--- Log opened Fri Jan 21 00:00:13 2005
00:18 < _Ragnar_> *grrr8
00:18 < _Ragnar_> == 22:14:44 =[5]=> Building base/uclibc [0.9.27 2.1.0-beta3].
00:18 < _Ragnar_> ...
00:18 < _Ragnar_> -> Compiler Cache Hits while pkg build: 0.58% (6 hits, 1031 misses)
00:18 < _Ragnar_> /bin/bash: Can't resolve symbol '__moddi3'
00:33 < rxr> hi
00:35 < _Ragnar_> hi rxr
00:38 < rxr> 2cu later
00:38 < _Ragnar_> cya
00:41 < CIA-7> rene * r5741 /trunk/package/base/ncurses/ncurses.conf:
00:41 < CIA-7> * improved ncurses to build with widechar support (contain some typo
00:41 < CIA-7> but I wuickly want to get it onto my iBook - sorry)
00:41 < _Ragnar_> *lol*
00:42 < rxr> so
00:42 * rxr leaving this public wlan and jumping in a bus in some seconds ...
00:43 < rxr> cu all - maybe - a bit later
00:49 < _Ragnar_> laters ;)
00:54 < mnemoc> re
00:55 < _Ragnar_> re
00:55 < mnemoc> _Ragnar_: i got that on 5-uclibc of a mnemosyne-minimal but not on uboot... i don't know what now dependecy is breaking bash, like as ncurses did
00:56 < _Ragnar_> I'd assume libm
00:57 < _Ragnar_> __moddi3 looks like math related
01:00 < mnemoc> why may it be part of 1-uclibc but not part of 5-uclibc
01:03 < _Ragnar_> no clue :/
01:03 < _Ragnar_> libm is still there tho
01:04 < mnemoc> readelf ?
01:05 < _Ragnar_> have to rebuild 1-uclibc first
01:08 < CIA-7> mnemoc * r5742 /trunk/package/base/ncurses/ncurses.conf: * fixed typo on ncurses.conf. (and why prefix='/' and not prefix= ?)
01:08 < mnemoc> Build-Target -job 1-uclibc does the trick
01:08 < _Ragnar_> currently building 5-uclibc
01:19 < mnemoc> no need to rebuild bash?
01:23 < _Ragnar_> n
01:23 < _Ragnar_> no*
01:25 < mnemoc> who owned __moddi3 on 1-uclibc ?
01:26 < _Ragnar_> how do I check that?
01:27 < mnemoc> readelf -s /lib/libm.so.0 | grep __moddi3
01:27 < mnemoc> readelf -s lib/libm.so.0 | grep __moddi3
01:29 < mnemoc> or something like: # for x in /lib/*.so.0; do echo $x; readelf -s $x | grep __moddi3; done
01:29 < _Ragnar_> libm is binary cmp the same before/after
01:29 < mnemoc> /lib/libc.so.6
01:29 < mnemoc> 713: 000156cd 101 FUNC GLOBAL DEFAULT 11 __moddi3@GLIBC_2.0
01:29 < _Ragnar_> libc.so.0 1186: 0003d510 567 FUNC GLOBAL DEFAULT 7 __moddi3
01:29 < _Ragnar_> (before)
01:30 < _Ragnar_> none after
01:30 < mnemoc> interesting
01:30 < _Ragnar_> so for some reason moddi3 doesn't get compiled into uclibc on pass 5
01:31 < mnemoc> moddi3 doesn't exist on uclibc source!
01:34 < _Ragnar_> then how can it be defined in 1-uclibc?
01:34 < mnemoc> 1-uclibc is cross compiled
01:34 < mnemoc> 5-uclibc is natively
01:34 < _Ragnar_> oh, yeah
01:35 < mnemoc> my question now is why bash depends on that foreign symbol
01:35 < _Ragnar_> lets build 1-bash statically ;)
01:35 < mnemoc> :)
01:36 < mnemoc> you can send a mail to t2ml doing the sugestion ;)
01:36 < _Ragnar_> ?
01:36 < mnemoc> ignore me
01:37 < _Ragnar_> no, you're way too helpful for that :)
01:37 < _Ragnar_> btw is there a way to compile some packages statically by default (without changing their .conf)
01:38 < mnemoc> with your target you can inject code to package builds
01:39 < mnemoc> including corresponding GCC_WRAPPER_APPEND and GCC_WRAPPER_REMOVE
01:39 < _Ragnar_> hmmm
01:39 < mnemoc> the code on GCC{,3}_WRAPPER_APPEND envvar is added to gcc command line
01:39 < mnemoc> and patterns on GCC_WRAPPER_REMOVE are removed from it
01:40 < mnemoc> look at cmd_wrapper.log
01:40 < _Ragnar_> how would I do this just for certain packages tho?
01:41 < mnemoc> create a target and add a pkg_$pkg.conf file there
01:41 < _Ragnar_> ah;)
01:41 < mnemoc> that file will _replace_ $confdir/$pkg.conf on buildtime
01:41 < _Ragnar_> ooh
01:42 < _Ragnar_> that sounds like a nightmare to keep up to date
01:42 < mnemoc> of course you can source $confdir/$pkg.conf on target/$target/pkg_$pkg.conf
01:42 < _Ragnar_> oh ... okay ... duh
01:42 * _Ragnar_ hits head on table
01:43 < mnemoc> take a look into target/mnemosyne/pkg_openssh.conf
01:44 < mnemoc> sourcing you can 'choose' if you add code before or after $pkg.conf
01:45 < _Ragnar_> btw ... Create statically linked binaries <-- this is misleading
01:45 < _Ragnar_> it does much more then just create statically linked binaries ... it also does not generate any dynamic ones (or tries not to)
01:46 < mnemoc> a side efect ;)
01:46 < mnemoc> you can always mail a proposal to t2ml :p
01:47 < _Ragnar_> *hrhr*
01:47 < mnemoc> brb
01:47 < mnemoc> :)
01:48 < _Ragnar_> I think I'll be trying myself on an own target first
01:50 < _Ragnar_> ldd build/a32-2.1.0-beta3-x86-athlon-xp-32-generic-expert/bin/bash not a dynamic executable
02:02 < _Ragnar_> mnemoc: okay compiling bash statically took care of the immediate problem
02:09 < _Ragnar_> wth?
02:09 < _Ragnar_> CC kernel/power/swsusp.o
02:09 < _Ragnar_> kernel/power/swsusp.c: In function `init_header':
02:09 < _Ragnar_> kernel/power/swsusp.c:331: error: parse error before ';' token
02:09 < _Ragnar_> kernel/power/swsusp.c: In function `sanity_check':
02:09 < _Ragnar_> kernel/power/swsusp.c:1084: error: parse error before ')' token
02:09 < _Ragnar_> 5-linux26 .... this gets buggier by the minute
02:16 < _Ragnar_> ah same error ... expr fails due to the unknown symbol
02:16 -!- ideal [~idealm@211.100.227.107] has joined #t2
02:16 < _Ragnar_> mnemoc: I thought you had that going for your system? what are you doing differently?
02:18 < mnemoc> re
02:19 < _Ragnar_> re
02:19 < mnemoc> try to build an uboot, that built perfectly here
02:19 < mnemoc> a very basic package selection
02:20 < mnemoc> that error must be a missing header, on the begining of the error
02:20 < _Ragnar_> but how can the package selection make a symbol suddenly appear in a lib?
02:20 < mnemoc> good question
02:21 < _Ragnar_> which? the linux26? no, it's an expr failing due to the same missing symbol
02:21 < mnemoc> oh
02:21 < mnemoc> i'll start a uboot build and see
02:21 < mnemoc> fresh build over trunk
02:25 < _Ragnar_> nm /usr/lib/gcc-lib/i386-pc-linux-gnu/3.2.3/libgcc.a | grep moddi3
02:25 < _Ragnar_> _moddi3.oS:
02:25 < _Ragnar_> 00000000 T __moddi3
02:25 < _Ragnar_> the symbol comes from gcc ;p
02:27 < mnemoc> uhm
02:34 < _Ragnar_> any idea?
02:37 < mnemoc> no
02:38 < mnemoc> do you have NLS enabled?
02:38 < mnemoc> that may be a difference
02:39 < mnemoc> the build that failed here had NLS enabled, and the one that worked, disabled
02:39 < _Ragnar_> where's the switch?
02:39 < _Ragnar_> anyway, diff between 1-uclibc and 5-uclibc
02:39 < _Ragnar_> < i686-unknown-linux-uclibc-ld -z now -Bsymbolic -shared --warn-common --export-dynamic --sort-common -z combreloc --discard-locals --discard-all --no-undefined -s -e _dl_boot -s
02:39 < _Ragnar_> oname=ld-uClibc.so.0 \
02:39 < _Ragnar_> < -o ld-uClibc-0.9.27.so i386/resolve.o ldso.o /export/t2-trunk/build/a32-2.1.0-beta3-x86-athlon-xp-32-generic-expert/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/i686-unknown-linux-u
02:40 < _Ragnar_> clibc/3.4.3/libgcc.a;
02:40 < _Ragnar_> ---
02:40 < _Ragnar_> > ld -z now -Bsymbolic -shared --warn-common --export-dynamic --sort-common -z combreloc --discard-locals --discard-all --no-undefined -s -e _dl_boot -soname=ld-uClibc.so.0 \
02:40 < _Ragnar_> > -o ld-uClibc-0.9.27.so i386/resolve.o ldso.o /usr/lib/gcc/i686-unknown-linux-uclibc/3.4.3/libgcc.a;
02:40 < _Ragnar_> and /usr/lib/gcc/i686-unknown-linux-uclibc/3.4.3/libgcc.a does contain the symbol
02:40 < _Ragnar_> so why doesn't it get bound into the .so ?
