Portage 2.1_pre5-r2 (default-linux/x86/2005.1, gcc-4.1.0, glibc-2.3.6-r3, 2.6.16-rc5-vs2.1.1-rc10-gentoo i686) ================================================================= System uname: 2.6.16-rc5-vs2.1.1-rc10-gentoo i686 Intel(R) Pentium(R) M processor 2.00GHz Gentoo Base System version 1.12.0_pre16 dev-lang/python: 2.4.2-r1 sys-apps/sandbox: 1.2.17 sys-devel/autoconf: 2.13, 2.59-r7 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1 sys-devel/binutils: 2.16.91.0.5 sys-devel/libtool: 1.5.22 virtual/os-headers: 2.6.11-r3 ACCEPT_KEYWORDS="x86 ~x86" AUTOCLEAN="yes" CBUILD="i686-pc-linux-gnu" CFLAGS="-O2 -march=pentium-m -pipe" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/kde/3/share/config /usr/share/X11/xkb /usr/share/config /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-O2 -march=pentium-m -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig collision-protect cvs sandbox sfperms sign strict" GENTOO_MIRRORS=" ftp://ftp-stud.fht-esslingen.de/Mirrors/gentoo http://distfiles.gentoo.org" LANG="C" LC_ALL="C" LINGUAS="en" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/var/svn/gentoo/gentoo-dev" SYNC="rsync://rsync.de.gentoo.org/gentoo-portage" USE="x86 X alsa hal nptl nptlonly opengl pam perl pic python readline ssl truetype elibc_glibc kernel_linux linguas_en userland_GNU" Unset: ASFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, LDFLAGS I just tried to compile dev-libs/beecrypt with gcc-4.1.0 and it fails (obviously due to beecrypt using --with-arch=pentium-m).
Created attachment 81261 [details] beecrypt-4.1.2-compile.log
why do you think the failure is due to -march=pentium-m ? at first glance, looks like the bug is in beecrypt, not gcc ...
(In reply to comment #2) > why do you think the failure is due to -march=pentium-m ? at first glance, > looks like the bug is in beecrypt, not gcc ... > Because if I drop the --with-arch=pentium-m to --with-arch=i686 everything is fine (as in beecrypt is compiling)
This is not gcc-4.1's fault. There are asm files in there that obviously cause some problems. Whose package is this? There is no metadata. :)
sanchan: vapier said you might be interested in maintaining this?
I'll take a look at it.
mmm there is too much assmebler here... On my amd64 system dev-libs/beecrypt-4.1.2-r1 fails compiling with the following error: x86_64-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2 -I. -I/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2/.. -DOPTIMIZE_I386 -march=athlon64 -Os -pipe -fforce-addr -ffast-math -msse2 -ftracer -m3dnow -Wall -Wa,--noexecstack -c aesopt.s -fPIC -DPIC -o .libs/aesopt.o aesopt.s: Assembler messages: aesopt.s:106: Error: suffix or operands invalid for `push' aesopt.s:107: Error: suffix or operands invalid for `push' aesopt.s:108: Error: suffix or operands invalid for `push' aesopt.s:109: Error: suffix or operands invalid for `push' aesopt.s:837: Error: suffix or operands invalid for `pop' aesopt.s:838: Error: suffix or operands invalid for `pop' aesopt.s:839: Error: suffix or operands invalid for `pop' aesopt.s:840: Error: suffix or operands invalid for `pop' aesopt.s:857: Error: suffix or operands invalid for `push' aesopt.s:858: Error: suffix or operands invalid for `push' aesopt.s:859: Error: suffix or operands invalid for `push' aesopt.s:860: Error: suffix or operands invalid for `push' aesopt.s:1596: Error: suffix or operands invalid for `pop' aesopt.s:1597: Error: suffix or operands invalid for `pop' aesopt.s:1598: Error: suffix or operands invalid for `pop' aesopt.s:1599: Error: suffix or