T2 IRC Log: 2007-06-21

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.

« prev | next »

--- Log opened Thu Jun 21 00:00:17 2007
--- Day changed Thu Jun 21 2007
00:00 < kic-> and 'generic'
00:00 < kic-> ?
00:02 < sepp_> yes
00:04 < kic-> hrm
00:04 < kic-> i don't know why mine failed then
00:04 < kic-> had other failures to
00:04 < sepp_> neither do i
00:04 < kic-> on modules
00:04 < kic-> i've ended up reverting back to i486
00:04 < kic-> and building this way
00:05 < sepp_> which module?
00:05 < kic-> infiniband
00:05 < sepp_> hm
00:07 < kic-> and even the i486 build of minimal xorg
00:07 < kic-> failed when it got to package 'parted'
00:07 < kic-> so i had to remove that package
00:07 < kic-> and rebuilding now
00:07 < kic-> maybe something is screwy in my host system
00:07 < kic-> host system is a minimal t2 build without xorg or desktop
00:07 < kic-> and an svn co of t2-trunk
00:07 < sepp_> that should be fine
00:08 < kic-> for the scripts
00:11 < sepp_> my box is not that fast ...
00:13 < Stelz> re
00:13 < sepp_> hi Stelz
00:14 < Stelz> hi sepp_ ;)
00:41 < kic-> what stage does it build the binary packages in?
00:41 < kic-> or is that in multiple stages
00:41 < sepp_> most in stage 5
00:42 < kic-> and what is the last stage?
00:43 < sepp_> 6 & 7 or if you are paranoid you can rebuild everything in stage 9
00:43 < sepp_> 6 & 7 are < 10 packages - iirc
00:45 < kic-> ah
00:46 < sepp_> there is a Config option to rebuild everything at stage 9 - that is nice if you want to release isos and you care about circular deps ...
00:46 < kic-> but i don't care
00:46 < kic-> hah
00:46 < kic-> these isos is just for a minimal t2 desktop environment
00:46 < kic-> for our developers to use
00:47 < kic-> as their workstations
00:47 < sepp_> neither do i, i never used that
00:47 -!- roman [n=roman@85-127-134-3.dynamic.adsl-line.inode.at] has quit [Read error: 104 (Connection reset by peer)]
01:20 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
01:30 -!- mtr_ [n=Michael@pD9E10C97.dip0.t-ipconnect.de] has joined #t2
01:44 -!- mtr [n=Michael@pD9E13447.dip0.t-ipconnect.de] has quit [Read error: 110 (Connection timed out)]
01:44 -!- mtr_ is now known as mtr
01:47 < kic-> wow
01:47 < kic-> you can pass -cfg system
01:48 < kic-> to Emerge-Pkg
01:48 < kic-> ?!
02:25 -!- sparc-kly_ [n=mubex@adsl-64-237-241-185.prtc.net] has joined #t2
02:42 -!- sparc-kly__ [n=mubex@adsl-64-237-248-104.prtc.net] has quit [Connection timed out]
02:52 -!- sparc-kly_ [n=mubex@adsl-64-237-241-185.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
03:38 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
04:51 -!- sparc-kly [n=mubex@adsl-64-237-241-185.prtc.net] has joined #t2
08:54 < mtr> moin
08:54 < CIA-9> susan * r24676 /trunk/package/database/sqlite/sqlite.desc: * updated sqlite (3.3.17 -> 3.4.0)
08:54 < CIA-9> susan * r24677 /trunk/package/gnome2/libglade/libglade.desc: * updated libglade (2.6.0 -> 2.6.1)
08:55 < CIA-9> susan * r24678 /trunk/package/gnome2/libglademm/libglademm.desc: * updated libglademm (2.6.3 -> 2.6.4)
08:55 < CIA-9> susan * r24679 /trunk/package/gnome2/libgsf/libgsf.desc: * updated libgsf (1.14.3 -> 1.14.4)
08:55 < CIA-9> susan * r24680 /trunk/package/audio/ladspa-caps/ladspa-caps.desc: * updated ladspa-caps (0.3.0 -> 0.4.2)
08:56 < CIA-9> susan * r24681 /trunk/package/base/lvm2/lvm2.desc: * updated lvm2 (2.02.25 -> 2.02.26)
08:57 < rxr> re
08:57 < rxr> moin mtr
08:59 < mtr> hi rxr
09:09 < CIA-9> susan * r24682 /trunk/package/games/scummvm/scummvm.desc: * updated scummvm (0.9.1 -> 0.10.0)
09:20 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
09:21 < tri> moin
09:22 < rxr> moin moin tri
09:24 < CIA-9> rene * r24683 /trunk/package/network/ntp/ntp.desc: * updated ntp (4.2.4 -> 4.2.4p2)
09:28 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
09:28 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
09:34 < CIA-9> rene * r24684 /trunk/package/games/scummvm/ (scummvm.conf scummvm.desc): * rewrote scummvm.conf to cross build
09:43 < tr1> if i now a package is only for x86 what to type in the [F] Flag ?
09:43 < tr1> s/now/know
09:43 < rxr> [R] + x86
09:43 < rxr> [R] as short for [ARCH]
09:44 < tr1> an [F] ?
09:44 < rxr> no F
09:44 < tr1> thx ;-)
09:45 < rxr> http://www.t2-project.org/handbook/html/t2.package.desc.html
09:45 < CIA-9> rene * r24685 /trunk/package/develop/libatomic_ops/libatomic_ops.desc: * updated libatomic_ops (1.1 -> 1.2)
09:45 < tr1> the handbook bot ;-)
09:45 < CIA-9> rene * r24686 /trunk/package/develop/php/php.desc: * updated php (5.2.2 -> 5.2.3)
09:46 -!- Stelz [n=stealth@unaffiliated/stelz] has quit ["There is intelligent life out there..."]
09:46 < tr1> nice is the openmoko channel if you ask your first question you always get a link to the faq :-)
09:46 < tr1> from a "real" bot
09:47 < rxr> lol
09:50 -!- Stelz [n=stealth@unaffiliated/stelz] has joined #t2
09:50 < tr1> hi stelz
09:50 < tr1> hm which dir fits for the kexec-tools ?
09:51 < Stelz> moin moin tr1
09:51 < Stelz> hello sepp_
09:51 < Stelz> nice to c u all :)
09:52 < CIA-9> rene * r24687 /trunk/package/develop/ (ccache/ccache.desc distcc/distcc.desc):
09:52 < CIA-9> * enabled ccache and distcc in the toolchain stage, so the build does
09:52 < CIA-9> not depend on in-host-system ones
09:54 < CIA-9> rene * r24688 /trunk/package/ (5 files in 5 dirs):
09:54 < CIA-9> * marked libpeak, vbetool, suspend, libieee1284, suspend, and
09:54 < CIA-9> id-utils to cross build
09:55 < tr1> woh! the CkPkg
09:55 < tr1> wups
09:55 < tr1> ch Create-CkSumPatch is a nice one :-)
09:55 < rxr> Check-PkgFormat is a nice one, too
09:57 < tr1> how to create a good: [P] [PRI] [PRIORITY] Flag ?
09:57 < tr1> should it be X or O ?
09:57 < tr1> and how to know which stage is needed
09:57 < tr1> and what Build Priority Nummber ?
