsn.de) got netsplit.
[12:36] BoS_ (n=BoS@dslb-088-072-035-084.pools.arcor-ip.net) got netsplit.
[12:36] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) got netsplit.
[12:36] toberkel (n=toberkel@toberkel.net) got netsplit.
[12:37] toberkel (n=toberkel@toberkel.net) returned to #rocklinux.
[12:38] BoS_ (n=BoS@dslb-088-072-035-084.pools.arcor-ip.net) returned to #rocklinux.
[12:38] th (n=th@montana.hbsn.de) returned to #rocklinux.
[12:45] <th> everything rejoined?
[12:46] <blindcoder> think so
[12:47] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) got lost in the net-split.
[12:49] <th> i bet i'm bending the defined policy a bit/
[12:49] <th> but i think we just need it.
[12:50] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) joined #rocklinux.
[12:51] <th> nano is NOT flagged CORE?
[12:51] <th> crystal does not include nano?
[12:51] <th> ahhhhh
[12:51] <th> nano != neon
[12:51] <th> damn
[12:54] <blindcoder> hehe
[12:54] <blindcoder> well, I would have waited for mathieu with the bitlbee update
[12:55] <th> that's what i meant
[12:55] <th> is he still active?
[12:56] <blindcoder> I don't know
[12:56] <blindcoder> but I know that esden isn't
[12:56] <th> how about sirkull?
[12:57] <blindcoder> no idea
[12:59] <th> who is "silicit"?
[13:00] <th> [M] Sebastian Knapp <rock@ccls-online.de>
[13:00] <th> is that giftnuss?
[13:01] <blindcoder> yes
[13:01] <th> so silicit == Sebastian Knapp == giftnuss?
[13:01] <blindcoder> yes
[13:02] <th> k
[13:02] Action: blindcoder suggests telling him to call the repo giftnuss
[13:02] <th> please do
[13:02] <th> at least the repository shall reflect the sm account name
[13:03] <th> svn: Commit failed (details follow):
[13:03] <th> svn: Can't convert string from native encoding to 'UTF-8':
[13:03] <th> haha
[13:03] <daja77> he could create a new sm account ;)
[13:06] <th> bsd-games is not flagged CORE either
[13:06] <th> so it should be moved to public
[13:07] <th> any votes for maintainership? (talking of repository)
[13:13] <th> i think if a submaster patch (e.g. for an update) is pending > 1.5 months.. then the original repository-owner shall not complain if patch get's applied without his/her vote.
[13:14] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) left irc: Read error: 110 (Connection timed out)
[13:15] <daja77> for th whole public?
[13:15] <blindcoder> tru
[13:15] <blindcoder> e
[13:15] <blindcoder> but I'd prefer not to see patches aging beyond a month anyway
[13:15] <daja77> +e
[13:16] <th> daja77: pardon?
[13:16] <th> daja77: ahhhhh no
[13:16] <th> daja77: for the packages i moved to public
[13:17] <blindcoder> hmm
[13:17] <blindcoder> why does 1-glibc23 hang here
[13:17] <blindcoder> make[2]: Entering directory `/root/0.5.2.1/src.glibc23.1142247841.20139.1091621689/glibc-2.3.6/localedata'
[13:17] <blindcoder> .././scripts/mkinstalldirs /root/0.5.2.1/build/pmx-TRUNK-x86-pentium-mmx-32-lvp/usr/lib/locale
[13:17] <blindcoder> mkdir /root/0.5.2.1/build/pmx-TRUNK-x86-pentium-mmx-32-lvp/usr/lib/locale
[13:17] <blindcoder> aa_DJ.UTF-8...
[13:18] <blindcoder> there the build sits and waits
[13:18] <blindcoder> 23899 root      25   0 54276  52m  752 R 99.2  5.9 102:50.59 ld.so
[13:19] <blindcoder> o_O
[13:19] <blindcoder> only when building for pmmx though
[13:37] <th> daja77: mind voting on 2006021615383712705?
[13:42] <blindcoder> th: want to upload the CLT iso to iso.rl.org?
[13:42] <th> no
[13:42] <daja77> haven't tried that ..
[13:42] <th> i'd prefer fixing errata first.