operands invalid for `pop' aesopt.s:1625: Error: suffix or operands invalid for `push' aesopt.s:1626: Error: suffix or operands invalid for `push' aesopt.s:1627: Error: suffix or operands invalid for `push' aesopt.s:1628: Error: suffix or operands invalid for `push' aesopt.s:2356: Error: suffix or operands invalid for `pop' aesopt.s:2357: Error: suffix or operands invalid for `pop' aesopt.s:2358: Error: suffix or operands invalid for `pop' aesopt.s:2359: Error: suffix or operands invalid for `pop' aesopt.s:2376: Error: suffix or operands invalid for `push' aesopt.s:2377: Error: suffix or operands invalid for `push' aesopt.s:2378: Error: suffix or operands invalid for `push' aesopt.s:2379: Error: suffix or operands invalid for `push' aesopt.s:3115: Error: suffix or operands invalid for `pop' aesopt.s:3116: Error: suffix or operands invalid for `pop' aesopt.s:3117: Error: suffix or operands invalid for `pop' aesopt.s:3118: Error: suffix or operands invalid for `pop' make[2]: *** [aesopt.lo] Error 1 make[2]: Leaving directory `/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2' make: *** [all] Error 2 the definition: -DOPTIMIZE_I386 looks like if it doesn't be there. amd64 team should take a look at this too. I'm not so good with assembler language, if anyone more skilled want take a look at this... my info: Portage 2.1_pre6-r1 (default-linux/amd64/2006.0, gcc-3.4.5, glibc-2.4-r0, 2.6.15-gentoo-r7 x86_64) ================================================================= System uname: 2.6.15-gentoo-r7 x86_64 AMD Athlon(tm) 64 Processor 3500+ Gentoo Base System version 1.12.0_pre16 distcc 2.18.3 x86_64-pc-linux-gnu (protocols 1 and 2) (default port 3632) [disabled] ccache version 2.4 [enabled] dev-lang/python: 2.4.2-r1 sys-apps/sandbox: 1.2.17 sys-devel/autoconf: 2.13, 2.59-r7 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1 sys-devel/binutils: 2.16.1-r2 sys-devel/libtool: 1.5.22 virtual/os-headers: 2.6.11-r3 ACCEPT_KEYWORDS="amd64 ~amd64" AUTOCLEAN="yes" CBUILD="x86_64-pc-linux-gnu" CFLAGS=" -march=athlon64 -Os -pipe -fforce-addr -ffast-math -msse2 -ftracer -m3dnow " CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/kde/3/share/config /usr/share/X11/xkb /usr/share/config /var/bind /var/qmail/control" CONFIG_PROTECT_MASK="/etc/env.d/java/ /etc/gconf /etc/java-config/vms/ /etc/revdep-rebuild /etc/splash /etc/terminfo /etc/texmf/web2c /etc/env.d" CXXFLAGS=" -march=athlon64 -Os -pipe -fforce-addr -ffast-math -msse2 -ftracer -m3dnow " DISTDIR="/home/ftp/distfiles" FEATURES="autoconfig ccache collision-protect confcache cvs distlocks elog metadata-transfer multilib-strict parallel-fetch sandbox sfperms sign strict stricter test userpriv" GENTOO_MIRRORS="ftp://mirror.switch.ch/mirror/gentoo/ http://gentoo.osuosl.org/ http://mirror.ovh.net/gentoo-distfiles/ ftp://ftp.heanet.ie/pub/gentoo/ http://www.die.unipd.it/pub/Linux/distributions/gentoo-sources/ ftp://ftp.unina.it/pub/linux/distributions/gentoo http://gentoo.ngi.it http://mirror.switch.ch/ftp/mirror/gentoo/ ftp://mirror.switch.ch/mirror/gentoo/ " LANG="it_IT.utf8" LINGUAS="it" PKGDIR="/home/ftp/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/overlays/migration-overlay /usr/local/overlays/tinyos /usr/local/overlays/gentoo-java-experimental" SYNC="rsync://rsync.europe.gentoo.org/gentoo-portage" USE="amd64 7zip X a52 aac aalib acl acpi alsa apache2 arts asf audiofile avi berkdb bitmap-fonts bzip2 cairo canvas ccache cdb cddb cdparanoia cdr chroot cpudetection crypt css cups curl dba dbus dga directfb dlloader dri dts dv dvd dvdr