--- Log opened Sun May 01 00:00:48 2005
00:14 -!- keinek [~Keinek@201.254.13.188] has quit [Remote closed the connection]
01:59 < mnemoc> where Times-Roman comes from?
03:36 -!- keinek [~Keinek@201.254.13.188] has joined #t2
03:36 < keinek> hi
03:53 < mnemoc> rehi
03:59 < sparc-kly> rerehi
03:59 < sparc-kly> lol
04:00 < mnemoc> :)
04:04 -!- sparc-kly [~sunultra@65-23-205-193.prtc.net] has quit ["Leaving"]
06:01 -!- keinek [~Keinek@201.254.13.188] has quit [Remote closed the connection]
08:18 -!- mipe [~mika@dsl10040.japo.fi] has joined #t2
08:34 < mnemoc> rxr: why usrgames/ instead of usr/games/ ?
08:36 < mipe> hi mnemoc
08:36 < mnemoc> hi mipe
08:36 < mnemoc> rxr: only xblast in there
08:41 < mipe> hmm why is fmod "disabled" from gawk with dietlibc?
08:41 < mnemoc> svn log ?
08:43 < mnemoc> uhm... why is the comment on dietlibc/pkg_patch/pkg_gawk.patch line _47_ ?
08:44 < mnemoc> mipe: where did you see it disabled?
08:45 < mipe> mnemoc:i guess its not disabled from awk, i guess its at dietlibc
08:45 < mipe> mnemoc:i just wondered why ncurses fails, gawk returns no fmod yet on build
08:47 < mnemoc> i guess it's just not yet implementend
08:47 < mnemoc> does dietlibc buildlog say something about it?
08:49 < mipe> nope,have to re-check it.
08:50 < mnemoc> i can be on fefe's cvs... that should be an small patch
08:52 < mnemoc> it*
08:52 < mnemoc> i have to go :(
08:52 < rxr> == 05/01/05 08:53:54 =[5]=> Finished building package qt.
08:52 < rxr> ^-sparc32
08:52 < rxr> cu mnemoc
08:52 < mnemoc> cu in the 'morning'
08:52 < mnemoc> rxr: :D
08:53 < rxr> yeah - overslept a bit
08:53 < rxr> mipe: because dietlibc has not a "full" libm
08:53 < rxr> why does ncurses fail?`never did here?
09:03 < rxr> 649 builds total, 286 completed fine, 2 with errors.
09:03 < rxr> ^- Error logs from default-2.2.0-alpha-sparc-v8-desktop:
09:05 < mipe> rxr: ncurses fails because:echo | gawk -f ./base/MKunctrl.awk >unctrl.c
09:06 < mipe> rxr: i suspect dietlibc doesnt have fmod for x86-64
09:07 < mipe> altought gawk gets fmod in -lm while configure
09:11 < mipe> well that isnt arch dependant i think so... i'm lost :(
09:15 < mipe> ah yes,no fmod in x86-64 dietlibc
09:20 < rxr> hm - for x86 it was present
09:20 < rxr> it was so far missing for powerpc and sparc
09:20 < mipe> yep its there for x86
09:20 < rxr> and gawk returns some error?
09:21 < rxr> like undefined referende fmod?
09:21 * rxr breakfast
09:21 < mipe> nope,it returns: no fmod yet while running
09:21 < mipe> on configure it finds it from libm but its not implemented for x86-64
09:29 < rxr> hm
09:29 < rxr> hack the configure a bit more to not find it ,-)
09:29 < rxr> e.g. completly skip it for dietlibc compilation for now
09:58 < mipe> var_append patchfiles ' ' 'no-fmod-x86-64.diff'
09:58 < mipe> whats wrong with this
09:59 < rxr> mipe: just add it to the existing dietlibc gawk patch ...
10:01 < mipe> hmm, my patch removes the check completly... since i couldnt figure out how to check if using dietlibc at configure
10:32 < CIA-9> rene * r8606 /branches/2.1/package/base/linux26/linux26.desc: * updated linux26 (2.6.11.7 -> 2.6.11.8)
10:33 < CIA-9> rene * r8607 /branches/2.1/architecture/share/kernel-common.conf.m4: * made sure usb-storage and usblp are modular
10:35 < CIA-9> rene * r8608 /trunk/architecture/share/ (kernel-block.conf.m4 kernel-common.conf.m4): * flagged usb-storage and usblp modular, likewise
10:36 < CIA-9> rene * r8609 /trunk/package/base/linux26/linux26.desc: * merged r8606 from banches/2.1: updated linux26 (2.6.11.7 -> 2.6.11.8)
10:37 < rxr> user 0m0.051s
10:38 < rxr> ^- parsing /sys/ and /proc for pci devices, parsing /proc for usb devices and matching the pci ones twice and the usb ones once ... ,-)!
