T2 IRC Log: 2004-11-05
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 Fri Nov 05 00:00:22 2004
01:28 < mnemoc> valentin: didn't helped, have any of you built a bootdisk this days?
03:42 -!- rxr_ [~rene@p213.54.238.53.tisdip.tiscali.de] has joined #t2
03:42 -!- Topic for #t2: T2 | The System Development Environment | http://www.exactcode.de/t2
03:42 -!- Topic set by rxr [] [Fri Oct 29 16:53:31 2004]
03:42 [Users #t2]
03:42 [ _Ragnar_ ] [ CIA-9 ] [ jsaw ] [ mnemoc] [ praenti] [ rxr_ ]
03:42 [ blindcoder] [ daja77] [ martin_] [ nzg ] [ rxr ] [ valentin]
03:42 -!- Irssi: #t2: Total of 12 nicks [0 ops, 0 halfops, 0 voices, 12 normal]
03:42 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
03:42 -!- [freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup
03:42 -!- Irssi: Join to #t2 was synced in 11 secs
03:44 -!- mnemoc_ [~amery@200.75.27.38] has joined #t2
03:56 -!- mnemoc [~amery@200.75.27.25] has quit [Read error: 110 (Connection timed out)]
03:56 -!- mnemoc_ is now known as mnemoc
03:59 -!- rxr [~rene@p213.54.220.138.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)]
04:03 -!- martin_ [~martin@brln-d9ba0251.pool.mediaWays.net] has quit [Read error: 110 (Connection timed out)]
04:05 -!- blindcod1r [~blindcode@dsl-082-082-101-072.arcor-ip.net] has joined #t2
04:05 -!- blindcoder [~blindcode@dsl-082-082-094-101.arcor-ip.net] has quit [Nick collision from services.]
04:05 -!- blindcod1r is now known as blindcoder
04:08 -!- martin_ [~martin@brln-d9ba034c.pool.mediaWays.net] has joined #t2
05:29 -!- nzg [~tschmidt@p508EA00E.dip.t-dialin.net] has quit [Read error: 104 (Connection reset by peer)]
05:30 -!- nzg [~tschmidt@p508EA00E.dip.t-dialin.net] has joined #t2
06:33 -!- CIA-9 [~CIA@to.je.spocco.com] has quit [Read error: 104 (Connection reset by peer)]
06:37 -!- CIA-9 [~CIA@to.je.spocco.com] has joined #t2
08:38 < blindcoder> moin
08:49 -!- CIA-9 [~CIA@to.je.spocco.com] has quit []
08:52 -!- CIA-9 [~CIA@to.je.spocco.com] has joined #t2
10:11 < rxr_> re
10:11 -!- You're now known as rxr
10:18 < martin_> -:- [invisible users on irc(9510)] 53%
10:39 < rxr> haha: ", running X.org version 6.8.0 on Gentoo. (Gentoo has not
10:39 < rxr> released 6.8.1 packages yet.)"
10:57 < CIA-9> rene * r4560 /trunk/package/base/procps/procps.desc: * updated procps (3.2.3 -> 3.2.4)
11:12 < CIA-9> rene * r4561 /trunk/package/www/firefox/firefox.desc: * updated firefox (1.0rc1 -> 1.0rc2)
11:12 < CIA-9> rene * r4562 /trunk/package/mail/thunderbird/thunderbird.desc: * updated thunderbird (0.8 -> 0.9)
11:14 < CIA-9> rene * r4563 /trunk/package/base/linux24/parse-config:
11:14 < CIA-9> * added a status message to linux24 that outputs a reminder about
11:14 < CIA-9> how to build 3rd party kernel modules ;-)
11:17 < CIA-9> rene * r4564 /trunk/package/base/curl/curl.desc: * updated curl (7.12.1 -> 7.12.2)
11:48 < valentin> mnemoc: i would be happy if i finished my sparc build
11:53 < rxr> .oO ( http://www.rocklinux.net/submaster/data/2004/11/0319152215459.patch )
11:55 < valentin> why exactly .oO ?
11:57 < CIA-9> rene * r4565 /trunk/package/x11/xcompmgr/xcompmgr.desc:
11:57 < CIA-9> * updated xcompmgr (initially I just wanted to fix the version typo
11:57 < CIA-9> ...) - tested
11:58 < CIA-9> rene * r4566 /trunk/package/x11/xautolock/xautolock.conf: * fixed xautolock to install the man-page
11:59 < rxr> valentin: I now downtaed binutils and still the same sparc issue ...
12:00 < rxr> valentin: so there is only gcc left ...
12:00 < rxr> valentin: causing that regression
12:00 < valentin> damn
12:02 < CIA-9> rene * r4567 /trunk/package/database/bdb/bdb_config.sh:
12:02 < CIA-9> * removed bdb chmod of .so files, shoudl be unneeded and collides with
12:02 < CIA-9> with the case no cxx wrapper is built ...
12:03 < CIA-9> rene * r4568 /trunk/package/audio/ardour/ardour.desc:
12:03 < CIA-9> * updated ardour (0.9beta11.2 -> 0.9beta12) - does not yet fix the
12:03 < CIA-9> build
12:06 < CIA-9> rene * r4569 /trunk/package/kde/qt/ (config.hlp config.in qt.conf): * only disable qt stl and exception support when requested
12:09 < CIA-9> rene * r4570 /trunk/target/desktop/config.in: * enabled tdb, qcad and thunderbird for the desktop target
12:09 < daja77> hm there was a binutils patch fixing regressions with gcc-3.4.x
12:19 < CIA-9> rene * r4571 /trunk/package/base/glibc/glibc.conf:
12:19 < CIA-9> * cleanup of glibc linuxthread tarball extraction - AFAICS it was even
12:19 < CIA-9> defect
12:23 -!- sparc-kly [~sparc-kly@65-23-204-82.prtc.net] has joined #t2
12:38 < CIA-9> rene * r4572 /trunk/package/www/firefox/firefox.conf:
12:38 < CIA-9> * oops - fixed typo in firefox.conf version overwrite that prevents
12:38 < CIA-9> the rc{1,2} update to build
12:46 < CIA-9> rene * r4573 /trunk/package/multimedia/dvdauthor/gcc34.patch: * fixed dvdauthor for gcc >= 3.4.x
12:55 < CIA-9> rene * r4574 /trunk/package/audio/soundtracker/gcc34.patch: * fixed soundtracker for gcc >= 3.4.x
13:06 < CIA-9> rene * r4575 /trunk/package/audio/streamtuner/gcc34.patch: * fixed streamtuner for gcc >= 3.4.x
13:17 -!- tschmidt_ [~tschmidt@p508EAF64.dip.t-dialin.net] has joined #t2
13:17 -!- tschmidt_ is now known as codegul
13:21 < CIA-9> rene * r4576 /trunk/package/audio/streamtuner/ (3 files):
13:21 < CIA-9> * oops, I comitted the streamtuner gcc fix too early, overlooked
13:21 < CIA-9> that it is a multi-tar package and one of the additional tars
13:21 < CIA-9> needed a fix, too
13:34 * rxr away
13:34 < rxr> cu
13:35 -!- nzg [~tschmidt@p508EA00E.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
13:38 -!- sparc-kly [~sparc-kly@65-23-204-82.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
13:39 -!- nzg [~tschmidt@p508EB090.dip.t-dialin.net] has joined #t2
13:41 -!- tschmidt_ [~tschmidt@pD95F8EEB.dip.t-dialin.net] has joined #t2
13:46 -!- tschmidt_ [~tschmidt@pD95F8EEB.dip.t-dialin.net] has quit [Read error: 232 (Connection reset by peer)]
13:49 -!- tschmidt_ [~tschmidt@pD95F8EEB.dip.t-dialin.net] has joined #t2
13:54 -!- codegul [~tschmidt@p508EAF64.dip.t-dialin.net] has quit [Read error: 238 (Connection timed out)]
14:02 -!- tschmidt_ is now known as codegul
14:03 -!- codegul [~tschmidt@pD95F8EEB.dip.t-dialin.net] has quit ["Verlassend"]
14:07 -!- nzg [~tschmidt@p508EB090.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
14:25 -!- CIA-9 [~CIA@to.je.spocco.com] has quit [Read error: 54 (Connection reset by peer)]
14:29 -!- CIA-9 [~CIA@to.je.spocco.com] has joined #t2
14:29 < mnemoc> moin
14:32 < valentin> moin mnemoc
14:36 < mnemoc> moin valentin
14:37 < mnemoc> i disabled all isomd5 stuff on bootdisk and diet still failing to build stuff
14:37 < mnemoc> (linuxrc.c in this case)
14:37 < mnemoc> i guess it is including glibc stuff
14:38 < jsaw> re
14:38 < mnemoc> re jsaw
14:39 * mnemoc fears to strace gcc :p
14:39 < jsaw> mnemoc: I'm not sure about exactly when I introduced the bootdisk diet patch, maybe you compare rock and t2 (sorry, I have no time to do it myself).