[13:42] <daja77> ack
[13:42] <th> daja77: there is no .conf in trunk
[13:44] <daja77> i have one here, perhaps due some other patches from netrunner
[13:44] <th> daja77: could you please clarify this somehow?
[13:45] <blindcoder> th: fixing errata meaning?
[13:45] <th> blindcoder: i dont want to upload the buggy CLT iso. i'd prefer to fix what's in the errata first.
[13:46] <th> daja77: 2006021416201427153 that's adding the .conf
[13:47] <th> daja77: could you perhaps create a new patch and vote agains all others, regarding qemu?
[13:48] <th> blindcoder: 2006021715383022606  which packages? i had no shared files in a crystal build.
[13:48] <blindcoder> th: I see, okay.
[13:48] <blindcoder> YAY! < 100 patches left *hugs th*
[13:49] <th> how many did i apply?
[13:50] <th> what did we decide regarding sane-backends? shall it stay CORE?
[13:50] <th> it must not be CORE but in public repository
[13:51] <th> how does 2006022108000419325 correspond to 2006031219303102828??
[13:51] <th> the latter states it is WIP
[13:51] <th> (this is for livecd target)
[13:51] <daja77> th: iirc yes
[13:52] <th> daja77: so i move it to base?
[13:52] <daja77> hm
[13:53] <daja77> the problems i had with qemu were not related to the avm patches ... just noticed
[13:54] <th> blindcoder: could you clarify the livecd target issue?
[13:54] <th> blindcoder: we have two patches here
[13:56] <th> okay!
[13:56] <th> i think everything else is unclear or CORE
[13:56] <th> blindcoder: perhaps you would like to create a diff between community and trunk and see if you can create some non-CORE-related sub-patches from it
[14:16] <blindcoder> th: shared files with quota
[14:16] <blindcoder> th: the livecd was a duplicate patch
[14:18] <blindcoder> th: all the community changes are in submaster so there's no point in creating another set of duplicate patches
[14:29] <th> blindcoder: ok so could we mark one as duplicate?
[14:29] <th> blindcoder: then i'd apply the other
[14:32] <blindcoder> th: the discarded one is the duplicate
[14:33] <th> 2006022108000419325 2006031219303102828 are both NOT discarded
[14:33] <blindcoder> huh?
[14:33] <blindcoder> just a moment
[14:33] <blindcoder> I'm confused
[14:34] <blindcoder> ah
[14:34] <blindcoder> both are needed
[14:34] <blindcoder> they're incremental
[14:34] <th> they apply in any order?
[14:34] <th> else there should be deps set
[14:34] <th> and the latter is commented to be WIP
[14:36] <blindcoder> well, I can't set deps on stfs patches
[14:36] <blindcoder> first: 2006022108000419325
[14:36] <blindcoder> second: 2006031219303102828
[14:36] <th> did you test a build with both patches?
[14:36] <blindcoder> no
[14:37] <th> hmm
[14:37] <blindcoder> because one is my patch and the other is stf
[14:37] Action: blindcoder != stf
[14:37] <daja77> true
[14:37] <th> well
[14:37] <th> it's not core
[14:37] <th> who's 'owner' of livecd target?
[14:38] <th> and another thing...
[14:38] <th> blindcoder: 2006022714301919554 please vote
[14:39] <blindcoder> th: fake is livecd owner
[14:39] <blindcoder> th: voted
[14:40] <th> fake: ping
[14:40] <th> blindcoder: applied
[14:40] <blindcoder> th: thanks :)
[14:43] <th> clifford: ping
[14:43] <th> clifford: please move mysql from your repository to base.
[14:43] <th> clifford: it's CORE so it MUST not be in personal repository.
[14:47] <clifford> yup.
[14:48] <th> i can do this
[14:48] <th> just ack it
[14:48] <blindcoder> daja77: 2006031314472619288
[14:48] <clifford> as you like. I just have to wait for my 'svn up' to complete. then I can move it..
[14:49] <th> clifford: k - just do it.
[14:51] <clifford> done.
[14:51] <th> ok - so i'll add the mysql update to my next journal
[14:53] <daja77> blindcoder: isn't that src line that you modified first missing now?