10:38 < rxr> (In C++ of course ..
10:38 < rxr> )
10:38 < rxr> user 0m0.047s
10:38 < rxr> ^- With -O2 instead of -O0 -ggdb ,-)
10:41 < rxr> ok - isapnp and firewire are waiting ...
11:35 < CIA-9> sebastian * r8610 /branches/2.1/package/graphic/ (hugin/hugin.desc libpano/libpano.conf libpano/libpano.desc): * changed libpano to use $JAVA_HOME, changed [P] to build after jdk
11:39 < rxr> gcc list:
11:39 < rxr> > Now, unless I've done something drastically wrong, it appears as if we
11:39 < rxr> > are spending 2/3 of our time in the libtool script.
11:39 < rxr> Yes, that's right. That's similar to what my oprofile experiments suggest.
11:39 < rxr> isn't that crazzy ?!?!
11:39 < rxr> ths damn libtool and auto* crap ..
11:41 < CIA-9> sebastian * r8611 /trunk/package/graphic/ (hugin/hugin.desc libpano/libpano.conf libpano/libpano.desc): * merged r8610 from banches/2.1: changed libpano to use $JAVA_HOME, changed [P] to build after jdk
11:43 < mipe> call for replacement of libtool...
11:44 < mipe> +if [ "${build_cpu}" = "x86_64" ] &&
11:44 < mipe> + [ "${build_os}" = "linux-gnudietlibc" ]; then
11:44 < mipe> + ac_cv_lib_m_fmod=no
11:44 < mipe> +fi
11:44 < mipe> good or bad? :)
11:45 < rxr> I would just set ot always to no
11:45 < mipe> x86 has fmod at dietlibc
11:45 < rxr> so we have the same behaviour on all platforms and know that dietlibc has to be fixed
11:45 < mipe> ah ok
11:45 < rxr> mipe: it makes no sense to use s.th. on x86 when it works on no other arch ...
11:46 < rxr> when we need it we need to fix all supported architectures ...
11:46 < mipe> so i'll set it to no, for every arch when using dietlibc
11:47 < rxr> yep
11:58 < rxr> so - I think I go refactor the hotplug++ code now - and extend my EmbeddedSTL until it has all the features currently utilized by hotplug++
12:08 < mipe> http://koti.japo.fi/~pmika/svn-2.1.diff
12:09 < mipe> testing finished... contains dietlibc x86-64 stat fix and gawk no fmod on dietlibc fix
12:11 < rxr> mipe: bootdisk works now ?
12:11 -!- sepp [~sepp@p213.54.215.174.tisdip.tiscali.de] has joined #t2
12:11 < mipe> rxr:i'm not that far yet,its building at stage 3
12:11 < sepp> hi all
12:12 < mipe> rxr: those are the fixes that i needed to get pass stage 2
12:12 < mipe> hi sepp
12:14 < rxr> mipe: the dietlibc diff makes tar work fine?
12:14 < rxr> glibc has
12:14 < rxr> /usr/include/bits/typesizes.h:#define __TIME_T_TYPE __SLONGWORD_TYPE
12:14 < rxr> so I signed time_t should be ok
12:14 < rxr> however dietlibc might be better off using a time_t for internally ...
12:15 < rxr> hi sepp
12:15 < rxr> mipe: you have my ok to commit those changes if you feel to
12:15 < mipe> rxr:yep, i looked over what glibc had for data types, but i didnt notice that dietlibc used time_t, so i just putted signed long,which is time_t
12:16 < mipe> well signed 16bit
12:16 < rxr> I did not say dietlibc has time_t - I saied it might be better off with one ...
12:17 < mipe> yep
12:17 < rxr> it has time_t
12:17 < rxr> /usr/dietlibc/include/utime.h: time_t actime; /* access time */
12:17 < rxr> /usr/dietlibc/include/utime.h: time_t modtime; /* modification time */
12:17 < rxr> it even does use it ...
12:17 < sepp> is it ok if i commit a sun-jdk update to 2.1? i updated jdk-1_5_0_02 to jdk-1_5_0_03 because they removed the 02 ...