14:39 < mnemoc> bootdisk or dietlibc?
14:40 < jsaw> the patch basically avoids the system diet and uses the one in the build environment
14:40 < jsaw> bootdisk
14:40 < mnemoc> i have no system diet
14:47 < mnemoc> jsaw: the only real difference between rock-trunk's and t2's is that we remove --enable-cxx from bdb
14:52 < daja77> O_o
14:55 < mnemoc> strace -o diet.strace -vvfF diet $CC linuxrc.o -o linuxrc :'(
15:07 < jsaw> mnemoc: don't forget to use DIETHOME="$base/build/${ROCKCFG_ID}/usr/dietlibc"
15:07 < jsaw> and
15:07 < jsaw> PATH="$base/build/${ROCKCFG_ID}/ROCK/tools.cross/diet-bin:$PATH"
15:08 < jsaw> otherwise your system diet is taken.
15:13 * mnemoc added some debug flags and tryed again
15:14 < mnemoc> btw, i don't have diet locally installed
15:14 < mnemoc> DIETHOME: /usr/src/rock/t2-trunk/build/boot-2.1.0-beta-x86-pentium-mmx-32-bootdisk-expert/usr/dietlibc
15:15 < mnemoc> 308 execve("/usr/src/rock/t2-trunk/build/boot-2.1.0-beta-x86-pentium-mmx-32-bootdisk-expert/ROCK/tools.cross/diet-bin/diet", ["diet", "i586-unknown-linux-gnu-cc", "linuxrc.o", "-o", "linuxrc"], [/* 230 vars */]) = 0
15:25 < mnemoc> In file included from /usr/src/rock/t2-trunk/build/boot-2.1.0-beta-x86-pentium-mmx-32-bootdisk-expert/usr/include/stdio.h:72,
15:25 < mnemoc> from /usr/src/rock/t2-trunk/target/bootdisk/linuxrc.c:32:
15:26 < mnemoc> uhmm
15:32 < mnemoc> according to strace gcc i looking on dietlibc/include on forth try, i guess cmd_wrapper is doing something
16:22 < rxr> re
16:23 < mnemoc> re rxr
16:23 < mnemoc> can you comment my last 5 lines?
16:23 < mnemoc> please
16:24 < rxr> ouhm
16:25 < rxr> can I track this when I'm at home and I simply remove the system diet on my box ?
16:25 < mnemoc> i hope so :)
16:39 < jsaw> hmmm, i have no idea.
16:41 < mnemoc> jsaw: can you build a bootdisk to see if you have this problem too? please :)
16:44 < rxr> mnemoc: I build one at home
16:44 < mnemoc> OT: how can i do xf86 to attach the mouse if it is plugged _after_ xserver has started?
16:44 < mnemoc> rxr: no problems?
16:45 < valentin> mnemoc: use /dev/mice
16:46 < mnemoc> uhm?
16:46 < valentin> what kind of mouse do you want to hotplug ?
16:46 < mnemoc> usb
16:47 < valentin> then use /dev/mice and imps2 in your x config
16:47 < mnemoc> i don't have an /dev/mice
16:47 < valentin> dev/input/mice ?
16:48 < mnemoc> neither
16:48 < daja77> mnemoc: doesn't matter?
16:48 < daja77> -?
16:48 < mnemoc> InputDevice "Mouse0" "CorePointer"
16:49 < mnemoc> InputDevice "Mouse1" "SendCoreEvents"
16:49 < daja77> just write /dev/input/mice, if it got listed atm or not
16:49 < mnemoc> Mouse0
16:49 < mnemoc> Option "Protocol" "PS/2"
16:49 < mnemoc> Option "Device" "/dev/misc/psaux"
16:49 < mnemoc> oh
16:49 < mnemoc> devfsd will create it?
16:49 < valentin> think so
16:49 < valentin> Section "InputDevice"
16:49 < valentin> Identifier "Mouse0"
16:49 < valentin> Driver "mouse"
16:49 < valentin> Option "Protocol" "IMPS/2" # "BusMouse"
16:49 < valentin> Option "Device" "/dev/input/mice"
16:49 < valentin> Option "Buttons" "5"
16:49 < valentin> Option "ZaxisMapping" "4 5"
16:49 < valentin> EndSection
16:50 < valentin> mixes touchpad/usb/ps2 mouses
16:50 < mnemoc> OMG
16:50 < mnemoc> ok... i'll trust
16:51 < valentin> hey - NO WARANTY
16:51 < mnemoc> too late
16:52 < valentin> does it work ?
16:52 < mnemoc> mom
16:54 < valentin> seems like it did not work :/
16:54 < valentin> mnemoc: still alive *g*
16:55 < mnemoc> :)
16:55 < mnemoc> screen
16:55 < mnemoc> usb replug worked
16:55 < mnemoc> but touchpad doesn't
16:57 < valentin> but your touchpad did before ?
16:58 < mnemoc> as PS/2 yes
16:58 < mnemoc> i had two mice set, as i pasted before
16:58 < mnemoc> one PS/2 at psaux for touchpad
16:58 < mnemoc> and one ImPS/2 for usb
16:58 < valentin> normaly that should mix into /dev/mice
16:59 < valentin> but i am not sure if you need to set up sth
17:00 < mnemoc> what happens if i set Mouse0 as PS/2 and Mouse1 to dev/mice?
17:00 < mnemoc> as ImPS/2?
17:00 < valentin> try it
17:01 < jsaw> is the "$arch" variable identical with the architecture/$arch directories?
17:06 < valentin> jsaw: within the buidscripts arch should be set to
17:06 < valentin> arch="${ROCKCFG_ARCH:-none}
17:08 < valentin> and that is set by the config script to the choosen architecture, so it should be compatible to the existing architecture dirs, shouldn't it ?
17:09 < jsaw> ah, thanks
17:09 < mnemoc> i loose my keyboard and i had to reboot :|
17:15 < mnemoc> but i'm happy with two mouse devices considering now i can {un,}plug usb :)
17:16 < mnemoc> http://www.compass.com/synaptics/ <--- is this needed? (lx 2.4.28 here)
17:16 < mnemoc> it looks quite old
17:28 * valentin watering some pasta
17:36 < jsaw> mv self /dev/home/meal
17:36 < jsaw> cu l8r
18:13 < rxr> rehi
18:43 < valentin> re
18:46 < valentin> rxr: did i answer jsaw's question correctly ?
18:46 < rxr> the arch variable? I think so
19:03 < rxr> valentin: but sparc results would be more interesting ... ;-)
19:04 < valentin> as soon as i manage to compile a glibc
19:08 < valentin> rxr: the last thing that is executed is glibc_prepatch
19:08 < valentin> no custmain
19:14 < valentin> hm, something here is strange:
19:14 < valentin> in build_this_package:
19:14 < valentin> if [ $autopatch = 1 ]; then
19:14 < valentin> hook_eval prepatch
19:14 < valentin> apply_patchfiles
19:14 < valentin> hook_eval postpatch
19:14 < valentin> fi
19:15 < valentin> after returning from prepatch hook, the execution stops
19:16 < rxr> post the exexutiong of glibc_prepatch in the private query ...