dvdread dynagraph emboss encode erandom esd exif expat extrafilters fame ffmpeg firefox flac fltk foomaticdb gd gdbm ggi gif gimp gimpprint glibc-omitfp glut gmp gnokii gnome gnutls gphoto2 gpm graphviz gstreamer gtk gtk2 gtkhtml guile hal idn ieee1394 imagemagick imlib ipv6 ithreads java jbig jikes jpeg jpeg2k justify kde kdeenablefinal lcms ldap libcaca linuxthreads-tls lm_sensors logitech-mouse logrotate lzo lzw lzw-tiff mad mailwrapper mbox mdb mhash mikmod mjpeg mmap mng motif mozsvg mp3 mpeg mpm-worker musicbrainz mysql ncurses nls nptl nptlonly numeric nvidia odbc offensive ogg on-the-fly-crypt openexr opengl oss pam pcre pcsc-lite pdf pdflib perl php png povray ppds python qt quicktime rdesktop readline rtc samba scanner sdl slang smp snmp speex spell ssl subversion svg sysfs tcltk tcpd test tetex tga theora threads tidy tiff timidity truetype truetype-fonts type1-fonts unicode usb userlocales utf8 v4l v4l2 vcd vda vorbis wmf xcomposite xine xinerama xml2 xmms xpm xscreensaver xv xvid yv12 zlib elibc_glibc input_devices_mouse input_devices_evdev input_devices_keyboard kernel_linux linguas_it userland_GNU video_cards_nvidia" Unset: ASFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, LC_ALL, LDFLAGS, MAKEOPTS
(In reply to comment #7) > the definition: -DOPTIMIZE_I386 looks like if it doesn't be there. amd64 team > should take a look at this too. I'm not so good with assembler language, if > anyone more skilled want take a look at this... configure.ac or the ebuild should be patched. configure.ac row 504 says: athlon*) CPPFLAGS="$CPPFLAGS -DOPTIMIZE_I386" but the ebuild use local myarch=$(get-flag march) that on my system is athlon64, so it matches the case above. It should be x86_64 in order to match x86_64) CPPFLAGS="$CPPFLAGS -DOPTIMIZE_X86_64" ;; pentium-m should match anything so no optimization will be triggered by the configure.ac
Created attachment 82300 [details, diff] update x86(_64) archs in configure.ac this patch updates the x86 and x86_64 archs in configure.ac based on the gcc-4.0.3 man page
(In reply to comment #9) > Created an attachment (id=82300) [edit] > update x86(_64) archs in configure.ac > > this patch updates the x86 and x86_64 archs in configure.ac based on the > gcc-4.0.3 man page > The submitted patch can't be used as is. -DOPTIMIZE_MMX -DOPTIMIZE_SSE -DOPTIMIZE_SSE2 should be specified as CPPFLAGS when USE="mmx, sse, sse2" are enabled. What is the purpose of having a custom cflags patch if we append new cflags that changes the ones preferred by the user? The patch also collides with custom cflag patch.
Created attachment 82561 [details, diff] updated patch noticed that too, here's the updated patch.. forgot to attach it..
(In reply to comment #11) > Created an attachment (id=82561) [edit] > updated patch > > noticed that too, here's the updated patch.. forgot to attach it.. > Thanks for the patch. Please, next time you submit a patch, don't include another existing patch changing the patch name. It took 15 minutes to figure out that the attached patch was merged with the existing custom-cflags and removing duplicate parts. I wasn't the maintainer of the ebuild when those patch was submitted, and since this is a compile fix the ebuild doesn't need a bump. I prefer to keep the existing patch and add your patch with some changes in order to avoid the collisions. I'm cleaning up the ebuild, I'll commit the update in a few minutes. Phreak please test the update with your pentium-m if this doesn't solve the problem I'll substitute the CFLAG pentium-m with i686.