02:42 < mnemoc> no clue
02:42 < mnemoc> is that too hard to patch on Makefile?
02:43 < _Ragnar_> huh
02:43 < _Ragnar_> ?
02:43 < _Ragnar_> if I knew what to patch ...
02:44 < mnemoc> the line that produces ld-uClibc-0.9.27.so to -lgcc instead of libgcc.a
02:47 < _Ragnar_> how would that help?
02:48 < mnemoc> but i think the way to go is why 0-gcc's libgcc.a has that symbol
02:48 < _Ragnar_> nono, both libgcc.a have the symbol
02:49 < mnemoc> oh
02:49 < _Ragnar_> just for the first case, the symbol gets linked into the ld-uClibc-0.9.27.so, but in the second case (on stage 5) it doesn't
02:54 < mnemoc> do you have the files to see what .o adds the dependency?
02:59 -!- rxr_ [~rene@p213.54.194.221.tisdip.tiscali.de] has joined #t2
02:59 -!- Topic for #t2: T2 | 2.1.0-beta2 RELEASED | The next generation of System Development Enviroments (SDE) | http://www.t2-project.org/ | CIA, don't ever leave us again!
02:59 -!- Topic set by mnemoc [] [Wed Jan 12 03:04:43 2005]
02:59 [Users #t2]
02:59 [ _Ragnar_] [ ideal ] [ nzg ] [ rxr ] [ valentin]
02:59 [ CIA-7 ] [ jsaw ] [ Postal ] [ rxr_ ]
02:59 [ daja77 ] [ mnemoc] [ praenti] [ sparc-kly]
02:59 -!- Irssi: #t2: Total of 13 nicks [0 ops, 0 halfops, 0 voices, 13 normal]
02:59 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
02:59 -!- [freenode-info] why register and identify? your IRC nick is how people know you. http://freenode.net/faq.shtml#nicksetup
03:00 -!- Irssi: Join to #t2 was synced in 11 secs
03:00 < Postal> hey rxr
03:00 < mnemoc> Postal: daily disconection
03:01 < Postal> ah
03:01 < Postal> mnemoc, what is afaik?
03:01 < mnemoc> that's why you see two rene-s
03:01 < mnemoc> as far as i know
03:01 < Postal> ah
03:04 -!- rxr [~rene@p83.129.4.226.tisdip.tiscali.de] has quit [Read error: 60 (Operation timed out)]
03:07 < _Ragnar_> mnemoc: is there a way to tell the linker to be very verbose about which symbols it gets and which it throws away?
03:09 < mnemoc> that's not my field... i can try to get the answer on #uclibc, but they use to answer only gentoo and buildroot answers
03:09 < _Ragnar_> great ;p
03:10 < _Ragnar_> 'okay, we don't help you to get a whole system with our lib installed, since it doesn't work anyway' ?:)
03:13 < mnemoc> the linker doesnt pull in symbols, it pulls in libraries
03:13 < mnemoc> SpanKY: .a files are merged complete?
03:13 < mnemoc> mnemoc: yes
03:13 < mnemoc> o_O
03:14 -!- sparc-kly [~sparc-kly@65-23-206-237.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
03:17 < _Ragnar_> okay update:
03:18 < _Ragnar_> if I cp build/a32-2.1.0-beta3-x86-athlon-xp-32-generic-expert/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/i686-unknown-linux-uclibc/3.4.3/libgcc.a to /usr/lib/gcc/i686-unknown-linux-uclibc/3.4.3/libgcc.a in the build root, uClibc _does_ contain the symbol
03:19 < _Ragnar_> (after rebuilding 5-uclibc)
03:19 < mnemoc> but native libgcc.a had __moddi3 right?
03:19 < _Ragnar_> yes
03:20 < mnemoc> considering our problem with ssp, i don't really trust our toolchain-gcc
03:21 < _Ragnar_> :/
03:22 < _Ragnar_> but the toolchain-libgcc seems to work whereas the native one didn't
03:22 < _Ragnar_> btw: diff -u toolchain/libgcc.a native/libgcc.a:
03:22 < _Ragnar_>
03:22 < _Ragnar_> -_moddi3.o:
03:22 < _Ragnar_> +_moddi3.oS: 00000000 T __moddi3
03:23 < mnemoc> i think the native was the right behaviour
03:23 < mnemoc> uhm?
03:23 < _Ragnar_> (and don't ask me why it cuts out newlines sometime)
03:23 < _Ragnar_> I think not ... this symbol has to come from somewhere after all
03:24 < rxr_> re
03:24 < mnemoc> wb rxr_!
03:24 < rxr_> oh - someone touched the ncurse.conf
03:24 < rxr_> sorry
03:24 < mnemoc> :)
03:24 < rxr_> wanted to have it on my iBook in linux for my bus ride ...
03:24 * rxr_ showering
03:25 < mnemoc> rxr_: why did you prefix='/'?
03:25 < rxr_> we got a bit cold waiting for various public transports ...
03:25 < rxr_> for fun
03:25 < rxr_> was not needed ...
03:25 -!- You're now known as rxr
03:25 < Postal> Created file outside basedir: /usr/lib/perl5/site_perl/5 ..
03:25 < Postal> !> Created file outside basedir: /usr/lib/perl5/site_perl/5 ..
03:25 < Postal> !> Created file outside basedir: /usr/lib/perl5/site_perl/5 ..
03:25 < Postal> !> Created file outside basedir: /usr/lib/perl5/site_perl/5 ..
03:25 < Postal> !> Created file outside basedir: /usr/lib/perl5/site_perl/5 ..
03:25 < Postal> !> base #1: /mnt/t2/t2-trunk
03:25 < Postal> !> base #2: /mnt/t2/t2-trunk/build/system-2.1.0-beta3-x86-a ..
03:25 < mnemoc> i thought you public transport system was always 'on time'
03:25 < Postal> -> $root/var/adm/logs/1-perl.out -> 1-perl.err
03:26 < rxr> mnemoc: *lol*
03:26 < mnemoc> rxr: does 1-perl build somewhere?
03:27 < rxr> on my ultrasparc
03:27 < CIA-7> rene * r5743 /trunk/package/base/ncurses/ncurses.conf:
03:27 < CIA-7> * adapted the include pathes the way other packages expect and fedora
03:27 < CIA-7> core sets up - I hope it causes no regressions - and thanks for
03:27 < CIA-7> fixing the stray char before ,-)
03:28 < rxr> but I have not review what it did :-(
03:28 < rxr> can do so now ..
03:28 < rxr> after showring
03:28 < rxr> with the ncurses tweaking aspell now builds with unicode support
03:28 < rxr> so - back soon
03:28 < mnemoc> cu rxr
03:28 < Postal> any ideas, cause the pearl error is halting the build cause its a stage 1 package
03:29 < mnemoc> Postal: vi package/perl/perl/perl.desc and change [P] -1-... for [P] ---...
03:29 < Postal> change it to 5?
03:30 < mnemoc> change the 1 with a -
03:31 < mnemoc> rxr: includedir=$includedir/ncursesw" <-- you need a new keyboard ;)
03:31 < _Ragnar_> *g*
03:31 < _Ragnar_> - Runtime Function: long long __modti3 (long long A, long long B)
03:31 < _Ragnar_> These functions return the remainder of the signed division of A
03:31 < _Ragnar_> and B.
03:31 < Postal> what is the command in vi to remove a single letter
03:31 < _Ragnar_> so the existence of that function is definitely required
03:32 < _Ragnar_> Postal: 'r' and then '-'
03:33 < Postal> usr/share/man/man8/depmod.8: module-init-tools modutils
03:33 < Postal> !> usr/share/man/man8/insmod.8: module-init-tools modutils
03:33 < Postal> !> usr/share/man/man8/lsmod.8: module-init-tools modutils
03:33 < Postal> !> usr/share/man/man8/modinfo.8: module-init-tools modutils
03:33 < Postal> !> usr/share/man/man8/modprobe.8: module-init-tools modutils
03:33 < Postal> !> usr/share/man/man8/rmmod.8: module-init-tools modutils
03:33 < Postal> !> Due to previous errors.
03:34 < mnemoc> o_O
03:34 < mnemoc> years without seen that error
03:35 < _Ragnar_> *g*
03:35 < Postal> im going to set up an ftp server so i can put the error messages on it
03:37 < CIA-7> mnemoc * r5744 /trunk/package/base/ncurses/ncurses.conf: * fixed ncurses.conf way to use $includedir (and fix a typo too)
03:38 < mnemoc> Postal: what is you host system? did you enabled module-init-tools using the option on Config?
03:44 < _Ragnar_> okay that explains a little: ... from gcc/mklibgcc.in: # .oS objects will have all non-local symbol definitions .hidden
03:45 < _Ragnar_> the first lib had .o files, the second (malfunctioning) had .oS (where those symbols supposedly are hidden and therefore not linked in)
03:47 < rxr> re
03:47 < mnemoc> oh
03:47 < mnemoc> rxr: that wasn't a typo
03:47 < mnemoc> rxr: damn
03:47 < mnemoc> don't update unitl i revert
03:49 < mnemoc> done
03:50 * mnemoc should keep his mind on one topic when he is sleepy
03:50 < CIA-7> mnemoc * r5745 /trunk/package/base/ncurses/ncurses.conf: * reverted 5744, mnemoc' stupidity
03:51 < rxr> lol
03:51 < rxr> rehi
03:51 < rxr> so - susan in bed ...
03:51 < rxr> 3:51 here ...
03:51 < rxr> and you know what? my ibook is under linux again - yeah!
