T2 IRC Log: 2004-11-28
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 Sun Nov 28 00:00:13 2004
00:03 < rxr> hi praenti
01:58 -!- sparc-kly [~sparc-kly@65-23-199-247.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
03:32 -!- sparc-kly [~sparc-kly@65-23-199-247.prtc.net] has joined #t2
03:36 -!- mtross [~michael@H99e6.h.pppool.de] has joined #t2
03:41 -!- mtr [~michael@H900c.h.pppool.de] has quit [Read error: 60 (Operation timed out)]
03:50 -!- _martin_ [~martin@brln-d9ba1e94.pool.mediaWays.net] has quit [Read error: 110 (Connection timed out)]
04:10 -!- _martin_ [~martin@brln-d9ba2709.pool.mediaWays.net] has joined #t2
05:01 -!- praenti [~praenti@mail.obster.org] has quit [Read error: 104 (Connection reset by peer)]
05:01 -!- praenti [~praenti@mail.obster.org] has joined #t2
08:52 -!- mtross [~michael@H99e6.h.pppool.de] has quit ["using sirc version 2.211+KSIRC/1.3.11"]
10:52 -!- rxr_ [~rene@p213.54.245.215.tisdip.tiscali.de] has joined #t2
10:52 -!- Topic for #t2: T2 | the system development environment | http://www.exactcode.de/t2
10:52 -!- Topic set by valentin [] [Sun Nov 7 16:41:27 2004]
10:52 [Users #t2]
10:52 [ _martin_] [ CIA-9 ] [ jsaw ] [ praenti] [ rxr_ ] [ valentin]
10:52 [ _Ragnar_] [ daja77] [ mnemoc] [ rxr ] [ sparc-kly]
10:52 -!- Irssi: #t2: Total of 11 nicks [0 ops, 0 halfops, 0 voices, 11 normal]
10:52 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
10:52 -!- [freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup
10:53 -!- Irssi: Join to #t2 was synced in 11 secs
11:01 -!- rxr [~rene@p213.54.211.224.tisdip.tiscali.de] has quit [Read error: 60 (Operation timed out)]
12:03 < _martin_> moin
12:10 < mnemoc> moin _martin_
12:43 < jsaw> re
12:43 < jsaw> 21:50 < mnemoc> where is jsaw when i need him? :(
12:43 < jsaw> ?
12:52 < valentin> moin
12:53 < valentin> spelling error: where is rxr when we need him ? ;)
12:53 < jsaw> :)
12:53 < jsaw> hi valentin
12:57 < valentin> btw: managed to understand the relevant parts of analytical mechanics - remaining question is: how to learn and remember it ?
12:57 < jsaw> good question...
13:24 < CIA-9> jsaw * r4870 /trunk/package/gnome2/evolution-webcal/evolution-webcal.desc:
13:24 < CIA-9> * roll back evolution-webcal (2.1.0 -> 2.0.1)
13:24 < CIA-9> (-> 2.1 is development version and won't
13:24 < CIA-9> fit with evolution-data-server/evolution)
13:34 -!- mtr [~michael@H99e6.h.pppool.de] has joined #t2
13:35 < mnemoc> re
13:35 < mtr> hi
13:35 < jsaw> hi mnemoc, mtr
13:35 < mnemoc> jsaw: i told that the last time mtr was here :)
13:36 < jsaw> mnemoc: yes, why?
13:37 < mnemoc> mtr discovered why glibc and uclibc get mixed when entering chroot. it's because fl_wrapper.so :p
13:38 < mnemoc> btw, hi mtr
13:38 < mtr> hi mnemoc
13:39 < mnemoc> mtr, did you build finished?
13:39 < jsaw> ah, ok. My 2cents on it are the following: link fl_wrapper.so statically against dietlibc, add an environment variable like "FLWRAPPER_LIBC" where the name of the libc is stored that is to be wrapped... (if NULL -> use /lib/libc.so)
13:41 < jsaw> btw, have you read the thread on ROCK ml concerning the varargs thing of open?
13:41 < mtr> mnemoc: no, still running, building kde pkgs
13:42 < mnemoc> jsaw: not yet :(
13:43 * mnemoc reading "flist wrapper patch"
13:44 < mtr> jsaw: mnemoc had basically the same idea yesterday. possibly an evironment variable would be a more flexible solution, mnemoc?
13:44 < jsaw> me somehow guesses, if args are passed in registers, "..." is still passed on stack, so probably, the implementation has to change (as reported for x86_64) ...?
13:45 < mnemoc> jsaw: highly probable
13:46 < jsaw> mtr,mnemoc: in contrast to glibc, dietlibc can be built statically, so are commands launched via a statically built command detected by fl_wrapper or not?
13:46 < jsaw> or file-io
13:46 < mnemoc> i don't think so
13:46 < jsaw> hmmm....
13:47 < jsaw> uclibc? is it dynamic by default, mtr?
13:47 < mtr> my understanding was, that the problem is not the linkage of fl_wrapper.so itself, rather then the method of resolving symbols
13:48 < mnemoc> but you said you missed entries using the other methid
13:49 < mtr> yes, the install_init funtion bypasses the uclibc-wrappers in any way
13:49 < mtr> basically i am missing the .init files, other seems to be ok
13:51 < mnemoc> i guess wil happen to every parse-config function
13:58 < mnemoc> i'll build a runit based minimal, which does extensive use of parse-config functions, against uclibc which your change to fl_wrapper
13:58 < mnemoc> s/which/with/
14:17 < mtr> i cannot locate the code for install_init, which does the .init processing, can someone help?
14:23 < mnemoc> sysvinit's parse-config
14:24 < mnemoc> mtr: sysvinit's parse-config
14:25 -!- You're now known as rxr
14:25 < rxr> moin
14:28 < mtr> hmm, install_init calls m4 and redirected stdout creates the .init file
14:28 < mtr> so the shells syscalls have to be intercepted, but the executing shell is the host systems shell using glibc, right?