09:59 < rxr> if the packages does not conflict with something else use X
10:00 < rxr> normal packages are just to be built for stage 5 and 9 (by default)
10:00 < tr1> if i use O it doesn't build ?
10:00 < rxr> if in doubt, copy the Priority from a simillar package
10:00 < rxr> yes, O in the [P] tag means disabled by default
10:00 < rxr> of course a target can switch it on
10:01 < tr1> ah ok
10:01 < rxr> we have quite some stuff disabled by default, alternative init systems, alternative print spoolers, alternative login utilities, alternative C libraries, ...
10:01 < tr1> simillar to kexec-tools ?
10:01 < rxr> normal packages should be on by defult
10:01 < rxr> kexec-tools should be on by default
10:02 < tr1> mkinitcpio is only an other mkinit script and doesn't conflict so it is ok if there is a X ?
10:02 < tr1> i mean which package is simillar to kexec-tools ?
10:02 < rxr> numatrl
10:02 < tr1> and where does it fit best "base" ?
10:02 < rxr> yeah - base for now
10:02 < rxr> though we certainly want to re-sort base
10:03 < rxr> when we suppot more kernels soon kexec-tools might also be best palce in a linux repository
10:03 < rxr> but for now I think base suits best
10:05 < tr1> how to enable kexec in the kernel build ? is it there by default or do i have to use my own config or "what ever" ?
10:06 < tr1> kexec-tools http://imi.st/kexec-tools.desc ;-)
10:07 < rxr> should I commit ?
10:07 < CIA-9> rene * r24689 /trunk/package/x11/glew/glew.conf: * fixed $root$prefix in glew.conf to be $root/$prefix, ...
10:09 < tr1> here is a nice articel about kexec: http://www.ibm.com/developerworks/linux/library/l-kexec.html?ca=dgr-lnxw07RebootFast
10:09 < rxr> !> kexec-tools: Unknown package category: base/tools
10:10 < tr1> ah soory !
10:10 < rxr> just base/tool
10:10 < rxr> fixed that
10:10 < rxr> btw. where did you got kexec only works on x86?
10:10 < tr1> yes
10:10 < rxr> should at least also cover powerpc
10:10 < rxr> probably x86-64 and powerp64, likewise
10:11 < tr1> hm not sure
10:11 < rxr> at least works on the PS3 here in the office :-)
10:11 < tr1> Kexec is currently available on the x86 32-bit platform only Having it on other architecture platforms such as PPC 64 and AMD 64 would be helpful.
10:11 < rxr> PS3 is PowerPC64
10:11 < tr1> from the ibm articel but it is an old one
10:11 < tr1> i don't have anythink else then x86
10:11 < rxr> heh, ok
10:11 < tr1> right now to test it
10:12 < rxr> I just leave the mark and test it on our exotic boxes, later wen we need it :-)
10:13 < CIA-9> rene * r24690 /trunk/package/base/kexec-tools/ (. kexec-tools.cache kexec-tools.desc):
10:13 < CIA-9> Tobias Richter :
10:13 < CIA-9> * added kexec-tools (20070330) - kexec-tools
10:13 < rxr> btw. t2 is now in the top-100 on distrowatch for the last quater
10:13 < rxr> http://distrowatch.com/stats.php?section=popularity
10:14 < tr1> :-)
10:15 < tr1> 3 PCLinuxOS ??? who want to use this...
10:15 < rxr> not me
10:16 < rxr> but I have the feeling this stat is very biased anyway
10:16 < rxr> probably some homepages have prominent links to distrowatch somewhere in their site-menu or so
10:16 < tr1> 59 Linspire :-) - yeah that is my favorite
10:18 < CIA-9> rene * r24691 /trunk/package/base/kexec-tools/kexec-tools.desc: * marked kexec-tools to just cross build as well as to build on powerpc64
10:22 < CIA-9> rene * r24692 /trunk/package/base/kexec-tools/kexec-tools.desc: * marked kexec-tools for x86-64, likewise
10:22 < tr1> i was searching right know which architecture is supported by kexec-tools but i'm not able to find anything - but i see you have found something ?
10:22 < tr1> or you have done test builds ?
10:22 -!- Stelz [n=stealth@unaffiliated/stelz] has quit ["There is intelligent life out there..."]
10:22 -!- Stelz [n=stealth@unaffiliated/stelz] has joined #t2
10:29 < rxr> tr1: I did test-builds on those two boxes
10:30 < rxr> and I knew it worked on the PS3
10:32 < tr1> ah ok - now i have to reorganize my mailbox....
10:32 < tr1> what about xen support in t2 ?
10:34 < CIA-9> rene * r24693 /trunk/package/base/ (7 files in 5 dirs):
10:34 < CIA-9> * fixed numactl and libdaemon to cross build
10:34 < CIA-9> * marked attr, acl, and termutils to just cross build, likewise
10:37 < CIA-9> rene * r24694 /trunk/package/java/jetty/jetty.desc: * updated jetty (6.1.3 -> 6.1.4)
10:39 < Stelz> moin moin rxr
10:42 < rxr> tr1: xen support is a little-bit there
10:42 < rxr> but due to xen complexity and kvm it was somewhat left lingering around the last time
10:42 < rxr> further xen and openvz support wanted
10:42 < rxr> moin moin Stelz
10:57 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
10:57 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
10:58 < tri> re
10:58 < tri> what doesis "xen support is a little-bit there"
10:58 < tri> can t2 run as a xen host
10:58 < rxr> run anyway
10:59 < rxr> I would see kernel patches more integrated ans also dom0 support be more automtic
10:59 < rxr> I'm out of office for some extended hour - cu then
10:59 < tri> cu rxr ;-)
11:00 < rxr> btw you can even built t2 optimized for Xen
11:00 < rxr> it's in the x86 options on the top
11:04 < tri> yes saw this already
11:04 < tri> xen-tools package isn't there ?
11:05 < tri> the t2 kernel is able to be a Dom0 (the XEN-Host) ?
11:05 < tri> ah xen-tools is available
11:07 -!- write_erase [n=olivier@pleiades4.data-mobiles.com] has joined #t2
11:07 < write_erase> hi
11:07 < tri> hi write_erase
11:07 < tri> how are you ?
11:07 < write_erase> fine, a bit busy ...
11:08 < write_erase> not enough time to have fun with t2
11:09 < tri> hm xen-tools are a bit outdated...
11:09 < tri> hm and argos and xen-shell would be nice
11:15 -!- Stelz [n=stealth@unaffiliated/stelz] has quit ["There is intelligent life out there..."]
11:20 < tri> hm xen-man is also nice
11:29 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
11:29 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
11:29 < tr1> nice: http://arstechnica.com/articles/paedia/IPv6.ars
11:30 < tr1> virt-manager looks also nice
11:31 < tr1> openqrm also
11:31 < tr1> ConVirt too
11:31 < tr1> hm and xen-tools need a update :-)
11:32 < tr1> please let this go into the irc-logs so i can remembe iit soon... :-)
11:44 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
11:45 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
11:48 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
12:13 < CIA-9> sebastian * r24695 /trunk/package/audio/libmpd/libmpd.desc: * updated libmpd (0.12.0 -> 0.13.0)
12:20 < sepp_> moin :)
12:35 -!- patoh [n=patoh@unaffiliated/pat/x-10101] has joined #t2
12:43 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
12:47 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
12:47 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
12:53 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
12:56 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
12:57 < tr1> re
12:57 < tr1> what about *.init files ?