03:52 < rxr> mnemoc: s.th. I should read right now ?
03:53 < rxr> susan and I have been there this evnening / night ...:
03:53 < rxr> http://www.my-tropical-islands.com/
03:54 < rxr> http://www.my-tropical-islands.com/engl/index.htm
03:54 < mnemoc> o_O
03:55 < mnemoc> is that reality or future project?
03:55 < rxr> reality ...
03:55 < rxr> we have been there today ..
03:55 < rxr> this building is from a company that wanted to build some cargo lifter
03:57 < mnemoc> _Ragnar_: oh, great... this you foind the malfunctoning part that injected that .oS-s inside libgcc.a?
03:57 < rxr> cargo lifter == some high tech project to create a next generation zeppelin; airship - to transport heavy cargo ..
03:57 < rxr> but they got bankrupt ...
03:57 < mnemoc> :)
03:57 < rxr> so some japanese bought the hall and made this ocean thing out of it ...
03:57 < rxr> - well I prevered the high tech stuff more ...
03:58 < _Ragnar_> mnemoc: they don't get hidden because the first one is a cross compiler, so libgcc assumes the as/ld can't handle .hidden, and therefore they are left visible
03:58 < rxr> this paradise think jsut open in december ...
03:59 < _Ragnar_> the second one is native, so it can detect that the as/ld supports .hidden, so it uses it
04:00 < mnemoc> _Ragnar_: can we convince as/ld our crosscompiler can handle .hidden symbols?
04:00 < mnemoc> rxr: i would love to have a place like that here
04:01 < mnemoc> rxr: this must be the cause of ssp thing too, getting hidden symbols out
04:01 < _Ragnar_> looks likely
04:03 < _Ragnar_> just have to patch the configure to set libgcc_visibility=yes
04:03 < Postal> mnemoc, to answer your question my host system is Suse 9.1
04:03 < mnemoc> _Ragnar_: --libgcc_visibility=yes ?
04:04 < rxr> mnemoc: the cargo lifter or this tropical islands thingi ?
04:04 < mnemoc> _Ragnar_: that's easy, [ $stagelevel -le 1 ] && var_append extraconfopt ' ' '--libgcc_visibility=yes'
04:04 < _Ragnar_> mnemoc: no I don't think it's that easy
04:05 < mnemoc> rxr: tropical islands on the middle of the city :)
04:05 < _Ragnar_> sorry, it's libgcc_visibility=no
04:05 < Postal> mnemoc, any ideas on the module-init-tools error?
04:06 < _Ragnar_> we don't want it to be visible
04:06 < mnemoc> _Ragnar_: [ $stagelevel -le 1 ] && var_append patchfiles ' ' $confdir/libgcc_novisibility.diff
04:06 < rxr> mnemoc: *lol*
04:06 < rxr> a) it is not in the middle of the citey
04:06 < rxr> it takes some subway, train and then bus shutle to reach it
04:06 < mnemoc> :)
04:06 < rxr> it is about 60km out of berlin ...
04:07 < rxr> about 1 1/4 hour at least
04:07 < Postal> rxr, what are you all talking about...
04:07 < rxr> and the way back - was ,... wel don't ask ..
04:07 < mnemoc> Postal: http://www.my-tropical-islands.com/engl/index.htm
04:07 < rxr> b) it is not for free - read rather expensive ...
04:07 < mnemoc> i never said free
04:08 < rxr> it it would not be susan's birthday we would most probablxy not be there (at least not today)
04:08 < rxr> there must be some lovely places ofer in chile, too ... ;-)
04:08 < _Ragnar_> mnemoc: would that mean I have to start my build again completely?
04:09 < mnemoc> rxr: ours are natural :p
04:09 < Postal> Thats wierd
04:09 < rxr> yeah - ours normally, too
04:09 < rxr> I'm also not such a contructed entertainment type - I perver good old nature ...
04:09 < rxr> e.g this was real fuN:
04:09 < rxr> http://gsmp.tfh-berlin.de/rene/photos/ireland/gallery.html
04:10 < rxr> ^- 3 weeks Ireland cycling - too bad in 1998 (my last real holiday)
04:10 < mnemoc> my last real^W holiday was on that year too
04:11 < rxr> arrg - my spar cbuild stopped 15:32
04:11 < rxr> !> File not found: download/mirror/m/man-1.5p.tar.bz2
04:11 < rxr> !> Did you run ./scripts/Download for this package?
04:11 < mnemoc> :D
04:11 * rxr could kick /me
04:11 < mnemoc> that was me
04:11 < rxr> I could kikc me for not running Downlaod after svn up ...
04:12 < rxr> at least the box did not yet freeze ...
04:12 < rxr> but I suspect the power supply anyway - e.g. it most often crashes when I have many other boxen and e.g. a 21" CRT powered on and my fridge starts recooling ...
04:14 < mnemoc> OT: _Ragnar_ discovered 0-libgcc.a keeps .hidden symbols visible because it thinks as/ld can't handle it, injecting not desired symbols; 1-perl doesn't build on anyone's; Postal gets shares between modutils and module-init-tools
04:15 < mnemoc> i have to go now, greetings to susan :)
04:15 < rxr> she is already sleeping ...
04:15 < rxr> cu mnemoc
04:15 < mnemoc> tomorroD[Dw
04:15 < rxr> yep
04:15 < mnemoc> i guessed she wasn't awake :)
04:15 < rxr> greetings over there, too
04:16 < mnemoc> same case, tomorrow :)
04:16 < mnemoc> cu rxr, _Ragnar_ and Postal. :)
04:16 < Postal> cu
04:16 < Postal> rxr, how have you been?
04:19 < Postal> i noticed that glibc was checked off on your to do list on the wiki
04:19 < rxr> well - the public transport chaos was quite exhausting ...
04:19 < rxr> the island was mostly fine
04:19 < Postal> i imagine
04:19 < Postal> The island look like lots of fun
04:20 < Postal> question for you
04:20 < Postal> I moved perl from stage one to stage 5, on mnemoc's advice, cause it was halting the build on stage one, will this cause problems?
04:20 < _Ragnar_> cu mnemoc
04:21 < rxr> not for a normal build
04:22 < Postal> since i snv up, i cannot get past stage 1
04:22 < rxr> I'll anyway fix this perl stage-1 build up since you need it if you cross build and want to use the result as a base for a native buildf
04:22 < rxr> what is the new problem ?
04:22 < rxr> anything beside 1-perl?
04:22 < Postal> usr/share/man/man8/depmod.8: module-init-tools modutils
04:22 < Postal> usr/share/man/man8/insmod.8: module-init-tools modutils
04:22 < Postal> usr/share/man/man8/lsmod.8: module-init-tools modutils
04:22 < Postal> usr/share/man/man8/modinfo.8: module-init-tools modutils
04:22 < Postal> usr/share/man/man8/modprobe.8: module-init-tools modutils
04:22 < Postal> usr/share/man/man8/rmmod.8: module-init-tools modutils
04:22 < rxr> what kind of build is that ?
04:22 < _Ragnar_> rxr: could you look at this .hidden problem, too?
04:22 < rxr> generic minimal
04:23 < Postal> minimal, with x.org
04:23 < rxr> ouh - hiden not tonight - after a sleep cycle I can ...
04:23 < rxr> Postal: maybe you get the shared bacause:
04:23 < rxr> formerly the pkg sel template had a bug not enabling module-init-tools
04:23 < _Ragnar_> k
04:23 < Postal> rxr, ok, well it seems thats fixed now lol
04:24 < _Ragnar_> bbl, heading home
04:24 < rxr> I fixed this -and when you svn up'ed you got module-init-tools overwriting modutils file because modutils was only built and "thought" it is the only one and should use the default - e.g. not .old names
04:24 * _Ragnar_ really really contemplates gentoo
04:24 < Postal> rxr, ok, so to fix this i...
04:24 < rxr> _Ragnar_: oh - no
04:25 < rxr> Postal: this is jur a warning of the build system that the files are shared which normaly default to an error
04:25 < rxr> just :
04:25 < rxr> moment
04:25 < Postal> ok
04:26 < Postal> what id really like to do is get inside this thing and really start testing it for you
04:26 < rxr> hm - just do it manuall:
04:27 < rxr> rm build/your-config/var/adm/flist/module-init-tools
04:27 < rxr> rm build/your-config/var/adm/flist/modutils
04:27 < rxr> rm build/your-config/var/adm/logs/*modutils
04:27 < rxr> rm build/your-config/var/adm/logs/*module-init-tools
04:27 < rxr> that will wipe the file list of the two packages and rebuild them ...
04:28 < rxr> _Ragnar_: Where do you get .hidden for wich arch and target?
04:28 < Postal> the second two in the list i cannot seem to find
04:29 < rxr> err
04:29 < rxr> Postal: .* to them ..
04:29 < rxr> they end with .err or .log ...
04:29 < Postal> ok
04:29 < rxr> Postal: youare on x86-64, too ?
04:29 < Postal> yes
04:30 < Postal> rxr, i need the stage numbers... give me a mom
04:31 < Postal> ok removed, now Build-Target?
04:33 < Postal> ..
04:35 < rxr> re
04:35 < rxr> yes
04:35 < Postal> re=??
04:35 < rxr> returned or simillar vairant ;-)
04:36 < Postal> rxr, i dont suppose the glibc problem is fixed?
04:36 < rxr> wich one ?
04:36 < rxr> the segfault ?
04:36 < Postal> thats the one lol
04:36 < rxr> should be fixed
04:36 < Postal> the only reason i ask is cause its building now
04:36 < rxr> as we guessed it was a bog in our flist tracing wraper
04:36 < rxr> it is tested to be fixed on sparc64, powerpc and x86 ...