14:29 < mtr> moin rxr
14:30 < rxr> damn - my home line is that saturated that a ssh into the world takes 20s or so ..
14:30 < rxr> I need to setup traffic shaping ...
14:30 < rxr> ouhm - latency is around 8s or so ...
14:37 -!- sparc-kly [~sparc-kly@65-23-199-247.prtc.net] has quit ["Leaving"]
14:40 < mnemoc> mtr: on chroot it's not :)
14:45 < jsaw> can somebody tell me the file size of the octave-forge package?
14:46 < rxr> binary or source
14:46 < rxr> ?
14:46 < rxr> I'll reboot my rs/6k soon ...
14:48 < jsaw> binary
14:48 < jsaw> usr/libexec/octave/2.1.59/site/oct/i586-unknown-linux-gnu/octave-forge/
14:48 < jsaw> contains 59 files of size
14:49 < jsaw> -> 39M <-
14:55 < rxr> 258 files on my iBook ...
14:55 < rxr> bu many symlink
14:55 < rxr> s
14:57 < rxr> .oo
14:57 < rxr> 594M total
14:57 < rxr> Package Size: 601.24 MB, 1059 files
14:59 -!- Sparc-kly|G4 [~sparc@65-23-199-247.prtc.net] has joined #t2
15:00 < jsaw> rxr: okay, so you seem to have the same problem.
15:01 < rxr> I seem to have a bigger problem .-(
15:01 < rxr> hm - if yaboot is able to load this kernel:
15:01 < rxr> 6516893 Nov 28 14:54 vmlinux_2.6.9-rock
15:01 < rxr> ^- .oO
15:01 < jsaw> unpacked, octave-forge is 3.5G
15:01 < jsaw> rxr: ohoh
15:01 < rxr> we ned a new name for the kernel images ;-)
15:02 < rxr> what about -precompiled ? or just -pc
15:02 < rxr> ?
15:02 < jsaw> -dist
15:03 < rxr> a - dist is nice ;-)
15:04 < rxr> I should do other things - and just because my brother saturates my home line I setup a 2.6 kernel in my rs6k just to get traffic shaping ...
15:04 < jsaw> hehe
15:06 < rxr> ok - cu - /me rebooting the router
15:06 < rxr> and hopefully it boots this fat kernel ...
--- Log closed Sun Nov 28 15:06:37 2004
--- Log opened Sun Nov 28 15:48:21 2004
15:49 < rxr> test log
15:49 < rxr> good ;-)
16:16 < CIA-9> rene * r4871 /trunk/architecture/share/kernel-common.conf:
16:16 < CIA-9> * space reduction for the default kernels: no included .config nor
16:16 < CIA-9> PCI name database (we have user-space tools for this anyway)
16:17 < CIA-9> rene * r4872 /trunk/package/base/linux24/ (linux24.conf lx_config.sh):
16:17 < CIA-9> * fixed linux* packages to install the various PPC kernels with
16:17 < CIA-9> version appended and changed the default kernel ending from -rock
16:17 < CIA-9> to -dist
16:27 < rxr> http://www.petitiononline.com/3dc4rdlb/petition.html
16:28 < mnemoc> *click
16:29 < rxr> http://lists.duskglow.com/mailman/listinfo/open-graphics
16:29 < rxr> http://developers.slashdot.org/developers/04/11/28/018227.shtml
16:36 < CIA-9> rene * r4873 /trunk/target/bootdisk/ (powerpc/build.sh sparc/build.sh x86/build.sh): * migrated bootdisk boot loader setup to new -dist kernel name ending
16:41 < CIA-9> jsaw * r4874 /trunk/package/scientific/octave/octave.conf:
16:41 < CIA-9> * create shared libs for octave
16:41 < CIA-9> (reduces size of octave-forge package from 800 to 15Mb!!!)
16:42 < rxr> ui!
16:42 < rxr> thanks!
16:46 < mnemoc> oh
16:47 -!- CorLis [~wehrlin@p3E9E6D1C.dip.t-dialin.net] has joined #t2
16:47 < CorLis> guggug
16:47 < jsaw> <- gone, cu l8r
16:47 < CorLis> laters jsaw
16:47 < rxr> cu jsaw
16:49 < CorLis> hrm. i see updates happen quite regularly...
16:49 < CorLis> May i ask how patches are chosen? or versions of apps?
16:49 < rxr> CorLis: yes - I hope this is good ;-)
16:49 < rxr> version are normally most up-to-date stable releases
16:50 < rxr> patches are only applied if needed for compile or security - normally no features or t2 specific hacks are included
16:50 < CorLis> good thing to have up-to-date-versions in a security-manner
16:51 < CorLis> I rather don't like patching a bleeding-edge app to another bleeding-more app, tho :P
16:52 < CorLis> so it's mostly patches to common, or better: known issues
16:52 < CorLis> like fixing a samba-security problem, ... (and the patch is something official)
16:52 < rxr> I prefer stable versions in t2 - however some packages are at -rc or -pre due to known issues in the former - or the maintainer just wanting it ...
16:53 < rxr> e.g. icewm is often at the latest -pre release just because the maintainer at rock submitted any version update possible ...
16:53 < CorLis> Ok, i understand that, if an app is still in a very evolving state
16:54 < mnemoc> or if it's almost releasing a new major version which will deprecate current stable
16:55 < mnemoc> e.g. mono
16:55 < mnemoc> or xfce
16:56 < CorLis> hrm. nod, i understand
16:57 < mnemoc> that's because they will be released before we tag stable :)
16:58 < mnemoc> or a few days after
16:59 < mnemoc> that allow us to update the stable, and not getting froze on a deprecated tree, without breaking stability
17:00 < CorLis> urm
17:00 < CorLis> Config config/default/config doesn't exist...