12:18 < rxr> sepp: sure
12:19 < mipe> rxr:yep i know it has time_t but i just compared to the other stat structs in other archs there, which they didnt have it. and i thought it would be nicely consisted with others if i put signed. i'll commit if i remember my password
12:20 < rxr> mipe: it seems you need to commit more often in order to not forget your passwd ,-)
12:20 < mipe> :)
12:20 < CIA-9> sebastian * r8612 /branches/2.1/package/java/sun-jdk-150/sun-jdk-150.desc: * updated sun-jdk-150 (1.5.0_02 -> 1.5.0_03)
12:24 < CIA-9> mika * r8613 /branches/2.1/package/base/dietlibc/ (pkg_patch/pkg_gawk.patch x86_64-stat-fix.patch):
12:24 < CIA-9> * dietlibc, changed stat to use
12:24 < CIA-9> signed long in mtime,atime and ctime.
12:24 < CIA-9> fix gawk not to use dietlibc fmod in
12:24 < CIA-9> any arch
12:27 < valentin> moin
12:27 < valentin> who is mika ? :)
12:27 < rxr> sepp: you do not forget to merge into trunk, right ?
12:27 < rxr> valentin: mipe
12:27 < rxr> moin valentin
12:28 < valentin> aha
12:28 < sepp> rxr: no :)
12:33 -!- minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has joined #t2
12:33 < minto> hi
12:34 < valentin> hi
13:14 < valentin> puh, 23 C
13:15 < valentin> they say it will be up to 29 C tomorrow
13:26 < mipe> where? in Germany? its barely 10 C (fin), cold as ....
13:28 < valentin> yes, here in Berlin
13:28 < valentin> can you send me some cold air for my room ?
13:29 < mipe> sure,i'll bottle some cold air and send it to you
13:29 < valentin> thank you :)
13:30 < valentin> maybe i'll return the bottle after i exchanged the air. then you'll have something warm.
13:30 < valentin> but beware - berlin air is not too good at all.
13:31 < rxr> what! it is so lovely over here
13:31 * rxr taking a deep breath
13:31 < valentin> rxr: feinstaub...
13:31 < rxr> .oO(uh - huh - huestel)
13:31 * rxr wasn't the best idea ,-)
13:31 < rxr> valentin: just joking ..
13:32 < valentin> and ozone
13:32 < valentin> ah, feinstaub == particulate matter
13:33 < valentin> at least no riot here today (like last year on may, 1st)
13:38 < rxr> 649 builds total, 297 completed fine, 2 with errors.
13:38 < rxr> ^- U5 progressing sparc32
13:43 < valentin> rxr: nice
13:53 < CIA-9> msluis * r8614 /trunk/package/develop/ltmm/ (. ltmm.desc): * added libtool-- a.k.a. ltmm (1.5)
13:56 < rxr> minto: what is that ?!?
13:56 < minto> rxr: a C++ wrapper around libltdl
13:57 < minto> libtool dynamic loading stuff
13:59 < valentin> the name is confusing - when i hear libtool i think about that command line utility ...
14:00 < minto> libtool is more than just a command line utility. Look at http://www.gnu.org/software/libtool/manual.html#Dlopened-modules
14:02 < minto> Also have look at the next paragraph ( http://www.gnu.org/software/libtool/manual.html#Using-libltdl )
14:03 < minto> This dynamic library loading stuff comes for free with the libtool package. Libtool-- is just a C++ wrapper around to make things easier.
14:05 < valentin> maybe i take a lool when i do some dynamic loading in the next c++ project
14:05 < valentin> look even
14:13 -!- sparc-kly [~sparc@65-23-205-193.prtc.net] has joined #t2
14:34 -!- sparc-kly [~sparc@65-23-205-193.prtc.net] has quit ["Leaving"]
14:34 < CIA-9> msluis * r8615 /trunk/package/network/jabberd/jabberd.conf: * fixed bug #20 by using package --with-extra-xx options.
14:36 * minto is going outside to enjoy the weather.
14:36 < minto> bbl
14:36 < rxr> cu minto
14:37 < rxr> Cpu(s): 0.7% user, 4.6% system, 0.0% nice, 0.0% idle, 94.7% IO-wait
14:37 < rxr> ^- 94% IO-wait ! ...
14:37 < rxr> building kdelibs on teh 128MB U5 ...