04:36 < Postal> i have no idea what that means
04:37 < rxr> we need to track file modifications
04:37 < rxr> historically to tar the package up
04:37 < rxr> but today also to track dependencies automatically ...
04:37 < Postal> ok..
04:37 < Postal> any news on mipes mulitlib build?
04:37 < rxr> you can use find (slow, write only (mostly), and not parallel builds save), strace (medium slow)
04:38 < rxr> yes - I fixed the mutilib stuff up
04:38 < Postal> so does grub compile?
04:38 < rxr> I use it on my sparc64 sucessfully ...
04:38 < rxr> I think he said yes
04:38 < rxr> my sparc64 build is in 3-libtool right now ..
04:38 < Postal> ok, stick around for about 20 min and ill have all your answers lol
04:39 < rxr> so - what is that .hidden thing about ...
04:39 * rxr scrollback
04:39 < rxr> damn - i should be in bed ...
04:39 < rxr> where my gf already is ...
04:39 < rxr> good that I have the logs on my FS
04:39 * rxr grepping
04:40 * Postal waites with bated breath
04:41 < rxr> Postal: I thin I'm in bed quite soon
04:41 < _Ragnar_> rxr: uclibc based system
04:41 < Postal> what ever floats your boat
04:41 < rxr> it is 4.41 here, Im exhausted from traveling around and - well I have to write an article in the morning - and well - scanner work in the affternoon ...
04:41 < rxr> _Ragnar_: ouhm - uclibc ...
04:42 < rxr> _Ragnar_: you can not start with s.th. easy, can you? ;-)
04:42 < rxr> _Ragnar_: uclibc for x86 or x86-64 ?
04:42 < _Ragnar_> x86
04:42 < rxr> and it is not yet known my either mnemoc or jsaw ?
04:42 < _Ragnar_> ?
04:42 < rxr> a known error I meant
04:43 < _Ragnar_> mnemoc knows about it
04:43 < rxr> for what package do you get that fore ?
04:43 < rxr> _Ragnar_: I meant in a sense that he got the same error during his uclibc builds, too ...
04:44 < _Ragnar_> no aparently not
04:44 < _Ragnar_> okay lets start from the beginning:
04:45 < _Ragnar_> uclibc ist first built with a cross compiler. this cross compiler assumes that the as/ld don't know about hidden symbols, therefore they put everything into libgcc.a as a global symbol
04:46 < _Ragnar_> uclibc then links libgcc.a to itself, and gets these symbols
04:46 < _Ragnar_> for example __moddi3
04:46 < _Ragnar_> then gcc gets built natively, and then it knows 'okay my as/ld do know about .hidden!' therefore it makes the symbols in libgcc.a hidden
04:47 < _Ragnar_> then uclibc links that new libgcc.a in ... and suddenly all programs lack those __moddi3 function
04:47 < rxr> ah ha
04:48 < rxr> does not sound like s.th. I can fix in a minute
04:48 < Postal> glibc errored on a very long and complicated sed command...
04:48 < rxr> I need to first build a uclibc myself to test it a bit ...
04:48 < _Ragnar_> okay
04:48 < rxr> Postal: ouhm
04:48 < rxr> what is your build system?
04:48 < rxr> the t2 you cross compiled previously ?
04:48 < _Ragnar_> I can try to get it to work myself, but I'd need to know how to get gcc's configure patched
04:48 < Postal> no, im still in suse 9.1
04:49 < rxr> _Ragnar_: maybe we can discuss this tomorrow with jsaw
04:49 < _Ragnar_> and gcc.conf ... is less then easily understandable :)
04:49 < _Ragnar_> kay
04:49 < rxr> Postal: you cross build again ?
04:49 < Postal> we had a problem with the move remember
04:49 < rxr> _Ragnar_: I removed some historic junk from it - just today
04:49 < rxr> Postal: yes :-(
04:49 < Postal> but no im not cross building, im doing a native build
04:50 < rxr> hm
04:50 < _Ragnar_> ah okay, I'll svn up
04:50 < Postal> let me get the error, i doubt it is somthing you can fix now but... prehaps you want to look at it
04:50 < rxr> yeah - post the defect log region
04:52 < Postal> sed -e 's,[{}],@&,g' \
04:52 < Postal> -e 's,/\*\(@.*\)\*/,\1,g' \
04:52 < Postal> -e 's,/\* *,/* @r{,g' -e 's, *\*/,} */,' \
04:52 < Postal> -e 's/\(@[a-z][a-z]*\)@{\([^}]*\)@}/\1{\2}/g'\
04:52 < Postal> examples/termios.c | expand > termios.c.texi.new
04:52 < Postal> mv -f termios.c.texi.new termios.c.texi
04:52 < Postal> pwd=`pwd`; \
04:52 < Postal> no libm-err-tab.pl $pwd/.. > libm-err-tmp
04:52 < Postal> /bin/sh: no: command not found
04:52 < Postal> make[2]: *** [stamp-libm-err] Error 127
04:52 < Postal> make[2]: Leaving directory `/TOOLCHAIN/src.glibc.1106278502.19357.0x7f0200/libc/manual'
04:52 < Postal> make[1]: *** [manual/subdir_install] Error 2
04:52 < Postal> make[1]: Leaving directory `/TOOLCHAIN/src.glibc.1106278502.19357.0x7f0200/libc'
04:52 < Postal> make: *** [install] Error 2
04:52 < rxr> Postal: easy I had this on my ultra sparc, too
04:52 < rxr> the glibc configure is crap and does this:
04:52 < rxr> detecting perl: no
04:52 < rxr> ...
04:52 < rxr> and then just using "no" as perl ..
04:52 < rxr> very smart
04:53 < rxr> I looked at it but did not feel like fixing this auto* dung
04:53 < rxr> Postal: have you perl installed locally
04:53 < Postal> ok
04:53 < Postal> yes
04:53 < Postal> of course lol
04:53 < rxr> which glibc build stage was that? 1-glibc?
04:53 < Postal> 5
04:53 < rxr> oh - 5 already ?
04:53 < Postal> yeah
04:53 < rxr> you have no perl yet, hm ?
04:53 < Postal> it only errors on 5
04:54 < rxr> so the build is continouing ?
04:54 < rxr> just counting the errors, not halting ?
04:54 < Postal> yes, we have no perl cause we moved it to stage 5 cause there was an error remember
04:54 < rxr> svn revert package/perl/perl/perl.desc
04:54 < Postal> it just built perl in stage 5 right now
04:55 < Postal> no need to revert it
04:55 < rxr> ok ok .
04:55 < Postal> Binary file for vim-6.3.gem not present. Skipped in package database.
04:55 < Postal> !> Binary file for grub-0.95.gem not present. Skipped in package database.
04:55 < Postal> !> Binary file for glibc-2.3.4-2004-12-24.gem not present. Skipped in package database.
04:55 < Postal> !> Binary file for bdb-4.2.52.gem not present. Skipped in package database.
04:55 < rxr> ? what did you do ?
04:55 < Postal> it finished
04:56 < rxr> oh - some minmual ...
04:56 < rxr> it built perl in a matter of seconds ?
04:56 < Postal> its already mostly built from before my exams
04:56 < Postal> and yes it did
04:56 < rxr> my god
04:56 < rxr> my sparc64 needs multiple hours to build the gcc ...
04:56 < rxr> perl was - no idea 30 minutes ..
04:57 < Postal> amd64 3400 2gb of ram with a 2gb swap space
04:57 < Postal> sata hard drive
04:57 < Postal> its also a -j4
04:57 < Postal> it built pearl in 5 min
04:57 < rxr> Sun Ultra SPARC ł 295Mhz ... 3xx MB of RAM ...
04:57 < rxr> some old IBM SCSI discs ...
04:58 < Postal> i can built glibc in 10-15 min
04:58 < rxr> .oO
04:58 < rxr> I need to update my main server ...
04:58 < Postal> gcc ion 30
04:58 < Postal> in*
04:58 < Postal> ;-)
04:58 < Postal> now about this glibc
04:58 < rxr> you could schedule glibc for a rebuild
04:58 < rxr> it will just build then now you have perl
04:58 < Postal> will it be fixed due to pearl?
04:59 < rxr> yes
04:59 < rxr> I jsut reverted the 1-perl thing - so until I fix perl it is build in stage3 just before glibc ...
05:00 < Postal> ok its building now
05:00 < rxr> so - much fun with your new x86-64 T2 ;-)(
05:00 < CIA-7> rene * r5746 /trunk/package/perl/perl/perl.desc: * disable perl for stage 1 again - needs some DESTDIR awareness
05:00 < Postal> should be done with all in less than 30
05:00 < rxr> you might like to announce it on the t2 list if you like ,-)
05:00 < rxr> well I'm off now
05:00 < Postal> is there a way i can clock the build?
05:00 < rxr> time ./scripts/Build-Target
05:01 < Postal> ok
05:01 < Postal> its clocking now
05:01 -!- praenti [~praenti@82.139.198.99] has quit [Read error: 104 (Connection reset by peer)]
05:01 -!- praenti [~praenti@82.139.198.99] has joined #t2
05:01 < Postal> ill give you the time tomorrow it building mostly glib
05:01 < Postal> glibc*
05:01 < _Ragnar_> [ $stagelevel -le 1 ] && sed -i -e "s/libgcc_visibility=no/libgcc_visibility=yes/g" gcc/configure
05:01 < Postal> vim grub and bdb
05:01 < _Ragnar_> does that look like it would work?
05:03 < rxr> Postal: well in this big custmain function?