17:01 < mnemoc> that's created after you do a config
17:01 < mnemoc> Config*
17:02 < CorLis> nod, file exists...
17:02 < mnemoc> ./scripts/Config -cfg corlis ---> config/corlis/{config,packages}
17:03 < CorLis> hrm, k, didn't specify a cfg-name when configuring
17:03 < mnemoc> ./scripts/Config --> config/default/{config,packages}
17:04 < CIA-9> rene * r4875 /trunk/package/audio/jack/jack.conf: * enabled jacks startjack wrapper to gain realtime capabilities
17:06 < CorLis> Hrm. config/corlis/config doesn't exist...
17:06 < CorLis> but file exists...
17:06 < CorLis> doh
17:07 < mnemoc> o_O
17:07 < mnemoc> using what script?
17:07 < CorLis> ./scripts/Download -cfg config/corlis/config -required
17:08 < CorLis> ./scripts/Download -cfg config/corlis/config does a successful mirror-choosing, first
17:08 < mnemoc> ./scripts/Download -cfg corlis -required
17:08 < CorLis> ah, that working
17:08 < mnemoc> :)
17:08 < CorLis> hrm. someone gotta rewrite that help then! :P
17:09 < CorLis> is t2 somewhat based on lfs, btw?
17:09 < mnemoc> nope
17:09 < CorLis> or was it a totally new approach
17:10 < mnemoc> LFS is a book to build your own linux machine
17:10 < mnemoc> T2 is a kit to crate an maintain distributions
17:10 < CorLis> that's why i've said "somewhat" :P
17:11 < mnemoc> both build source :)
17:11 < CIA-9> rene * r4876 /trunk/package/audio/bio2jack/bio2jack.conf:
17:11 < CIA-9> * fixed bio2jack to actually install it's bio2jack.h ... (well - many
17:11 < CIA-9> packages include a private copy for this reasons - but I do not want
17:11 < CIA-9> to copy the file into any project ...)
17:12 < rxr> CorLis: no - t2 is not based or LFS - not inspiried or ao ..
17:12 < CorLis> thanks :P
17:12 < mnemoc> we are more based on 'man gcc' than on LFS :)
17:12 < rxr> e.g. LFS does build the initial toolchain statically linked - we do not - and the rest of LFS is just an instruction for normal package builds and dependencies list - nothing we used
17:14 < rxr> CorLis: which doc does list incorrect Config or Download instructions? I would be happy to fix it ;-)
17:16 < CorLis> Documentation/BUILD
17:16 < rxr> ouh - we should remove those ;_)
17:16 < rxr> CorLis: did you notice the handbook?
17:16 < CorLis> *cough* It's sundays :P
17:17 < CorLis> ok, gotta smoke a cig :P
17:17 < rxr> http://www.exactcode.de/products/t2-handbook/
17:21 < CorLis> yup, have it :P
17:21 < CorLis> and laaaaters
17:21 -!- CorLis [~wehrlin@p3E9E6D1C.dip.t-dialin.net] has quit ["leaving"]
17:31 < rxr> mnemoc: I'm just fixing the the bootdisk stuff ...
17:32 < mnemoc> :D
19:40 * rxr rebooting the rs6k a few times - cu
--- Log closed Sun Nov 28 19:40:41 2004
--- Log opened Sun Nov 28 21:00:20 2004
21:00 < rxr> test
21:01 < mnemoc> 0
21:01 < rxr> seems yaboot does not want to load kernels bigger than 4.x MB or so ..
21:01 < rxr> but now I have a 2.6.9 running on the rs6k instead of the previous quite ancient 2.4.21 ;-)
21:01 < rxr> never touch a running syssttem
21:01 < rxr> system even
21:02 < mnemoc> how much downtime?
21:02 < rxr> especially if it is a rs6k needing 4 minutes or so for the power on self test ...
21:02 < mnemoc> oh
21:02 < rxr> oh - quite much - it is just my home router - so I even ocationally power it off when I leave for longer ...
21:03 < rxr> yeah - but the selftest is cool - showing progress on a 4 element dot matrix LED display ;-)
21:03 < rxr> and when it passes you can be sure the box is ok ;-)
21:03 * mnemoc wants one on those
21:03 < mnemoc> of*
21:03 < rxr> not this poor mans x86 BIOS toys ;-)
21:03 < rxr> POST ok - but heavy memory failurs *grin*
21:04 < rxr> mnemoc: the rs6k has a graphical OpenFirmware - showing icons for found cpu and ram and such - as well as playing a fanfare.wav or so when it is up ,-)
21:05 < mnemoc> does it support remote admin in case of failure?
21:05 < rxr> "only" via serial console - as most OpenFirmware implementations you can assing the primary console to a serial port ..
21:12 < rxr> does anyone here do real traffic shaping in linux?
21:24 < rxr> hm - so far those seem to suck:
21:24 < rxr> Stochastic Fairness Queueing
21:24 < rxr> Token Bucket Filter
21:35 -!- sparc-kly [~sparc-kly@65-23-199-247.prtc.net] has joined #t2
21:36 < rxr> linux24: boot/System.map_2.4.28-dist
21:36 < rxr> linux24: boot/initrd-2.4.28-dist.img
21:36 < rxr> linux24: boot/kconfig_2.4.28-dist
21:36 < rxr> linux24: boot/vmlinuz_2.4.28-dist
21:36 < rxr> linux24: lib/modules/2.4.28-dist
21:36 < rxr> linux26: boot/System.map_2.6.9-dist
21:36 < rxr> linux26: boot/initrd-2.6.9-dist.img
21:36 < rxr> linux26: boot/kconfig_2.6.9-dist
21:36 < rxr> linux26: boot/vmlinuz_2.6.9-dist
21:36 < rxr> linux26: lib/modules/2.6.9-dist
21:36 < rxr> nist, isn't it?