14:37 < rxr> Mem: 125160k total, 124128k used, 1032k free, 96k buffers
14:37 < rxr> Swap: 262128k total, 131064k used, 131064k free, 3560k cached
14:37 < rxr> 130MB of swap in use ...
14:38 < rxr> guess organzing some RAM for the box would be a good idea ..
14:39 -!- sparc-kly [~sunultra@65-23-205-193.prtc.net] has joined #t2
14:44 -!- sparc-kly [~sunultra@65-23-205-193.prtc.net] has quit ["Leaving"]
14:47 < rxr> 168pin EDO ECC DIMM
14:48 < rxr> http://www-1.ibm.com/servers/eserver/pseries/hardware/entry/b50_specs.html
14:48 < rxr> I wonder if the IBM B50 RAM will fit in the U5 ...
14:48 < rxr> I haev 768 or so MB in it - and it is barely needed ...
14:49 < rxr> dropping that to 512MB and plugging the reminder into the U5 would be a nice deal ...
14:51 -!- sparc-kly [~sunultra@65-23-205-193.prtc.net] has joined #t2
14:53 -!- minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has quit [Read error: 110 (Connection timed out)]
14:56 < rxr> interesting the RS6000 B50 handbook from IBM already listed that Linux (2.2. at that time) is available for it ...
14:58 * rxr offline - to take out RAM off the RS6K router ... *g*
14:58 < rxr> cu
--- Log opened Sun May 01 16:15:44 2005
16:24 < CIA-9> mika * r8616 /branches/2.1/package/network/portmap/portmap.conf: * portmap, pass WRAP_DIR=$libdir to makeopt
16:26 < mipe> well that was wierd...
16:26 < mipe> kiss just builded when i included debug info there
16:28 < rxr> 60ns 168-pin EDO 3-Volt JEDEC buffered DIMM (32MB, 64MB, or 128MB and only in pairs)
16:28 < rxr> mipe: hm
16:28 < mipe> when is 2.1-final(or next iso release) date
16:29 < mipe> i can provide this minimal-minimal thing, unless i have time to do desktop build
16:30 < mipe> export SDECFG_DEBUG='1'
16:30 < mipe> it builds but with 0, it doesnt
16:36 < rxr> well - test the resulting ISO before you investigate kiss I would say ...
16:39 -!- Netsplit saberhagen.freenode.net <-> irc.freenode.net quits: jsaw, CIA-9
16:39 -!- Netsplit over, joins: CIA-9, jsaw
16:40 < mipe> at bootdisk target build_stage2.sh what does those - and + signs mean at package_map list? i need to add glibc32 to it...
16:40 -!- minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has joined #t2
16:41 < minto> Hi, I am back. pffff it's just too hot outside.
16:44 < mipe> hmm, what emulator can support x86-64?
16:48 < rxr> bochs
16:48 < rxr> maybe qemu ..
16:48 < rxr> mipe: just try booting it *grr* ?
16:49 < mipe> rxr: yep
17:29 < rxr> mipe: and? succees??
17:35 < valentin> re
17:45 < rxr> valentin: btw. regarding the not too nice cout.setf (std::ios_base::hex, ...) there are of course manipulators that make that easier
17:45 < rxr> std::cout /* or your favourite stream */ << std::hex << your_value << std::endl;
17:46 < rxr> I'll have slightly cleaner ostream and istream soon
17:46 < rxr> (not yet as template ...)
17:46 < rxr> that will follow when I have too much time
17:46 < rxr> after the {i,o}stream revisiting I'll do the fstream
17:47 < valentin> rxr: i know about manipulators
17:48 < valentin> but i still think printf ("abc: %.4f %04d %s", f, i, s); is way shorter than the c++ variant
17:49 < valentin> i would like to have sth like cout << format (" some weird format string") << a << b << c
17:49 < valentin> that would be nice
17:50 < valentin> esp. separation strings like << " " << you have between each ouput variable are quite anoying
17:52 < mnemoc> .oO( did anyone tested that stuff merged to 2.1? )o
17:52 < valentin> hi mnemoc
17:53 < rxr> valentin: maybe it is possible to hack such kind of state machine manipulator that pops to the next token on each input is possible
17:53 < rxr> should I try one day to implement such a beast ?
17:53 < valentin> if you want :)
17:54 < valentin> the question is: what to do if the argument type does not match the format ?
17:54 < mnemoc> hi valentin, rxr
17:54 < rxr> valentin: throw logic_exception ,-)
17:54 < rxr> and put the stream in the bad state ...