05:03 < rxr> yes - should
05:04 < _Ragnar_> thx
05:04 < rxr> ok - I'm in bed ...
05:05 < _Ragnar_> goodnight & sleep well
05:05 < rxr> if you wankt to post some notes: t2@exactcode.de (t2 list no reg required) or rene@exactcode.de
05:05 < rxr> n8 all
05:05 < rxr> have fun!
05:11 < Postal> cu
06:47 < jsaw> re
06:51 < _Ragnar_> hi
06:54 < jsaw> hi _Ragnar_
06:56 < jsaw> _Ragnar_: how's life?
06:56 < _Ragnar_> uhm
06:56 < _Ragnar_> don't ask ...
06:56 < _Ragnar_> ;p
06:57 < jsaw> hehe
06:57 < _Ragnar_> do you happen to know why 0-gcc has --disable-shared ?
06:57 < jsaw> no
06:58 < _Ragnar_> okay
06:58 < jsaw> I can only imagine that this way it's easier to seperate host and t2 toolchain
06:59 < jsaw> but I don't know really.
06:59 < _Ragnar_> okay
06:59 < _Ragnar_> if you want, scroll back and look at what I said to rxr a few hours ago
07:00 < jsaw> time frame?
07:00 < jsaw> you mean the libgcc_visibility?
07:00 < _Ragnar_> yeah
07:02 < _Ragnar_> the respective .oS are only generated when the shared libs gets built (can't figure out why)
07:09 < jsaw> _Ragnar_: there's so much to read, would you mind giving me a short abstract what the problem is? :)
07:09 < _Ragnar_> basically, it seems uclibc links to libgcc.a
07:10 < _Ragnar_> in the cross-built, this imports certain functions from libgcc.a into uclibc
07:11 < _Ragnar_> later on, when libgcc is built natively, those symbols are hidden, and therefore not imported into uclibc any more, and therefore a lot of programs just fail with missing symbols
07:11 < jsaw> ah ok.
07:11 < jsaw> btw, the linking to libgcc is always done unless you specifiy -nostdlib
07:12 < jsaw> on what architecture?
07:12 < _Ragnar_> x86-32
07:13 < _Ragnar_> yeah ... but this way, the uclibc symbol overrides the symbol in libgcc
07:14 < _Ragnar_> and then, when the symbol suddenly disappears from uclibc, well, where should it come from? libgcc.a is linked to, not libgcc.so
07:14 < _Ragnar_> I think mnemoc is right in saying that the symbols should not be in uclibc at all
07:16 < jsaw> It's more of an chicken'n'egg problem. If uclibc uses functions (builtins of gcc), libgcc.a should provide them. But uclibc should not expose them to the world then.
07:18 < _Ragnar_> uclibc does not directly use them, but the compiler might insert calls to them ... one of them is __moddi3 which is a integer division
07:20 < jsaw> another thing I do not get yet is, how does this show up, what's the error you encounter?
07:21 < _Ragnar_> bash, expr, whatever showing 'could not resolve symbol __moddi3' ...
07:21 < jsaw> ah, ok. What processor are you optimizing for?
07:22 < _Ragnar_> athlon xp
07:23 < jsaw> on a athlong xp I guess.
07:27 < _Ragnar_> ?
07:27 < jsaw> your host system?
07:27 < _Ragnar_> yes
07:33 < jsaw> I don't know why this is not inlined. Does libgcc_visiblity solve this problem?
07:33 < _Ragnar_> for stage 0 the limiting factor was --disable-shared
07:34 < _Ragnar_> I'm currently trying stage 1
07:34 < jsaw> what compiler/distro do you have on your host?
07:34 < _Ragnar_> mostly rock, but I compiled t2's gcc on that system
07:35 < _Ragnar_> the problems manifest themselves at stage 5 when uclibc gets rebuilt
07:38 < _Ragnar_> hmm...
07:38 < _Ragnar_> why would the configure on 1-gcc be called twice with different parameters?
07:39 < jsaw> o_O
07:40 < jsaw> oh, that can happen. The automatic configure run and an additional explicit configure call in the .conf script.
07:40 < _Ragnar_> the second one does naughty things;p
07:40 < jsaw> o_O
07:42 < _Ragnar_> ouch
07:42 < _Ragnar_> the second one goes very much haywire
07:44 < _Ragnar_> /export/t2-trunk/src.gcc.1106289593.4462.fad8824d/gcc-3.4.3/gcc/configure: line
07:44 < _Ragnar_> 1: /export/t2-trunk/build/a32-2.1.0-beta3-x86-athlon-xp-32-generic-expert/usr/i686-unknown-linux-uclibc/bin/as: No such file or directory
07:44 < _Ragnar_> of course! this is in chroot already
07:45 < jsaw> stage 1? no shouldn't be.
07:45 < jsaw> stage 1 is cross compiler
07:45 < _Ragnar_> oh okay
07:46 < jsaw> oh, I see. ../../configure without arguments
07:47 < _Ragnar_> should I comment that out?
07:47 < jsaw> no that's only libiberty.
07:50 < jsaw> but you can try changing this ../../configure line to
07:50 < jsaw> eval $configprefix ../../configure $confopt
07:50 < _Ragnar_> but it seems to affect the other parts, like for example resetting libgcc_visibility
07:51 < jsaw> so try the "eval..." in this line
07:51 < jsaw> no
07:51 < _Ragnar_> okay
07:52 < jsaw> that will not work. It needs a native libiberty.
07:52 < jsaw> Hmmm.
07:52 < jsaw> Please report this to the mailing list.
07:52 < _Ragnar_> ?
07:52 < _Ragnar_> *g*
07:52 < jsaw> I have no more ideas currently.
07:52 < jsaw> the libiberty seems to be need natively.
07:53 < _Ragnar_> hmmm
07:54 < _Ragnar_> natively for the host environment, or natively for the cross chroot?
07:54 < jsaw> the latter one (not that I understand the reason...)
07:55 < _Ragnar_> hmmm
07:55 < _Ragnar_> then we should redo the main config after the lib is built
07:56 < jsaw> Good idea. Maybe doing it in reverse order does it.
07:58 < jsaw> I have to go to work in a few minutes, cu later
07:58 < _Ragnar_> cu laters thanks for your insights
08:02 < jsaw> I don't think I could give to many insights, still, welcome... (now I'm really off)
08:02 < CIA-7> jsaw * r5747 /trunk/scripts/Check-PkgFormat: * REGISTER is not parsed.... must add "Restricted" explicitly
08:51 < CIA-7> jsaw * r5748 /trunk/scripts/functions: * source_file needs the url to decide mirror|local (in populate_archdir)
08:52 < CIA-7> jsaw * r5749 /trunk/package/network/samba/samba.desc: * add NOPARALLEL to samba
08:56 < CIA-7> jsaw * r5751 /trunk/package/java/blackdown-jdk/blackdown-jdk.conf: * do not clutter up /opt too much: use /opt/java as prefix
08:56 < CIA-7> jsaw * r5750 /trunk/package/java/blackdown-jre/blackdown-jre.conf:
08:56 < CIA-7> * do not clutter up /opt too much: use /opt/java as prefix
08:56 < CIA-7> (due to incompatibilities in conjunction with commercial apps,
08:56 < CIA-7> eg. Matlab, multiple jre may be necessary...)
08:58 < CIA-7> jsaw * r5752 /trunk/package/graphic/ghostscript/ghostscript.desc: * add NOPARALLEL to ghostscript package
09:00 < jsaw> re
09:02 < _Ragnar_> re
09:04 < jsaw> hi _Ragnar_
09:11 < _Ragnar_> heya
09:11 < _Ragnar_> I'm about to head to bed
09:12 < jsaw> what time in eh, s.diego?
09:15 < _Ragnar_> midnight
09:16 < jsaw> sleep well, _Ragnar_ (9am here btw)
09:16 < _Ragnar_> thanks, I know
09:17 < jsaw> :)
10:09 -!- CIA-temp688 [~CIA@flapjack.navi.cx] has joined #t2
10:10 -!- CIA-temp688 is now known as CIA-12
10:10 -!- CIA-7 [~CIA@flapjack.navi.cx] has quit [Read error: 60 (Operation timed out)]
11:17 -!- ideal [~idealm@211.100.227.107] has quit ["Leaving"]
11:25 -!- Netsplit saberhagen.freenode.net <-> irc.freenode.net quits: jsaw, valentin, Postal, mnemoc
11:25 -!- Netsplit over, joins: valentin, Postal, mnemoc, jsaw
12:24 < rxr> moin
12:27 < CIA-12> rene * r5753 /trunk/package/base/ncurses/ncurses.conf:
12:27 < CIA-12> * reverted the ncurses header relocated from Fedora Core - causes
12:27 < CIA-12> regressions ...
13:32 < CIA-12> jeru * r5754 /trunk/package/network/pound/config.hlp: * added config help entries
14:35 < mnemoc> rehi
14:46 < CIA-12> jeru * r5755 /trunk/package/develop/clearsilver/ (clearsilver.conf python24-configure.patch):
14:46 < CIA-12> * disabling html compression support (linking 'neo_cgi' against zlib doesn't work)
14:46 < CIA-12> * added patch for configure script because of changes in Python 2.4
14:48 < CIA-12> jeru * r5756 /trunk/package/develop/trac/ (. trac.desc): * added Trac package
14:55 < CIA-12> jsaw * r5758 /trunk/package/perl/perl-archive-zip/ (. perl-archive-zip.desc): * add perl-archive-zip (1.33)
14:56 < CIA-12> jsaw * r5757 /trunk/package/perl/perl-compress-zlib/ (. perl-compress-zlib.desc): * add perl-compress-zlib (1.14)
14:59 < mnemoc> rxr: config/minimal/linux.cfg is merged or used to replace?
