T2 IRC Log: 2009-04-02
This is the log as captured by an IRC bot in the channel. The statements are those of the individual people and might not neccessarily reflect the policy and legal rules as set forth by the T2 SDE Project.
--- Log opened Thu Apr 02 00:00:13 2009
00:57 < the_rainbow0> Actually, I found out that: ./script/Build-Pkg -cfg myconfig package, installs the package into the actual system, and doesn't update the filesystem I made previously using Build-Target, I guess I should indicate the root directory to Build-Pkg like this: ./script/Build-Pkg -cfg myconfig -root ./buid/myconfig package, I'm trying this out
01:54 -!- koan [n=koan@unaffiliated/koan] has quit ["Lost terminal"]
01:54 -!- koan [n=koan@unaffiliated/koan] has joined #t2
05:16 -!- the_rainbow0 [i=ee@41.209.167.146] has quit [Read error: 110 (Connection timed out)]
05:56 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: dsoul
05:57 -!- Netsplit over, joins: dsoul
07:56 -!- synchris [n=synchris@athedsl-192134.home.otenet.gr] has joined #t2
07:56 -!- T-One [n=T-One@62.218.122.66] has joined #t2
07:57 -!- synchris_odium [n=synchris@pdpc/supporter/professional/synchris] has quit [Read error: 60 (Operation timed out)]
09:12 < rxr> re
09:14 < rxr> http://t2-project.org/packages/samba.html
09:38 < CIA-41> rene * r32794 /trunk/package/database/sqlite/sqlite.desc: * updated sqlite (3.6.10 -> 3.6.12)
09:38 < CIA-41> rene * r32795 /trunk/package/network/samba/samba.desc: * updated samba (3.3.2 -> 3.3.3)
09:39 < CIA-41> rene * r32796 /trunk/package/graphic/gocr/gocr.desc: * updated gocr (0.46 -> 0.47)
09:39 < CIA-41> rene * r32797 /trunk/package/xorg/xrandr/xrandr.desc: * updated xrandr (1.2.3 -> 1.3.0)
09:45 < CIA-41> rene * r32798 /trunk/package/archiver/zziplib/ (install.patch zziplib.desc): * updated zziplib (0.13.49 -> 0.13.50)
10:21 -!- RayMan72 [n=radjis@ip245-128-173-82.adsl2.static.versatel.nl] has joined #t2
11:25 -!- the_rainbow0 [n=ee@41.209.167.135] has joined #t2
11:28 < the_rainbow0> Hello guys
11:30 < the_rainbow0> I finally buit my T2 target, and got the ISO for the livecd, I used kernel 2.6.29, but after it boots, it can't mount the live.squash image, because of the version mismatch, I guess it needs squshfs 4.0?
11:32 < rxr> yeah - inded
11:32 < rxr> indeed
11:33 < rxr> you need the not yet released version 4 tools
11:33 < rxr> and also either unionfs or aufs if you want a working livecd
11:33 < rxr> you disabled unionfs yesterday due to 2.6.29, I hope aufs still built?
11:55 < the_rainbow0> thanks for your answer, aufs still builds yes
11:55 < the_rainbow0> I disabled squshfs and unionfs
11:56 < the_rainbow0> tell me rxr, can't I download the dev version of squashfs 4 and unionfs and build them with T2? You know I'm getting tired of waiting for the squashfs 4
11:56 < the_rainbow0> final version
11:57 -!- EnTeQuAk [n=EnTeQuAk@static.142.84.46.78.clients.your-server.de] has left #t2 []
12:07 < rxr> the_rainbow0: maybe some cvs snapshot
12:13 < the_rainbow0> cool, then I think I'm gonna try it out
12:14 < the_rainbow0> buit I don't understand how some linux distro could include the kernel 2.6.29, within their livecds like Fedora 11
12:17 < rxr> most linux distributions are highly modified
12:17 < rxr> especially redhat ships unreleased stuff for quarters until some maintainer of theirs get to push their changes upstreamm
12:19 < the_rainbow0> ahaaa, okay okay I see
12:23 -!- the_rainbow0 is now known as kernel
12:23 -!- kernel is now known as the_rainbow
12:23 < rxr> I commit the squashfs-tools cvs update in some seconds
12:24 < the_rainbow> ah that's greaaat, I'm waiting for it
12:24 < rxr> done
12:24 < rxr> CIA is lagging
12:24 < CIA-41> rene * r32799 /trunk/package/filesystem/squashfs-tools/ (11 files):
12:24 < CIA-41> * updated squashfs-tools (3.4 -> 4.0-cvs), removed previous
12:24 < CIA-41> kernel patch glue, it's upstream since 2.6.29
12:25 < rxr> only lightly tested with: mksquashfs scripts sq ; mount -o loop sq /mnt/; l /mnt ; # ...
12:25 < the_rainbow> coool, I'm gonna try it out
12:25 < rxr> have fun - cu then - have to concentrate on some work
12:25 < the_rainbow> will I have a working livecd without unionfs?
12:25 < the_rainbow> Ah thanks man, and good luck in your work
12:29 < rxr> yes, the t2 scripts try aufs and as fallback unionfs (I think)
12:29 < rxr> (if aufs works with 2.6.29, not tested yet on my side)
12:29 < rxr> you know - they break so much "APIs" with each kernel release ...