19:16 < valentin> ok
19:41 * rxr testing a new valgrind for ppc
19:47 < rxr> cool - the 2.2.0 ppc snapshots of valgrind now works
19:50 < mnemoc> re
19:53 < rxr> hi mnemoc
19:54 < mnemoc> hi rxr
20:05 < CIA-9> rene * r4577 /trunk/package/develop/valgrind/ (valgrind.conf valgrind.desc): * added the 2.2.0-ppc valgrind - tested on powerpc
20:10 < mnemoc> rxr: do you think /sbin/init should work as MTAs and allow 'changes on the fly' or selecting one INIT as _the_ init should disable others?
20:12 < rxr> interesting that you bring up the /sbin/init on-the-fly change
20:13 < rxr> I just thought about that in the lecture I was in the uni for, today
20:13 < mnemoc> i don't want to
20:13 < mnemoc> i want to select init on Config, once and forever
20:13 < mnemoc> currently i use runit and disable sysvinit package, but i lack of the other tools provided by sysvinit. on my target now i patch sysvinit Makefile to not install {tel,}init. i guess this can be done automaticly if sysvinit is not choosen as init on Config, and init' stage-9 would never write /sbin/init, write /sbin/init.$pkg if
20:13 < mnemoc> ROCKCONFIG_INIT is not set
20:13 < rxr> don't you think it would be interesting to change the init system in the system (e.g. to test s.th. new) without the need for a full rebuild?
20:13 < mnemoc> or points to me
20:14 < mnemoc> i guess that would bring _lot_ of chaos
20:14 < rxr> yes, potentinally it could
20:14 < mnemoc> another thing is to decouple init_style of rc.d' scripts
20:14 < rxr> this is why I wrote "thought", not plan to ;-)
20:15 < rxr> please explain
20:15 < mnemoc> something like RH's /etc/sysconfig/functions
20:15 < mnemoc> rc.d/$foo source that and call clean functions
20:16 < mnemoc> i.e. init_style can change on the fly
20:16 < mnemoc> btw, init_schema != init_style
20:16 < mnemoc> an init mandates an init_schema
20:16 < mnemoc> an init_style is just the way to show output
20:16 < rxr> I had this in drock
20:17 < mnemoc> i know
20:17 < rxr> clifford did not liked that, because he wanted "freestanding" init files
20:17 < mnemoc> but clifford is not here :)
20:17 < rxr> that you can copy from on sys to another and still have working init files ...
20:17 < rxr> hehe
20:17 < rxr> this was just the explanation why we have those ...
20:17 < rxr> I have nothing again clean functions ...
20:17 < mnemoc> i had a 'fight' with clifford about that two years ago
20:18 < rxr> but in the overal concept, we have init depended init scripts anyway, since e.g. minit does need s.th. totally different ...
20:18 < rxr> but ripping out the inlined functions from the init scripts get's a vote from me, too
20:18 < rxr> mnemoc: another thing, have you read my mail weeks ago about ripping coded out of /etc?
20:19 < mnemoc> we have runit, which use ./run ans supervision
20:19 < rxr> e.g. rockplug, rocknet and the init scripts?
20:19 < mnemoc> yes
20:19 < mnemoc> /etc is for config, not for scripts
20:19 < rxr> .oO(so much stuff going on in t2 land)
20:19 < rxr> do this proposal has an ack from your side?
20:20 * rxr is a bit puzzled about the general feeling regarding this due to the lack of replies ...
20:20 < mnemoc> we are an small comunity :)
20:21 < rxr> but not long
20:21 < rxr> mnemoc: can you push some init proposal on the list?
20:21 < rxr> and vote for /etc code relocation? ;-)
20:21 < mnemoc> for 2.2? :)
20:22 < rxr> btw. let's first get 2.1 out before we do more restructuring
20:22 < rxr> yep
20:22 < rxr> I would like to keep 2.1 as stable as possible and get some PR done with releasing it ...
20:22 < mnemoc> can you take a look into bootdisk problem i reported? :)
20:22 < rxr> and getting a full ISO set for all of our current 3 main architectures (x86, ppc, sparc) out for convienience
20:22 < rxr> )
20:23 < rxr> mnemoc: you mean the diet problem?
20:23 < rxr> with a good 2.1 release we should be able to gether some wieder audience and thus reduce our own's workload
20:23 < mnemoc> yes
20:23 < rxr> at least I hope so ..
20:23 < rxr> yep, diet debugging already scheduled
20:23 < rxr> but first I need to review some glibc crap to get our sparc builds on-tracik
20:23 < mnemoc> i think diet is ok
20:24 < rxr> just not the includes, i know ...
20:24 < rxr> I still hope to get to diet tonight
20:24 < mnemoc> strace shows it pass through cmd_wrapper which (i guess) is adding unproper -includes
20:24 < rxr> ah, thanks for the hint I'll see
20:25 < mnemoc> i'll do some tiny changes on sysvinit/config.in to select init and do non-intrusive stuff to improve runit
20:25 < CIA-9> valentin * r4578 /trunk/package/base/glibc/glibc.conf: * fixed prepatch hook - do not abort anymore just because grep matches nothing
20:25 < mnemoc> a bigger change is needed but i'll write a proposal for 2.2
20:28 < rxr> I hate the glibc's Makefiles
20:29 < mnemoc> they reflect gnu policies
20:29 < rxr> huh?
20:29 < rxr> they reflect braindamage
20:29 < rxr> writing ugly seperate build in objdir crap with undebugable criptic shit is not GNU policy
20:30 < mnemoc> not written in paper but written in monk's damaged brain
20:31 < rxr> no in ullrich dreppers ones ... :-(
20:31 < rxr> hey sneakers on pro7
20:32 < mnemoc> no pro7 here
20:32 < rxr> to bad it already started so it is too late for cat > sneakers.mmg
20:38 < rxr> damn - even 16:9 and dolby digital ...
20:38 < rxr> damn - and I missed it by 16 minutes
20:43 < CIA-9> mnemoc * r4579 /trunk/package/network/dhcpcd/ (dhcpcd.conf rocknet_dhcpcd.sh): * Add rocknet module for dhcpcd
20:46 < rxr> well - I need a new harddisc for all the high-res mpeg2's anyway ...
20:53 < mnemoc> i have a 20G disk on the laptop and 10 at home :p
21:47 < rxr> boring job of the day: reverse engeneering glibc makefiles
21:51 -!- rxr changed the topic of #t2 to: T2 | From the folks that get the GNU/Junk working for you | http://www.exactcode.de/t2
21:51 -!- rxr changed the topic of #t2 to: T2 | From the folks that get the GNU/Junk working for you ...
21:59 < rxr> oh my god
21:59 < rxr> the glibc makefiles even call other directories Makefiles during the build
22:00 < rxr> this is in the elf subdir:
22:00 < rxr> make -C ../io objdir=/home/t2-trunk/src.glibc.1099679636.17339.799171250/glibc-2.3.3/objdir -f Makefile -f ../elf/rtld-Rules rtld-all rtld-modules='rtld-xstat64.os rtld-fxstat64.os rtld-open.os rtld-close.os rtld-read.os rtld-write.os rtld-lseek.os rtld-access.os rtld-fcntl.os rtld-getcwd.os rtld-readlink.os rtld-xstatconv.os rtld-lxstat.os'
22:00 < rxr> notice the ../io :-(((((((((((((((((
22:01 < rxr> yeah yeah - reproduced the crap error
22:28 -!- CIA-9 [~CIA@to.je.spocco.com] has quit [Read error: 104 (Connection reset by peer)]
22:29 < rxr> valentin: good news: it is not gcc-3.4.x alone, the same preprocessed assembler junk does not compile with the in-system 3.2.x one either
22:29 < rxr> so it is some bug in the glibc compile process ...