15:00 < CIA-12> jsaw * r5759 /trunk/package/office/ooo/ (6 files): * given up on OOo 1.1, now trying 1.9-m70
15:00 -!- _Ragnar__ [loki@66-146-166-62.skyriver.net] has joined #t2
15:00 < mnemoc> wb _Ragnar__
15:01 -!- _Ragnar_ [loki@66-146-166-62.skyriver.net] has quit [Read error: 101 (Network is unreachable)]
15:15 < mnemoc> bbl
15:36 -!- _Ragnar_ [loki@66-146-166-62.skyriver.net] has joined #t2
15:36 -!- _Ragnar__ [loki@66-146-166-62.skyriver.net] has quit [Read error: 104 (Connection reset by peer)]
15:37 < Postal> Hey all
16:23 -!- mipe [~mipe@dsl10040.japo.fi] has joined #t2
16:23 < rxr> re
16:23 < rxr> hi Postal
16:27 < rxr> sparc64 build:
16:27 < rxr> # ./scripts/Create-ErrList -cfg system
16:27 < rxr> Error logs from system-2.1.0-beta3-sparc-v9-64-desktop-expert:
16:27 < rxr> [5] base/gcc [5] base/strace
16:27 < rxr> [5] develop/gnupth [5] base/gmp
16:27 < rxr> [5] shells/ash [5] database/bdb
16:27 < rxr> [5] archiver/unshield [5] x11/xorg
16:27 < rxr> [5] mail/exim [5] audio/cdparanoia
16:27 < rxr> [5] network/openldap
16:27 < rxr> 1021 builds total, 217 completed fine, 11 with errors.
16:29 < rxr> hm - should fix the xorg mesa assembly - since otherwise a lot of stuff will fail soon ...
16:39 < CIA-12> rene * r5760 /trunk/package/base/modutils/modutils.desc: * marked modutils aas NOPARALLEL (failed here once out of 6 builds ...)
16:40 < mipe> rxr:gnupth,cdparanoia as noparallel too,forgot to send patch to mailing list,sorry
16:41 < mipe> +ash
16:49 < rxr> thansk
16:52 < mipe> hmm,how do i add patch manually in package.conf, i mean i'm checking with if arch is x86_64 then add patch,otherwise it doesnt, with var_append?
16:54 < mnemoc> rehi
16:56 < rxr> hi
16:56 < mnemoc> hi rxr
16:57 < mnemoc> why is your 5-gcc failing?
16:58 < mnemoc> OT: how went your osx thing?
16:58 < rxr> needs to continue hacking it soon ...
16:58 < rxr> i slept quite a lot today - due to exhausting due to yesterdays traveling ...
16:58 < rxr> mnemoc: do you know mipe name and email for a changelog entry
16:58 < rxr> ?
16:59 < mnemoc> nope
17:00 < CIA-12> rene * r5761 /trunk/package/ (3 files in 3 dirs):
17:00 < CIA-12> mipe<@IRC>:
17:00 < CIA-12> * marked ash, gnupth and cdparanoia as NOPARALLEL
17:03 < mnemoc> *G*
17:06 < mnemoc> rxr: do you know how to know the io of an isa card?
17:07 < rxr> if it is an classic card, no way
17:07 < rxr> open the case and take a look at the jumpers
17:07 < rxr> if it is a isapnp card use the 2.4 and up PnP code ...
17:07 < rxr> it should setup it automatically
17:07 < rxr> there was userspace tools for that (for 2.0 and 2.2) but they sucked and are not needed anymore ...
17:08 < rxr> you can trial and error modprobing the kernel module with different values until it works ...
17:10 < mnemoc> trial and error worked :)
17:10 < mnemoc> i rebuilt lx2.6 to include isapnp but that didn't worked
17:13 < mnemoc> rxr: where should i add it to work on boot, /etc/conf/kernel ?
17:14 < rxr> yep
17:22 < rxr> mnemoc: gcc in the above list failed due to sparc64 oddities ...
17:22 < rxr> nothing serious to bother you ;-)
17:22 < rxr> i'm running a x86 ref build right now
17:22 < mnemoc> bother me? :p
17:22 < mnemoc> fresh or additive ref build?
17:23 < rxr> nothing serious to be of interest for you ;-)
17:23 < mnemoc> :D
17:23 < rxr> fresh
17:23 < rxr> Cleanup -full
17:24 < mnemoc> ec-outpost.dyndns.org/t2 ?
17:24 < rxr> when there is new stuff, yes
17:24 < rxr> but right now it has zero erros ...
17:26 < mnemoc> OT: why do we use sysvinit's killall5 instead of killall on scripts?
17:26 < rxr> no idea
17:29 < mnemoc> did you read my RFC about Config's /choice/ ?
17:29 < rxr> yes
17:29 < mnemoc> and?
17:30 < mnemoc> i don't know if someone use 'unexistant default value' as a feature
17:32 < rxr> well - I would say do the modification
17:32 < rxr> maybe even warn on the terminal that the default value was not found
17:35 < mnemoc> an echo?
17:35 < rxr> s.th. that will show up on the terminal
17:35 < mnemoc> ok
17:35 < rxr> ich stdout is not redirected there, an echo
17:57 < _Ragnar_> moin
17:58 < mnemoc> moin _Ragnar_
17:58 < _Ragnar_> hi mnemoc
17:59 < mnemoc> how was your investigation?
18:00 < _Ragnar_> I _think_ I got the symbols out of uclibc
18:00 < mnemoc> how?
18:00 < _Ragnar_> ugly
18:00 < _Ragnar_> ;)
18:00 < mnemoc> sed -e '....' /lib/libc.so.0 ?
18:00 < mnemoc> :)
18:01 < _Ragnar_> no;)
18:01 < mnemoc> then it's not ugly :)
18:01 < _Ragnar_> [ $stagelevel -le 1 ] && (cd ..; sed -i -e "s/libgcc_visibility=no/echo
18:01 < _Ragnar_> 'override visibility'; libgcc_visibility=yes/g" gcc/configure; grep visibility gcc/configure )
18:01 < _Ragnar_> (okay, the echo and grep can go now)
18:02 < _Ragnar_> and then --enable-shared was needed for 0-gcc
18:02 < mnemoc> if you need more space: [ $stagelevel -le 1 ] && hook_add preconf 5 'gcc_override_visibility'
18:03 < mnemoc> uhm
18:03 < _Ragnar_> ?
18:04 < mnemoc> ? <--- to the uhm or or hook_add ?
18:04 < _Ragnar_> to the uhm
18:05 < mnemoc> i was thinking in side effects of shared libraries on the cross-compiler
18:05 < mnemoc> libgcc_visibility=no <--- is that the default value?
18:05 < _Ragnar_> no
18:05 < _Ragnar_> the default depends on whether as/ld supports it
18:06 < mnemoc> and removing that check out of configure?
18:07 < _Ragnar_> I'm doing that, libgcc_visibility is forced to yes
18:08 < _Ragnar_> another test is in mklibgcc, which tests for --enable-shared
18:08 < mnemoc> --disable-shared forces libgcc_visibility=no ?
18:09 < _Ragnar_> in a way .. the test is if libgcc_visibility=yes && shlib_enabled then ...
18:10 < mnemoc> and removing that " && shlib_enabled" ?
18:10 < _Ragnar_> could do that
18:10 < mnemoc> looks cleaner to me
18:10 < _Ragnar_> what side effects do you think there can be?
18:11 < mnemoc> there should be a reason to have "--disable-shared" there
18:11 < _Ragnar_> legacy?:)
18:11 < mnemoc> maybe
18:12 < _Ragnar_> it didn't hurt my build ... how do you ask svn to see when it was added, and who added it?
18:13 < mnemoc> svn blame
18:13 < _Ragnar_> oh, and why's there a completely new configure run for libibery?
18:14 < _Ragnar_> just looked, it seems to come from rock
18:14 < mnemoc> t2 has less than a year ;)
18:14 < _Ragnar_> I know ;)
18:16 < _Ragnar_> brb, breakfast
18:24 < mnemoc> _Ragnar_: where do you live?
18:40 < CIA-12> mnemoc * r5762 /trunk/scripts/config.func: * improved config's choice function to take first element as default if it's not part of the list of choices.
18:43 < CIA-12> mnemoc * r5763 /trunk/package/base/sysvinit/config.in: * cleaned init config because now we don't need to worry if default is available or not
18:50 < _Ragnar_> california
18:54 < mnemoc> what time is it there?
19:00 < _Ragnar_> 10am
19:01 < mnemoc> uhm, i have 5 hours to you, and 4 to rene :)
19:01 < mipe> where do you live mnemoc?
19:02 < mipe> its 8pm here :)
19:03 < daja77> huh where are you?
19:03 < mipe> .fi
19:04 < daja77> ah :)
19:11 < mnemoc> Chile
19:11 < mnemoc> hi daja77
19:12 < mipe> argh gcc doesnt produce 32bit libs too
19:14 < _Ragnar_> too?
19:15 < mipe> grub needs 32bit libs to compile on x86_64
19:16 < _Ragnar_> ah
19:17 < _Ragnar_> nice ;) 2-gcc is built fine without patching :)
19:18 < _Ragnar_> but anyway .... the longer I think about it, the less clear it is to me why uclibc links to libgcc.a at all
19:21 < mnemoc> _Ragnar_: how are you patching it now?
19:22 < mnemoc> _Ragnar_: glibc links to libgcc.a too
19:22 < _Ragnar_> wierdness
19:34 < mnemoc> _Ragnar_: how are you patching it? /libgcc_visibility.*/libgcc_visibility=yes/ ?