12:29 < the_rainbow> yep I know, anyway, I'm gonna try that out and let you know
12:30 < rxr> have fun
12:30 * rxr lunch and then a pile of work - cu
12:49 -!- the_rainbow [n=ee@41.209.167.135] has quit [Read error: 104 (Connection reset by peer)]
12:49 -!- the_rainbow [i=ee@41.209.167.135] has joined #t2
13:16 < rxr> http://edsjunk.net/tutorials.html
13:57 -!- mpp [n=mpp@2001:470:9986:340:0:0:0:3] has joined #t2
13:57 < mpp> moinsen
14:55 < CIA-41> aldas * r32800 /trunk/package/base/coreutils/ (coreutils.conf coreutils.desc): * updated coreutils (7.1 -> 7.2)
15:08 < rxr> moinsen moinsen mpp
15:09 < mpp> hey rxr
15:35 -!- the_rainbow0 [i=ee@41.209.167.135] has joined #t2
15:35 -!- the_rainbow [i=ee@41.209.167.135] has quit [Read error: 104 (Connection reset by peer)]
15:37 < the_rainbow0> failed to loop mount live.squash
15:38 < the_rainbow0> it seems that the latest squashfs-tools snapshot isn't mature enough to run a livecd
15:42 -!- mqueiros [n=mqueiros@217.70.75.198] has quit ["Ex-Chat"]
15:45 -!- mqueiros [n=mqueiros@217.70.75.198] has joined #t2
15:45 < rxr> maybe the wrong squashfs tools where used and you got a version 3 file
15:45 < rxr> you can check with the file utility
15:46 < rxr> file ...path-to-the-mounted-iso/live.squash
16:00 < the_rainbow0> okay, I'm gonna check it out
16:01 < the_rainbow0> version 3.1
16:01 -!- whorrel [n=whorrel@122.173.27.235] has joined #t2
16:02 < the_rainbow0> but, why isn't it using the squshfs-tools 4, that I downloaded from trun a while ago?
16:03 < rxr> maybe it accidently used it from the build system
16:03 < the_rainbow0> just a note: I'm only updating the kernel in my build, to keep all the other packages built, I'm doing so like that: ./script/Build-Target -cfg myconfig -job 5-linux26
16:03 < rxr> can you install (e.g. Emerge-Pkg) the squashfs tools into the host system for a test ?
16:03 < rxr> you do not need to mess with the linux26 package
16:03 < rxr> only squashfs-tools
16:03 < the_rainbow0> so I don't need to rebuild the whole kernel?
16:03 < rxr> ok - before you emerge, rm buid/.../var/adm/logs/*-squashfs-tools*
16:04 < rxr> and re Build-Target
16:04 < the_rainbow0> okay
16:04 < rxr> no, it#s in the kernel now
16:04 < rxr> you only need the matching user-space tools to create the FS
16:04 < rxr> version 4.0 squashfs
16:04 < the_rainbow0> okay, I'm gonna try to emerge it to test it locally
16:05 < rxr> please also rebuild the squashfs-tools in your targe
16:05 < rxr> t
16:05 < the_rainbow0> like that: ./script/Build-Target -cfg myconfig -job 5-squashfs-tools
16:05 < rxr> if there are no bugs t2 build scripts should use the tools build along the new toolchain in your new sandbox'ed system
16:05 < rxr> yeah - just that the toolchain is build in stage 0
16:05 < rxr> so also at least -job 0-squashfs-tools
16:05 < rxr> but best just remove the log files as advised above
16:06 < the_rainbow0> okay I'm going to do that
16:06 < rxr> and re-run Build-Target -cfg ... without a -job so t2 will build the removed log file packages
16:06 < the_rainbow0> so I put the stage 0 only?
16:06 < the_rainbow0> okay okay
16:06 < rxr> in theory that should be enough - but cleaned is to wipe the package log and rebuild all of the package
16:06 < the_rainbow0> you mean the squashfs-tools package?
16:06 < the_rainbow0> only this one need to be built
16:07 < rxr> yes, it's bulit at least twice
16:07 < rxr> in stage 0 for the toolchain (also used to build the CD ISO)
16:07 < rxr> and in stage 5 for the new, sandboxed system
16:07 < rxr> in theory the stage0 rebuild should be enough to get a matching ISO for your new kernel
16:07 < the_rainbow0> so after I remove the logs, I'm just gonna build the target without a job switch
16:07 < rxr> but to have the new tools in the installed system the stage 5 rebuild is necessary, too
16:07 < rxr> yes
16:07 < the_rainbow0> okay I'm gonna try that out
16:08 < rxr> the Build-Target uses the package logs to know what needs to be build and what's already there
16:08 < the_rainbow0> ah okay great
16:10 < the_rainbow0> by the way rxr, by reading the t2-handbook I found some mistakes, I should send them to t2 in order to correct them
16:10 < the_rainbow0> also, it's good to add some new sections in the book
16:10 < the_rainbow0> for advanced use
16:10 < rxr> yes, please send
16:10 < rxr> notw that some typos where fixed since the last publishing
16:10 < rxr> in the handbooks SVN tree
16:10 < the_rainbow0> okay
16:10 < rxr> http://svn.exactcode.de/t2-handbook/trunk/src/en/book/
16:11 < rxr> be sure to base on that for sending in fixes
16:11 < the_rainbow0> sure
16:11 < rxr> (yes, we'll re-generate the web/html view and pdf's soon)
16:12 < the_rainbow0> man, when I build the target there's a build error for squashfs-tools, it says that: No file to patch
16:12 < the_rainbow0> 2 out of 2 hunks ignored
16:12 < rxr> hm
16:13 < the_rainbow0> wait, I'm gonna send you the log
16:13 < rxr> maybe your cvs snapshot is differing from mine
16:13 < rxr> which patch?
16:13 < rxr> I guess the compression level option patch ?
16:14 < the_rainbow0> I'm sending you the log
16:16 < the_rainbow0> Removing old error logs ...
16:16 < the_rainbow0> -> Removing var/adm/logs/0-squashfs-tools.err ...