22:31 < rxr> nope
22:31 < rxr> it is already enought to preprocess the junk with the old gcc
22:37 < rxr> -.text; .global __libc_open; .align 4; __libc_open:; .type __libc_open,@function;; mov 5, %g1; ta 0x10; bcs __syscall_error_handler; nop; .section .gnu.linkonce.t.__sparc.get_pic.l7,"ax",@progbits; .globl __sparc.get_pic.l7; .hidden __sparc.get_pic.l7; .type __sparc.get_pic.l7,@function; __sparc.get_pic.l7: retl; add %o7, %l7, %l7; .previous; .subsection 3; __syscall_error_handler: save %sp,-96,%sp; sethi %hi(_GLOBAL_OFFSET_TABLE_-4), %l7; call __sparc.get_pic.l7; add %l7, %lo(_GLOBAL_OFFSET_TABLE_+4), %l7; ld [%l7 + errno], %l0; st %i0, [%l0]; jmpl %i7+8, %g0; restore %g0, -1, %o0; .previous;
22:37 < rxr> +.text; .global __libc_open; .align 4; __libc_open:; .type __libc_open,@function;; mov 5, %g1; ta 0x10; bcs __syscall_error_handler; nop; .section .gnu.linkonce.t. 1 .get_pic.l7,"ax",@progbits; .globl 1 .get_pic.l7; .hidden 1 .get_pic.l7; .type 1 .get_pic.l7,@function; 1 .get_pic.l7: retl; add %o7, %l7, %l7; .previous; .subsection 3; __syscall_error_handler: save %sp,-96,%sp; sethi %hi(_GLOBAL_OFFSET_TABLE_-4), %l7; call 1 .get_pic.l7; add %l7, %lo(_GLOBAL_OFFSET_TABLE_+4), %l7; ld [%l7 + errno], %l0; st %i0, [%l0]; jmpl %i7+8, %g0; restore %g0, -1, %o0; .previous;
22:39 < rxr> this is diff -u working defect ...
22:47 < rxr> hm
22:47 < rxr> maybe this one
22:47 < rxr> http://sources.redhat.com/bugzilla/show_bug.cgi?id=467
22:47 < rxr> http://sources.redhat.com/ml/libc-alpha/2004-05/msg00125.html
22:47 < rxr> ?
22:48 -!- CIA-10 [~CIA@to.je.spocco.com] has joined #t2
22:49 < rxr> yep, here:
22:49 < rxr> http://sources.redhat.com/ml/libc-alpha/2004-05/msg00123.html
22:49 < rxr> ^- our bug
22:51 < jsaw> oh my, I was battling with this in openoffice for a while (more specific in STLport, where they used "linux" somewhere in a path...)
22:52 * mnemoc hides to commit heretical patch
22:53 < jsaw> ?
22:53 < rxr> ?
22:53 < CIA-10> mnemoc * r4581 /trunk/package/network/dhcpcd/nodetach_option.patch: * Add option (-a) to dhcpcd to not detach after getting dhcp's answer.
22:54 < rxr> jsaw: what was your exact problem?
22:55 < jsaw> I do not remember exactly what they did, I only remember that "linux" was expanded to "1",
22:55 < jsaw> and some includes ended up with looking for ../1/../header-file.h
22:56 < rxr> ah so ;-) I like the preprocessor
22:56 < jsaw> yeah, me too
23:01 < jsaw> oh, btw, I just started a bootdisk build and have no other tree to test a shadow update (4.0.3 -> 4.0.5), somebody should do it, from the NEWS file:
23:01 < jsaw> fixed securirty bug in libmisc/pwdcheck.c which allow unauthorized
23:01 < jsaw> account properties modification.
23:02 < jsaw> Probably the "hotfix.patch", "mallocdef.patch" and "libpam-fix.patch" may also be removed.
23:03 < jsaw> <- or should I simply commit without testing?
23:05 < rxr> jsaw: just building the package? or rebuilding all?
23:05 < rxr> I would rather avoid doing a system build on the athlon, since I want to release a CD-set of the current build ...
23:06 < rxr> btw, and old screen or so on sparcseg-faulted today:
23:06 < rxr> Suddenly the Dungeon collapses!! - You die...
23:06 < rxr> ^- is what that yielded ...
23:06 < jsaw> just building the package, but... I also hasitate to a ./scripts/Build-Pkg here...
23:06 < rxr> ok - i do
23:06 < rxr> (my four boxes eating energy need workload)
23:07 < jsaw> :)
23:08 < jsaw> oh, just checked cvs Changelog:
23:08 < jsaw> b* libmisc/pwdcheck.c:
23:08 < jsaw> one word fix .. ommited "else" was removed. Now security bug which allow
23:08 < jsaw> unauthorized account properties modification is realy closed.
23:10 < jsaw> let's wait for 4.0.6 if it's anyway not fixed ...
23:10 < rxr> ouhm
23:12 < jsaw> (or create the patch from cvs...)
23:14 < rxr> no, thanks
23:14 < rxr> we can wait a day or so for it ...
23:14 < jsaw> yep.
23:14 < rxr> 971 builds total, 922 completed fine, 49 with errors.
23:14 < rxr> if anyone wanna help fixing ...
23:14 < jsaw> is linux 2.4 needed in the bootdisk target?
23:14 < rxr> nope
23:15 < rxr> but we still have to twek the scripts to use one or both or so of the kernels and modules ...
23:15 < jsaw> == 11/05/04 23:13:36 =[1]=> Aborted building package linux24.
23:15 < rxr> historically the bootdisk used this default kernel thing, since it is gone a new scheme needs to be choosen
23:15 < rxr> aborted?
23:16 < mnemoc> my bootdisk built both kernels
23:16 < jsaw> rwsem-spinlock.h:57: error: previous declaration of 'init_rwsem' was here
23:16 < jsaw> rwsem-spinlock.c:36: error: conflicting types for 'init_rwsem'
23:16 < rxr> yes, that is ok and intended
23:16 * rxr did not got that ..
23:16 < rxr> jsaw: what are you optimizing for?
23:16 < mnemoc> jsaw: that happend with x86-generic here
23:17 < mnemoc> jsaw: x86-pmmx worked perfect
23:17 < jsaw> yes, x86-generic
23:17 < rxr> yep - I built athlon without error ;-)
23:17 < jsaw> <- has an old 386
23:17 < rxr> jsaw: coudl you fix this?
23:17 < jsaw> will try to...
23:18 < jsaw> oh damn, the FASTCALL thing
23:20 < rxr> the build errors over here for an athlon desktop build:
23:20 < rxr> ardour avview boost dvgrab efsd elicit entrance firefox gedit-plugins
23:20 < rxr> ggrab gkrellm-plugins glade gnomemeeting gnucash gsmp gtkhtml10 jamin
23:20 < rxr> kahakai kaxtv koffice ladspa-swh-plugins ladspa-vst libgnomedb linphone
23:20 < rxr> mergeant mod_mono mono-develop mplayer-fonts mysql-perl mythtv nvrec ocs
23:20 < rxr> perl-dbd-postgresql pixieplus qcad quota ripperx rivatv stlport
23:20 < rxr> streamtuner superkaramba transcode vdr vls vnc vstserver xmp zapping zgv
23:20 < mnemoc> jsaw: /me installing elks on a 286 now :p
23:20 < jsaw> mnemoc: nice!
23:22 < jsaw> bbl
23:34 < rxr> for whom does this happen regularly?:
23:34 < rxr> root:/usr/src/t2-trunk/src.qcad.1099693663.25973.3102287048# ./debug.sh
23:34 < rxr> debug-qcad:[src.qcad.1099693663.25973.3102287048]# cd qcad*
23:34 < rxr> ./debug.sh: line 3: 25830 Segmentation fault (core dumped) chroot "/usr/src/t2-trunk/build/default-2.1.0-beta-x86-athlon-32-desktop" /bin/bash ROCK/src.qcad.1099693663.25973.3102287048/debug_x.sh
23:38 < rxr> aside from that qcad fix on the way
23:38 * rxr showering
--- Log closed Sat Nov 06 00:00:37 2004
01:28 < mnemoc> valentin: didn't helped, have any of you built a bootdisk this days?