17:54 < rxr> hi mnemoc
17:55 < mnemoc> sepp: can you share with me those sun-java updated files?
17:57 < mipe> rxr:not tested yet,i'll test them in a hour or so. need to pack my things,will be out of town for couple of days
17:57 < sepp> hi mnemoc
17:57 < sepp> share what?
17:58 < mnemoc> * updated sun-jdk-150 (1.5.0_02 -> 1.5.0_03) <-- these files
17:59 < sepp> the the jdk*.bin ?
18:00 < mnemoc> yes :)
18:01 < sepp> hmm why, you can download it on java.sun.com
18:01 < mnemoc> doesn't work with links
18:01 < sepp> ohh
18:01 < mnemoc> and downloading to home to upload it there is hard with a 64k link
18:02 < mnemoc> but if you can't, don't worry
18:03 < sepp> i can - how?
18:03 < mnemoc> http:// or ftp:// it somewhere?
18:04 < mnemoc> where i could curl/wget them
18:04 < sepp> you can wget http://192.18.97.107/ECom/EComTicketServlet/BEGIN5EF24F950A82720DFD40CA6323198354/-2147483648/840192075/1/615302/615254/840192075/2ts+/westCoastFSEND/jdk-1.5.0_03-oth-JPR/jdk-1.5.0_03-oth-JPR:11/jdk-1_5_0_03-linux-amd64-rpm.bin
18:04 < sepp> ;)
18:04 < mnemoc> let see
18:04 < mnemoc> nice
18:04 < sepp> and http://192.18.97.107/ECom/EComTicketServlet/BEGIN5EF24F950A82720DFD40CA6323198354/-2147483648/840192075/1/615302/615170/840192075/2ts+/westCoastFSEND/jdk-1.5.0_03-oth-JPR/jdk-1.5.0_03-oth-JPR:4/jdk-1_5_0_03-linux-i586.bin
18:05 < sepp> 1st one was wrong, sorry
18:05 < sepp> should be http://192.18.97.107/ECom/EComTicketServlet/BEGIN5EF24F950A82720DFD40CA6323198354/-2147483648/840192075/1/615302/615266/840192075/2ts+/westCoastFSEND/jdk-1.5.0_03-oth-JPR/jdk-1.5.0_03-oth-JPR:12/jdk-1_5_0_03-linux-amd64.bin
18:06 < mnemoc> i tought i need to auth before :)
18:06 < mnemoc> uhm....rpm
18:06 < mnemoc> abort abort abort
18:06 < sepp> i have only 9kb/s upstream so this would take ages
18:06 < mnemoc> :)
18:07 < sepp> > 90mb both things together
18:09 < mnemoc> jdk-1_5_0_03-linux-i586.bin' at 2637368 (5%) 28.9K/s eta:26m [Receiving data]
18:09 < sepp> :)
18:10 < mnemoc> :)
18:21 < mnemoc> sepp: libpano way to get JAVA_HOME is not ok
18:21 < mnemoc> rxr: does chroot load profile.d/ stuff?
18:22 < sepp> not the debug.sh
18:22 < mnemoc> sepp: what about pkgprefix ?
18:22 -!- mipe [~mika@dsl10040.japo.fi] has quit ["testing"]
18:23 < sepp> i thought this way any jdk should work
18:24 < sepp> ibm blackdown sun ...
18:24 < mnemoc> any? you ask directly for profile.d/jdk
18:26 -!- mipe [~mika@dsl10040.japo.fi] has joined #t2
18:26 < sepp> yes, all jdks do . $base/package/*/ibm-jre-142/ibm-jre-142.conf
18:26 < sepp> and so set profile.d/jdk
18:27 < sepp> set $JAVA_HOME ...
18:27 < mnemoc> so all jdks share the same profile.d entry now?
18:28 < sepp> minto: ?
18:30 < mnemoc> is would prefer a shared java-conf.in
18:30 < mnemoc> with: if pkginstalled ; then
18:31 < mnemoc> pkgprefix -t
18:31 < mnemoc> JAVA_HOME=$( pkgprefix )
18:31 < mnemoc> elif pkginstalled ; then
18:31 < mnemoc> ...
18:31 < mnemoc> everything well defined, nothing vague
18:32 < mnemoc> and we source that java-conf.in from every java dependent package
18:37 < rxr> not too many errors left in order to get my module matching linked with EmbeddedSTL ,-)
18:38 < mnemoc> :)
18:49 < rxr> there is no C equivalent for the std::istream::peek? is there ?