16:16 < the_rainbow0> == 16:13:17 =[0]=> Building filesystem/squashfs-tools [4.0-cvs 8.0-trunk].
16:16 < the_rainbow0> -> Reading package config: package/filesystem/squashfs-tools/squashfs-tools.conf
16:16 < the_rainbow0> -> Preparing build in src.squashfs-tools.beeserver.20090402.161314.7412.localhost
16:16 < the_rainbow0> -> Building. Writing output to $root/var/adm/logs/0-squashfs-tools.out
16:16 < the_rainbow0> Command Wrapper Debug: running 'gcc --version' ..
16:16 < the_rainbow0> gcc is /root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/wrapper/gcc
16:16 < the_rainbow0> Old Command: gcc --version
16:16 < the_rainbow0> ENVPREFIX = 'CMD'
16:16 < the_rainbow0> OTHERS = ''
16:16 < the_rainbow0> INSERT = '--help?:-pipe -Os'
16:16 < the_rainbow0> REMOVE = '-O -O[0-9s] -mtune* -march* -mcpu* -g*'
16:16 < the_rainbow0> APPEND = '-s'
16:16 < the_rainbow0> FILTER = ''
16:16 < the_rainbow0> conditonal arg: --help?:-pipe.
16:16 < the_rainbow0> conditonal: --help, lhs: , rhs: -pipe.
16:16 < the_rainbow0> Insert: -pipe
16:16 < the_rainbow0> Insert: -Os
16:16 < the_rainbow0> Copy: --version
16:16 < the_rainbow0> Append: -s
16:16 < the_rainbow0> Old PATH: /root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/wrapper:/root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/bin:/root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/crosscc:/opt/samba/sbin:/opt/qt3/sbin:/opt/openldap/sbin:/opt/kde3/sbin:/opt/gnome2/sbin:/opt/samba/bin:/opt/qt3/bin:/opt/prelude/bin:/opt/openldap/bin:/opt/kde3/bin:/opt/gnome2/bin:/usr/games:/
16:16 < the_rainbow0> usr/X11/sbin:/usr/X11/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/opt/kde3/bin
16:16 < the_rainbow0> New PATH: /root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/bin:/root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/crosscc:/opt/samba/sbin:/opt/qt3/sbin:/opt/openldap/sbin:/opt/kde3/sbin:/opt/gnome2/sbin:/opt/samba/bin:/opt/qt3/bin:/opt/prelude/bin:/opt/openldap/bin:/opt/kde3/bin:/opt/gnome2/bin:/usr/games:/usr/X11/sbin:/usr/X11/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/
16:17 < the_rainbow0> opt/kde3/bin
16:17 < the_rainbow0> New Command: gcc -pipe -Os --version -s
16:17 < the_rainbow0> gcc (GCC) 4.2.2
16:17 < the_rainbow0> Copyright (C) 2007 Free Software Foundation, Inc.
16:17 < the_rainbow0> This is free software; see the source for copying conditions. There is NO
16:17 < the_rainbow0> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
16:17 < the_rainbow0> [ writing debug log to /root/t2-trunk/src.squashfs-tools.beeserver.20090402.161314.7412.localhost/cmd_wrapper.log ]
16:17 < the_rainbow0> `/root/t2-trunk/src.squashfs-tools.beeserver.20090402.161314.7412.localhost/archdir/squashfs-tools-2009-04-02.tar.bz2' -> `/root/t2-trunk/download/mirror/s/squashfs-tools-2009-04-02.tar.bz2'
16:17 -!- the_rainbow0 [i=ee@41.209.167.135] has quit [Excess Flood]
16:17 -!- the_rainbow0 [i=ee@41.209.167.135] has joined #t2
16:17 < rxr> excess flood, sorry
16:18 < rxr> the full log is in build/.../var/adm/logs/...
16:18 < the_rainbow0> oops, I'm sorry
16:18 < rxr> can you look yourself which patch failed to apply ?
16:18 < the_rainbow0> this one: /squashfs3.4-patch.orig 2008-08-27 07:52:07.000000000 +0200
16:18 < rxr> yeah - it's not me who kicked you, that was the feenode.net service ...
16:18 < rxr> hm?
16:18 < the_rainbow0> and: squashfs3.4/kernel-patches/linux-2.6.23/squashfs3.4-patch 2008-10-20 20:40:53.000000000 +0200
16:18 < rxr> what does svn st package/filesystem/squashfs-tools say ?
16:19 < the_rainbow0> wait
16:19 < rxr> have you svn up'ed package/filesystem/squashfs-tools ?
16:20 < the_rainbow0> M package/filesystem/squashfs-tools/compressionLevel.patch
16:20 < the_rainbow0> M package/filesystem/squashfs-tools/squashfs-tools.desc
16:20 < rxr> why did you mod them ?
16:20 < rxr> svn revert those and svn up ...
16:21 < the_rainbow0> I'm not really an exprerienced svn user, I just use svn to download source files and such stuff
16:21 -!- whorrel [n=whorrel@122.173.27.235] has quit ["Lost terminal"]
16:21 < rxr> svn revert package/filesystem/squashfs-tools/*
16:22 < the_rainbow0> ok
16:22 < rxr> svn up package/filesystem/squashfs-tools/
16:22 < the_rainbow0> done
16:23 < the_rainbow0> now, I'm trying to build the target again
16:23 < rxr> try to build the package again
16:23 < rxr> I cross my fingers and press my thumbs for you and go back to actual coding work, ...
16:23 < the_rainbow0> lol thanks a billion man :D, it seems that it's building fine !