[14:53] <th> clifford: what was our decision for sane-backends?
[14:54] <th> clifford: shall it stay CORE and be moved to public?
[14:54] <blindcoder> daja77: huh?
[14:54] <th> clifford: shall it stay CORE and be moved to BASE of course?
[14:55] <blindcoder> daja77: root@ceres:/usr/src/rock-clifford# ./scripts/Create-ISO
[14:55] <blindcoder> daja77: ...
[14:55] <blindcoder> daja77:         -source         include package sources on the images
[14:55] <daja77> the reason it is core atm is that it triggers scanning software in kdegraphics which can't be added later without recompiling
[14:56] <th> sounds like it should be moved then
[14:56] <clifford> th: I'm not sure. I'd move it to base because I'm not sure what will be broken if we kick it out of CORE..
[14:57] <th> k
[14:57] <th> will do
[15:06] [raphael] (n=raphael@raphael.netpark.at) joined #rocklinux.
[15:12] <th> ok - i defined my next journal
[15:54] <th> 1Downloading download/mirror/s/sane-backends-1.0.17.tar.gz ...
[15:54] <th> curl: (22) The requested URL returned error: 404
[15:54] <blindcoder> humm
[15:55] <blindcoder> time to renew my mobile phone contract
[15:56] <daja77> oh dl url changed
[15:56] <daja77> ftp://ftp.sane-project.org/pub/sane/sane-backends-1.0.17/sane-backends-1.0.17.tar.gz is that now
[15:56] <daja77> they've moved away from mostang now i guess
[15:57] <th> yea
[15:57] <daja77> it worked when I made the update, it is not on the mirrors?
[15:57] <th> it wont hurt if i use the tu-wien mirror i guess
[15:58] <th> daja77: it wont reach the mirrors until committed to trunk
[15:58] <daja77> yes
[15:59] <th> for x in boot bbs crystal; do ./scripts/Build-Target -cfg $x ; done
[15:59] <blindcoder> th: good luck
[16:00] <th> thanks
[16:29] <fake> hi
[16:29] <fake> th: pong
[16:29] <owl> hi fake 
[16:29] <blindcoder> moin fake 
[16:29] <fake> moin!
[16:30] Action: fake rebuilding sysfiles...
[16:30] <fake> th: cool apply run ;(
[16:30] <fake> ;) i mean
[16:30] <th> fake: policy enforcer ;)
[16:30] <fake> i have a 64 bit build stuck at the bootloader btw
[16:30] <th> fake: i pinged you cause you are livecd owner
[16:30] <fake> i 0nwnz0r it
[16:31] <th> fake: there are two patches in submaster
[16:31] <th> i think you did not vote yet
[16:31] <fake> hrm
[16:31] <fake> 'needs a positive vote from (fake)' is not a good search query
[16:32] <th> fake: 2006022108000419325 2006031219303102828
[16:32] <fake> th: the one from stf modifies the old package selection sets
[16:33] <th> fake: could you rediff all of it in one big patch?
[16:35] <fake> th: later - i need to fix this sysfiles stone_mod_install thing first
[16:35] <th> sure
[16:35] <fake> i wanted to do that apply run today ;P
[16:35] <th> fake: i've kde update, pam, pppd and such in my current journal
[16:36] <fake> th: kay
[16:36] <fake> th: i have pam, shadow, cracklib etc in again, too
[16:36] <fake> th: i mailed my journal to the list, actually)
[16:36] <th> +gnupg, firefox, cracklib, gpm things
[16:43] <th> fake: i'm waiting for our new fancy, eyecandy, smng
[16:47] <blindcoder> there are way too many next generations these days
[16:48] daja77_ (n=daja77@dslb-088-072-047-063.pools.arcor-ip.net) joined #rocklinux.
[16:48] <daja77_> grmpf
[16:50] daja77__ (n=daja77@dslb-088-072-047-063.pools.arcor-ip.net) joined #rocklinux.
[16:50] daja77__ (n=daja77@dslb-088-072-047-063.pools.arcor-ip.net) left irc: Client Quit
[16:52] daja77 (n=daja77@dslb-088-072-033-076.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out)
[17:20] <fake> damn provider...
[17:20] <daja77_> yep ..