12:58 < sepp_> .init ?
13:00 < tr1> for example in the xen-tools package
13:00 < tr1> there is a file xend.init
13:01 < tr1> this is the script for xen-tools which lives in /etc/init.d/ - so *init files are automaticliy installed by t2 (like *patch files are automatily applied)
13:02 < tr1> ?
13:03 < sepp_> if there is a make install rule
13:03 < tr1> hm
13:03 < sepp_> if not look at package/network/apache or so
13:03 < tr1> a thx a take a look
13:05 < tr1> hm at the apache and at the xen-tools are no make install rule to install the init scripts
13:05 < tr1> but they are there
13:06 < tr1> i think they are copyied automaticly
13:06 < tr1> i will have a look at the log file
13:09 < tr1> if i have changed a available package how to create a nice diff ?
13:09 < tr1> for the whole package folder ?
13:10 < tr1> wups sorry http://www.t2-project.org/handbook/html/t2-book.html#t2.package.diff
13:10 < sepp_> svn diff package/foo/bar > blah.patch
13:14 < tr1> http://imi.st/xen-tools.patch
13:15 < tr1> xen-tools wasn't build for me - with this patch it is building and it is also updated to version 3.5
13:15 < tr1> maybe somebody have look ?
13:18 < sepp_> not me :p
13:19 < tr1> np ;-)
13:21 < sepp_> but the checksum is missing
13:24 < tr1> hm
13:25 < sepp_> ./scripts/Create-CkSumPatch xen-tools | patch -p1
13:25 < sepp_> :)
13:27 < tr1> http://imi.st/xen-tools.patch updated
13:28 < tr1> and you have already looked at it ;-)
13:33 < sepp_> that is much easier to apply than the complete changed file :)
13:37 < tr1> yes, rene told me to make patch files - that was why i was asking how to make this but
13:37 < tr1> (13:06:22) sepp_: svn diff package/foo/bar > blah.patch
13:37 < tr1> helped me - thx ;-)
13:39 < sepp_> np
13:46 < sepp_> bbl
13:51 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
13:54 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
13:57 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
14:19 -!- PAVLINUX [n=pavel@ppp85-140-164-188.pppoe.mtu-net.ru] has joined #t2
14:20 < PAVLINUX> Good mornig Europe :)
14:21 -!- Stelz [n=stealth@unaffiliated/stelz] has joined #t2
14:31 < sepp_> GOOD MORNING VIETNAM
14:31 < sepp_> from some stupid film :)
14:35 -!- PAVLINUX [n=pavel@ppp85-140-164-188.pppoe.mtu-net.ru] has quit ["using sirc version 2.211+KSIRC/1.3.12"]
15:19 -!- kic_ [i=kicstah@adsl-074-238-088-201.sip.bct.bellsouth.net] has joined #t2
15:19 -!- kic- [i=kicstah@adsl-074-238-088-201.sip.bct.bellsouth.net] has quit [Read error: 104 (Connection reset by peer)]
15:20 -!- kic_ is now known as kic-
15:43 < kic-> arg
15:43 < kic-> glibc32 is failing to build for me
15:44 < kic-> on a generic x86_64, minimal xorg target
15:45 < kic-> if my hardware supports 64bit, but my OS is a 32bit environment, and i'm trying to use T2 to build a 64bit target, all on the same machine..
15:45 < kic-> is that going to be a problem if i don't specify 'cross compile'
15:45 < kic-> ?
15:49 -!- somethingbad [i=vladi@unicorn.vladi.biz] has joined #t2
15:49 < somethingbad> hello *
15:52 < sepp_> hello somethingbad
15:52 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
15:52 < tri> re
15:53 < sepp_> kic-: if you are running a 32bit linux on a x86-64 you can not execute the resulting binaries
15:53 < tri> ping rxr
15:56 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
15:56 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
15:57 < sepp_> wb tr1 :)
15:57 < tr1> hm
15:57 < tr1> damm internet connection
15:57 < tr1> or proxy or firewall or whatever...
15:59 < kic-> is there a prebuilt generic x86_64 iso on the t2 mirrors somewhere?
15:59 -!- tr1 [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has quit [Read error: 104 (Connection reset by peer)]
15:59 < somethingbad> :)
16:00 < somethingbad> i have 1 problem with building T2
16:00 < kic-> i suppose i could install one, svn the build scripts, and then build a 64bit target, eh?
16:00 < somethingbad> can I ask here for help?
16:00 < sepp_> yes
16:00 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
16:01 < somethingbad> problem is on stage 5 with building kernel (2.6.17.14), i'm using router config
16:01 < sepp_> kic-: http://osiris.tfh-berlin.de/mirror/t2/stable/6.0.1/minimal-x86-64-athlon64/t2-minimal-6.0.1-x86-64-athlon64-mmx_cd1.iso
16:01 < somethingbad> it says "/sbin/mkinitrd: line 120: file: command not found"... and searching for switch_root at /usr/embutils
16:02 < sepp_> you need to enable the embutils package
16:02 < somethingbad> i installed it manualy, but there is now switch_root
16:02 < sepp_> the router target is outdated, i think
16:02 < kic-> no
16:02 < kic-> cannot
16:02 < kic-> i don't think
16:02 < kic-> sepp, its an intel x86_64
16:02 < kic-> my machine
16:02 < kic-> not an AMD
16:02 < kic-> ?
16:03 -!- tri [n=Richter@pd95b2df8.dip0.t-ipconnect.de] has left #t2 []
16:03 < somethingbad> there is no*
16:04 < sepp_> no usr/bin/file?
16:04 < kic-> sepp, that iso install will run on a Xeon 64bit?
16:04 < somethingbad> sepp_: nothing... only pivot_root exist
16:05 < kic-> sepp's popular today :D
16:05 < sepp_> kic-: i dont know
16:05 < sepp_> i lost track following intels cpus
16:05 < kic-> ok
16:05 < kic-> fair enough
16:06 < sepp_> if it is a real x86-64 you can run it
16:06 < kic-> k
16:07 < kic-> for $1200 for the CPU alone
16:07 < kic-> it better be a real x86-64
16:07 < kic-> :D
16:08 -!- Stelz [n=stealth@unaffiliated/stelz] has quit ["There is intelligent life out there..."]
16:09 -!- Stelz [n=stealth@unaffiliated/stelz] has joined #t2
16:10 < sepp_> somethingbad: better do not use the outdated router target
16:10 < sepp_> i think it even has a comment in config that it is incomplete
16:12 < somethingbad> thank you...
16:12 < somethingbad> :)
16:14 < sepp_> you can use generic embedded
16:22 -!- Stelz is now known as Enqlave
16:24 < somethingbad> I think to use 7.0-rc src
16:28 < rxr> re
16:31 -!- Enqlave [n=stealth@unaffiliated/stelz] has quit [Read error: 104 (Connection reset by peer)]
16:31 -!- SLZ [n=stealth@80.240.210.253] has joined #t2
16:32 < kic-> re
16:32 < kic-> Rene
16:33 < kic-> have you had a successfull build of minimal xorg with x86_64
16:33 < kic-> ?
16:33 < kic-> 7.0
16:36 < rxr> oehm - probably
16:36 < rxr> you get a failure ?
