--- Log opened Sat Nov 14 00:00:17 2009
01:06 -!- mqueiros [n=mqueiros@217.70.75.151] has joined #t2
04:00 -!- synchris_ [n=synchris@adsl148-230.kln.forthnet.gr] has joined #t2
04:13 -!- synchris [n=synchris@pdpc/supporter/professional/synchris] has quit [Read error: 110 (Connection timed out)]
10:14 < mtr> moin
10:20 < mtr> gcc 4.4 targets tested, svn trunk:
10:21 < mtr> x86 -> x86: minimal-x
10:22 < mtr> x86 -> x86_64: minimal (cross)
10:22 < mtr> x86 -> arm: custom target (cross)
10:23 < mtr> builds are all fine, except an issue with the cross compiled perl which segfaults on x86_64 ...
10:23 -!- mqueiros [n=mqueiros@217.70.75.151] has quit ["Ex-Chat"]
10:24 < mtr> mpp: can you post your config and build logs?
10:26 < mj-netbits> moin
10:42 < mj-netbits> mtr can you test the i586 cross build ?
10:43 < mj-netbits> i found an error on building glibc
10:43 < mj-netbits> the patchfile x86-i586.patch
10:44 < mj-netbits> must be removed !!
10:44 < mj-netbits> if i remove the patch the build was successfully
10:45 < mj-netbits> he makes a double include in memcpy_chk.S! This is the error. I think with the new glibc this was incorrect.
11:50 < mtr> hi mj-netbits
11:51 < mtr> strange, on trunk I have in 1-glibc.log:
11:51 < mtr> Apply patch /work/T2/t2-clean/package/base/glibc/x86-i586.patch ...
11:51 < mtr> patching file sysdeps/i386/i586/memcpy_chk.S
11:51 < mtr> patching file sysdeps/i386/i586/mempcpy_chk.S
11:51 < mtr> patching file sysdeps/i386/i586/memset_chk.S
11:51 < mtr> Apply patch /work/T2/t2-clean/package/base/glibc/x86-i686.patch ...
11:51 < mtr> patching file csu/defs.awk
11:51 < mtr> patching file sysdeps/i386/bp-asm.h
11:52 < mtr> mj-netbits: which target do you test?
11:53 < mtr> mj-netbits: generic, desktop? custom?
11:55 -!- mpp [n=mpp@2a01:48:251:107:203:dff:fe13:6c9e] has joined #t2
12:41 -!- mpp [n=mpp@2a01:48:251:107:203:dff:fe13:6c9e] has quit ["good night - good fight"]
12:45 -!- LMJ [n=serwou@laf31-4-82-236-42-164.fbx.proxad.net] has quit ["Leaving"]
13:41 < mj-netbits> i use generic/bootstrap
13:54 < mj-netbits> i have tested now /x86 viac3, viac32, i586. Now i modify the gcc-options for viac7 (i think here are the wrong gcc options (he was only for older gcc-version)) and test it again.
14:12 < mtr> mj-netbits: yes, you're right,
14:12 < mtr> mj-netbits: the patch applies does lead to double inclusion of the memcpy*.S
14:13 < mj-netbits> yea :)
14:13 < mj-netbits> can you remove this from the svn ?
14:13 < mj-netbits> or must i make a new patchfile
14:14 < mj-netbits> i think we dosn't need this patch file. i think it is deprecated.
14:14 < mtr> yes, patch is obsolete and should be removed
14:15 < mj-netbits> perfekt
14:15 < mj-netbits> perfect :)
14:16 < mtr> would you like your name/email in the svn commit log?
14:16 < mj-netbits> no you can add yours
14:19 < CIA-17> mtr * r34886 /trunk/package/base/ (eglibc/x86-i586.patch glibc/x86-i586.patch):
14:19 < CIA-17> * removed obsolete glibc/eglibc i586 patches
14:19 < CIA-17> (as reported on irc)
14:20 < mj-netbits> :)
14:20 < mtr> mj-netbits: as you build on trunk - does the gcc-4.4 work on your side?