16:24 < the_rainbow0> and yes it finished building it
16:24 < the_rainbow0> Good luck dude, I'll let you know about the final results
16:43 -!- the_rainbow0 is now known as the_rainbow
17:10 -!- T-One [n=T-One@62.218.122.66] has quit ["Verlassend"]
17:10 -!- RayMan72 [n=radjis@ip245-128-173-82.adsl2.static.versatel.nl] has quit ["i've to Leave"]
17:31 < the_rainbow> what's the default login for the T2 livecd?
17:35 < the_rainbow> Confirmation: The squashfs-tools snapshot works great, and the livecd boots fine
17:35 < the_rainbow> Just can't figure out the default root password for the livecd
17:42 < rxr> no default, we hat this on the mailing list some days ago
17:44 < rxr> http://article.gmane.org/gmane.comp.t2.devel/4055
17:45 < rxr> http://article.gmane.org/gmane.comp.t2.devel/4056
18:38 -!- mpp [n=mpp@2001:470:9986:340:0:0:0:3] has quit ["Verlassend"]
19:03 < the_rainbow> then, there's something I have to make in the config
19:03 < the_rainbow> because even with UN: root and PW: root it doesn't work
19:10 < rxr> as the links to the mailing list indicate you have to changes some lines of code to make the trunk livecd login work
19:10 < rxr> what livecd did you build? minimal? or + X.org graphic ?
19:11 < the_rainbow> minimal
19:15 < rxr> without any X ?
19:15 < the_rainbow> yep
19:15 < the_rainbow> I just need the minimum, no X needed
19:15 < rxr> ok - in the stable tree we setup kde to auto-login the guest
19:15 < rxr> user
19:16 < the_rainbow> what about bash?
19:16 < rxr> how would you want such a text live system to allow login ?
19:16 < rxr> we could generetate a random password on startup and set it for the guest user and print it to the screen
19:16 < the_rainbow> default login, like for example the Fedora LiveCD
19:16 < rxr> or would you like an fixed passwd ?
19:16 < rxr> I bet fedora has a graphic login with auto-login ?
19:16 < the_rainbow> yes, just a fixed one
19:17 < rxr> fixed passworD ?
19:17 < the_rainbow> yes it has a graphic login and Bash login as well
19:17 < rxr> a little insecure if 50% of the dau users do not change it
19:17 < the_rainbow> yes fixed password
19:17 < rxr> what password does fedora use ?
19:17 < the_rainbow> no password, just the user name, root
19:17 < the_rainbow> the root has no password on Fedora
19:18 < the_rainbow> okay, if I know how to do it with a random password, I would also know how to make it fixzd
19:18 < the_rainbow> fixzd*
19:18 < the_rainbow> fixed*
19:18 < rxr> not really - for a fixed password you could simply inject a predefined etc/shadow file
19:18 < rxr> for a random password some lines of code are required
19:18 < the_rainbow> okay okay I see
19:19 < the_rainbow> how can t2 run post scripts after the build in a chrooted environment?
19:20 < rxr> you can
19:20 < rxr> but for a fixed password you do not need that
19:20 < rxr> and for random passwords I would suggest generating it at boot time
19:20 < the_rainbow> okay, I'm gonna start with a fixed password
19:20 < the_rainbow> where should I start?
19:21 < the_rainbow> to inject my predefined line into the build shadow
19:21 < the_rainbow> I just edit the shadow file in the build directory?
19:21 < the_rainbow> and after that I start building the target?
19:23 < rxr> you could copy a fixed file into target/.../rootfs/etc/shadow and Build-Target would iclude it into the squashfs
19:23 < rxr> a etc/shadow with the fixed password hash
19:23 < rxr> but while we are at it we could implement a random one at boot and print it on the screen to solve this once-and-for-all
19:24 < the_rainbow> yes that would be cool to generate a random password, and print it out on the shell
19:25 -!- the_rainbow [i=ee@41.209.167.135] has left #t2 []
19:26 -!- the_rainbow [i=ee@41.209.167.135] has joined #t2
19:29 < the_rainbow> Can't I add the file to build/.../etc/ ?