21:37 < rxr> far better than -rock ;-)
21:50 < rxr> ok - this gives at least better than defaults results:
21:50 < rxr> tc qdisc add dev ppp0 root tbf rate 120kbit latency 70ms burst 2500
21:58 < rxr> FreeBSD 5.3-STABLE (GENERIC) #0: Fri Nov 5 18:16:52 CET 2004
21:58 < rxr> To do a fast search for a file, try
21:58 < rxr> locate filename
21:58 < rxr> locate uses a database that is updated every Saturday (assuming your computer
21:58 < rxr> is running FreeBSD at the time) to quickly find files based on name only.
21:58 < rxr> ;-)
22:06 < rxr> hey - cool - my old ROCK 1.7 based server has a higher thruput than the freebsd box:
22:06 < rxr> INFO: Testing ... ok 69860.000
22:06 < rxr> INFO: Testing ... ok 57887.000
22:06 < rxr> connected on the same witch at the same uni ...
22:27 < rxr> the gsmp box:
22:27 < rxr> 22:56:13 up 417 days, 11:05, 1 user, load average: 0.09, 0.07, 0.10
22:50 < rxr> hm - my ireland photo gallery is quite often accessed:
22:50 < rxr> 138 hits this month so far
23:07 < rxr> haha - 4 people got to http://www.exactcode.de by googling for "download suse 9.2 professional"
23:09 < valentin> ??
23:09 < rxr> yeah - quite cool, isn't it ;-)
23:09 < rxr> take a look at the stats for dl.exactcode - last line at the buttom I think
23:10 < rxr> ah - no - firt line in the top of the search keywords box :-)
23:11 < valentin> i cannot reproduce that with google
23:13 < rxr> from the stat page:
23:13 < rxr> download suse 9.2 professional
23:13 < rxr> 4
23:13 < rxr> 2.6 %
23:15 < rxr> hm - can't find us using the above neither ...
23:15 < valentin> strange
23:15 < rxr> mabye people clicked to page 56 in order to find a download location
23:15 < rxr> and there somewhere http://www.exactcode.de was due to suse linux and download beeing in the irc log ...
23:15 < rxr> who knows ...
23:16 < rxr> ok - let's take a look on my todo
23:16 < rxr> ah ja - bootdisk target diet issue ...
23:17 < rxr> valentin: my rs6k now runs 2.6.9 - and dump traffic shaping ;-)
23:18 < rxr> valentin: and have you seen the gsmp uptime? 417 days
23:20 < valentin> jep
23:20 < valentin> btw: lang == en
23:20 < valentin> :)
23:22 < rxr> where did I wrote lang != en ?
23:22 < rxr> the "ja" ?
23:23 < valentin> y
23:23 < valentin> just were afraid you mistook the channels
23:24 < rxr> nope
23:24 < rxr> damn - I not a bit motivated for the diet issue .. - but the sooner I track it the sooner I can do more fun things ..
23:26 < valentin> http://www.google.com/search?hl=en&lr=&ie=ISO-8859-1&q=download+suse+9.2+professional+exactcode&btnG=Search
23:27 < rxr> yep ;-)
23:27 < rxr> maybe I should download Fedora Core 3 or so later on and give it a try
23:28 < valentin> thanks sparc-kly :)
23:28 < sparc-kly> ? im not
23:28 < sparc-kly> lol
23:28 < rxr> and maybe we should download Yellowdog Linux and see if all the bugs of previous versions got fixed ;-)
23:28 < valentin> hehe
23:29 < rxr> and has anyone ever downloaded Mandrake Linux? ...
23:29 < rxr> I heared they are at Mandrake Linux 10.1 ;-)
23:30 < rxr> ok - enough - off to linuxrc ...
23:30 < sparc-kly> anyone run linux in NEC mobilepro 790 ?
23:30 < rxr> nope - but on a Psion Revo ;-) with just 16MB of Ram and 37.xyz Mhz ;-)
23:34 < valentin> rxr: btw - my xorg is still running without a crash - thanks !
23:35 < rxr> cool
23:35 < rxr> http://www.rocklinux.net/lurker/message/20041128.200039.4dc49259.html
23:35 < rxr> ^- is this relevant for us?
23:35 < valentin> for the google log: in case of xorg freezing with ATI Radeon Mobility try Option "DynamicClocks" "false"
23:36 < valentin> hm, do not thing we got that error ...
23:39 < rxr> nope - never saw this - maybe yet another funny regression rock folks hide in their make it overclicated code ...
23:40 < rxr> valentin: for the sake of google: If you have X.org 6.8.0 or 6.8.1 and your Radeon > 7500 freezes the machine randomly ...
23:40 < rxr> might be fixed in later versions of X.org - since it is already under investigation ...
23:41 < rxr> hm - should I simply make this change:
23:41 < rxr> http://www.rocklinux.net/lurker/message/20041126.161555.c881f3c2.html
23:41 < rxr> I think I already fixed -fPIC due to spart64 ...
23:41 < rxr> ah - wrong mail ID ...
23:42 < rxr> I meant to import this code change:
23:42 < rxr> http://www.rocklinux.net/lurker/message/20041126.161003.be9b1380.html
23:43 < valentin> why not ?
23:46 < rxr> ok - the patch is 99% noise ...
23:46 < rxr> I'll not even consider taking any further look into it ...