[17:33] <fake> bind-mounting dev to /mnt/target/dev doesn't work
[17:34] <fake> blindcoder: see above
[17:42] <fake> or does it?
[17:42] <fake> if i do it manually it works... grml..
[17:44] <fake> ... because a ramfs is mounted over it. wee ;)
[18:31] Action: fake thinking about the missing-native-64bit-bootloader-problem
[18:31] <fake> proposal: in the bootdisk, build a x86_64 (cross-)toolchain and crosscompile a 64 bit kernel along with the 32 bit kernel, as we do (read: will do) on sparc32/64
[18:32] <fake> that way, one can choose to boot that 64 bit kernel for a 64 bit install (chroot will work)
[18:36] <fake> http://www.rocklinux.net/submaster/smadm.cgi?i=2006031318351814466
[19:10] clifford (n=clifford@213-229-1-138.sdsl-line.inode.at) left irc: Read error: 110 (Connection timed out)
[19:13] <[raphael]> th: what's the status of udev, is it already "functional"?
[19:14] <[raphael]> (status in ROCK that is, not by itself)
[19:15] <th> [raphael]: seems a bit flaky when you have scsi
[19:15] <th> [raphael]: thats because the udev-helper scripts in rock are horribly broken
[19:16] <th> [raphael]: i'll work in more detail on it after my current journal
[19:16] <[raphael]> i see, thx for updating me
[19:17] <th> no problem
[19:18] <th> how is kolab status?
[19:18] <[raphael]> kolab 2.1 not yet released
[19:18] <[raphael]> but, the updating process in kolab is "supported"
[19:19] <[raphael]> that is, you usually can update any installation to the next
[19:19] <[raphael]> version
[19:19] <th> cool
[19:19] mnemoc (n=amery@user4-2.tutopia-dialup.ifxnw.cl) joined #rocklinux.
[19:19] <[raphael]> so, if I install a beta now, the update procedure from the beta to the final release is documented
[19:19] <th> and it would work with rock-trunk, right?
[19:20] <[raphael]> yes :)
[19:20] <[raphael]> which is nice actually
[19:20] <[raphael]> or, it works almost
[19:20] <th> even with current trunk?
[19:20] <[raphael]> it needs some patching
[19:20] <th> heavy patching needed?
[19:20] <th> what sort of changes?
[19:20] <[raphael]> honestly, I didn't test current trunk
[19:20] <[raphael]> the one from, say, last september
[19:20] <[raphael]> usually there are only package updates, sometimes also configuration updates and hopefully no LDAP changes
[19:21] <[raphael]> actually I don't know if 2.0 -> 2.1 will be supported - so far there never was any second release of kolab2 ;)  but I guess it is
[19:21] <[raphael]> I'm sure they will document the steps from 2.0 to 2.1, but it might require a bit of work (or care)
[19:22] <[raphael]> th, I also tested a trunk from mid-december
[19:22] <[raphael]> that worked as well
[19:22] <[raphael]> but the build itself acted a bit weird because devfs support was already removed from 2.6 kernel, so I removed the install again from my laptop
[19:23] <th> yea
[19:26] <fake> blindcoder: i will add /bin/cat to the rockinitrd now. we need sed & find now, anyways, so cat is 'grad wurscht'
[19:26] <th> well
[19:26] <th> there is a patch pending
[19:26] <th> which is required for scsi udev scripts
[19:26] <fake> th: one?
[19:26] <th> 2006022810451609842
[19:27] <fake> th: i'm just completeing the 2nd additional one that builds ontop of 3 others or so
[19:27] <fake> stinky mess
[19:27] <th> sed, cut, and basename were missing.
[19:27] <fake> tr, find and cat are missing, too
[19:27] <th> tr is added already
[19:27] <th> as basename is
[19:27] <fake> no
[19:27] <th> i did that today
[19:27] <fake> not to package/blindcoder/rockinitrd/initrd_base.txt
[19:28] <th> oh
[19:28] <fake> patch is in the works...