19:29 < the_rainbow> to be included in the final build
19:29 < rxr> you could change the file there
19:30 < rxr> but that would be a shortcut quick hack, and the next fresh build would not have it anymore
19:30 < the_rainbow> yes, because there's no target/generic/rootfs/etc
19:31 < rxr> you could create it on your side
19:31 < rxr> still more premanent then ad-hoc tweaking the build sandbox root
19:32 < the_rainbow> ah cool, so, I just create entires there, and they get included in the final build, great
19:33 < rxr> yes
19:33 < rxr> and normally, if you do long term stuff with t2, you create your own target dir with all the definitions you want
19:34 < rxr> see target/archivista for a comprehensive example
19:34 < the_rainbow> okay, great
19:34 < the_rainbow> thanks for the great advices, there's also something I noticed with the livecd in the init process: "starting kernel log daemon" hangs
19:34 < the_rainbow> until I kill it with Ctrl-c
19:36 < rxr> hm
19:36 < rxr> maybe something wrong with the writable in-memory aufs overlay
19:37 < the_rainbow> yes maybe
19:37 < the_rainbow> when I googled it, I found it might be caused by the presence of ldap line in the file /etc/nsswitch.conf
19:48 < rxr> btw. you should be able to get a pre-init shell by passing init=/bin/bash to the livecd kernel bootup in grub
19:52 < the_rainbow> cool, thanks
21:47 -!- RayMan72 [n=radjis@e227069.upc-e.chello.nl] has joined #t2
22:29 -!- bdheeman [n=bdheeman@122.173.27.235] has joined #t2
23:43 -!- bdheeman [n=bdheeman@122.173.27.235] has left #t2 []
--- Log closed Fri Apr 03 00:00:13 2009
00:57 < the_rainbow0> Actually, I found out that: ./script/Build-Pkg -cfg myconfig package, installs the package into the actual system, and doesn't update the filesystem I made previously using Build-Target, I guess I should indicate the root directory to Build-Pkg like this: ./script/Build-Pkg -cfg myconfig -root ./buid/myconfig package, I'm trying this out
01:54 -!- koan [n=koan@unaffiliated/koan] has quit ["Lost terminal"]
01:54 -!- koan [n=koan@unaffiliated/koan] has joined #t2
05:16 -!- the_rainbow0 [i=ee@41.209.167.146] has quit [Read error: 110 (Connection timed out)]
05:56 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: dsoul
05:57 -!- Netsplit over, joins: dsoul
07:56 -!- synchris [n=synchris@athedsl-192134.home.otenet.gr] has joined #t2
07:56 -!- T-One [n=T-One@62.218.122.66] has joined #t2
07:57 -!- synchris_odium [n=synchris@pdpc/supporter/professional/synchris] has quit [Read error: 60 (Operation timed out)]
09:12 < rxr> re
09:14 < rxr> http://t2-project.org/packages/samba.html
09:38 < CIA-41> rene * r32794 /trunk/package/database/sqlite/sqlite.desc: * updated sqlite (3.6.10 -> 3.6.12)
09:38 < CIA-41> rene * r32795 /trunk/package/network/samba/samba.desc: * updated samba (3.3.2 -> 3.3.3)
09:39 < CIA-41> rene * r32796 /trunk/package/graphic/gocr/gocr.desc: * updated gocr (0.46 -> 0.47)
09:39 < CIA-41> rene * r32797 /trunk/package/xorg/xrandr/xrandr.desc: * updated xrandr (1.2.3 -> 1.3.0)
09:45 < CIA-41> rene * r32798 /trunk/package/archiver/zziplib/ (install.patch zziplib.desc): * updated zziplib (0.13.49 -> 0.13.50)
10:21 -!- RayMan72 [n=radjis@ip245-128-173-82.adsl2.static.versatel.nl] has joined #t2
11:25 -!- the_rainbow0 [n=ee@41.209.167.135] has joined #t2
11:28 < the_rainbow0> Hello guys
11:30 < the_rainbow0> I finally buit my T2 target, and got the ISO for the livecd, I used kernel 2.6.29, but after it boots, it can't mount the live.squash image, because of the version mismatch, I guess it needs squshfs 4.0?
11:32 < rxr> yeah - inded
11:32 < rxr> indeed
11:33 < rxr> you need the not yet released version 4 tools
11:33 < rxr> and also either unionfs or aufs if you want a working livecd
11:33 < rxr> you disabled unionfs yesterday due to 2.6.29, I hope aufs still built?
11:55 < the_rainbow0> thanks for your answer, aufs still builds yes
11:55 < the_rainbow0> I disabled squshfs and unionfs
11:56 < the_rainbow0> tell me rxr, can't I download the dev version of squashfs 4 and unionfs and build them with T2? You know I'm getting tired of waiting for the squashfs 4
11:56 < the_rainbow0> final version
11:57 -!- EnTeQuAk [n=EnTeQuAk@static.142.84.46.78.clients.your-server.de] has left #t2 []
12:07 < rxr> the_rainbow0: maybe some cvs snapshot
12:13 < the_rainbow0> cool, then I think I'm gonna try it out
12:14 < the_rainbow0> buit I don't understand how some linux distro could include the kernel 2.6.29, within their livecds like Fedora 11
12:17 < rxr> most linux distributions are highly modified
12:17 < rxr> especially redhat ships unreleased stuff for quarters until some maintainer of theirs get to push their changes upstreamm
12:19 < the_rainbow0> ahaaa, okay okay I see
12:23 -!- the_rainbow0 is now known as kernel
12:23 -!- kernel is now known as the_rainbow
12:23 < rxr> I commit the squashfs-tools cvs update in some seconds
12:24 < the_rainbow> ah that's greaaat, I'm waiting for it
12:24 < rxr> done
12:24 < rxr> CIA is lagging
12:24 < CIA-41> rene * r32799 /trunk/package/filesystem/squashfs-tools/ (11 files):
12:24 < CIA-41> * updated squashfs-tools (3.4 -> 4.0-cvs), removed previous
12:24 < CIA-41> kernel patch glue, it's upstream since 2.6.29
12:25 < rxr> only lightly tested with: mksquashfs scripts sq ; mount -o loop sq /mnt/; l /mnt ; # ...
12:25 < the_rainbow> coool, I'm gonna try it out
12:25 < rxr> have fun - cu then - have to concentrate on some work
12:25 < the_rainbow> will I have a working livecd without unionfs?
12:25 < the_rainbow> Ah thanks man, and good luck in your work
12:29 < rxr> yes, the t2 scripts try aufs and as fallback unionfs (I think)
12:29 < rxr> (if aufs works with 2.6.29, not tested yet on my side)
12:29 < rxr> you know - they break so much "APIs" with each kernel release ...