18:50 < mipe> can svn handle symlinks?
18:51 < rxr> mipe: only 1.1 IIRC
18:51 < rxr> however I never tried
18:51 < rxr> why would you need it?
18:51 < mipe> target/bootdisk/x86 --> x86-64
18:51 < mipe> fastest way,but not the best way
18:52 < mipe> i guess its time to add a check to build.sh
18:53 < valentin> rxr: there is
18:54 < valentin> or at least the functionality with getc/ungetc
18:56 < valentin> btw: is there a complete and exact online stl reference ?
18:56 < valentin> like this: http://www.cplusplus.com/ref/ (<- some parts like iostream only)
18:59 < rxr> valentin: why not the ANSI standard as we have in PDF ,-)?
19:00 -!- mipe [~mika@dsl10040.japo.fi] has quit ["testing"]
19:00 < minto> mnemoc sepp: Hi
19:00 < minto> Regarding the profile.d all jdk's set profiled./jdk and all jre's profile.d/jre
19:01 < valentin> rxr: because "browsing" a pdf sucks
19:01 < minto> pkginstalled should be fine but what package to check for? I don't like a bunch of pkginstalled checks, one for every jdk installed.
19:02 < rxr> with kpdf 3.4 searching is quite convienent
19:02 < rxr> just that the old PDF has no idex is a bit annoying ...
19:03 < minto> Having a java-conf.in is also fine, but where should it go? I don't like the idea of adding it to one to the jdk's since this would have all other jdk's depend on the first one.
19:05 < rxr> some state class improvements and it the match code will build - just that I have a lot to implement then ...
19:05 < rxr> so far mostly protyping ...
19:06 < rxr> (on the EmbeddedSTL site)
19:06 -!- sparc-kly [~sunultra@65-23-205-193.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
19:07 < rxr> 11 files changed, 648 insertions(+), 225 deletions(-)
19:07 < valentin> http://www.sgi.com/tech/stl/
19:09 < mnemoc> minto: hi
19:10 < mnemoc> minto: for ${foo}-conf.in we have to choose one, i think sun-jre-142.
19:10 < mnemoc> minto: if all jdks write into profile.d/jdk you can install two because the second will fail
19:12 -!- mipe [~mika@dsl10040.japo.fi] has joined #t2
19:12 < minto> mnemoc: You want a separate profile.d/xxx for every jdk and jre?
19:12 < mnemoc> minto: we should do the same for jdk-s sourcing ibm's .conf. that is ilegal
19:13 < mnemoc> minto: you _can't_ share a file between two packages which can be installed at the same time
19:15 < minto> mnemoc: installing two jdk with separate profile.d/xxx would make things unpredicatable. Both profile.d/xxx set JAVA_HOME. But which one is last? This is why I want some alternative system.
19:15 < sepp> they all need JAVA_HOME ?
19:16 < minto> It is not that the all need JAVA_HOME. Other java package need it to point to the jdk to be used.
19:16 -!- sepp [~sepp@p213.54.215.174.tisdip.tiscali.de] has quit [Read error: 104 (Connection reset by peer)]
19:18 -!- sepp [~sepp@p213.54.224.184.tisdip.tiscali.de] has joined #t2
19:18 < mnemoc> this whole thing is ugly
19:18 < minto> sepp: It is not that the all need JAVA_HOME. Other java package need it to point to the jdk to be used.
19:18 < minto> mnemoc: What is the alternative?
19:18 < mnemoc> minto: we can't allow shares
19:19 < mnemoc> minto: we choose explicitly one for buildings
19:19 < mnemoc> minto: on runtime, the user have to decide what JAVA_HOME to use if he wants to use especificly one
19:20 < mnemoc> if not, the last JAVA_HOME set will win
19:20 < minto> mnemoc: build time needs JAVA_HOME too, I am afraid?
19:20 < mnemoc> minto: java-conf.in solves that
19:20 < mnemoc> minto: using explict priorization
19:21 < mnemoc> we can do it this way:
19:21 < minto> Should all java packages (not only jre and jdk) use java-conf.in?
19:21 < mnemoc> yes
19:22 < mnemoc> like all mono packages use mono-conf.in
19:22 < mnemoc> and all gnome2 packages use gnome2-conf.in
19:22 < minto> Okay, let's do it.