23:49 -!- mnemoc_ [~amery@200.75.27.58] has joined #t2
23:50 < rxr> hi mnemoc_
23:52 -!- mnemoc [~amery@200.75.27.36] has quit [Read error: 110 (Connection timed out)]
23:52 -!- mnemoc_ is now known as mnemoc
23:56 -!- CIA-9 [~CIA@to.je.spocco.com] has quit [Remote closed the connection]
--- Log closed Mon Nov 29 00:00:48 2004
00:03 < rxr> hi praenti
01:58 -!- sparc-kly [~sparc-kly@65-23-199-247.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
03:32 -!- sparc-kly [~sparc-kly@65-23-199-247.prtc.net] has joined #t2
03:36 -!- mtross [~michael@H99e6.h.pppool.de] has joined #t2
03:41 -!- mtr [~michael@H900c.h.pppool.de] has quit [Read error: 60 (Operation timed out)]
03:50 -!- _martin_ [~martin@brln-d9ba1e94.pool.mediaWays.net] has quit [Read error: 110 (Connection timed out)]
04:10 -!- _martin_ [~martin@brln-d9ba2709.pool.mediaWays.net] has joined #t2
05:01 -!- praenti [~praenti@mail.obster.org] has quit [Read error: 104 (Connection reset by peer)]
05:01 -!- praenti [~praenti@mail.obster.org] has joined #t2
08:52 -!- mtross [~michael@H99e6.h.pppool.de] has quit ["using sirc version 2.211+KSIRC/1.3.11"]
10:52 -!- rxr_ [~rene@p213.54.245.215.tisdip.tiscali.de] has joined #t2
10:52 -!- Topic for #t2: T2 | the system development environment | http://www.exactcode.de/t2
10:52 -!- Topic set by valentin [] [Sun Nov 7 16:41:27 2004]
10:52 [Users #t2]
10:52 [ _martin_] [ CIA-9 ] [ jsaw ] [ praenti] [ rxr_ ] [ valentin]
10:52 [ _Ragnar_] [ daja77] [ mnemoc] [ rxr ] [ sparc-kly]
10:52 -!- Irssi: #t2: Total of 11 nicks [0 ops, 0 halfops, 0 voices, 11 normal]
10:52 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
10:52 -!- [freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup
10:53 -!- Irssi: Join to #t2 was synced in 11 secs
11:01 -!- rxr [~rene@p213.54.211.224.tisdip.tiscali.de] has quit [Read error: 60 (Operation timed out)]
12:03 < _martin_> moin
12:10 < mnemoc> moin _martin_
12:43 < jsaw> re
12:43 < jsaw> 21:50 < mnemoc> where is jsaw when i need him? :(
12:43 < jsaw> ?
12:52 < valentin> moin
12:53 < valentin> spelling error: where is rxr when we need him ? ;)
12:53 < jsaw> :)
12:53 < jsaw> hi valentin
12:57 < valentin> btw: managed to understand the relevant parts of analytical mechanics - remaining question is: how to learn and remember it ?
12:57 < jsaw> good question...
13:24 < CIA-9> jsaw * r4870 /trunk/package/gnome2/evolution-webcal/evolution-webcal.desc:
13:24 < CIA-9> * roll back evolution-webcal (2.1.0 -> 2.0.1)
13:24 < CIA-9> (-> 2.1 is development version and won't
13:24 < CIA-9> fit with evolution-data-server/evolution)
13:34 -!- mtr [~michael@H99e6.h.pppool.de] has joined #t2
13:35 < mnemoc> re
13:35 < mtr> hi
13:35 < jsaw> hi mnemoc, mtr
13:35 < mnemoc> jsaw: i told that the last time mtr was here :)
13:36 < jsaw> mnemoc: yes, why?
13:37 < mnemoc> mtr discovered why glibc and uclibc get mixed when entering chroot. it's because fl_wrapper.so :p
13:38 < mnemoc> btw, hi mtr
13:38 < mtr> hi mnemoc
13:39 < mnemoc> mtr, did you build finished?
13:39 < jsaw> ah, ok. My 2cents on it are the following: link fl_wrapper.so statically against dietlibc, add an environment variable like "FLWRAPPER_LIBC" where the name of the libc is stored that is to be wrapped... (if NULL -> use /lib/libc.so)
13:41 < jsaw> btw, have you read the thread on ROCK ml concerning the varargs thing of open?
13:41 < mtr> mnemoc: no, still running, building kde pkgs
13:42 < mnemoc> jsaw: not yet :(
13:43 * mnemoc reading "flist wrapper patch"
13:44 < mtr> jsaw: mnemoc had basically the same idea yesterday. possibly an evironment variable would be a more flexible solution, mnemoc?
13:44 < jsaw> me somehow guesses, if args are passed in registers, "..." is still passed on stack, so probably, the implementation has to change (as reported for x86_64) ...?
13:45 < mnemoc> jsaw: highly probable
13:46 < jsaw> mtr,mnemoc: in contrast to glibc, dietlibc can be built statically, so are commands launched via a statically built command detected by fl_wrapper or not?
13:46 < jsaw> or file-io
13:46 < mnemoc> i don't think so
13:46 < jsaw> hmmm....
13:47 < jsaw> uclibc? is it dynamic by default, mtr?
13:47 < mtr> my understanding was, that the problem is not the linkage of fl_wrapper.so itself, rather then the method of resolving symbols
13:48 < mnemoc> but you said you missed entries using the other methid
13:49 < mtr> yes, the install_init funtion bypasses the uclibc-wrappers in any way
13:49 < mtr> basically i am missing the .init files, other seems to be ok
13:51 < mnemoc> i guess wil happen to every parse-config function
13:58 < mnemoc> i'll build a runit based minimal, which does extensive use of parse-config functions, against uclibc which your change to fl_wrapper
13:58 < mnemoc> s/which/with/
14:17 < mtr> i cannot locate the code for install_init, which does the .init processing, can someone help?
14:23 < mnemoc> sysvinit's parse-config
14:24 < mnemoc> mtr: sysvinit's parse-config
14:25 -!- You're now known as rxr
14:25 < rxr> moin
14:28 < mtr> hmm, install_init calls m4 and redirected stdout creates the .init file
14:28 < mtr> so the shells syscalls have to be intercepted, but the executing shell is the host systems shell using glibc, right?