19:34 < _Ragnar_> still have the old line in
19:35 < mnemoc> including grep? ;)
19:35 < _Ragnar_> yes
19:35 < _Ragnar_> still debugging ;)
19:36 < _Ragnar_> when I've reached stage 5 and no programs fail, I'll take those out
19:36 < mnemoc> :D
19:36 < _Ragnar_> what are the build requirements for t2 to the host binutils?
19:41 < mnemoc> we build one replacement on stage-0, i guess it doesn't build on redhat's 2.96.
19:42 < mnemoc> on native builds we do a pseudo-crosscompiler which should solve any problem
19:43 < mnemoc> but on cross builds we use host cc/ld on some stuff, as BUILD{,-}CC and HOSTCC, which can make troubles
19:51 < _Ragnar_> just wondering if it's really right to force libgcc_visibility to yes
19:51 < mnemoc> everything depends why gcc/configure is setting it to no
20:19 < rxr> glibc list:
20:19 < rxr> from the gcc release maintainer:
20:19 < rxr> Mark Mitchell wrote:
20:19 < rxr> > How about this one?
20:19 < rxr> Of course not. The code is not tested, the arch maintainer hasn't agreed, and you yourself will never touch the cvs archive.
20:20 < rxr> ^- the answer is from my beloved Ullrich Drepper ...
20:20 < rxr> isn't that communication stlye crazzy ?
20:21 < mnemoc> friendly
20:22 < rxr> definetly ...
20:23 < rxr> HEAD:
20:23 < rxr> Error logs from reference-2.1.0-beta3-x86-athlon-32-reference:
20:23 < rxr> 1393 builds total, 196 completed fine, 0 with errors.
20:24 < rxr> mnemoc: we should freeze soon for the next beta, hm?
20:27 < rxr> maybe after my reference build and and a trivial regressoin fixup round ?
20:27 < rxr> like monday ?
20:27 < rxr> valentin: you sparc64 stage-5 begins to materialize on my U30 ...
20:27 < rxr> s/you/a/
20:31 < mnemoc> re
20:33 < mnemoc> rxr: i waiting for this libgcc issue, for taging, and COPYRIGHT-NOTE and maybe stone -> passwd to release, comments? there is stuff on top of the TODO
20:33 < rxr> sure
20:33 < rxr> the two stone bugs must be fixed forrelease
20:34 < mnemoc> any other 'showstoper'?
20:34 < rxr> not so far
20:35 < rxr> at least not if we want to make zeroerror sparc64 a release criteria ...
20:35 < rxr> don't want that is ...
20:35 < mnemoc> i think we are going very straightforward
20:35 < rxr> yes absolutly ...
20:35 < rxr> t2 progress is amazing ...
20:36 < mnemoc> is the stone/passwd thing pam related or dialog?
20:36 < rxr> no pam or so
20:37 < rxr> there is some error flashing up on the terminal
20:37 < rxr> just too short to read it - one need to debug that soon ...
20:37 < rxr> and we should check the return value of the passwd command ...
20:37 < rxr> maybe missing library sym-link or so ...
20:37 < rxr> altough we already do run ldconfig somewhere ...
20:38 < rxr> that we can announce immense x86-64 and sparc64 progress is very interesting - no need to be zero error for the next beta ...
20:39 < mnemoc> no need to be zero error
20:42 < mnemoc> http://svn.exactcode.de/big.png <--- cliff died?
20:43 < rxr> don't hope and think so - he seems to have marked <100 patches in submaster and most probably runs a reference build with tons of errors or so ..
20:43 < rxr> or the master in cheif and only commitee is busy @ work ...
20:45 < mnemoc> i will never understand him
20:46 < rxr> me neither
20:47 < rxr> wow - i entered my flat and the sparc died ...
20:48 < rxr> I somehow think this mainboard / cpu combo from sun is very supply voltage instable ...
20:48 < rxr> a starting fridge or light makes this damn thing freeze ...
20:48 < rxr> damn
20:52 < mnemoc> Error logs from default-2.1.0-beta3-x86-pentium-32-uboot-expert:
20:52 < mnemoc> 102 builds total, 102 completed fine, 0 with errors.
20:52 < valentin> rxr: oh
20:52 < mnemoc> i don't understand why i don't have libgcc.a problems on uboot, but i do on minimal
20:53 < mnemoc> uhm, i don't have libgcc.a on uboot!
20:53 < mnemoc> only .so
21:02 < mnemoc> rxr: have you cleaned that hardware reciently?
21:02 < rxr> nope ...
21:03 < rxr> I fear to depart that beast ...
21:05 < jsaw> re
21:06 < rxr> hi jsaw
21:06 < jsaw> hi rxr
21:06 < jsaw> ooo failed due to python just like blender did... damn
21:06 < rxr> will be a long night over here - I hope someone will be around here *g*?
21:07 < jsaw> not really, must wake up early to go shopping with family...
21:08 < rxr> ;-)
21:08 < mnemoc> hi jsaw
21:08 < jsaw> btw, is blender compiling now or still failing due to tcl/tk linkage of python?
21:08 < jsaw> hi mnemoc
21:09 < mnemoc> freenode
21:09 < mnemoc> [15:44:06] 7 #geekpreview freenode
21:09 < mnemoc> [15:44:06] 7 #geekpreview freenode
21:10 < mnemoc> err
21:10 < mnemoc> sorry
21:10 < mnemoc> jsaw: rene is running a new reference build
21:10 < jsaw> oh, ic
21:11 < jsaw> mnemoc: what comes now (uboot)?
21:11 < mnemoc> 102 builds total, 102 completed fine, 0 with errors.
21:11 < jsaw> that's why I ask, what's the next step?
21:12 < mnemoc> i'm designing a bit how will i make it work
21:12 < mnemoc> finish my CV :p
21:12 < rxr> 1393 builds total, 210 completed fine, 0 with errors.
21:12 < rxr> == 20:33:09 =[5]=> Building x11/xorg [6.8.1 2.1.0-beta3].
21:13 < mnemoc> jsaw: making it bootable over cd, floppy and usb-storage, configuring a complete busybox enviroment and adding missing tools
21:13 < mnemoc> and then writing an optional 'installer' over it
21:14 < rxr> hm - 225V in the outlet ...
21:14 < jsaw> okay - not that much. should be finished tomorrow I guess.
21:15 < jsaw> 220-230V is usual, or?
21:15 < rxr> should hook my oscilloscope in to see what happens when the fridge stars or the light toggled ...
21:15 < rxr> jsaw: yes
21:16 < rxr> valentin: could you come over and power on the water cooker while I watch the osc. ?
21:16 < rxr> e.g. no sample osc. here ... :-(
21:16 < rxr> mnemoc: you think cleaning the box might help ?
21:17 < mnemoc> a blower will safe that poor machine
21:17 < jsaw> oh, problems with your sparc...?
21:17 < rxr> yesh - the usual freezes ...
21:17 < rxr> first I thought some compiler bug causing miscomplied kernels
21:17 < jsaw> redo the thermal paste
21:17 < rxr> but it works reliable on my U5 ...
21:18 < jsaw> and clean all fans and fins
21:19 < rxr> then I asked David S. Miller and he said 3.3 should be fine now (formerly he only wanted his beloved egcspatch to be used for kernel builds)- I was usign 3.2 at that time - so I used the official debian kernel - and it freezes in equal intervals ...
21:19 < rxr> fins what ?
21:19 < jsaw> Kuehlerfinnen
21:19 < rxr> ah ;-)
21:19 < rxr> I do not think it is a thermal problem ..
21:20 < jsaw> did you run a memory checker?
21:20 < rxr> it most often freezes when some other huge power sucker is switched on ...
21:20 < rxr> jsaw: I do not have a SPARC memory checker
21:20 < rxr> e.g. when my athlon is power on the fridge or light is enough to get it down ...
21:20 < jsaw> when the thing gets to hot it's more likely to react on peaks in the supply voltage
21:21 < jsaw> too hot even
21:21 < rxr> I was away for over 24 hours - it did not freezed in that time perdion - but before I left 3 times in a row ...
21:21 < rxr> ouhm ...
21:21 < rxr> freeze even ..
21:21 < jsaw> I'd try cleaning first.
21:22 < jsaw> DAMN. libpython contains tcl symbols but is not linked to it...
21:24 < mnemoc> btw, i'll be away until monday
21:25 < rxr> oh - much fun then
21:25 < mnemoc> you will have three days to fix the stone bugs :p
21:26 < mnemoc> _Ragnar_: how did your patch worked?
21:26 < _Ragnar_> I'm at 5-linux26 now
21:27 < mnemoc> go go go
21:28 < _Ragnar_> of course the error doesn't appear is only libgcc.so is used ... for that --enable-shared has to be present, and gcc will also link the .so to all the binaries, so it doesn't matter where the symbols come from
21:28 < _Ragnar_> if*
21:29 < rxr> jsaw: I switched python to build a shared library a week ago or so ...
21:29 < rxr> jsaw: is that an in-system build that misses that change ?
21:30 < jsaw> no, the shared lib is fine. But ldd shows not tcl, while you can see lots of Tcl_* with objdump. That's why blender failed before.
21:30 < mnemoc> rxr: do you know why 0-gcc is --disable-shared?
21:30 < rxr> mnemoc: no
21:31 < mnemoc> that's what triggers our libgcc.a problem
21:31 < _Ragnar_> (one of the triggers)
21:31 < _Ragnar_> I think the main one is the configure mess in 1-gcc tho
21:31 < rxr> mnemoc: you could dive thru our history and see if it is present at the revision 1 initial import (which I think)
21:32 < _Ragnar_> I think it came from rock
21:32 < rxr> can you define which part of the mess ?