14:20 < mj-netbits> yes
14:20 < mj-netbits> no problem at the moment
14:21 < mtr> you're on x86 as build host?
14:21 < mj-netbits> but i only test cross for x86 and 64Bit
14:21 < mj-netbits> yes
14:21 < mj-netbits> i use ubuntu server 9.10
14:21 < mj-netbits> with 32Bit x86
14:22 < mtr> no x86_64 as build system?
14:22 < mj-netbits> in this days i test all via settings and add settings for the atom 230/270/330
14:22 < mj-netbits> no
14:22 < mj-netbits> not at home
14:22 < mj-netbits> on work i use 64Bit
14:22 < mtr> I guess there may be some glitches with the gcc update on 64 bit
14:23 < mj-netbits> hmm i can test on Monday morning
14:23 < mtr> same here, I still run 32-Bit on an athlon64
14:24 < mtr> rxr's reference builder is x86_64 too
14:24 < mj-netbits> i have a intel notebook (hp hdx 9350) with 32Bit ubuntu 9.10. at work i use a i7 with 4 Intel SSD so the test is a little faster :)
14:25 < mj-netbits> so if you need test on a x86 system i can help.
14:25 < mtr> :)
14:27 < mtr> rxr plan was to branch off 8.0 soon, and the gcc-4.4 update should not delay it too much
14:27 < mtr> if we got too many regressions he'll revert the gcc update on the 8.0 branch later on
14:27 < mj-netbits> hmm
14:28 < mj-netbits> gcc-4.4 would be nice
14:28 < mj-netbits> where is rxr?
14:28 < mtr> my last dektop build from scratch had >300 package build errors ...
14:28 < mj-netbits> oh
14:29 < mj-netbits> with gcc 4.4 ?
14:29 < mtr> no, before the update ...
14:29 < mj-netbits> so i think we need one more patch :)
14:30 < mtr> rxr is usually most time present here, if not he's likely abroad
14:31 < mj-netbits> next week i think i can spend more time for t2-project. so i can help you if you want.
14:36 < mtr> i guess rxr will tag a 8.0-RC first, and we should try to fix as many packages then
14:36 < mtr> any test build ressources are welcome ;-)
14:36 < mtr> there is nice report script: misc/archive/RegressionsReport.sh
14:36 < mtr> to generate package build statistics
14:40 < mj-netbits> oh nice
14:53 -!- mpp [n=mpp@2a01:48:251:107:203:dff:fe13:6c9e] has joined #t2
14:59 < mpp> gmp build error
14:59 < mpp> checking C++ compiler g++ ... no
14:59 < mpp> configure: error: C++ compiler not available, see config.log for details
15:00 < mpp> Due to previous errors, no 0-gmp.log file!
15:00 < mpp> (Try enabling xtrace in the config to track an error inside the build system.)
15:00 < mpp> --- BUILD ERROR ---
15:00 < mpp> Creating file list and doing final adaptions ...
15:00 < mpp> Searching for orphaned files ...
15:00 < mpp> Found 5 files for this package.
15:00 < mpp> Found 0 orphaned files for this package.
15:00 < mpp> Clear (old) md5sums ...
15:00 < mpp> Creating md5sum files ... done.
15:00 < mpp> Creating package description ...
15:00 < mpp> Making post-install adaptions.
15:00 < mpp> it's rather stage 0 than 1 as i reported previously
15:00 < mpp> so wth
15:00 < mpp> no c++ ?
15:00 < mpp> config is generic minimal without x
15:00 < mpp> host is x86_64
15:02 < mtr> maybe you do not have g++ ...
15:02 < mtr> g++ --version ?
15:03 < mpp> g++ -v
15:03 < mpp> Es werden eingebaute Spezifikationen verwendet.