14:29 < mtr> moin rxr
14:30 < rxr> damn - my home line is that saturated that a ssh into the world takes 20s or so ..
14:30 < rxr> I need to setup traffic shaping ...
14:30 < rxr> ouhm - latency is around 8s or so ...
14:37 -!- sparc-kly [~sparc-kly@65-23-199-247.prtc.net] has quit ["Leaving"]
14:40 < mnemoc> mtr: on chroot it's not :)
14:45 < jsaw> can somebody tell me the file size of the octave-forge package?
14:46 < rxr> binary or source
14:46 < rxr> ?
14:46 < rxr> I'll reboot my rs/6k soon ...
14:48 < jsaw> binary
14:48 < jsaw> usr/libexec/octave/2.1.59/site/oct/i586-unknown-linux-gnu/octave-forge/
14:48 < jsaw> contains 59 files of size
14:49 < jsaw> -> 39M <-
14:55 < rxr> 258 files on my iBook ...
14:55 < rxr> bu many symlink
14:55 < rxr> s
14:57 < rxr> .oo
14:57 < rxr> 594M total
14:57 < rxr> Package Size: 601.24 MB, 1059 files
14:59 -!- Sparc-kly|G4 [~sparc@65-23-199-247.prtc.net] has joined #t2
15:00 < jsaw> rxr: okay, so you seem to have the same problem.
15:01 < rxr> I seem to have a bigger problem .-(
15:01 < rxr> hm - if yaboot is able to load this kernel:
15:01 < rxr> 6516893 Nov 28 14:54 vmlinux_2.6.9-rock
15:01 < rxr> ^- .oO
15:01 < jsaw> unpacked, octave-forge is 3.5G
15:01 < jsaw> rxr: ohoh
15:01 < rxr> we ned a new name for the kernel images ;-)
15:02 < rxr> what about -precompiled ? or just -pc
15:02 < rxr> ?
15:02 < jsaw> -dist
15:03 < rxr> a - dist is nice ;-)
15:04 < rxr> I should do other things - and just because my brother saturates my home line I setup a 2.6 kernel in my rs6k just to get traffic shaping ...
15:04 < jsaw> hehe
15:06 < rxr> ok - cu - /me rebooting the router
15:06 < rxr> and hopefully it boots this fat kernel ...
--- Log closed Sun Nov 28 15:06:37 2004
--- Log opened Sun Nov 28 15:48:21 2004
15:49 < rxr> test log
15:49 < rxr> good ;-)
16:16 < CIA-9> rene * r4871 /trunk/architecture/share/kernel-common.conf:
16:16 < CIA-9> * space reduction for the default kernels: no included .config nor
16:16 < CIA-9> PCI name database (we have user-space tools for this anyway)
16:17 < CIA-9> rene * r4872 /trunk/package/base/linux24/ (linux24.conf lx_config.sh):
16:17 < CIA-9> * fixed linux* packages to install the various PPC kernels with
16:17 < CIA-9> version appended and changed the default kernel ending from -rock
16:17 < CIA-9> to -dist
16:27 < rxr> http://www.petitiononline.com/3dc4rdlb/petition.html
16:28 < mnemoc> *click
16:29 < rxr> http://lists.duskglow.com/mailman/listinfo/open-graphics
16:29 < rxr> http://developers.slashdot.org/developers/04/11/28/018227.shtml
16:36 < CIA-9> rene * r4873 /trunk/target/bootdisk/ (powerpc/build.sh sparc/build.sh x86/build.sh): * migrated bootdisk boot loader setup to new -dist kernel name ending
16:41 < CIA-9> jsaw * r4874 /trunk/package/scientific/octave/octave.conf:
16:41 < CIA-9> * create shared libs for octave
16:41 < CIA-9> (reduces size of octave-forge package from 800 to 15Mb!!!)
16:42 < rxr> ui!
16:42 < rxr> thanks!
16:46 < mnemoc> oh
16:47 -!- CorLis [~wehrlin@p3E9E6D1C.dip.t-dialin.net] has joined #t2
16:47 < CorLis> guggug
16:47 < jsaw> <- gone, cu l8r
16:47 < CorLis> laters jsaw
16:47 < rxr> cu jsaw
16:49 < CorLis> hrm. i see updates happen quite regularly...
16:49 < CorLis> May i ask how patches are chosen? or versions of apps?
16:49 < rxr> CorLis: yes - I hope this is good ;-)
16:49 < rxr> version are normally most up-to-date stable releases
16:50 < rxr> patches are only applied if needed for compile or security - normally no features or t2 specific hacks are included
16:50 < CorLis> good thing to have up-to-date-versions in a security-manner
16:51 < CorLis> I rather don't like patching a bleeding-edge app to another bleeding-more app, tho :P
16:52 < CorLis> so it's mostly patches to common, or better: known issues
16:52 < CorLis> like fixing a samba-security problem, ... (and the patch is something official)
16:52 < rxr> I prefer stable versions in t2 - however some packages are at -rc or -pre due to known issues in the former - or the maintainer just wanting it ...
16:53 < rxr> e.g. icewm is often at the latest -pre release just because the maintainer at rock submitted any version update possible ...
16:53 < CorLis> Ok, i understand that, if an app is still in a very evolving state
16:54 < mnemoc> or if it's almost releasing a new major version which will deprecate current stable
16:55 < mnemoc> e.g. mono
16:55 < mnemoc> or xfce
16:56 < CorLis> hrm. nod, i understand
16:57 < mnemoc> that's because they will be released before we tag stable :)
16:58 < mnemoc> or a few days after
16:59 < mnemoc> that allow us to update the stable, and not getting froze on a deprecated tree, without breaking stability
17:00 < CorLis> urm
17:00 < CorLis> Config config/default/config doesn't exist...