03:42 -!- rxr_ [~rene@p213.54.238.53.tisdip.tiscali.de] has joined #t2
03:42 -!- Topic for #t2: T2 | The System Development Environment | http://www.exactcode.de/t2
03:42 -!- Topic set by rxr [] [Fri Oct 29 16:53:31 2004]
03:42 [Users #t2]
03:42 [ _Ragnar_ ] [ CIA-9 ] [ jsaw ] [ mnemoc] [ praenti] [ rxr_ ]
03:42 [ blindcoder] [ daja77] [ martin_] [ nzg ] [ rxr ] [ valentin]
03:42 -!- Irssi: #t2: Total of 12 nicks [0 ops, 0 halfops, 0 voices, 12 normal]
03:42 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
03:42 -!- [freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup
03:42 -!- Irssi: Join to #t2 was synced in 11 secs
03:44 -!- mnemoc_ [~amery@200.75.27.38] has joined #t2
03:56 -!- mnemoc [~amery@200.75.27.25] has quit [Read error: 110 (Connection timed out)]
03:56 -!- mnemoc_ is now known as mnemoc
03:59 -!- rxr [~rene@p213.54.220.138.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)]
04:03 -!- martin_ [~martin@brln-d9ba0251.pool.mediaWays.net] has quit [Read error: 110 (Connection timed out)]
04:05 -!- blindcod1r [~blindcode@dsl-082-082-101-072.arcor-ip.net] has joined #t2
04:05 -!- blindcoder [~blindcode@dsl-082-082-094-101.arcor-ip.net] has quit [Nick collision from services.]
04:05 -!- blindcod1r is now known as blindcoder
04:08 -!- martin_ [~martin@brln-d9ba034c.pool.mediaWays.net] has joined #t2
05:29 -!- nzg [~tschmidt@p508EA00E.dip.t-dialin.net] has quit [Read error: 104 (Connection reset by peer)]
05:30 -!- nzg [~tschmidt@p508EA00E.dip.t-dialin.net] has joined #t2
06:33 -!- CIA-9 [~CIA@to.je.spocco.com] has quit [Read error: 104 (Connection reset by peer)]
06:37 -!- CIA-9 [~CIA@to.je.spocco.com] has joined #t2
08:38 < blindcoder> moin
08:49 -!- CIA-9 [~CIA@to.je.spocco.com] has quit []
08:52 -!- CIA-9 [~CIA@to.je.spocco.com] has joined #t2
10:11 < rxr_> re
10:11 -!- You're now known as rxr
10:18 < martin_> -:- [invisible users on irc(9510)] 53%
10:39 < rxr> haha: ", running X.org version 6.8.0 on Gentoo. (Gentoo has not
10:39 < rxr> released 6.8.1 packages yet.)"
10:57 < CIA-9> rene * r4560 /trunk/package/base/procps/procps.desc: * updated procps (3.2.3 -> 3.2.4)
11:12 < CIA-9> rene * r4561 /trunk/package/www/firefox/firefox.desc: * updated firefox (1.0rc1 -> 1.0rc2)
11:12 < CIA-9> rene * r4562 /trunk/package/mail/thunderbird/thunderbird.desc: * updated thunderbird (0.8 -> 0.9)
11:14 < CIA-9> rene * r4563 /trunk/package/base/linux24/parse-config:
11:14 < CIA-9> * added a status message to linux24 that outputs a reminder about
11:14 < CIA-9> how to build 3rd party kernel modules ;-)
11:17 < CIA-9> rene * r4564 /trunk/package/base/curl/curl.desc: * updated curl (7.12.1 -> 7.12.2)
11:48 < valentin> mnemoc: i would be happy if i finished my sparc build
11:53 < rxr> .oO ( http://www.rocklinux.net/submaster/data/2004/11/0319152215459.patch )
11:55 < valentin> why exactly .oO ?
11:57 < CIA-9> rene * r4565 /trunk/package/x11/xcompmgr/xcompmgr.desc:
11:57 < CIA-9> * updated xcompmgr (initially I just wanted to fix the version typo
11:57 < CIA-9> ...) - tested
11:58 < CIA-9> rene * r4566 /trunk/package/x11/xautolock/xautolock.conf: * fixed xautolock to install the man-page
11:59 < rxr> valentin: I now downtaed binutils and still the same sparc issue ...
12:00 < rxr> valentin: so there is only gcc left ...
12:00 < rxr> valentin: causing that regression
12:00 < valentin> damn
12:02 < CIA-9> rene * r4567 /trunk/package/database/bdb/bdb_config.sh:
12:02 < CIA-9> * removed bdb chmod of .so files, shoudl be unneeded and collides with
12:02 < CIA-9> with the case no cxx wrapper is built ...
12:03 < CIA-9> rene * r4568 /trunk/package/audio/ardour/ardour.desc:
12:03 < CIA-9> * updated ardour (0.9beta11.2 -> 0.9beta12) - does not yet fix the
12:03 < CIA-9> build
12:06 < CIA-9> rene * r4569 /trunk/package/kde/qt/ (config.hlp config.in qt.conf): * only disable qt stl and exception support when requested
12:09 < CIA-9> rene * r4570 /trunk/target/desktop/config.in: * enabled tdb, qcad and thunderbird for the desktop target
12:09 < daja77> hm there was a binutils patch fixing regressions with gcc-3.4.x
12:19 < CIA-9> rene * r4571 /trunk/package/base/glibc/glibc.conf:
12:19 < CIA-9> * cleanup of glibc linuxthread tarball extraction - AFAICS it was even
12:19 < CIA-9> defect
12:23 -!- sparc-kly [~sparc-kly@65-23-204-82.prtc.net] has joined #t2
12:38 < CIA-9> rene * r4572 /trunk/package/www/firefox/firefox.conf:
12:38 < CIA-9> * oops - fixed typo in firefox.conf version overwrite that prevents
12:38 < CIA-9> the rc{1,2} update to build
12:46 < CIA-9> rene * r4573 /trunk/package/multimedia/dvdauthor/gcc34.patch: * fixed dvdauthor for gcc >= 3.4.x
12:55 < CIA-9> rene * r4574 /trunk/package/audio/soundtracker/gcc34.patch: * fixed soundtracker for gcc >= 3.4.x
13:06 < CIA-9> rene * r4575 /trunk/package/audio/streamtuner/gcc34.patch: * fixed streamtuner for gcc >= 3.4.x
13:17 -!- tschmidt_ [~tschmidt@p508EAF64.dip.t-dialin.net] has joined #t2
13:17 -!- tschmidt_ is now known as codegul
13:21 < CIA-9> rene * r4576 /trunk/package/audio/streamtuner/ (3 files):
13:21 < CIA-9> * oops, I comitted the streamtuner gcc fix too early, overlooked
13:21 < CIA-9> that it is a multi-tar package and one of the additional tars
13:21 < CIA-9> needed a fix, too
13:34 * rxr away
13:34 < rxr> cu
13:35 -!- nzg [~tschmidt@p508EA00E.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
13:38 -!- sparc-kly [~sparc-kly@65-23-204-82.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
13:39 -!- nzg [~tschmidt@p508EB090.dip.t-dialin.net] has joined #t2
13:41 -!- tschmidt_ [~tschmidt@pD95F8EEB.dip.t-dialin.net] has joined #t2
13:46 -!- tschmidt_ [~tschmidt@pD95F8EEB.dip.t-dialin.net] has quit [Read error: 232 (Connection reset by peer)]
13:49 -!- tschmidt_ [~tschmidt@pD95F8EEB.dip.t-dialin.net] has joined #t2
13:54 -!- codegul [~tschmidt@p508EAF64.dip.t-dialin.net] has quit [Read error: 238 (Connection timed out)]
14:02 -!- tschmidt_ is now known as codegul
14:03 -!- codegul [~tschmidt@pD95F8EEB.dip.t-dialin.net] has quit ["Verlassend"]
14:07 -!- nzg [~tschmidt@p508EB090.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
14:25 -!- CIA-9 [~CIA@to.je.spocco.com] has quit [Read error: 54 (Connection reset by peer)]
14:29 -!- CIA-9 [~CIA@to.je.spocco.com] has joined #t2
14:29 < mnemoc> moin
14:32 < valentin> moin mnemoc
14:36 < mnemoc> moin valentin
14:37 < mnemoc> i disabled all isomd5 stuff on bootdisk and diet still failing to build stuff
14:37 < mnemoc> (linuxrc.c in this case)
14:37 < mnemoc> i guess it is including glibc stuff
14:38 < jsaw> re
14:38 < mnemoc> re jsaw
14:39 * mnemoc fears to strace gcc :p
14:39 < jsaw> mnemoc: I'm not sure about exactly when I introduced the bootdisk diet patch, maybe you compare rock and t2 (sorry, I have no time to do it myself).