15:03 < mpp> Ziel: x86_64-t2-linux-gnu
15:03 < mpp> Konfiguriert mit: ../configure --prefix=/usr --bindir=/usr/bin --sbindir=/usr/sbin --libdir=/usr/lib64 --datadir=/usr/share --includedir=/usr/include --infodir=/usr/info --mandir=/usr/man --sysconfdir=/etc --localstatedir=/var --disable-debug --with-libpam --with-pam --enable-libpam --enable-pam --build=x86_64-t2-linux-gnu --host=x86_64-t2-linux-gnu --target=x86_64-t2-linux-gnu --enable-__cxa_atexit --disable-checking --disable-bootstrap --disabl
15:03 < mpp> e-libstdcxx-pch --enable-multilib --with-gnu-as --with-gnu-ld --enable-threads=posix --enable-libgcj --with-ecj-jar=/TOOLCHAIN/download/mirror/e/ecj-4.3.jar --enable-languages=c,c++,objc,java,fortran --enable-shared
15:03 < mpp> Thread-Modell: posix
15:03 < mpp> gcc-Version 4.3.4 (GCC)
15:03 < mpp> g++ --version
15:03 < mpp> g++ (GCC) 4.3.4
15:03 < mpp> Copyright (C) 2008 Free Software Foundation, Inc.
15:03 < mpp> Dies ist freie Software; die Kopierbedingungen stehen in den Quellen. Es
15:03 < mpp> gibt KEINE Garantie; auch nicht für MARKTGÄNGIGKEIT oder FÜR SPEZIELLE ZWECKE.
15:04 < mpp> svn co is fresh
15:04 < mpp> no previous builds / ccaches
15:08 < mtr> can you check or upload the config.log in src.gcc..../gcc-4.4.2/objdir ?
15:09 < mtr> maybe we got a hint from src.gcc..../cmd_wrapper.log, check the last lines
15:10 < mpp> k - i'll do a bit later on
15:26 < mpp> but the gcc package does not fail ...
15:26 < mpp> or did i misunderstand something ?
15:29 < mtr> to build gcc-4.4 in stage-0 the host system gcc and g++ are required
15:29 < mpp> i get that
15:29 < mtr> in stage-1 the freshly build cross compiler from stage-0 is used
15:29 < mpp> but the build process fails on gmp not gcc or g++
15:30 < mpp> @stage 0
15:31 < mtr> gmp build priority is lower than gcc/g++, gcc depends on it
15:31 < mtr> gmp and its c++ wrapper libs are also build using the host system compiler
15:32 < mpp> ...
15:32 < mtr> my guess is that we inject some compile flags with the t2 wrappers which are insane for x86_64
15:32 < mpp> k
15:32 < mtr> unfortunately i do not have such a test environment running ...
15:33 < mpp> k - what do you suggest with my problem ?
15:33 < mpp> i have a x86_64 host
15:33 < mpp> building native
15:35 -!- LMJ [n=serwou@laf31-4-82-236-42-164.fbx.proxad.net] has joined #t2
15:35 < mtr> a. for further urgent builds: revert rev 34864-34867
15:36 < mtr> b. maybe rxr has this issue on his builder too
15:36 < mtr> c. we could get some hints from config.log's and cmd_wrapper.log
15:37 -!- LMJ [n=serwou@laf31-4-82-236-42-164.fbx.proxad.net] has left #t2 []
15:37 < mpp> k
15:38 < mtr> hmm, maybe you could change the -march setting in gmp.conf, line 24
15:38 < mtr> var_append SYSCXX_WRAPPER_APPEND ' ' "-march=$HOSTTYPE"
15:39 < mtr> you $HOSTTYPE should be: x86_64
15:40 < mtr> but the correct gcc option may be "-march=x86-64"
15:40 -!- LMJ [n=serwou@laf31-4-82-236-42-164.fbx.proxad.net] has joined #t2
15:40 < mtr> you may give it a try
15:42 < mtr> if it works you have to change package/develop/ppl/ppl.conf the same way
16:06 < mpp> k tried those
16:06 < mpp> no lock
16:06 < mpp> luck :-)
16:55 -!- mpp [n=mpp@2a01:48:251:107:203:dff:fe13:6c9e] has quit ["good night - good fight"]
--- Log closed Sun Nov 15 00:00:17 2009