16:36 < kic-> yeah, different misc ones
16:36 < kic-> yesterday with 1-perl
16:36 < kic-> and -lgdbm
16:37 < kic-> then get glibc32 build errors
16:37 < kic-> and then most recently bash exec format error
16:37 < kic-> chroot: bash exec format error
16:37 < rxr> .oO
16:37 < rxr> do you have a 64bit system on the box ?
16:38 < kic-> the hardware is, the environment was the 6.0.0 x86
16:38 < kic-> not 64 i don't think
16:38 -!- SLZ is now known as Stelz
16:38 < rxr> the exec format error you will get when you try running a 64bit binary on a 32bit kernel
16:39 < kic-> figures :D
16:39 < kic-> so now i downloaded the desktop-x86-64-athlon64
16:39 < kic-> from a mirror
16:39 < kic-> and installing that now
16:39 < kic-> i have to get something running today for the developers, they haven't been working in the past month from problems with timesys
16:40 < kic-> i'll install that iso that i just downloaded, svn a 7.0 build set
16:40 < kic-> and build a new generic, minimal xorg, x86_64, no optimizations, for them
16:40 < rxr> you should have stayed on the 6.0 install you had ... and just use the 7.0-trunk for the product development
16:40 < kic-> that should succeed right?
16:42 < kic-> couldn't
16:42 < kic-> the 6.0 install was 32bit
16:42 < kic-> all dev machines are 64 bit, and we'll be building 64bit targets all the time
16:43 < kic-> this desktop-x86-64-athlon64 is from /mirror/t2/stable/6.0.0/desktop-x86-64-athlon64
16:43 < kic-> so its also 6.0.0
16:47 < somethingbad> is there any problem with mirrors for 7.0?
16:47 < rxr> I hope not
16:47 < rxr> just hte nexus appears to be done, but that starts to be a constant of this FreeBSD box
16:48 -!- root [n=root@adsl-074-238-088-206.sip.bct.bellsouth.net] has joined #t2
16:48 < somethingbad> "INFO: download from mirror failed, trying original URL. unrecognized!"
16:48 < root> hi all
16:49 -!- root is now known as keztoo
16:49 < keztoo> .
16:50 < keztoo> rxr, are you there?
16:51 < rxr> yes
16:51 < rxr> somethingbad: that might be the case even if the mirror is fine and the update is just new
16:51 < keztoo> i downloaded trunk yesterday
16:51 < keztoo> updated this morning
16:51 < keztoo> changed nothing in config
16:52 < keztoo> and my build is breaking in stage 1 on glibc
16:52 < keztoo> i'm not quite sure what to do next
16:53 < keztoo> im going 64b to 64b so no cross compile
16:54 < keztoo> undefined reference to __sync_fetch_and_add_4
16:54 < rxr> x86-64 ?
16:54 < sepp_> enable optimization >386
16:54 < sepp_> != generic
16:55 < keztoo> yes
16:55 < keztoo> AMD64, Intel EM64T option in cft
16:56 < keztoo> the only thing i changed in cfg was the number of jobs from 2 -> 5
16:57 < keztoo> generic , no special optimization
16:57 < sepp_> select something other
16:58 < sepp_> not generic ...
16:58 < rxr> sepp_: are you sure ?
16:58 < rxr> sepp_: on x86 generic == i386
16:58 < rxr> sepp_: but on x86-64 generic means first amd64 and that should build
16:58 < sepp_> yes, and that does not work for glibc32
16:59 < rxr> oh
16:59 < rxr> that must be fixed
16:59 < sepp_> if you build multilib on x86-64 with 'No special optimisation' glibc32 gets no -march
16:59 < rxr> didn't I hardwired glibc32 to build for i686 for x86-64 ?
16:59 < rxr> oh - damn
16:59 < rxr> that is a bug
16:59 < rxr> thanks sepp_ and keztoo to point this out, did not notice so far
17:00 < rxr> I guess I always build either athon64 ore core2 optimized
17:00 < sepp_> just found it out last night
17:00 < keztoo> error log says auto detection for optimization impossible, then has a bunch of flags like disable-debug-passed-thru , etc
17:00 < sepp_> because i forgot to set optimization to anything for an minimal xorg test
17:01 < rxr> ok - I stopped my minimal64 test-build and selected gneric and go fixing ...
17:01 < sepp_> :)
17:01 < kic-> if you want ssh access to my clean system to build one rene
17:01 < kic-> lemme know
17:01 < kic-> :D
17:01 < kic-> i'll happily take the isos that you build on it
17:01 < kic-> :D
17:01 < rxr> kic-: :_)
17:02 < rxr> is it faster than the core2 MacBook I type on ?
17:02 < kic-> its a Xeon 3ghz, core2 duo
17:02 < rxr> but maybe I just continue testing on the macbook as I can carry it around later and commit on-road then
17:02 < rxr> kic-: btw. see query
17:03 < keztoo> ok set optimize to core2 and restarted the build
17:03 < keztoo> ty sepp
17:04 [Users #t2]
17:04 [@ChanServ] [ emte ] [ LMJ ] [ rxr ] [ Stelz ]
17:04 [ cbrake ] [ Fatal ] [ mtr ] [ sepp_ ] [ TobiX ]
17:04 [ CIA-9 ] [ keztoo] [ patoh ] [ somethingbad] [ valentin ]
17:04 [ Dallur ] [ kic- ] [ Ragnarin] [ sparc-kly ] [ write_erase]
17:04 -!- Irssi: #t2: Total of 20 nicks [1 ops, 0 halfops, 0 voices, 19 normal]
17:07 < kic-> rene
17:07 < sepp_> svn diff
17:07 < sepp_> lol
17:07 < kic-> if i do scripts/Config to set a config for the base/host system i'm running
17:07 < sepp_> damn mouse focus
17:07 < kic-> and add kernel modules as O modulename
17:07 < kic-> in the package list
17:07 < kic-> and then do an Emerge-Pkg -missing=only linux26
17:08 < kic-> will that build the latest 2.6.21 kernel minus the modules i listed?
17:09 < rxr> you mean like sky2 or so?
17:09 < rxr> kernel modules included in the kernel ?
17:10 < kic-> yeah
17:10 < kic-> like infiniband
17:10 < kic-> etc
17:10 < kic-> hrm
17:10 < kic-> maybe i should try the import my own .config option
17:10 < kic-> for kernel config
17:10 < rxr> no
17:10 < kic-> o
17:10 < rxr> the package selection will not influence that
17:10 < kic-> ahh, k
17:10 < rxr> you need to provide your own custon kernel config for that
17:11 < kic-> ok, good deal
17:11 < rxr> the package selection is just for the packages to build
17:11 < kic-> maybe they meant only 3rd party modules then
17:11 < kic-> last night
17:11 < kic-> when they said that
17:11 < rxr> yes, third party modules
17:11 < kic-> ahh, gotcha
17:11 < rxr> as they are add-on packages
17:12 < kic-> yeah
17:14 < CIA-9> rene * r24696 /trunk/package/x86/ (18 files in 10 dirs):
17:14 < CIA-9> * fixed spicctrl, longrun, memtest86, x86info, microcode_ctl, and
17:14 < CIA-9> cpuburn to cross build
17:14 < CIA-9> * marked acpid, sonypid, dmidecode, and yasm to cross build
17:15 < kic-> Checked out revision 24695
17:15 < kic-> yay
17:15 < kic-> base 64bit desktop install
17:16 < kic-> latest t2-trunk from svn
17:16 < kic-> time to get started
17:17 < kic-> erm
17:17 < kic-> so no generic?