12:29 < the_rainbow> yep I know, anyway, I'm gonna try that out and let you know
12:30 < rxr> have fun
12:30 * rxr lunch and then a pile of work - cu
12:49 -!- the_rainbow [n=ee@41.209.167.135] has quit [Read error: 104 (Connection reset by peer)]
12:49 -!- the_rainbow [i=ee@41.209.167.135] has joined #t2
13:16 < rxr> http://edsjunk.net/tutorials.html
13:57 -!- mpp [n=mpp@2001:470:9986:340:0:0:0:3] has joined #t2
13:57 < mpp> moinsen
14:55 < CIA-41> aldas * r32800 /trunk/package/base/coreutils/ (coreutils.conf coreutils.desc): * updated coreutils (7.1 -> 7.2)
15:08 < rxr> moinsen moinsen mpp
15:09 < mpp> hey rxr
15:35 -!- the_rainbow0 [i=ee@41.209.167.135] has joined #t2
15:35 -!- the_rainbow [i=ee@41.209.167.135] has quit [Read error: 104 (Connection reset by peer)]
15:37 < the_rainbow0> failed to loop mount live.squash
15:38 < the_rainbow0> it seems that the latest squashfs-tools snapshot isn't mature enough to run a livecd
15:42 -!- mqueiros [n=mqueiros@217.70.75.198] has quit ["Ex-Chat"]
15:45 -!- mqueiros [n=mqueiros@217.70.75.198] has joined #t2
15:45 < rxr> maybe the wrong squashfs tools where used and you got a version 3 file
15:45 < rxr> you can check with the file utility
15:46 < rxr> file ...path-to-the-mounted-iso/live.squash
16:00 < the_rainbow0> okay, I'm gonna check it out
16:01 < the_rainbow0> version 3.1
16:01 -!- whorrel [n=whorrel@122.173.27.235] has joined #t2
16:02 < the_rainbow0> but, why isn't it using the squshfs-tools 4, that I downloaded from trun a while ago?
16:03 < rxr> maybe it accidently used it from the build system
16:03 < the_rainbow0> just a note: I'm only updating the kernel in my build, to keep all the other packages built, I'm doing so like that: ./script/Build-Target -cfg myconfig -job 5-linux26
16:03 < rxr> can you install (e.g. Emerge-Pkg) the squashfs tools into the host system for a test ?
16:03 < rxr> you do not need to mess with the linux26 package
16:03 < rxr> only squashfs-tools
16:03 < the_rainbow0> so I don't need to rebuild the whole kernel?
16:03 < rxr> ok - before you emerge, rm buid/.../var/adm/logs/*-squashfs-tools*
16:04 < rxr> and re Build-Target
16:04 < the_rainbow0> okay
16:04 < rxr> no, it#s in the kernel now
16:04 < rxr> you only need the matching user-space tools to create the FS
16:04 < rxr> version 4.0 squashfs
16:04 < the_rainbow0> okay, I'm gonna try to emerge it to test it locally
16:05 < rxr> please also rebuild the squashfs-tools in your targe
16:05 < rxr> t
16:05 < the_rainbow0> like that: ./script/Build-Target -cfg myconfig -job 5-squashfs-tools
16:05 < rxr> if there are no bugs t2 build scripts should use the tools build along the new toolchain in your new sandbox'ed system
16:05 < rxr> yeah - just that the toolchain is build in stage 0
16:05 < rxr> so also at least -job 0-squashfs-tools
16:05 < rxr> but best just remove the log files as advised above
16:06 < the_rainbow0> okay I'm going to do that
16:06 < rxr> and re-run Build-Target -cfg ... without a -job so t2 will build the removed log file packages
16:06 < the_rainbow0> so I put the stage 0 only?
16:06 < the_rainbow0> okay okay
16:06 < rxr> in theory that should be enough - but cleaned is to wipe the package log and rebuild all of the package
16:06 < the_rainbow0> you mean the squashfs-tools package?
16:06 < the_rainbow0> only this one need to be built
16:07 < rxr> yes, it's bulit at least twice
16:07 < rxr> in stage 0 for the toolchain (also used to build the CD ISO)
16:07 < rxr> and in stage 5 for the new, sandboxed system
16:07 < rxr> in theory the stage0 rebuild should be enough to get a matching ISO for your new kernel
16:07 < the_rainbow0> so after I remove the logs, I'm just gonna build the target without a job switch
16:07 < rxr> but to have the new tools in the installed system the stage 5 rebuild is necessary, too
16:07 < rxr> yes
16:07 < the_rainbow0> okay I'm gonna try that out
16:08 < rxr> the Build-Target uses the package logs to know what needs to be build and what's already there
16:08 < the_rainbow0> ah okay great
16:10 < the_rainbow0> by the way rxr, by reading the t2-handbook I found some mistakes, I should send them to t2 in order to correct them
16:10 < the_rainbow0> also, it's good to add some new sections in the book
16:10 < the_rainbow0> for advanced use
16:10 < rxr> yes, please send
16:10 < rxr> notw that some typos where fixed since the last publishing
16:10 < rxr> in the handbooks SVN tree
16:10 < the_rainbow0> okay
16:10 < rxr> http://svn.exactcode.de/t2-handbook/trunk/src/en/book/
16:11 < rxr> be sure to base on that for sending in fixes
16:11 < the_rainbow0> sure
16:11 < rxr> (yes, we'll re-generate the web/html view and pdf's soon)
16:12 < the_rainbow0> man, when I build the target there's a build error for squashfs-tools, it says that: No file to patch
16:12 < the_rainbow0> 2 out of 2 hunks ignored
16:12 < rxr> hm
16:13 < the_rainbow0> wait, I'm gonna send you the log
16:13 < rxr> maybe your cvs snapshot is differing from mine
16:13 < rxr> which patch?
16:13 < rxr> I guess the compression level option patch ?
16:14 < the_rainbow0> I'm sending you the log
16:16 < the_rainbow0> Removing old error logs ...
16:16 < the_rainbow0> -> Removing var/adm/logs/0-squashfs-tools.err ...
16:16 < the_rainbow0> == 16:13:17 =[0]=> Building filesystem/squashfs-tools [4.0-cvs 8.0-trunk].