14:39 < mnemoc> bootdisk or dietlibc?
14:40 < jsaw> the patch basically avoids the system diet and uses the one in the build environment
14:40 < jsaw> bootdisk
14:40 < mnemoc> i have no system diet
14:47 < mnemoc> jsaw: the only real difference between rock-trunk's and t2's is that we remove --enable-cxx from bdb
14:52 < daja77> O_o
14:55 < mnemoc> strace -o diet.strace -vvfF diet $CC linuxrc.o -o linuxrc :'(
15:07 < jsaw> mnemoc: don't forget to use DIETHOME="$base/build/${ROCKCFG_ID}/usr/dietlibc"
15:07 < jsaw> and
15:07 < jsaw> PATH="$base/build/${ROCKCFG_ID}/ROCK/tools.cross/diet-bin:$PATH"
15:08 < jsaw> otherwise your system diet is taken.
15:13 * mnemoc added some debug flags and tryed again
15:14 < mnemoc> btw, i don't have diet locally installed
15:14 < mnemoc> DIETHOME: /usr/src/rock/t2-trunk/build/boot-2.1.0-beta-x86-pentium-mmx-32-bootdisk-expert/usr/dietlibc
15:15 < mnemoc> 308 execve("/usr/src/rock/t2-trunk/build/boot-2.1.0-beta-x86-pentium-mmx-32-bootdisk-expert/ROCK/tools.cross/diet-bin/diet", ["diet", "i586-unknown-linux-gnu-cc", "linuxrc.o", "-o", "linuxrc"], [/* 230 vars */]) = 0
15:25 < mnemoc> In file included from /usr/src/rock/t2-trunk/build/boot-2.1.0-beta-x86-pentium-mmx-32-bootdisk-expert/usr/include/stdio.h:72,
15:25 < mnemoc> from /usr/src/rock/t2-trunk/target/bootdisk/linuxrc.c:32:
15:26 < mnemoc> uhmm
15:32 < mnemoc> according to strace gcc i looking on dietlibc/include on forth try, i guess cmd_wrapper is doing something
16:22 < rxr> re
16:23 < mnemoc> re rxr
16:23 < mnemoc> can you comment my last 5 lines?
16:23 < mnemoc> please
16:24 < rxr> ouhm
16:25 < rxr> can I track this when I'm at home and I simply remove the system diet on my box ?
16:25 < mnemoc> i hope so :)
16:39 < jsaw> hmmm, i have no idea.
16:41 < mnemoc> jsaw: can you build a bootdisk to see if you have this problem too? please :)
16:44 < rxr> mnemoc: I build one at home
16:44 < mnemoc> OT: how can i do xf86 to attach the mouse if it is plugged _after_ xserver has started?
16:44 < mnemoc> rxr: no problems?
16:45 < valentin> mnemoc: use /dev/mice
16:46 < mnemoc> uhm?
16:46 < valentin> what kind of mouse do you want to hotplug ?
16:46 < mnemoc> usb
16:47 < valentin> then use /dev/mice and imps2 in your x config
16:47 < mnemoc> i don't have an /dev/mice
16:47 < valentin> dev/input/mice ?
16:48 < mnemoc> neither
16:48 < daja77> mnemoc: doesn't matter?
16:48 < daja77> -?
16:48 < mnemoc> InputDevice "Mouse0" "CorePointer"
16:49 < mnemoc> InputDevice "Mouse1" "SendCoreEvents"
16:49 < daja77> just write /dev/input/mice, if it got listed atm or not
16:49 < mnemoc> Mouse0
16:49 < mnemoc> Option "Protocol" "PS/2"
16:49 < mnemoc> Option "Device" "/dev/misc/psaux"
16:49 < mnemoc> oh
16:49 < mnemoc> devfsd will create it?
16:49 < valentin> think so
16:49 < valentin> Section "InputDevice"
16:49 < valentin> Identifier "Mouse0"
16:49 < valentin> Driver "mouse"
16:49 < valentin> Option "Protocol" "IMPS/2" # "BusMouse"
16:49 < valentin> Option "Device" "/dev/input/mice"
16:49 < valentin> Option "Buttons" "5"
16:49 < valentin> Option "ZaxisMapping" "4 5"
16:49 < valentin> EndSection
16:50 < valentin> mixes touchpad/usb/ps2 mouses
16:50 < mnemoc> OMG
16:50 < mnemoc> ok... i'll trust
16:51 < valentin> hey - NO WARANTY
16:51 < mnemoc> too late
16:52 < valentin> does it work ?
16:52 < mnemoc> mom
16:54 < valentin> seems like it did not work :/
16:54 < valentin> mnemoc: still alive *g*
16:55 < mnemoc> :)
16:55 < mnemoc> screen
16:55 < mnemoc> usb replug worked
16:55 < mnemoc> but touchpad doesn't
16:57 < valentin> but your touchpad did before ?
16:58 < mnemoc> as PS/2 yes
16:58 < mnemoc> i had two mice set, as i pasted before
16:58 < mnemoc> one PS/2 at psaux for touchpad
16:58 < mnemoc> and one ImPS/2 for usb
16:58 < valentin> normaly that should mix into /dev/mice
16:59 < valentin> but i am not sure if you need to set up sth
17:00 < mnemoc> what happens if i set Mouse0 as PS/2 and Mouse1 to dev/mice?
17:00 < mnemoc> as ImPS/2?
17:00 < valentin> try it
17:01 < jsaw> is the "$arch" variable identical with the architecture/$arch directories?
17:06 < valentin> jsaw: within the buidscripts arch should be set to
17:06 < valentin> arch="${ROCKCFG_ARCH:-none}
17:08 < valentin> and that is set by the config script to the choosen architecture, so it should be compatible to the existing architecture dirs, shouldn't it ?
17:09 < jsaw> ah, thanks
17:09 < mnemoc> i loose my keyboard and i had to reboot :|
17:15 < mnemoc> but i'm happy with two mouse devices considering now i can {un,}plug usb :)
17:16 < mnemoc> http://www.compass.com/synaptics/ <--- is this needed? (lx 2.4.28 here)
17:16 < mnemoc> it looks quite old
17:28 * valentin watering some pasta
17:36 < jsaw> mv self /dev/home/meal
17:36 < jsaw> cu l8r
18:13 < rxr> rehi
18:43 < valentin> re
18:46 < valentin> rxr: did i answer jsaw's question correctly ?
18:46 < rxr> the arch variable? I think so
19:03 < rxr> valentin: but sparc results would be more interesting ... ;-)
19:04 < valentin> as soon as i manage to compile a glibc
19:08 < valentin> rxr: the last thing that is executed is glibc_prepatch
19:08 < valentin> no custmain
19:14 < valentin> hm, something here is strange:
19:14 < valentin> in build_this_package:
19:14 < valentin> if [ $autopatch = 1 ]; then
19:14 < valentin> hook_eval prepatch
19:14 < valentin> apply_patchfiles
19:14 < valentin> hook_eval postpatch
19:14 < valentin> fi
19:15 < valentin> after returning from prepatch hook, the execution stops
19:16 < rxr> post the exexutiong of glibc_prepatch in the private query ...