17:17 < rxr> looks so
17:17 < kic-> i can do generic system though right
17:17 < rxr> will hunt that bug
17:17 < kic-> instead of archivista, or wireless router, psion, etc
17:17 < rxr> ah no
17:17 < kic-> ah
17:17 < rxr> choose genreic target
17:17 < kic-> erm
17:18 < kic-> doesn't have one
17:18 < rxr> but in the cpu optmization x86-64 and then not the generic sub-arch, but core2
17:18 < kic-> :D
17:18 < kic-> has generic embedded, and generic system
17:18 < kic-> sega dreamcast
17:18 < kic-> exact desktop
17:18 < kic-> etc
17:18 < kic-> oh
17:18 < kic-> i see
17:18 < kic-> gneric target, but optimizations can't be generic or 'default' type
17:18 < kic-> k
17:32 < keztoo> i am attempting to build a new package , where is a good place to look to see how to do one besides looking at existing .conf and .desc files?
17:32 < keztoo> i think i need a specific directory structure in my tar but i'm not sure
17:32 < keztoo> where everything goes
17:32 < CIA-9> rene * r24697 /trunk/package/x86/ (5 files in 2 dirs): * fixed nasm and cpuid to cross build
17:33 < rxr> keztoo: no, unlike on debian you need nothing in your tar-ball
17:33 < rxr> in T2 all meta info is in the package dir
17:33 < rxr> so working with the meta data and updates are very easy
17:33 < rxr> the package details:
17:33 < keztoo> i think i want it to be part of network extra
17:34 < rxr> http://www.t2-project.org/handbook/html/t2.package.html
17:34 < rxr> and the pages thereafter
17:34 < rxr> http://www.t2-project.org/handbook/html/t2.package.desc.html
17:34 < rxr> and co
17:34 < rxr> normal t2 packages just need a .desc file
17:34 < rxr> you can just copy an existing one and exchange the meta data and off you go
17:35 < keztoo> ok, ill give it a try
17:35 < keztoo> :-)
17:35 < rxr> if it is a package on freshmeat.net you can even use misc/archive/fmnewpackage.sh to help you get it created
17:35 < keztoo> after i read up a bit
17:36 < keztoo> i shoulkd be able to point to localhost for the download, right?
17:36 < kic-> rene, keztoo is talking about creating our own package
17:37 < kic-> our application developed in-house is a loadable kernel module
17:37 < keztoo> same here
17:37 < kic-> keztoo, this robert
17:37 < kic-> i know
17:37 < kic-> :D
17:37 < keztoo> lol
17:38 < rxr> file:// might work
17:38 < rxr> I would need to take a look
17:38 < rxr> such features are seldomely asked
17:38 < keztoo> i'll try both
17:39 < keztoo> so just so i understand what i just read
17:39 < kic-> we will stress test the t2 features
17:39 < kic-> don't worry
17:39 < kic-> :D
17:39 -!- sepp [n=sepp@Z5335.z.pppool.de] has joined #t2
17:40 < keztoo> i create my .desc and .conf (in my case conf is empty) and add those along with my make file to the tar
17:40 < kic-> i've had 4 guys working with T2 since friday
17:40 < kic-> all 4 doing multiple config types, multiple builds, multiple package slections, multiple archs
17:40 < kic-> etc
17:40 < kic-> all day every day, since friday
17:40 < kic-> heh
17:40 < kic-> wow
17:41 < rxr> keztoo: to your source tar ?
17:41 < kic-> emerge on firefox
17:41 < rxr> no
17:41 < kic-> takes a lont time :D
17:41 < keztoo> then i just include it in the packages file and kick off the build
17:41 < rxr> you need the .desc/.conf in some package/contrib(whatever/package directory
17:41 < kic-> keztoo, i think you need to create a package directory for it
17:41 < keztoo> yes
17:41 < kic-> in the t2 sub directories
17:41 < kic-> those are where your .conf and .desc files will go
17:42 < kic-> you probably won't have to change your tarball
17:42 < rxr> yes, tarballs are kept untouched
17:42 < kic-> omg
17:42 < kic-> i got something right
17:42 < kic-> how cool am I
17:42 < kic-> ?
17:43 < kic-> :P
17:43 < kic-> alright, new cisco pix is here, will be idling while i configure it
17:43 < keztoo> so root/package/network is where i create a new subdir?
17:44 < keztoo> and stick my files in there?
17:44 < rxr> root ?
17:44 < keztoo> root meaning the root of my t2 buold file system
17:44 < rxr> if root == t2-tree root then yes
17:44 < kic-> t2-trunk/package/network/our_pkg_name
17:44 < keztoo> yes, sorry
17:44 < rxr> yes
17:44 < kic-> k
17:45 < keztoo> ok, soon as my build completes i'll give it a try
17:54 < keztoo> oops, build broke again
17:54 < rxr> which erorr ?
17:54 < keztoo> linux 24 is the pkg
17:54 < keztoo> let me look at thge err log
17:55 < rxr> hm hm
17:56 < keztoo> err is 'no rule to make target 'include/linux/utsrelease.h
17:56 < keztoo> wait, that was up high
17:57 < keztoo> let me look lower in the log file
17:59 < keztoo> well it kept going but i think that was the err that broke the build
18:01 -!- sepp_ [n=sepp@Z46ef.z.pppool.de] has quit [Read error: 110 (Connection timed out)]
18:01 < keztoo> .
18:02 < rxr> keztoo: this should not really broke the build as the script is testing two ways to grab the version for older and newer kerneles
18:03 < keztoo> well later it also got a 'conflicting types for sys_wait4' from context.c
18:03 -!- somethingbad [i=vladi@unicorn.vladi.biz] has quit ["Lost terminal"]
18:04 < keztoo> i think the conflicting types err is what actually broke it
18:04 < rxr> I'm not sure if the linux24 package should be built for x86-64 at all
18:04 < rxr> we mostly have it for 2.4 forcing embedded setups
18:04 < rxr> while I can look at it after my glibc32 fix you might just want to disable it anyway :-)
18:05 < rxr> and just use linux26 instead
18:05 < keztoo> the break is in asm/unistd.h
18:05 < keztoo> function execve
18:06 < keztoo> i will go disable it in package and retart the build
18:06 < rxr> (.text+0x105d3): undefined reference to `__sync_fetch_and_add_4'
18:06 < rxr> indeed, ...
18:08 < keztoo> is 26 included by default or do i have to turn it on
18:09 < rxr> 26 is on by default
18:09 < keztoo> excellent, then the build is already started and will hopefully be done soon
18:11 < kic-> nice function name
18:13 < kic-> scripts/Config
18:13 < kic-> and specifying custom kernel config file
18:13 < kic-> will this be used when i do an Emerge-Pkg -missing=only linux26
18:13 < kic-> ?