[19:28] <th> i was was talking of build_stage1.sh
[19:28] <fake> build_stage1 is for bootdisk ;)
[19:28] <fake> build_stage1 is mkinitrd on steroids
[19:29] <fake> (with a completely different code path to make things more interesting)
[19:32] <fake> http://store.ethereal.com/cheatshirt.html
[19:32] <fake> lol
[19:34] <th> hehe
[19:41] <fake> I AM SO SICK OF IT
[19:42] <fake> GNAAAR
[19:42] <daja77_> of the shirt?
[19:42] <fake> no, UDEV!
[19:42] <daja77_> ah, like blindy
[19:42] <th> well it is a big change. that's for sure
[19:43] <fake> the /dev/sda{1,2,3,4} files are created, but /dev/scsi/host0/bus0/target0/lun0 contains only 'generic'!
[19:43] <fake> *sigh*
[19:44] <th> that's because of our terrible scsi script
[19:44] <th> we should steal one somewhere
[19:44] Action: mnemoc miss devfs :(
[19:44] <fake> that one is stolen
[19:44] <fake> ;-)
[19:44] <fake> it was part of udev once
[19:44] <daja77_> mnemoc: me too
[19:44] <fake> devfs was a bliss
[19:45] <[raphael]> well, a year ago you had plans to incorporate BSD kernels (and base systems?)
[19:45] <[raphael]> FreeBSD has devfs
[19:45] <mnemoc> :)
[19:45] <daja77_> *shudder*
[19:46] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) joined #rocklinux.
[19:46] <mnemoc> [raphael]: some plans, even been doables are quite hard due to the lack of time and other priorities
[19:47] <[raphael]> yes, I understand this one, I'm also a victim of this disease
[19:47] <fake> ah, now the devices are there
[19:47] <fake> nice.
[19:47] <fake> we need syslog running in the initrd
[19:47] <fake> so we can debug them crappy udev scripts
[19:48] <fake> GRRR!
[19:57] <fake> (of course the devices are created once i re-run the pseudo-udevstart in the readily booted system)
[20:10] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) left #rocklinux.
[20:10] <[raphael]> mnemoc: the issue with incorporating BSD kernels (and base systems) is that if you do this, you probably get something very similar (but potentially much less evolved) system such as pkgsrc, which is available for Linux and BSD base systems already. The difference ROCK makes is only that ROCK can be used to build complete systems out of one box, this is the only (or most notable) thing that makes it special, everything else already exists
[20:10] <[raphael]> like cross platform package systems and such
[20:11] <[raphael]> these things are actually coming into existience like .... whatever, plentiful at least
[20:12] <[raphael]> so I hope that ROCK really focuses on this DBK thing, which is after all its speciality
[20:15] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) joined #rocklinux.
[21:48] mnemoc (n=amery@user4-2.tutopia-dialup.ifxnw.cl) left irc: Read error: 110 (Connection timed out)
[21:51] <daja77_> w00t apply storm in my mailbox ;)
[21:52] <fake> jup
[21:53] <fake> i think we finally get a grip at udev
[21:53] Action: fake installing regression testing machine
[21:55] <daja77_> nice
[23:05] <SMP> _we_? you should say _you_ are finally able to look beyond what you have known ...
[23:11] <fake> SMP: 'we' in terms of 'our distribution build kit'
[23:11] <fake> SMP: and 'get a grip at udev' as in 'with all the patches in submaster applied, it actually kind of works'
[23:11] <fake> SMP: and: hi!
[23:12] <SMP> it has always worked for me -- especially not using patches from submaster ;->
[23:12] <fake> SMP: you could have shared your wisdom, you know ;)
[23:16] <fake> SMP: can you give a comment on rockinitrd?
[23:16] <fake> SMP: the root=/dev/ram0 smells evil
[23:17] <th> i think the root=/dev/ram0 will not be needed with initramfs
[23:17] <SMP> yes
[23:18] <th> and then it's up to that first stage to decide about the root
[23:18] <th> that's the fun of userspace then.
[23:19] <fake> the problem was, if i don't set root=/dev/ram0, i can't exec /sbin/init without an argument, because /dev/initctl gets created...
[23:19] <SMP> eh?
[23:19] <fake> exactly what i thought...
[23:19] <SMP> what do you need init for?
[23:20] <th> fake: talking of the next stage?
[23:20] <th> fake: we dont use /sbin/init in rockinitrd, do we?