19:16 < valentin> ok
19:41 * rxr testing a new valgrind for ppc
19:47 < rxr> cool - the 2.2.0 ppc snapshots of valgrind now works
19:50 < mnemoc> re
19:53 < rxr> hi mnemoc
19:54 < mnemoc> hi rxr
20:05 < CIA-9> rene * r4577 /trunk/package/develop/valgrind/ (valgrind.conf valgrind.desc): * added the 2.2.0-ppc valgrind - tested on powerpc
20:10 < mnemoc> rxr: do you think /sbin/init should work as MTAs and allow 'changes on the fly' or selecting one INIT as _the_ init should disable others?
20:12 < rxr> interesting that you bring up the /sbin/init on-the-fly change
20:13 < rxr> I just thought about that in the lecture I was in the uni for, today
20:13 < mnemoc> i don't want to
20:13 < mnemoc> i want to select init on Config, once and forever
20:13 < mnemoc> currently i use runit and disable sysvinit package, but i lack of the other tools provided by sysvinit. on my target now i patch sysvinit Makefile to not install {tel,}init. i guess this can be done automaticly if sysvinit is not choosen as init on Config, and init' stage-9 would never write /sbin/init, write /sbin/init.$pkg if
20:13 < mnemoc> ROCKCONFIG_INIT is not set
20:13 < rxr> don't you think it would be interesting to change the init system in the system (e.g. to test s.th. new) without the need for a full rebuild?
20:13 < mnemoc> or points to me
20:14 < mnemoc> i guess that would bring _lot_ of chaos
20:14 < rxr> yes, potentinally it could
20:14 < mnemoc> another thing is to decouple init_style of rc.d' scripts
20:14 < rxr> this is why I wrote "thought", not plan to ;-)
20:15 < rxr> please explain
20:15 < mnemoc> something like RH's /etc/sysconfig/functions
20:15 < mnemoc> rc.d/$foo source that and call clean functions
20:16 < mnemoc> i.e. init_style can change on the fly
20:16 < mnemoc> btw, init_schema != init_style
20:16 < mnemoc> an init mandates an init_schema
20:16 < mnemoc> an init_style is just the way to show output
20:16 < rxr> I had this in drock
20:17 < mnemoc> i know
20:17 < rxr> clifford did not liked that, because he wanted "freestanding" init files
20:17 < mnemoc> but clifford is not here :)
20:17 < rxr> that you can copy from on sys to another and still have working init files ...
20:17 < rxr> hehe
20:17 < rxr> this was just the explanation why we have those ...
20:17 < rxr> I have nothing again clean functions ...
20:17 < mnemoc> i had a 'fight' with clifford about that two years ago
20:18 < rxr> but in the overal concept, we have init depended init scripts anyway, since e.g. minit does need s.th. totally different ...
20:18 < rxr> but ripping out the inlined functions from the init scripts get's a vote from me, too
20:18 < rxr> mnemoc: another thing, have you read my mail weeks ago about ripping coded out of /etc?
20:19 < mnemoc> we have runit, which use ./run ans supervision
20:19 < rxr> e.g. rockplug, rocknet and the init scripts?
20:19 < mnemoc> yes
20:19 < mnemoc> /etc is for config, not for scripts
20:19 < rxr> .oO(so much stuff going on in t2 land)
20:19 < rxr> do this proposal has an ack from your side?
20:20 * rxr is a bit puzzled about the general feeling regarding this due to the lack of replies ...
20:20 < mnemoc> we are an small comunity :)
20:21 < rxr> but not long
20:21 < rxr> mnemoc: can you push some init proposal on the list?
20:21 < rxr> and vote for /etc code relocation? ;-)
20:21 < mnemoc> for 2.2? :)
20:22 < rxr> btw. let's first get 2.1 out before we do more restructuring
20:22 < rxr> yep
20:22 < rxr> I would like to keep 2.1 as stable as possible and get some PR done with releasing it ...
20:22 < mnemoc> can you take a look into bootdisk problem i reported? :)
20:22 < rxr> and getting a full ISO set for all of our current 3 main architectures (x86, ppc, sparc) out for convienience
20:22 < rxr> )
20:23 < rxr> mnemoc: you mean the diet problem?
20:23 < rxr> with a good 2.1 release we should be able to gether some wieder audience and thus reduce our own's workload
20:23 < mnemoc> yes
20:23 < rxr> at least I hope so ..
20:23 < rxr> yep, diet debugging already scheduled
20:23 < rxr> but first I need to review some glibc crap to get our sparc builds on-tracik
20:23 < mnemoc> i think diet is ok
20:24 < rxr> just not the includes, i know ...
20:24 < rxr> I still hope to get to diet tonight
20:24 < mnemoc> strace shows it pass through cmd_wrapper which (i guess) is adding unproper -includes
20:24 < rxr> ah, thanks for the hint I'll see
20:25 < mnemoc> i'll do some tiny changes on sysvinit/config.in to select init and do non-intrusive stuff to improve runit
20:25 < CIA-9> valentin * r4578 /trunk/package/base/glibc/glibc.conf: * fixed prepatch hook - do not abort anymore just because grep matches nothing
20:25 < mnemoc> a bigger change is needed but i'll write a proposal for 2.2
20:28 < rxr> I hate the glibc's Makefiles
20:29 < mnemoc> they reflect gnu policies
20:29 < rxr> huh?
20:29 < rxr> they reflect braindamage
20:29 < rxr> writing ugly seperate build in objdir crap with undebugable criptic shit is not GNU policy
20:30 < mnemoc> not written in paper but written in monk's damaged brain
20:31 < rxr> no in ullrich dreppers ones ... :-(
20:31 < rxr> hey sneakers on pro7
20:32 < mnemoc> no pro7 here
20:32 < rxr> to bad it already started so it is too late for cat > sneakers.mmg
20:38 < rxr> damn - even 16:9 and dolby digital ...
20:38 < rxr> damn - and I missed it by 16 minutes
20:43 < CIA-9> mnemoc * r4579 /trunk/package/network/dhcpcd/ (dhcpcd.conf rocknet_dhcpcd.sh): * Add rocknet module for dhcpcd
20:46 < rxr> well - I need a new harddisc for all the high-res mpeg2's anyway ...
20:53 < mnemoc> i have a 20G disk on the laptop and 10 at home :p
21:47 < rxr> boring job of the day: reverse engeneering glibc makefiles
21:51 -!- rxr changed the topic of #t2 to: T2 | From the folks that get the GNU/Junk working for you | http://www.exactcode.de/t2
21:51 -!- rxr changed the topic of #t2 to: T2 | From the folks that get the GNU/Junk working for you ...
21:59 < rxr> oh my god
21:59 < rxr> the glibc makefiles even call other directories Makefiles during the build
22:00 < rxr> this is in the elf subdir:
22:00 < rxr> make -C ../io objdir=/home/t2-trunk/src.glibc.1099679636.17339.799171250/glibc-2.3.3/objdir -f Makefile -f ../elf/rtld-Rules rtld-all rtld-modules='rtld-xstat64.os rtld-fxstat64.os rtld-open.os rtld-close.os rtld-read.os rtld-write.os rtld-lseek.os rtld-access.os rtld-fcntl.os rtld-getcwd.os rtld-readlink.os rtld-xstatconv.os rtld-lxstat.os'
22:00 < rxr> notice the ../io :-(((((((((((((((((
22:01 < rxr> yeah yeah - reproduced the crap error
22:28 -!- CIA-9 [~CIA@to.je.spocco.com] has quit [Read error: 104 (Connection reset by peer)]
22:29 < rxr> valentin: good news: it is not gcc-3.4.x alone, the same preprocessed assembler junk does not compile with the in-system 3.2.x one either
22:29 < rxr> so it is some bug in the glibc compile process ...