18:14 < kic-> or is that only for Build-Target and Build-Pkg
18:14 < rxr> if you specifiy your custom kernel config it will also be used on Emerge
18:14 < kic-> k
18:14 < kic-> great
18:15 < rxr> Emerge-Pkg and Build-Target of course use the same internal package build script code
18:15 < kic-> figured
18:15 < kic-> :D
18:15 < kic-> testing it now
18:15 < kic-> just downloaded vanilla 2.6.21 src, ran make menuconfig, set all my options
18:15 < kic-> and copied over the .config i made there
18:16 -!- misl [n=chatzill@84-104-172-187.cable.quicknet.nl] has joined #t2
18:17 -!- write_erase [n=olivier@pleiades4.data-mobiles.com] has quit [Remote closed the connection]
18:22 < misl> hi folks
18:23 < keztoo> hi misl
18:23 < kic-> successful emerge
18:23 < kic-> booting it now
18:25 < kic-> success
18:25 < kic-> nice
18:25 < kic-> damn i like that
18:35 < keztoo> build broke again, this time on busybox
18:35 < keztoo> phase I
18:36 < rxr> damn
18:36 < rxr> why do you hit so many bugs
18:36 < rxr> what's the error this time ?
18:37 < keztoo> libbb/bb_strtonum line 103: conflicting types for bb_strtou
18:37 < keztoo> just lucky i guess
18:37 < rxr> ouhm
18:37 < rxr> maybe a recent update not yet tested for 64bit or so
18:38 < keztoo> also string to int (bb_strtou) at 140 and bb_strtoi at 115 and 141
18:39 < keztoo> i mean upper
18:39 < keztoo> at 140
18:40 < keztoo> i suppose i could remove busybox but i think we u;ltimately want it in our slimmed down target
18:41 < keztoo> think i should just remove it and move on rxr?
18:42 < rxr> yeah - maybe disable
18:43 < rxr> my test build should hit it soon
18:43 < keztoo> k
18:43 < kic-> i'm starting my build now too
18:44 < kic-> x86_64, generic target, minimal xorg, core2 opts
18:44 < kic-> customer kern.conf
18:44 < kic-> *custom
18:44 < kic-> no other changes, lets see what happens :D
18:46 < keztoo> breaks on ia64/gnu-efi now
18:46 < rxr> hm - that always builds here
18:47 < rxr> what's the error ?
18:48 < keztoo> boxdraw.c , line 103: expected '=', ; asm or __attribute__ before LiblsValTextGraphics
18:48 < keztoo> sValid
18:48 < keztoo> same for lines 153 and 164
18:48 < rxr> well - you probably do not boot on an EFI board, if you do not have fun debugging it disable it, likewise :.(
18:49 < keztoo> just different label names; isValidAscii and isValidEfiCtlChar
18:49 < keztoo> will do
18:49 < keztoo> build is restarted
18:57 < keztoo> fyi also had to remove elili (ia64)
18:58 < rxr> yes, needs the gnuefi thing
18:58 < keztoo> np
19:02 -!- tri_ [n=tri@p57ADCC66.dip0.t-ipconnect.de] has joined #t2
19:02 < tri_> re
19:02 < tri_> ping rxr
19:02 < rxr> yes
19:03 < tri_> http://imi.st/xen-tools.patch
19:03 < tri_> *.init files in the package dir are copy to /etc/init.d by t2 ?
19:04 < rxr> yes, expanded and copied
19:05 < tri_> ah ok the patch updates the tools to version 3.5 and i think the conf and the patch file isn't needed anymore
19:05 < rxr> you have not test-build ?
19:06 < tri_> yes it works here
19:06 < tri_> the build is succeful and the tools are running here
19:06 < tri_> without the conf and the patch file
19:09 < rxr> indeed built here now, as well
19:10 < tri_> ah fine :-)
19:10 < rxr> hm - on the other hand
19:10 < rxr> I'm not sure it they install as much as they installed before
19:10 < tri_> ?
19:11 < rxr> heh
19:11 < rxr> you install totally different tools than this package used to install before
19:12 < tri_> :-o
19:12 < rxr> the package used to built and install the xen tools as include in the xen sourcecode distribution
19:12 < tri_> which are the "right" ones ?
19:12 < tri_> the package which is in t2 doesn't build here so i was searching...
19:13 < tri_> and found this "xen-tools" :-)
19:22 < rxr> the official xen source release
19:29 -!- tri_ [n=tri@p57ADCC66.dip0.t-ipconnect.de] has left #t2 []
20:09 < keztoo> well made it to stage 3 this time but broke on sysfiles
20:12 < keztoo> got up to 'Creating user messagebus', then it failed with a GID 78 exists
20:14 < keztoo> it found the already existing group crontab and skipped that but then it broke trying to create user messagebus with GID 78 exists.
20:14 < sepp> it has gid 16
20:14 < keztoo> I'll look at the postsysfiles scripts to see if i can fix this manually
20:15 < keztoo> not sure what u mean sepp?
20:15 < sepp> messagebus is already group 16
20:16 < keztoo> oic
20:16 < sepp> does it tries to create a new group while running make?
20:17 < keztoo> i think so
20:17 < sepp> that is weird
20:18 < keztoo> dont know if its in make or post processing type script, checking now
20:18 < sepp> is that a cross build?
20:19 < keztoo> no std 64bit
20:19 < sepp> hmm
20:19 < keztoo> just verified that messagebus is indeed gid 16 in my grioups file also
20:26 < sepp> it broke on 3-sysfiles, that is weird
20:26 < sepp> that just cps things around
20:29 < keztoo> i see that
20:34 < keztoo> what is really strange is that it failed adding group messagebus but a quick grep shows no messagebus in any of the files in the package/base/sysfiles subdir
20:35 < sepp> nor gid 78
20:35 < sepp> if you enable xtrace you get a much more detailed logfile ...
20:36 < keztoo> where do i do that? i'll try anything
20:36 < sepp> in Config
20:36 < keztoo> if i change the config will that restart the build from phase 0 again?
20:36 < sepp> no
20:37 < sepp> that would be a bummer
20:37 < keztoo> ok, ill go set it on and restart the build
20:37 < sepp> "Create debug information (xtrace) for builds"
20:38 < keztoo> done and build started
20:38 < keztoo> does config reset the packages file?
20:39 < keztoo> i disabled linux24 earlier but now it is compiling it again
20:39 < sepp> it rebuilds it
20:39 < sepp> if you want to disable something there is also an option
20:39 < keztoo> np
20:39 < sepp> you must not manually edit the packages file ;)
20:40 < sepp> if you do a svn up and restart the build it will also be regenerated ...
20:40 < keztoo> i c
20:40 < keztoo> so use the config script to disable packages?
20:41 < sepp> Show expert and experimental options - Custom package selection - Edit package selection rules - enter 'O linux24'
20:43 < sepp> what t2 version do you use?
20:44 < sepp> btw, yes, Config is way too slow right now ... :p
20:45 < keztoo> currently running the default 64 bit version 2.6.17.9-dist #1 SMP Aug 22, x86_64
20:45 < sepp> i mean the thing you build
20:46 < keztoo> i grabbed the svn trunk yesterday, and updated this morning (about 5 hrs agho)
20:46 < sepp> and which target?
20:47 < sepp> i just wonder why there is linux24 in the list ...
20:48 < keztoo> generic:x86-64 (AMD64, Intel):opt=core2
20:48 < sepp> with some package selection template?