[23:20] <SMP> and what creates /dev/initctl beforehand?
[23:20] <fake> i want to load my modules in the initrd, and start udev and stuff, and then i want to exec /sbin/init on my root filesystem.. don't i?
[23:20] <SMP> no
[23:21] <SMP> wait, do you use /sbin/init or /linuxrc on the initrd?
[23:21] <fake> if i don't set init=/linuxrc root=/dev/ram0, linuxrc isnt' process ID 1, that's why i cant exec(/sbin/init)
[23:21] <fake> right.
[23:22] <fake> the /dev/initctl thing was something else.
[23:22] <fake> SMP: /linuxrc (which is a shellscript now.. yuck)
[23:23] <SMP> by the way -- you really really really want to have a common 2.4-initrd 2.6-initramfs framework based off klibc
[23:23] <th> gaaaaaaah
[23:23] <SMP> heh
[23:23] <th> i doubt that it is matured enough
[23:23] <th> for our needs
[23:23] <th> i'd prefer a busybox version for the time being
[23:23] <th> ;-)
[23:23] <fake> i don't care, as long as it works ;)
[23:24] <SMP> it's supposed to run udev now
[23:24] <th> SMP: but the shell...
[23:24] <th> SMP: is too... feature-less
[23:24] <th> for my taste
[23:24] <fake> so, if it's not intended to exec /sbin/init from the initrd, what should i do?
[23:24] <th> but hey - busybox started to support being built with klibc ;-)
[23:25] <SMP> oh, boy. the shell's only purpose is to run a few shell scripts
[23:25] <th> it has to load modules
[23:25] <th> it has to care dm/lvm/dmcrypt
[23:25] <th> nfsroot perhaps
[23:25] <SMP> yeah. what funky features does a shell have to have for that?
[23:26] <th> some string-ops to read kernel0cmdline
[23:26] <th> s/0/-/
[23:26] <th> i think that was what was driving me nuts
[23:27] <fake> th wants an bbs-initrd ;)
[23:27] <fake> *a
[23:27] <th> almost
[23:27] <th> i'd like a modular loadable system
[23:27] <th> which can load rescue overlays
[23:28] <th> why should i leave my stage1, when i'm comfortable in it?
[23:28] <th> i could implement a bootloader in stage1
[23:28] <th> kexec ;-)
[23:29] <SMP> kexec r0x0r
[23:29] <th> did you play with it?
[23:29] <fake> so i take it both of you can't help me with my conceptual problem with initramfs vs. /sbin/init...
[23:30] <SMP> # grep -c kexec /sbin/init.d/system
[23:30] <SMP> 2
[23:30] <th> it should make no different how the pid==1 program is called
[23:30] <th> 0
[23:30] <fake> th: it does for '/sbin/init'
[23:30] <th> SMP: damn what are you doing?
[23:31] <SMP> fake: as long as it's just a conceptual problem, we only have conceptual solutions ;>
[23:31] <th> fake: why? just put the right script in /sbin/init
[23:31] <fake> o_O
[23:31] <th> .   run_init_process("/sbin/init");
[23:31] <th> .   run_init_process("/etc/init");
[23:31] <th> .   run_init_process("/bin/init");
[23:31] <th> .   run_init_process("/bin/sh");
[23:31] <th> choose one name
[23:31] <SMP> th: damn why?
[23:31] <fake> define 'right script'
[23:31] <th> fake: your linuxrc
[23:31] <th> SMP: why what?
[23:32] <fake> th: so you're saying i should rename /linuxrc to /sbin/init?
[23:32] <th> fake: yes.
[23:32] <SMP> ? -> <th> SMP: damn what are you doing?
[23:32] <th> SMP: with kexec
[23:32] <fake> th: but /linuxrc gets called by the kernel, too, just like /sbin/init would...
[23:32] <SMP> hehe
[23:32] <th> fake: are you sure?
[23:32] <fake> th: well, almost
[23:33] <fake> th: but i'll try
[23:33] <th> fake: init/main.c
[23:33] <th> last lines
[23:33] <th> fake: basically this just saves a kernel-param
[23:33] <SMP> th: I have to use an external USB hdd for my system-vg on my laptop. but the bios hangs when it's connected to the laptop before the linux kernel has started ... se the point of kexec ..? ;-)
[23:33] <fake> th: did you put rockinitrd into your current journal?