22:31 < rxr> nope
22:31 < rxr> it is already enought to preprocess the junk with the old gcc
22:37 < rxr> -.text; .global __libc_open; .align 4; __libc_open:; .type __libc_open,@function;; mov 5, %g1; ta 0x10; bcs __syscall_error_handler; nop; .section .gnu.linkonce.t.__sparc.get_pic.l7,"ax",@progbits; .globl __sparc.get_pic.l7; .hidden __sparc.get_pic.l7; .type __sparc.get_pic.l7,@function; __sparc.get_pic.l7: retl; add %o7, %l7, %l7; .previous; .subsection 3; __syscall_error_handler: save %sp,-96,%sp; sethi %hi(_GLOBAL_OFFSET_TABLE_-4), %l7; call __sparc.get_pic.l7; add %l7, %lo(_GLOBAL_OFFSET_TABLE_+4), %l7; ld [%l7 + errno], %l0; st %i0, [%l0]; jmpl %i7+8, %g0; restore %g0, -1, %o0; .previous;
22:37 < rxr> +.text; .global __libc_open; .align 4; __libc_open:; .type __libc_open,@function;; mov 5, %g1; ta 0x10; bcs __syscall_error_handler; nop; .section .gnu.linkonce.t. 1 .get_pic.l7,"ax",@progbits; .globl 1 .get_pic.l7; .hidden 1 .get_pic.l7; .type 1 .get_pic.l7,@function; 1 .get_pic.l7: retl; add %o7, %l7, %l7; .previous; .subsection 3; __syscall_error_handler: save %sp,-96,%sp; sethi %hi(_GLOBAL_OFFSET_TABLE_-4), %l7; call 1 .get_pic.l7; add %l7, %lo(_GLOBAL_OFFSET_TABLE_+4), %l7; ld [%l7 + errno], %l0; st %i0, [%l0]; jmpl %i7+8, %g0; restore %g0, -1, %o0; .previous;
22:39 < rxr> this is diff -u working defect ...
22:47 < rxr> hm
22:47 < rxr> maybe this one
22:47 < rxr> http://sources.redhat.com/bugzilla/show_bug.cgi?id=467
22:47 < rxr> http://sources.redhat.com/ml/libc-alpha/2004-05/msg00125.html
22:47 < rxr> ?
22:48 -!- CIA-10 [~CIA@to.je.spocco.com] has joined #t2
22:49 < rxr> yep, here:
22:49 < rxr> http://sources.redhat.com/ml/libc-alpha/2004-05/msg00123.html
22:49 < rxr> ^- our bug
22:51 < jsaw> oh my, I was battling with this in openoffice for a while (more specific in STLport, where they used "linux" somewhere in a path...)
22:52 * mnemoc hides to commit heretical patch
22:53 < jsaw> ?
22:53 < rxr> ?
22:53 < CIA-10> mnemoc * r4581 /trunk/package/network/dhcpcd/nodetach_option.patch: * Add option (-a) to dhcpcd to not detach after getting dhcp's answer.
22:54 < rxr> jsaw: what was your exact problem?
22:55 < jsaw> I do not remember exactly what they did, I only remember that "linux" was expanded to "1",
22:55 < jsaw> and some includes ended up with looking for ../1/../header-file.h
22:56 < rxr> ah so ;-) I like the preprocessor
22:56 < jsaw> yeah, me too
23:01 < jsaw> oh, btw, I just started a bootdisk build and have no other tree to test a shadow update (4.0.3 -> 4.0.5), somebody should do it, from the NEWS file:
23:01 < jsaw> fixed securirty bug in libmisc/pwdcheck.c which allow unauthorized
23:01 < jsaw> account properties modification.
23:02 < jsaw> Probably the "hotfix.patch", "mallocdef.patch" and "libpam-fix.patch" may also be removed.
23:03 < jsaw> <- or should I simply commit without testing?
23:05 < rxr> jsaw: just building the package? or rebuilding all?
23:05 < rxr> I would rather avoid doing a system build on the athlon, since I want to release a CD-set of the current build ...
23:06 < rxr> btw, and old screen or so on sparcseg-faulted today:
23:06 < rxr> Suddenly the Dungeon collapses!! - You die...
23:06 < rxr> ^- is what that yielded ...
23:06 < jsaw> just building the package, but... I also hasitate to a ./scripts/Build-Pkg here...
23:06 < rxr> ok - i do
23:06 < rxr> (my four boxes eating energy need workload)
23:07 < jsaw> :)
23:08 < jsaw> oh, just checked cvs Changelog:
23:08 < jsaw> b* libmisc/pwdcheck.c:
23:08 < jsaw> one word fix .. ommited "else" was removed. Now security bug which allow
23:08 < jsaw> unauthorized account properties modification is realy closed.
23:10 < jsaw> let's wait for 4.0.6 if it's anyway not fixed ...
23:10 < rxr> ouhm
23:12 < jsaw> (or create the patch from cvs...)
23:14 < rxr> no, thanks
23:14 < rxr> we can wait a day or so for it ...
23:14 < jsaw> yep.
23:14 < rxr> 971 builds total, 922 completed fine, 49 with errors.
23:14 < rxr> if anyone wanna help fixing ...
23:14 < jsaw> is linux 2.4 needed in the bootdisk target?
23:14 < rxr> nope
23:15 < rxr> but we still have to twek the scripts to use one or both or so of the kernels and modules ...
23:15 < jsaw> == 11/05/04 23:13:36 =[1]=> Aborted building package linux24.
23:15 < rxr> historically the bootdisk used this default kernel thing, since it is gone a new scheme needs to be choosen
23:15 < rxr> aborted?
23:16 < mnemoc> my bootdisk built both kernels
23:16 < jsaw> rwsem-spinlock.h:57: error: previous declaration of 'init_rwsem' was here
23:16 < jsaw> rwsem-spinlock.c:36: error: conflicting types for 'init_rwsem'
23:16 < rxr> yes, that is ok and intended
23:16 * rxr did not got that ..
23:16 < rxr> jsaw: what are you optimizing for?
23:16 < mnemoc> jsaw: that happend with x86-generic here
23:17 < mnemoc> jsaw: x86-pmmx worked perfect
23:17 < jsaw> yes, x86-generic
23:17 < rxr> yep - I built athlon without error ;-)
23:17 < jsaw> <- has an old 386
23:17 < rxr> jsaw: coudl you fix this?
23:17 < jsaw> will try to...
23:18 < jsaw> oh damn, the FASTCALL thing
23:20 < rxr> the build errors over here for an athlon desktop build:
23:20 < rxr> ardour avview boost dvgrab efsd elicit entrance firefox gedit-plugins
23:20 < rxr> ggrab gkrellm-plugins glade gnomemeeting gnucash gsmp gtkhtml10 jamin
23:20 < rxr> kahakai kaxtv koffice ladspa-swh-plugins ladspa-vst libgnomedb linphone
23:20 < rxr> mergeant mod_mono mono-develop mplayer-fonts mysql-perl mythtv nvrec ocs
23:20 < rxr> perl-dbd-postgresql pixieplus qcad quota ripperx rivatv stlport
23:20 < rxr> streamtuner superkaramba transcode vdr vls vnc vstserver xmp zapping zgv
23:20 < mnemoc> jsaw: /me installing elks on a 286 now :p
23:20 < jsaw> mnemoc: nice!
23:22 < jsaw> bbl
23:34 < rxr> for whom does this happen regularly?:
23:34 < rxr> root:/usr/src/t2-trunk/src.qcad.1099693663.25973.3102287048# ./debug.sh
23:34 < rxr> debug-qcad:[src.qcad.1099693663.25973.3102287048]# cd qcad*
23:34 < rxr> ./debug.sh: line 3: 25830 Segmentation fault (core dumped) chroot "/usr/src/t2-trunk/build/default-2.1.0-beta-x86-athlon-32-desktop" /bin/bash ROCK/src.qcad.1099693663.25973.3102287048/debug_x.sh
23:38 < rxr> aside from that qcad fix on the way
23:38 * rxr showering
--- Log closed Sat Nov 06 00:00:37 2004