20:48 < keztoo> no
20:48 < keztoo> none
20:48 < sepp> ohh
20:48 < keztoo> i think it should not be pulling in 24
20:49 < sepp> generic without any template means build ~2700 packages
20:49 < keztoo> so far i have had to remove: gnu-efi, elilo, busybox and linux24
20:50 < keztoo> i was worried about dependencies so i tried to just build the default to get started
20:50 < keztoo> new build w xtrace enabled is done
20:50 < sepp> maybe you should use something like desktop to get started
20:52 < keztoo> just changed config to minimal w x and basic desktop
20:53 -!- pdinoto [n=pdinoto@200.117.246.103] has joined #t2
20:53 < sepp> there is also an extra 'Exact Desktop' target
20:53 < sepp> but i guess both are fine
20:53 < keztoo> no, i have none, or minimal, or minimal and x or minimal and x and desk or benchmark
20:54 < keztoo> set config and restarted build w xtrace enabled
20:54 < sepp> the 1st option '(generic) Generic System' you can select 'Exact Desktop' there
20:55 < sepp> that is way better to get started than generic with all packages
20:56 < keztoo> will do
20:56 < sepp> just a hint ...
20:56 < sepp> of course you can even make your own
20:57 < keztoo> i will but first i need to get a good build, then i can start removing packages
20:57 < keztoo> just 2 b safe
20:58 < sepp> :)
20:59 < sepp> better to remove something from 700 packages and not 2700 :)
20:59 < keztoo> those config changes caused the buiold to start again from phase 0
20:59 < sepp> yes, you changed the full target
20:59 < sepp> that is how it works, but if you have ccache it is not that bad
21:00 < keztoo> so it will just use the cache for already compiled pkgs, like gcc and such?
21:00 < sepp> i should have wanred you :/
21:00 < keztoo> np
21:00 < keztoo> its all part of the learning curve
21:01 < keztoo> i really appreciate your help
21:01 < sepp> compiler cache, it is not at 100% at the 1st time but it gets faster
21:02 < keztoo> that is good because my boss really needs me to get our module compiled against the kernel today
21:02 < keztoo> once this builds i still have to add my package and recompile and keep my fingers crossed
21:02 < sepp> i hope you have a fast pc
21:03 < keztoo> its decent , 2 dual cores
21:03 < keztoo> j5
21:03 < keztoo> but glibc and gcc take forever
21:03 < sepp> but with 2700 packages it would be running way over 24h
21:04 < keztoo> ouch
21:04 < keztoo> thats about what it took to download the trunk yesterday
21:04 < sepp> both bloatware :/
21:05 < keztoo> sepp, once the new target is built, what is the easiest way to get the kernel source on it?
21:05 < keztoo> the sources that were used to build it'
21:06 < sepp> to build a 3rd party module?
21:07 < keztoo> well yes, i am currently building a target which will most probably be .21.5 and i want to use THAT target to compile my modulke against
21:07 < keztoo> i will burd a cd and create a new system from the result of this build
21:07 < keztoo> but i then want to use that system to compile my modulke
21:08 < sepp> t2 has no kernel source package because it is a bad idea, you get version mismatches, other .configs and in the end it does not work
21:08 < sepp> t2 builds 3rd party mods as part of the kernel
21:08 < keztoo> but the current hiost is .17 and i would like to use .21 to buikld with
21:09 < sepp> the kernel you build with build-target
21:09 < keztoo> so i thought i would use the result of this build (a .21 kernel) to create a new desk top to develop with
21:09 < keztoo> and that is where i will need the kernel sources
21:10 < sepp> you can extend linux26 from an extra package to build your mod while the kernel source is still there - inside the linux26 package
21:11 < keztoo> so once i boot my new desktop and am running .21 you are sayiong that inside the linux26 pkg is the kernel sources?
21:12 < sepp> it is just a hook where you extend the linux26 build - once the main kernel is done it evaluates your code to build a 3rd party kernel mod
21:12 < sepp> maybe check package/network/acx100/postlinux.conf and package/network/acx100/pkg_linux26_post.conf
21:13 < keztoo> so how would i do that on the .21 system that i am building a target for now, or are you saying that i will always produce a target from the .17 i am running now?
21:13 < sepp> the host kernel does not matter, t2 builds a sandbox to compile things in
21:14 < keztoo> so basically i build our target on .17 with my module and it produces a .21 target but i can't easily build a target from the .21 target
21:15 < keztoo> i thought once i created the .21 target i could just copy the sources over to it and burn the CD with them in it
21:15 < keztoo> then i could build targets from a .21 system
21:15 < sepp> what?
21:15 < keztoo> ok, let me start again
21:16 < keztoo> my current system is a .17 kernel 2.6.17.9 specifically
21:16 < sepp> btw, there are 28 3rd part kernel mods already, if you want to take a look ...
21:16 < sepp> ok
21:16 < keztoo> now when i build a target from this it creates a .21 kernel
21:17 < keztoo> i think it is currently 21.5
21:17 < sepp> yes
21:17 < keztoo> now i would like to create a CD from the .21 target and create a new desktop from the .21 CD
21:17 < keztoo> so i can boot into a .21 kernel
21:17 < sepp> why create yet a new one?
21:18 < sepp> the cd will be fully working
21:18 < keztoo> well, here we go
21:18 < keztoo> so once i create a new desktop system from the .21 CD
21:18 < keztoo> i would like to then use that as my build host
21:20 < sepp> why not use the .21 CD as it is?
21:20 < sepp> why build with the new system again?
21:21 < keztoo> oic what your asking
21:21 < keztoo> well because i would like to create a build machine and would like to have the latest kernel running on nthe build machine
21:21 < keztoo> .21
21:22 < keztoo> i think what you are trying to tell me is
21:23 < keztoo> the version of build machine doesnt matter since it produce a target from different sources anyway, right?
21:23 < keztoo> i just thought it would be better to have a .21 machine as my build machine
21:23 < sepp> yes, as long as gcc/bash/awk and some other things work well it does not matter since it build a entire toolchain 1st
21:23 < sepp> to compile all the rest
21:24 < sepp> that is why there are several stages
21:24 < keztoo> right so even if i am running a -9 build host the target will always be the latest build as long as i keep my trunk up to date
21:24 < sepp> it does not matter what kernel or gcc version the host is running
21:24 < keztoo> got it
21:24 < sepp> yes
21:26 < sepp> that is why there are stage 0 & 1 - it builds a fresh toolchain, gcc binutils bash ............
21:26 < keztoo> well, that can be good or bad
21:26 < keztoo> lol
21:26 < keztoo> cross compiles are a bit of a pain
21:27 < keztoo> especially when it tries to run something from the target that the host doesn't support
21:27 < sepp> cross compiler support works not that way
21:27 < keztoo> like the x makekeys issue i ran across yesterday
21:27 < keztoo> it doesnt try to chroot and such?
21:28 < sepp> no, you can not run the final binaries anyway
21:29 < keztoo> so if i get an error that an intermediate program the make process created and then tried to run that is solved by configuration?