21:32 < rxr> or at least point to soem lines that "look messy" ?
21:33 < _Ragnar_> configure gets calls more then once ... the first call correctly identifies as/ld as .hidden-compatible and builds for a cross compiler
21:34 < _Ragnar_> the second call uses the as/ld of the target as test objects, and since those don't execute in the host system, the tests fail ... boom
21:35 < rxr> .oO
21:35 < _Ragnar_> I'm not sure I understand all of it
21:36 < _Ragnar_> anyway, --enable-shared for stage 0 and forcing libgcc_visibility=yes on stage 1 seems to do the job
21:36 < mnemoc> _Ragnar_: imagine that inside an 'ebuild'
21:36 < _Ragnar_> that's why I'm reluctant to switch to gentoo ;)
21:37 < rxr> ,-)
21:37 < rxr> gentoo would also not get you the "build systemwith uclibc", does it ?
21:38 < _Ragnar_> I thought they have that option
21:38 < mnemoc> afaik they use the toolchain created by buildroot
21:38 < _Ragnar_> ah okay ... so no pure uclibc system
21:39 < mnemoc> uclibc is not part of gentoo, but you can hack gentoo to use the official toolchain to build your gentoo box
21:41 < _Ragnar_> _Ragnar_:#rocklinux> anyone know why there's a --disable-shared in 0-gcc ?
21:41 < _Ragnar_> fake:#rocklinux> _Ragnar_: because 0-gcc only installs headers
21:42 < _Ragnar_> our 0-gcc does not seem to only install headers tho?
21:43 < rxr> *lol*
21:43 < rxr> that is not true ...
21:44 < _Ragnar_> ;D
21:44 < rxr> 0-gcc builds either a pseudo cross or real cross compiler ...
21:44 < rxr> in both t2 and rock ...
21:44 < rxr> 0-glibc only installs headers, ...
21:44 < rxr> _Ragnar_: there are reasons why I left rock .. ;-)
21:45 < _Ragnar_> *lol*
21:45 < _Ragnar_> yeah I just checked on a older rock-build, they also build stuff on 0-gcc
21:48 < mnemoc> _Ragnar_: how did you got here?
21:48 < _Ragnar_> huh?
21:49 < _Ragnar_> I read renes announcement on the rock ml
21:51 < rxr> svn cat -r2 http://www.rocklinux.net/svn/rock-linux/rock-trunk/package/base/gcc3/gcc3.conf | less
21:52 < mnemoc> cvs history is completely lost?
21:53 -!- mipe [~mipe@dsl10040.japo.fi] has quit ["party"]
21:53 < rxr> we never had a real
21:53 < rxr> CVS history
21:54 < rxr> it was only used for a few weeks ++
21:54 < rxr> before that clifford did it like the bsah maintainer
21:54 < rxr> private tree, collecting patches
21:54 < rxr> releases every now and then ...
21:54 < rxr> later that private snapshot was made available via rsync ..
21:55 < rxr> _Ragnar_: thinking about it, I think it is save to not specify --disable-shared for the 0-gcc
21:55 < _Ragnar_> :)
21:58 < mnemoc> Error logs from uminimal-2.1.0-beta3-x86-pentium-32-mnemosyne-expert:
21:58 < mnemoc> [5] network/iptables [5] network/lftp
21:58 < mnemoc> 142 builds total, 140 completed fine, 2 with errors.
21:58 < mnemoc> minimal/uclibc
21:58 < rxr> ;-)
22:00 < _Ragnar_> ;) nice
22:00 < mnemoc> no-NLS
22:03 < rxr> I consider patching binutils for sparc64 ...
22:03 < _Ragnar_> ?
22:03 < rxr> it does not accept s lot of inline assmebler used in X ...
22:04 < rxr> I just disabled all of it .. but I have no idea if DRI will work then ...
22:04 < _Ragnar_> O_o
22:04 < rxr> seems noone really cared about the SPARC assembly in mesa for a long time ...
22:05 < _Ragnar_> maybe most sparc64 are used as servers
22:05 < rxr> maybe not many sparc64 dists exists ... ;-)
22:05 < rxr> and users of those ;-)
22:06 < _Ragnar_> are they all running slowlaris?
22:06 < rxr> well not all ..
22:06 < rxr> there is this RedHat based audora
22:06 < rxr> 32bit userspace IIRC
22:06 < rxr> the others work even less - and not many use 64 bit userspace ...
22:09 < mnemoc> so you plain pure-64?
22:09 < mnemoc> do*
22:09 < mnemoc> err
22:09 < mnemoc> do you plan pure-64?
22:10 < rxr> I just do pure 64bit ...
22:10 < rxr> I worked on this all the time I sit in front of a SPARC
22:10 < rxr> just last week it materialized with all of the cleanup and bug hunting we did for t2 ...
22:11 < mnemoc> :D
22:11 < rxr> jsaw's flist wrapper fix was the last missing piece
22:11 < mnemoc> lftp: buffer.h:34:19: iconv.h: No such file or directory
22:12 < mnemoc> iptables: ip6tables.c: undefined reference to `in6addr_any'
22:25 < rxr> so - X mostly installed on the sparc64 via manually assembly disabling ...
22:25 < rxr> need to review what's wrong with the clobbers and registers used ...
22:26 < rxr> maybe it is also only 32bit assembly or so ..
22:26 < rxr> we'll see - for my first build some DRI disabled X is enought
22:26 < rxr> considering on my current 32bit user-space SPARCs the DRI for the Sun ffb it not working anyway
22:27 < rxr> (due to simple fact that noone so far bothers to port it to the latest DRI / DRM API ...
22:27 < rxr> it fails for the first version check as far as I can see ...
22:31 < _Ragnar_> hmmm
22:36 < rxr> >> The famous Pentium bug, which reportedly cost Intel 470 million dollars,
22:36 < rxr> >> was due to an incorrectly generated table (the table was originally
22:36 < rxr> >> correct, but the script to embed the table into a programmable logic array
22:36 < rxr> >> had a bug). Five positions in a lookup table used by the microcode for
22:36 < rxr> >> the division instruction had the wrong value.
22:41 < _Ragnar_> O_o
22:49 < _Ragnar_> mnemoc: is it possible that building with uclibc takes a lot longer then with normal glibc?
23:08 < jsaw> ccache?
23:09 < rxr> brrr
23:09 < rxr> cold
23:10 < rxr> brr 2.4°C
23:11 < rxr> maybe I should place the U30 on the window sill
23:12 < rxr> I think I'll fix mkdosfs for PowerPC to be able to create a fat on my usb stick in favour of doing the dos assembly exercies myself and use some copy from the internet ...
23:16 < _Ragnar_> *lol*
23:18 < rxr> hey - I DOS assembly was my fun stuff in 1990 - not in todays world ...
23:18 < rxr> have I mentioned that this university just bores me ...
23:18 < jsaw> not since a couple of days
23:20 < rxr> at least in system programming I'm ahead of the exercises, because the shell, awk or C hacking can always be done within minutes ...
23:20 < rxr> and I always nearly do them on the day he releases the next two for the next 14 days or so ..
23:21 < _Ragnar_> *hrhr*
23:22 < CIA-12> jsaw * r5764 /trunk/package/python/python/python.conf: * clean up python, tkinter stuff is not needed (anymore)
23:27 < rxr> does this fix your ooo and blender builds ?
23:28 < jsaw> he... we'll see what happens...
23:28 < jsaw> ... in a few hours...
23:28 < rxr> oh
23:28 < jsaw> Expected build time for ooo: >2h
23:29 < rxr> like I see in a few hours what 5-xorg and 5-gcc do on the sparc ...
23:29 < rxr> !> mkdosfs.c: undefined reference to `errno'
23:29 < jsaw> actually I should say >>2h
23:29 < jsaw> hehe
23:29 < jsaw> #include
23:29 < rxr> if it does not freeze again before ...
23:30 < rxr> ack
23:34 < rxr> I guess/hope 5-qt is also fixed on sparc due to your flist wrapper fix ..
23:35 < rxr> the last time it crashed mysteriously generting some .ui file or so, too ...
23:36 < jsaw> hmmm
23:44 < CIA-12> rene * r5765 /trunk/package/filesystem/mkdosfs/ (3 files):
23:44 < CIA-12> * updated mkdosfs (0.4.1) to dosfstools (2.9) - which contains a
23:44 < CIA-12> dosfsck and works slightly better on PowerPC - a rename will follow
23:45 < CIA-12> rene * r5766 /trunk/package/filesystem/mkdosfs/compile.patch:
23:45 < CIA-12> * added forgotten mkdosfs/compile.patch
23:45 < CIA-12> (never do a svn co without svn st)
23:45 < mnemoc> re
23:45 < rxr> hi mnemoc
23:47 < mnemoc> hi rxr
23:47 < jsaw> hi mnemoc
23:47 < mnemoc> hi jsaw
23:48 < mnemoc> _Ragnar_: i don't feel uclibc builds longer than glibc, but you can try the same pkgsel with both libc and compare the times on cache file
23:51 < rxr> or the time meassured with time ...
23:57 < CIA-12> rene * r5767 /trunk/package/filesystem/ (6 files in 2 dirs):
23:57 < CIA-12> * renamed mkdosfs -> dosfstools (remeber to rm the old flist to avoid
23:57 < CIA-12> shared file conflicts ...)
23:57 < CIA-12> * fixed dosfstools rebuild
23:57 < CIA-12> * updated the dosfstools.cache file
--- Log closed Sat Jan 22 00:00:13 2005