19:22 < mnemoc> we can do it looping
19:22 < mnemoc> for jdk in $1 ibm-150 sun-150 another-crap; do
19:22 < minto> But I still want an alternative system some day :-)
19:22 < mnemoc> if pkginstalled $jdk; then
19:23 < mnemoc> minto: sure
19:23 < mnemoc> pkgprefix -t $jdk
19:23 < mnemoc> JAVA_HOME=$( pkgprefix $jdk )
19:23 < mnemoc> break
19:23 < mnemoc> fi
19:23 < mnemoc> done
19:24 < mnemoc> and when sourceing java-conf.in we can tell if is supports gcj or kaffe
19:24 < mnemoc> . $base/package/java/*/java-conf.in kaffe
19:24 < mnemoc> so it will prefer kaffe if it's available
19:25 < mnemoc> $* instead of $1 obviously
19:26 < minto> mnemoc: I am still troubled by the profile.d stuff.
19:26 < mnemoc> we can use the same token than on prefix for profile.d/ entry
19:26 < mnemoc> minto: yes, runtime is another issue
19:26 -!- tilix [~tilix@212.116.139.141] has joined #t2
19:26 < mnemoc> minto: but the last that set JAVA_HOME will win
19:26 < mnemoc> minto: that wont produce errors
19:26 < minto> ack
19:27 < mnemoc> minto: user can explicitly source profile.d/myprefered on his .profile
19:28 < mnemoc> java stuff deserves large documentation on our CMS
19:29 < minto> ack
19:29 < mnemoc> good to see we have an agreement :)
19:29 < minto> Yepp
19:30 < tilix> I can translate the new website to Bulgarian. Is it possible?
19:30 < mnemoc> sure
19:30 < tilix> What should I do?
19:30 < mnemoc> wait until jeru revives and finish i18n support :)
19:31 < tilix> ok
19:33 < tilix> is it nessesary for the bootdisk target to build both kernels (2.4 and 2.6)?
19:33 < mnemoc> tilix: no, you can disable any using pkgsel
19:34 < mnemoc> tilix: Config -> Expert -> package sleection -> '- linux24'
19:34 < tilix> thanks
19:36 < mnemoc> - will kill it
19:36 < mnemoc> O just disable it
19:36 < mnemoc> and X enable
19:37 < mnemoc> i have to go
19:37 < mnemoc> cu
19:38 < minto> cu mnemoc
19:39 < tilix> Can't open input file /home/ilia/t2-trunk/download/mirror/s/skas-2.6.11-v9-pre1.patch.bz2: No such file or directory.
19:39 < tilix> How should I download it?
19:44 < tilix> i got it manually
19:48 -!- tilix [~tilix@212.116.139.141] has left #t2 ["Konversation terminated!"]
20:16 -!- mipe [~mika@dsl10040.japo.fi] has quit ["Leaving"]
21:02 -!- ojh [~omer@208-45-96-152.eugn.qwest.net] has joined #t2
21:04 < ojh> Quick question: How do I force the T2 install disk to load the 2.6 kernel. I have a laptop that locks up on 2.4.
21:05 < rxr> install disk or install CD ?
21:06 < rxr> which version of it ?
21:06 < rxr> the latest ones boot 2.6 by default
21:26 < sepp> hmmm what means: Kernel configuration found, performing sanity checks ... Kernel configuration lacks needed options, please correct! ABORTING.
21:26 < sepp> i use a custom kernel config
21:27 < valentin> sepp: is that error produced by the kernel or by t2 scripts ?
21:28 < valentin> s/kernel/kernel makefile
21:28 < sepp> looks like the kernel
21:29 < sepp> last line before the error is: build_this_package:810 (last $?=0)> make ARCH=i386 CROSS_COMPILE= KCC=kcc ...
21:31 < sepp> it does not abort without a custon kernel config
21:31 < rxr> sepp: most problably 3rd party modules from some other package are build
21:31 < rxr> scroll a bit up to see which sub-package kernel module build this is in ,-)
21:32 < rxr> valentin: Strousrtup 21.10 Exercises 16. Design and implement a pattern-matching input operatoin. Use printf-style format strings to specify a pattern. ...
21:33 < ojh> rxr: Install CD.
21:34 -!- sparc-kly [~sparc@64.237.128.15] has joined #t2
21:35 < ojh> rxr: Thanks, my new (to me) HP laptop hangs on the 2.4 kernel. Been able to boot Knoppix with the 2.6 option and it runs sweet. :)
21:36 < valentin> rxr: nice Exercise - is the Stroustrup's solution available somewhere ? :)
21:36 < ojh> rxr: So I can only install a Linux that uses 2.6 at the install stage.