21:29 < sepp> if you build for arm on x86 it would not work to chroot into the arm root
21:29 < keztoo> cause i had x try to run a cross build version of makekeys yesterday
21:30 < keztoo> which would not run cause i was on a 32 bit and building for a 64bit
21:30 < sepp> if it build an helper for a cross build you will need to compile that with the host compiler and run it
21:30 < keztoo> same thing happend last week when it tried to chroot and run bin/bash
21:30 < sepp> so the package needs a patch
21:30 < keztoo> ok, so we are on the same page
21:30 < keztoo> i understand
21:33 < sepp> if you would want to cross compile i.e. ruby that could be quite hard - it would want to execute itself - ofcourse that would not work in an cross build - so you would need to patch it to build a host interpreter 1st and use that to process whatever it does
21:34 < keztoo> yes i see and even in the case of some base pkgs, like x, they need help during the build phase so for example they dont try to build a helper program using the target compiler, but rather the host compiler
21:35 < keztoo> i think i am starting to get it
21:35 < keztoo> i am just a bit slow
21:35 < keztoo> :-)
21:35 < sepp> yes, many packages are not very croos build friendly
21:37 < sepp> but there are already 871 working
21:37 < sepp> so that is quite good
21:41 < sepp> 895 packages - so 1/3 :)
21:42 < keztoo> :-)
21:42 < keztoo> i think it is a good system and a good approach
21:43 < keztoo> but i am sure i will have issues adding my new module this afternoon
21:44 < keztoo> which is a good 3rd party module to look at as a template - you know a small simple one
21:44 < keztoo> i used the .conf and .desc files from python so far
21:45 < keztoo> and of course modified them a bit
21:45 < sepp> package/network/acx100 is quite easy to understand
21:46 < sepp> it does not do very much ...
21:47 < sepp> `ls -l package/*/*/pkg_linux26_post.conf` <- lists all linux26 'extenders'
21:48 < sepp> well, most
21:49 < keztoo> ok, ty
21:49 < keztoo> ill look at that one
21:50 < sepp> or package/network/adm8211
21:50 < keztoo> and the list of extenders
21:51 < sepp> if you are lucky all you need to do is to set some kerneldir makeopt and maybe patch a uname call away
21:54 < keztoo> i hope so, just have it kick off our make would probably be good enough
21:55 < sepp> if it has not /usr/src/linux hardcoded as expectation it could be
21:55 < keztoo> no we can configure that
22:04 < sepp> :)
22:08 < kic-> 1-perl
22:08 < kic-> *crosses fingers*
22:08 < kic-> yay
22:08 < kic-> success
22:08 < kic-> on to autoconf
22:08 < kic-> and this time, with -lgdbm
22:08 < kic-> :P
22:09 < sepp> hehe :)
22:09 < sepp> what caused the error last time, any clue?
22:09 < kic-> dunno, i ended up taking out the -lgdbm the way you suggested with var_append
22:10 < kic-> although
22:10 < kic-> what i'm doing differently now
22:10 < kic-> is building ExactDesktop
22:10 < kic-> instead of generic system
22:10 < sepp> ahh, ok
22:12 < sepp> guess there is some problem when you build 'everything'
22:19 < keztoo> maybe that was the problem i was having with sysfiles
22:19 < keztoo> some other package might have already set the GIDs
22:19 < keztoo> we'll see , that broke on stage 3 i'm only up to stagte 2 now
22:21 -!- Stelz [n=stealth@unaffiliated/stelz] has quit ["There is intelligent life out there..."]
22:21 < sepp> grep 78 package/*/*/*.in
22:21 < sepp> ohhh
22:22 < sepp> `grep 78 package/*/*/*.in | grep safe_groupadd` that does not look that good
22:23 < keztoo> o u mean re: the GID error?
22:23 < sepp> the last line shows a collision
22:24 < sepp> dovecot vs jabber ...
22:24 < sepp> both 78
22:25 < keztoo> what about sysfiles?
22:25 < keztoo> that is where the error actually occurs on my system
22:25 < keztoo> i can probably just remove jabber for my purposes
22:25 < keztoo> and probably dovecot too
22:26 -!- sparc-kly [n=mubex@adsl-64-237-241-185.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
22:26 < sepp> they are not in the desktop target ;)
22:27 < sepp> no need to remove anything, but it sgould be fixed ...
22:28 < keztoo> this is why build all is not a good idea :-)
22:29 < sepp> well, build all takes forever but it should be tested
22:29 < keztoo> take it from me, it won't build :-)
22:29 < sepp> but i never do
22:35 -!- idealm [n=ideal@c58-107-18-66.belrs2.nsw.optusnet.com.au] has joined #t2
22:36 -!- idealm [n=ideal@c58-107-18-66.belrs2.nsw.optusnet.com.au] has quit [Read error: 104 (Connection reset by peer)]
22:43 -!- Stelz [n=stealth@unaffiliated/stelz] has joined #t2
22:56 -!- keztoo---- [n=root@adsl-074-238-088-202.sip.bct.bellsouth.net] has joined #t2
22:57 < keztoo----> .
22:57 < sepp> . .
22:58 -!- misl [n=chatzill@84-104-172-187.cable.quicknet.nl] has quit ["ChatZilla 0.9.78.1 [Firefox 2.0.0.4/2007051502]"]
23:02 -!- kic_ [n=kicstah@adsl-074-238-088-202.sip.bct.bellsouth.net] has joined #t2
23:07 -!- kic- [i=kicstah@adsl-074-238-088-201.sip.bct.bellsouth.net] has quit [Nick collision from services.]
23:07 -!- kic_ is now known as kic-
23:14 -!- keztoo [n=root@adsl-074-238-088-206.sip.bct.bellsouth.net] has quit [Read error: 110 (Connection timed out)]
23:17 -!- keztoo-- [n=root@adsl-074-238-088-202.sip.bct.bellsouth.net] has joined #t2
23:18 < keztoo--> ..
23:18 -!- keztoo---- [n=root@adsl-074-238-088-202.sip.bct.bellsouth.net] has quit [Read error: 110 (Connection timed out)]
23:19 < keztoo--> ..
23:19 < keztoo--> sepp u still here?
23:19 < sepp> yes
23:20 < keztoo--> build broke instage 3 on filesystem/parted
23:20 < keztoo--> says it needs me to download gnu/readline
23:21 < sepp> i saw that today - but i was too lazy to check it, it is some ncurses fuckup :/
23:22 < sepp> i disabled it ...
23:23 < sepp> that is some new regression in trunk - maybe disable it right now
23:25 < sepp> i think its configure is too dumb, but i did not really check it :(
23:29 < sepp> it might works with: var_append GCC_WRAPPER_APPEND " " "-lncurses"
23:32 < sepp> or if you need parted right now confopt append --no-readline or whatever the error output suggests
23:38 -!- keztoo-- [n=root@adsl-074-238-088-202.sip.bct.bellsouth.net] has quit [Read error: 110 (Connection timed out)]
23:39 -!- tri_ [n=tri@p57ADCC66.dip0.t-ipconnect.de] has joined #t2
23:39 < tri_> join #edevelp
23:39 < sepp> hehe
23:40 -!- keztoo-- [n=root@adsl-074-238-088-202.sip.bct.bellsouth.net] has joined #t2
23:40 < tri_> hm moin ;-)
23:40 < sepp> hi tri :)
23:40 < tri_> hi sepp
23:43 < sepp> tri_: are you an e17 developer?
23:48 < tri_> no only interessted at evas efl etk ewl etc...
23:48 < tri_> for iwlwifi rx problems: http://bughost.org/iwlwifi/iwlwifi-txpower.patch
23:49 < sepp> ok
23:51 < tri_> so have to sleep now :-)
23:51 < tri_> gn8 !
23:51 < tri_> cu
23:51 -!- cbrake is now known as cbrake_away
23:51 < sepp> cu tri
23:51 -!- tri_ [n=tri@p57ADCC66.dip0.t-ipconnect.de] has left #t2 []
--- Log closed Fri Jun 22 00:00:24 2007