16:16 < the_rainbow0> -> Reading package config: package/filesystem/squashfs-tools/squashfs-tools.conf
16:16 < the_rainbow0> -> Preparing build in src.squashfs-tools.beeserver.20090402.161314.7412.localhost
16:16 < the_rainbow0> -> Building. Writing output to $root/var/adm/logs/0-squashfs-tools.out
16:16 < the_rainbow0> Command Wrapper Debug: running 'gcc --version' ..
16:16 < the_rainbow0> gcc is /root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/wrapper/gcc
16:16 < the_rainbow0> Old Command: gcc --version
16:16 < the_rainbow0> ENVPREFIX = 'CMD'
16:16 < the_rainbow0> OTHERS = ''
16:16 < the_rainbow0> INSERT = '--help?:-pipe -Os'
16:16 < the_rainbow0> REMOVE = '-O -O[0-9s] -mtune* -march* -mcpu* -g*'
16:16 < the_rainbow0> APPEND = '-s'
16:16 < the_rainbow0> FILTER = ''
16:16 < the_rainbow0> conditonal arg: --help?:-pipe.
16:16 < the_rainbow0> conditonal: --help, lhs: , rhs: -pipe.
16:16 < the_rainbow0> Insert: -pipe
16:16 < the_rainbow0> Insert: -Os
16:16 < the_rainbow0> Copy: --version
16:16 < the_rainbow0> Append: -s
16:16 < the_rainbow0> Old PATH: /root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/wrapper:/root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/bin:/root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/crosscc:/opt/samba/sbin:/opt/qt3/sbin:/opt/openldap/sbin:/opt/kde3/sbin:/opt/gnome2/sbin:/opt/samba/bin:/opt/qt3/bin:/opt/prelude/bin:/opt/openldap/bin:/opt/kde3/bin:/opt/gnome2/bin:/usr/games:/
16:16 < the_rainbow0> usr/X11/sbin:/usr/X11/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/opt/kde3/bin
16:16 < the_rainbow0> New PATH: /root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/bin:/root/t2-trunk/build/beeserver-8.0-trunk-generic-x86-i486/TOOLCHAIN/tools.cross/crosscc:/opt/samba/sbin:/opt/qt3/sbin:/opt/openldap/sbin:/opt/kde3/sbin:/opt/gnome2/sbin:/opt/samba/bin:/opt/qt3/bin:/opt/prelude/bin:/opt/openldap/bin:/opt/kde3/bin:/opt/gnome2/bin:/usr/games:/usr/X11/sbin:/usr/X11/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/
16:17 < the_rainbow0> opt/kde3/bin
16:17 < the_rainbow0> New Command: gcc -pipe -Os --version -s
16:17 < the_rainbow0> gcc (GCC) 4.2.2
16:17 < the_rainbow0> Copyright (C) 2007 Free Software Foundation, Inc.
16:17 < the_rainbow0> This is free software; see the source for copying conditions. There is NO
16:17 < the_rainbow0> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
16:17 < the_rainbow0> [ writing debug log to /root/t2-trunk/src.squashfs-tools.beeserver.20090402.161314.7412.localhost/cmd_wrapper.log ]
16:17 < the_rainbow0> `/root/t2-trunk/src.squashfs-tools.beeserver.20090402.161314.7412.localhost/archdir/squashfs-tools-2009-04-02.tar.bz2' -> `/root/t2-trunk/download/mirror/s/squashfs-tools-2009-04-02.tar.bz2'
16:17 -!- the_rainbow0 [i=ee@41.209.167.135] has quit [Excess Flood]
16:17 -!- the_rainbow0 [i=ee@41.209.167.135] has joined #t2
16:17 < rxr> excess flood, sorry
16:18 < rxr> the full log is in build/.../var/adm/logs/...
16:18 < the_rainbow0> oops, I'm sorry
16:18 < rxr> can you look yourself which patch failed to apply ?
16:18 < the_rainbow0> this one: /squashfs3.4-patch.orig 2008-08-27 07:52:07.000000000 +0200
16:18 < rxr> yeah - it's not me who kicked you, that was the feenode.net service ...
16:18 < rxr> hm?
16:18 < the_rainbow0> and: squashfs3.4/kernel-patches/linux-2.6.23/squashfs3.4-patch 2008-10-20 20:40:53.000000000 +0200
16:18 < rxr> what does svn st package/filesystem/squashfs-tools say ?
16:19 < the_rainbow0> wait
16:19 < rxr> have you svn up'ed package/filesystem/squashfs-tools ?
16:20 < the_rainbow0> M package/filesystem/squashfs-tools/compressionLevel.patch
16:20 < the_rainbow0> M package/filesystem/squashfs-tools/squashfs-tools.desc
16:20 < rxr> why did you mod them ?
16:20 < rxr> svn revert those and svn up ...
16:21 < the_rainbow0> I'm not really an exprerienced svn user, I just use svn to download source files and such stuff
16:21 -!- whorrel [n=whorrel@122.173.27.235] has quit ["Lost terminal"]
16:21 < rxr> svn revert package/filesystem/squashfs-tools/*
16:22 < the_rainbow0> ok
16:22 < rxr> svn up package/filesystem/squashfs-tools/
16:22 < the_rainbow0> done
16:23 < the_rainbow0> now, I'm trying to build the target again
16:23 < rxr> try to build the package again
16:23 < rxr> I cross my fingers and press my thumbs for you and go back to actual coding work, ...
16:23 < the_rainbow0> lol thanks a billion man :D, it seems that it's building fine !