[23:33] <th> fake: no
[23:34] <th> fake: i wanted to do the non-udev non-initrd things first
[23:34] <fake> th: then a few more things will pile up...
[23:34] <th> fake: what do you mean?
[23:34] <fake> th: i'm gonna try this now, the regression testing machine is installed...
[23:34] <th> fake: what would that be?
[23:34] <th> fake: do you have virtually infinite computing power for this?
[23:35] <fake> th: renaming rockinitrd's linuxrc to sbin/init
[23:35] <th> ahh cool
[23:35] <fake> th: p4 3ghz 1gig ram
[23:35] <th> btw - i found out that missing to set the timezone in your build-environment is fatal when building xorg
[23:35] <th> cause `date` will contain --
[23:35] <th> and that will end a xml comment too early
[23:35] <fake>  -- see zic manpage ;)
[23:35] <th> in some xorg font.conf
[23:36] <fake> *rofl*
[23:36] <SMP> holy crap
[23:36] <th> so all fonts are messed
[23:36] <th> took me a while to find this one
[23:38] <daja77_> yeah figured that out once with rxr, thought we had a fix for that
[23:39] <th> we should warn in our Build-Target script perhap
[23:39] <th> s
[23:40] <fake> th: scripts/Check-System
[23:40] <SMP> yeah, that crappy timezone pops up at some other places, too
[23:40] Action: fake remembers swatch 'the beat' 
[23:40] <th> hahaha
[23:41] <fake> th: i'm tempted to apply that whole udev/rockinitrd batch
[23:41] <th> fake: well.
[23:41] <fake> th: first thing after that is to move everything from /etc/udev/scripts to /lib/udev
[23:41] <th> fake: it would put trunk in a better state
[23:41] <th> fake: but it violates policy ;)
[23:41] <fake> th: after that, add the firmware loading helpers
[23:42] <fake> th: i'll move rockinitrd from package/blindcoder to package/base first ;0)
[23:42] <th> that's a required move
[23:42] <fake> th: then how would it be against policy?
[23:43] <fake> next thing after firmware loader would be remove that udev initscript... *shakes head*
[23:43] <th> fake: policy is that tests are required for CORE commites
[23:43] <th> s/tes/ts/
[23:43] <fake> th: i just installed my test build
[23:43] <fake> after re-building the iso 12 times
[23:44] <th> hmm
[23:44] <th> well
[23:44] <th> you would not want to move it to /lib/udev first?
[23:44] <fake> no, absolutely not
[23:44] <th> i see
[23:44] <th> lemme check my journal
[23:45] <fake> any more patches on top of that stack and i'll totaly loose it ;)
[23:45] <SMP> I have a feeling it would be best to throw away all udev patches, not apply them and then fix the mess later
[23:46] <fake> SMP: i started trying to 'fix the mess' 2 months ago
[23:46] <th> SMP: well there is some mess in trunk already
[23:46] <fake> exactly.
[23:47] <fake> th: another important thing is to not patch the debian rules file, but to create our own one
[23:47] Action: daja77_ nods sleepy
[23:47] <th> fake: we should define what we want first
[23:47] <th> what naming policy
[23:47] <th> i'd like to have /dev/discs/
[23:47] <SMP> nooo
[23:48] <th> hehe
[23:48] Action: SMP throws in a "pluggable!" and takes cover ;p
[23:48] <fake> th: like it is now?
[23:56] <th> what kind of pluggable?
[23:56] <th> fake: no. that's broken for scsi
[23:57] <fake> hey, how about /dev/${DISK_LABEL}
[23:57] <fake> then we can create funny disks called sda1 ;-)
[23:58] <th> /dev/label/...
[23:58] <th> /dev/label/MY_LABEL
[23:58] <SMP> it's there. /dev/disk/by-label 
[23:58] <fake> and when my label is "../../root" ?
[23:58] <fake> *gg*
[23:58] <fake> *if
[23:59] <th> then'l label my usb stick ../etc/shadow
[00:00] --- Tue Mar 14 2006