21:37 < rxr> ojh: most if not all of the T2 releases boot 2.6 by default
21:37 < sepp> ahh looks like acx100 makes problems
21:37 < rxr> which version do you have in front of you ?
21:37 < rxr> valentin: no idea - use google ?
21:37 < rxr> (or implement it - or wait for me to do it)=
21:38 < valentin> rxr: do not hurry
21:38 < ojh> rxr: Excellent then :)
21:39 < minto> valentin: maybe boost has something similar http://www.boost.org/libs/format/index.html
21:43 -!- ojh [~omer@208-45-96-152.eugn.qwest.net] has left #t2 []
21:44 < CIA-9> sebastian * r8617 /branches/2.1/package/java/sun-jre-150/sun-jre-150.desc: * updated sun-jre-150 (1.5.0_02 -> 1.5.0_03)
21:46 < minto> sepp: can you merge these jre and jdk changes into trunk as well.
21:46 < sepp> yes
21:46 < minto> thanx
21:46 < sepp> i only changed it because 02 is no longer online at java.sun.com
21:47 < minto> If I notices it changed I would have changed it.
21:48 < rxr> valentin: the first quick hack i/o and if/of -stream will directly use FILE* and c functions
21:49 < rxr> I'll implement the full streambuf and co framework sometime later
21:49 < rxr> at least not for the hotplug++ I want to have running with this embSTL by tomorrow ...
21:49 < CIA-9> sebastian * r8618 /trunk/package/java/ (sun-jdk-150/sun-jdk-150.desc sun-jre-150/sun-jre-150.desc): * merged r8617 and r8612, sun-jre-150 and sun-jdk-150 update
21:50 < minto> rxr valentin: Have a look at boost. I think it already does what you guys try to achieve.
21:50 < minto> example: std::cout << format("%|1$1| %|2$3|") % "Hello" % 3 << std::endl;
21:50 < minto> Look at http://www.boost.org/libs/format/index.html
21:50 < valentin> rxr: ok
21:51 < valentin> minto: *click*
21:53 < valentin> hm, they use the '%' operator, so they do not need to mess around with the stream
21:54 < valentin> okay, thanks
21:57 < sepp> hmm in the desktop target a custom kernel config is really useless if 3rd party mods need hamradio for example
21:59 < sepp> i guess there are not many users with hamradio support in the kernel conf
22:02 < rxr> 61644 May 1 22:01 linux-modmatch
22:02 < rxr> ouhm - not too tiny
22:02 < rxr> and not yet fully functional ...
22:03 < rxr> 57548 May 1 22:02 linux-modmatch
22:03 < rxr> ^- -Os -s
22:04 * sparc-kly need this http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=20328&item=5770454203&rd=1
22:10 -!- sparc-kly [~sparc@64.237.128.15] has quit [Read error: 54 (Connection reset by peer)]
22:24 -!- sparc-kly [~sunultra@64.237.128.15] has joined #t2
22:47 < sepp> rxr: i would remove the acx100 pkg from the desktop target, it breaks custom kernel config with strange hamradio deps ...
22:50 < rxr> I think built here
22:50 -!- sparc-kly [~sunultra@64.237.128.15] has quit ["Leaving"]
22:52 < sepp> it was the only thing i changed
22:53 -!- mtr [~michael@p54AFB1A0.dip0.t-ipconnect.de] has joined #t2
22:54 -!- sparc-kly [~sunultra@64.237.128.15] has joined #t2
22:55 -!- minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has quit ["Chatzilla 0.9.66 [Mozilla rv:1.7.6/20050318]"]
22:56 -!- sparc-kly [~sunultra@64.237.128.15] has quit [Client Quit]
23:00 -!- mtr_ [~michael@p54AFA12C.dip0.t-ipconnect.de] has quit [Read error: 145 (Connection timed out)]
23:09 -!- keinek [~Keinek@201.254.9.201] has joined #t2
23:09 < keinek> hi
23:17 * sepp found a old hdd and will test rc1 now ;)
23:18 < sepp> cu
23:18 -!- sepp [~sepp@p213.54.224.184.tisdip.tiscali.de] has quit ["leaving"]
23:33 -!- sparc-kly [~sunultra@64.237.128.15] has joined #t2
23:40 -!- sparc-kly [~sunultra@64.237.128.15] has quit ["Leaving"]
--- Log closed Mon May 02 00:00:44 2005