17:01 < mnemoc> that's created after you do a config
17:01 < mnemoc> Config*
17:02 < CorLis> nod, file exists...
17:02 < mnemoc> ./scripts/Config -cfg corlis ---> config/corlis/{config,packages}
17:03 < CorLis> hrm, k, didn't specify a cfg-name when configuring
17:03 < mnemoc> ./scripts/Config --> config/default/{config,packages}
17:04 < CIA-9> rene * r4875 /trunk/package/audio/jack/jack.conf: * enabled jacks startjack wrapper to gain realtime capabilities
17:06 < CorLis> Hrm. config/corlis/config doesn't exist...
17:06 < CorLis> but file exists...
17:06 < CorLis> doh
17:07 < mnemoc> o_O
17:07 < mnemoc> using what script?
17:07 < CorLis> ./scripts/Download -cfg config/corlis/config -required
17:08 < CorLis> ./scripts/Download -cfg config/corlis/config does a successful mirror-choosing, first
17:08 < mnemoc> ./scripts/Download -cfg corlis -required
17:08 < CorLis> ah, that working
17:08 < mnemoc> :)
17:08 < CorLis> hrm. someone gotta rewrite that help then! :P
17:09 < CorLis> is t2 somewhat based on lfs, btw?
17:09 < mnemoc> nope
17:09 < CorLis> or was it a totally new approach
17:10 < mnemoc> LFS is a book to build your own linux machine
17:10 < mnemoc> T2 is a kit to crate an maintain distributions
17:10 < CorLis> that's why i've said "somewhat" :P
17:11 < mnemoc> both build source :)
17:11 < CIA-9> rene * r4876 /trunk/package/audio/bio2jack/bio2jack.conf:
17:11 < CIA-9> * fixed bio2jack to actually install it's bio2jack.h ... (well - many
17:11 < CIA-9> packages include a private copy for this reasons - but I do not want
17:11 < CIA-9> to copy the file into any project ...)
17:12 < rxr> CorLis: no - t2 is not based or LFS - not inspiried or ao ..
17:12 < CorLis> thanks :P
17:12 < mnemoc> we are more based on 'man gcc' than on LFS :)
17:12 < rxr> e.g. LFS does build the initial toolchain statically linked - we do not - and the rest of LFS is just an instruction for normal package builds and dependencies list - nothing we used
17:14 < rxr> CorLis: which doc does list incorrect Config or Download instructions? I would be happy to fix it ;-)
17:16 < CorLis> Documentation/BUILD
17:16 < rxr> ouh - we should remove those ;_)
17:16 < rxr> CorLis: did you notice the handbook?
17:16 < CorLis> *cough* It's sundays :P
17:17 < CorLis> ok, gotta smoke a cig :P
17:17 < rxr> http://www.exactcode.de/products/t2-handbook/
17:21 < CorLis> yup, have it :P
17:21 < CorLis> and laaaaters
17:21 -!- CorLis [~wehrlin@p3E9E6D1C.dip.t-dialin.net] has quit ["leaving"]
17:31 < rxr> mnemoc: I'm just fixing the the bootdisk stuff ...
17:32 < mnemoc> :D
19:40 * rxr rebooting the rs6k a few times - cu
--- Log closed Sun Nov 28 19:40:41 2004
--- Log opened Sun Nov 28 21:00:20 2004
21:00 < rxr> test
21:01 < mnemoc> 0
21:01 < rxr> seems yaboot does not want to load kernels bigger than 4.x MB or so ..
21:01 < rxr> but now I have a 2.6.9 running on the rs6k instead of the previous quite ancient 2.4.21 ;-)
21:01 < rxr> never touch a running syssttem
21:01 < rxr> system even
21:02 < mnemoc> how much downtime?
21:02 < rxr> especially if it is a rs6k needing 4 minutes or so for the power on self test ...
21:02 < mnemoc> oh
21:02 < rxr> oh - quite much - it is just my home router - so I even ocationally power it off when I leave for longer ...
21:03 < rxr> yeah - but the selftest is cool - showing progress on a 4 element dot matrix LED display ;-)
21:03 < rxr> and when it passes you can be sure the box is ok ;-)
21:03 * mnemoc wants one on those
21:03 < mnemoc> of*
21:03 < rxr> not this poor mans x86 BIOS toys ;-)
21:03 < rxr> POST ok - but heavy memory failurs *grin*
21:04 < rxr> mnemoc: the rs6k has a graphical OpenFirmware - showing icons for found cpu and ram and such - as well as playing a fanfare.wav or so when it is up ,-)
21:05 < mnemoc> does it support remote admin in case of failure?
21:05 < rxr> "only" via serial console - as most OpenFirmware implementations you can assing the primary console to a serial port ..
21:12 < rxr> does anyone here do real traffic shaping in linux?
21:24 < rxr> hm - so far those seem to suck:
21:24 < rxr> Stochastic Fairness Queueing
21:24 < rxr> Token Bucket Filter
21:35 -!- sparc-kly [~sparc-kly@65-23-199-247.prtc.net] has joined #t2
21:36 < rxr> linux24: boot/System.map_2.4.28-dist
21:36 < rxr> linux24: boot/initrd-2.4.28-dist.img
21:36 < rxr> linux24: boot/kconfig_2.4.28-dist
21:36 < rxr> linux24: boot/vmlinuz_2.4.28-dist
21:36 < rxr> linux24: lib/modules/2.4.28-dist
21:36 < rxr> linux26: boot/System.map_2.6.9-dist
21:36 < rxr> linux26: boot/initrd-2.6.9-dist.img
21:36 < rxr> linux26: boot/kconfig_2.6.9-dist
21:36 < rxr> linux26: boot/vmlinuz_2.6.9-dist
21:36 < rxr> linux26: lib/modules/2.6.9-dist
21:36 < rxr> nist, isn't it?