16:24 < the_rainbow0> and yes it finished building it
16:24 < the_rainbow0> Good luck dude, I'll let you know about the final results
16:43 -!- the_rainbow0 is now known as the_rainbow
17:10 -!- T-One [n=T-One@62.218.122.66] has quit ["Verlassend"]
17:10 -!- RayMan72 [n=radjis@ip245-128-173-82.adsl2.static.versatel.nl] has quit ["i've to Leave"]
17:31 < the_rainbow> what's the default login for the T2 livecd?
17:35 < the_rainbow> Confirmation: The squashfs-tools snapshot works great, and the livecd boots fine
17:35 < the_rainbow> Just can't figure out the default root password for the livecd
17:42 < rxr> no default, we hat this on the mailing list some days ago
17:44 < rxr> http://article.gmane.org/gmane.comp.t2.devel/4055
17:45 < rxr> http://article.gmane.org/gmane.comp.t2.devel/4056
18:38 -!- mpp [n=mpp@2001:470:9986:340:0:0:0:3] has quit ["Verlassend"]
19:03 < the_rainbow> then, there's something I have to make in the config
19:03 < the_rainbow> because even with UN: root and PW: root it doesn't work
19:10 < rxr> as the links to the mailing list indicate you have to changes some lines of code to make the trunk livecd login work
19:10 < rxr> what livecd did you build? minimal? or + X.org graphic ?
19:11 < the_rainbow> minimal
19:15 < rxr> without any X ?
19:15 < the_rainbow> yep
19:15 < the_rainbow> I just need the minimum, no X needed
19:15 < rxr> ok - in the stable tree we setup kde to auto-login the guest
19:15 < rxr> user
19:16 < the_rainbow> what about bash?
19:16 < rxr> how would you want such a text live system to allow login ?
19:16 < rxr> we could generetate a random password on startup and set it for the guest user and print it to the screen
19:16 < the_rainbow> default login, like for example the Fedora LiveCD
19:16 < rxr> or would you like an fixed passwd ?
19:16 < rxr> I bet fedora has a graphic login with auto-login ?
19:16 < the_rainbow> yes, just a fixed one
19:17 < rxr> fixed passworD ?
19:17 < the_rainbow> yes it has a graphic login and Bash login as well
19:17 < rxr> a little insecure if 50% of the dau users do not change it
19:17 < the_rainbow> yes fixed password
19:17 < rxr> what password does fedora use ?
19:17 < the_rainbow> no password, just the user name, root
19:17 < the_rainbow> the root has no password on Fedora
19:18 < the_rainbow> okay, if I know how to do it with a random password, I would also know how to make it fixzd
19:18 < the_rainbow> fixzd*
19:18 < the_rainbow> fixed*
19:18 < rxr> not really - for a fixed password you could simply inject a predefined etc/shadow file
19:18 < rxr> for a random password some lines of code are required
19:18 < the_rainbow> okay okay I see
19:19 < the_rainbow> how can t2 run post scripts after the build in a chrooted environment?
19:20 < rxr> you can
19:20 < rxr> but for a fixed password you do not need that
19:20 < rxr> and for random passwords I would suggest generating it at boot time
19:20 < the_rainbow> okay, I'm gonna start with a fixed password
19:20 < the_rainbow> where should I start?
19:21 < the_rainbow> to inject my predefined line into the build shadow
19:21 < the_rainbow> I just edit the shadow file in the build directory?
19:21 < the_rainbow> and after that I start building the target?
19:23 < rxr> you could copy a fixed file into target/.../rootfs/etc/shadow and Build-Target would iclude it into the squashfs
19:23 < rxr> a etc/shadow with the fixed password hash
19:23 < rxr> but while we are at it we could implement a random one at boot and print it on the screen to solve this once-and-for-all
19:24 < the_rainbow> yes that would be cool to generate a random password, and print it out on the shell
19:25 -!- the_rainbow [i=ee@41.209.167.135] has left #t2 []
19:26 -!- the_rainbow [i=ee@41.209.167.135] has joined #t2
19:29 < the_rainbow> Can't I add the file to build/.../etc/ ?
19:29 < the_rainbow> to be included in the final build
19:29 < rxr> you could change the file there
19:30 < rxr> but that would be a shortcut quick hack, and the next fresh build would not have it anymore
19:30 < the_rainbow> yes, because there's no target/generic/rootfs/etc
19:31 < rxr> you could create it on your side
19:31 < rxr> still more premanent then ad-hoc tweaking the build sandbox root
19:32 < the_rainbow> ah cool, so, I just create entires there, and they get included in the final build, great
19:33 < rxr> yes
19:33 < rxr> and normally, if you do long term stuff with t2, you create your own target dir with all the definitions you want
19:34 < rxr> see target/archivista for a comprehensive example
19:34 < the_rainbow> okay, great
19:34 < the_rainbow> thanks for the great advices, there's also something I noticed with the livecd in the init process: "starting kernel log daemon" hangs
19:34 < the_rainbow> until I kill it with Ctrl-c
19:36 < rxr> hm
19:36 < rxr> maybe something wrong with the writable in-memory aufs overlay
19:37 < the_rainbow> yes maybe
19:37 < the_rainbow> when I googled it, I found it might be caused by the presence of ldap line in the file /etc/nsswitch.conf
19:48 < rxr> btw. you should be able to get a pre-init shell by passing init=/bin/bash to the livecd kernel bootup in grub
19:52 < the_rainbow> cool, thanks
21:47 -!- RayMan72 [n=radjis@e227069.upc-e.chello.nl] has joined #t2
22:29 -!- bdheeman [n=bdheeman@122.173.27.235] has joined #t2
23:43 -!- bdheeman [n=bdheeman@122.173.27.235] has left #t2 []
--- Log closed Fri Apr 03 00:00:13 2009