21:37 < rxr> far better than -rock ;-)
21:50 < rxr> ok - this gives at least better than defaults results:
21:50 < rxr> tc qdisc add dev ppp0 root tbf rate 120kbit latency 70ms burst 2500
21:58 < rxr> FreeBSD 5.3-STABLE (GENERIC) #0: Fri Nov 5 18:16:52 CET 2004
21:58 < rxr> To do a fast search for a file, try
21:58 < rxr> locate filename
21:58 < rxr> locate uses a database that is updated every Saturday (assuming your computer
21:58 < rxr> is running FreeBSD at the time) to quickly find files based on name only.
21:58 < rxr> ;-)
22:06 < rxr> hey - cool - my old ROCK 1.7 based server has a higher thruput than the freebsd box:
22:06 < rxr> INFO: Testing
22:06 < rxr> INFO: Testing
22:06 < rxr> connected on the same witch at the same uni ...
22:27 < rxr> the gsmp box:
22:27 < rxr> 22:56:13 up 417 days, 11:05, 1 user, load average: 0.09, 0.07, 0.10
22:50 < rxr> hm - my ireland photo gallery is quite often accessed:
22:50 < rxr> 138 hits this month so far
23:07 < rxr> haha - 4 people got to http://www.exactcode.de by googling for "download suse 9.2 professional"
23:09 < valentin> ??
23:09 < rxr> yeah - quite cool, isn't it ;-)
23:09 < rxr> take a look at the stats for dl.exactcode - last line at the buttom I think
23:10 < rxr> ah - no - firt line in the top of the search keywords box :-)
23:11 < valentin> i cannot reproduce that with google
23:13 < rxr> from the stat page:
23:13 < rxr> download suse 9.2 professional
23:13 < rxr> 4
23:13 < rxr> 2.6 %
23:15 < rxr> hm - can't find us using the above neither ...
23:15 < valentin> strange
23:15 < rxr> mabye people clicked to page 56 in order to find a download location
23:15 < rxr> and there somewhere http://www.exactcode.de was due to suse linux and download beeing in the irc log ...
23:15 < rxr> who knows ...
23:16 < rxr> ok - let's take a look on my todo
23:16 < rxr> ah ja - bootdisk target diet issue ...
23:17 < rxr> valentin: my rs6k now runs 2.6.9 - and dump traffic shaping ;-)
23:18 < rxr> valentin: and have you seen the gsmp uptime? 417 days
23:20 < valentin> jep
23:20 < valentin> btw: lang == en
23:20 < valentin> :)
23:22 < rxr> where did I wrote lang != en ?
23:22 < rxr> the "ja" ?
23:23 < valentin> y
23:23 < valentin> just were afraid you mistook the channels
23:24 < rxr> nope
23:24 < rxr> damn - I not a bit motivated for the diet issue .. - but the sooner I track it the sooner I can do more fun things ..
23:26 < valentin> http://www.google.com/search?hl=en&lr=&ie=ISO-8859-1&q=download+suse+9.2+professional+exactcode&btnG=Search
23:27 < rxr> yep ;-)
23:27 < rxr> maybe I should download Fedora Core 3 or so later on and give it a try
23:28 < valentin> thanks sparc-kly :)
23:28 < sparc-kly> ? im not
23:28 < sparc-kly> lol
23:28 < rxr> and maybe we should download Yellowdog Linux and see if all the bugs of previous versions got fixed ;-)
23:28 < valentin> hehe
23:29 < rxr> and has anyone ever downloaded Mandrake Linux? ...
23:29 < rxr> I heared they are at Mandrake Linux 10.1 ;-)
23:30 < rxr> ok - enough - off to linuxrc ...
23:30 < sparc-kly> anyone run linux in NEC mobilepro 790 ?
23:30 < rxr> nope - but on a Psion Revo ;-) with just 16MB of Ram and 37.xyz Mhz ;-)
23:34 < valentin> rxr: btw - my xorg is still running without a crash - thanks !
23:35 < rxr> cool
23:35 < rxr> http://www.rocklinux.net/lurker/message/20041128.200039.4dc49259.html
23:35 < rxr> ^- is this relevant for us?
23:35 < valentin> for the google log: in case of xorg freezing with ATI Radeon Mobility try Option "DynamicClocks" "false"
23:36 < valentin> hm, do not thing we got that error ...
23:39 < rxr> nope - never saw this - maybe yet another funny regression rock folks hide in their make it overclicated code ...
23:40 < rxr> valentin: for the sake of google: If you have X.org 6.8.0 or 6.8.1 and your Radeon > 7500 freezes the machine randomly ...
23:40 < rxr> might be fixed in later versions of X.org - since it is already under investigation ...
23:41 < rxr> hm - should I simply make this change:
23:41 < rxr> http://www.rocklinux.net/lurker/message/20041126.161555.c881f3c2.html
23:41 < rxr> I think I already fixed -fPIC due to spart64 ...
23:41 < rxr> ah - wrong mail ID ...
23:42 < rxr> I meant to import this code change:
23:42 < rxr> http://www.rocklinux.net/lurker/message/20041126.161003.be9b1380.html
23:43 < valentin> why not ?
23:46 < rxr> ok - the patch is 99% noise ...
23:46 < rxr> I'll not even consider taking any further look into it ...
23:49 -!- mnemoc_ [~amery@200.75.27.58] has joined #t2
23:50 < rxr> hi mnemoc_
23:52 -!- mnemoc [~amery@200.75.27.36] has quit [Read error: 110 (Connection timed out)]
23:52 -!- mnemoc_ is now known as mnemoc
23:56 -!- CIA-9 [~CIA@to.je.spocco.com] has quit [Remote closed the connection]
--- Log closed Mon Nov 29 00:00:48 2004