I've to wait for the commit. There are some major trouble with amd64. Using CFLAGS="-march=athlon64" The ebuild fails on aes assmebler code. Using CFLAGS="-march=k8" the ebuild compile fine but there's a multilib issue: File /var/tmp/portage/beecrypt-4.1.2-r1/image///usr/lib/libbeecrypt.so.6.4.0 matches a file type that is not allowed in /usr/lib 64 team, any idea? I'll attach to this bug the updated ebuild and the updated patch.
Created attachment 82703 [details, diff] Update of the arch-update patch Update of the arch-update patch, many thanks to Benedikt Boehm for the initial submission of the patch.
Created attachment 82704 [details] updated ebuild Updated ebuild: some cleanups and the new arch-update patch. Please test it for pentium-m issue. amd64 team, please test this with -march=athlon64 e -march=k8
Works for me on pentium-m
compile fine for me with -march=pentium3, but fails test phase on python modules. Making check in test make[2]: Entering directory `/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2/python/test' PYTHONPATH=.. python test_all.py verbose Traceback (most recent call last): File "test_all.py", line 58, in ? print_versions() File "test_all.py", line 19, in print_versions from _bc import mpw ImportError: No module named _bc make[2]: *** [check] Error 1 make[2]: Leaving directory `/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2/python/test' make[1]: *** [check-recursive] Error 1 make[1]: Leaving directory `/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2/python' make: *** [check-recursive] Error 1
fails compile phase with -march=athlon-xp on linux32 chroot with the following error: i686-pc-linux-gnu-g++ -shared -nostdlib /usr/lib/gcc/i686-pc-linux-gnu/3.4.5/../../../crti.o /usr/lib/gcc/i686-pc-linux-gnu/3.4.5/crtbeginS.o .libs/aes.o .libs/base64.o .libs/beecrypt.o .libs/blockmode.o .libs/blockpad.o .libs/blowfish.o .libs/dhaes.o .libs/dldp.o .libs/dlkp.o .libs/dlpk.o .libs/dlsvdp-dh.o .libs/dsa.o .libs/elgamal.o .libs/endianness.o .libs/entropy.o .libs/fips186.o .libs/hmac.o .libs/hmacmd5.o .libs/hmacsha1.o .libs/hmacsha256.o .libs/md5.o .libs/hmacsha384.o .libs/hmacsha512.o .libs/memchunk.o .libs/mp.o .libs/mpbarrett.o .libs/mpnumber.o .libs/mpprime.o .libs/mtprng.o .libs/pkcs1.o .libs/pkcs12.o .libs/rsa.o .libs/rsakp.o .libs/rsapk.o .libs/sha1.o .libs/sha256.o .libs/sha384.o .libs/sha512.o .libs/sha_k.o .libs/timestamp.o .libs/cppglue.o .libs/aesopt.o .libs/blowfishopt.o .libs/mpopt.o .libs/sha1opt.o -lrt -lpthread -ldl -L/usr/lib/gcc/i686-pc-linux-gnu/3.4.5 -L/usr/lib/gcc/i686-pc-linux-gnu/3.4.5/../../../../i686-pc-linux-gnu/lib -L/usr/lib/gcc/i686-pc-linux-gnu/3.4.5/../../.. /usr/lib/gcc/i686-pc-linux-gnu/3.4.5/libstdc++.so -L/usr/i686-pc-linux-gnu/bin -L/usr/i686-pc-linux-gnu/lib -lm -lc -lgcc_s /usr/lib/gcc/i686-pc-linux-gnu/3.4.5/crtendS.o /usr/lib/gcc/i686-pc-linux-gnu/3.4.5/../../../crtn.o -march=athlon-xp -msse2 -Wl,-soname -Wl,libbeecrypt.so.6 -o .libs/libbeecrypt.so.6.4.0 .libs/aesopt.o: In function `aesEncrypt': : multiple definition of `aesEncrypt' .libs/aes.o:aes.c:(.text+0x5a0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/3.4.5/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `aesEncrypt' changed from 4352 in .libs/aes.o to 3222 in .libs/aesopt.o .libs/aesopt.o: In function `aesDecrypt': : multiple definition of `aesDecrypt' .libs/aes.o:aes.c:(.text+0x16a0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/3.4.5/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `aesDecrypt' changed from 4350 in .libs/aes.o to 3222 in .libs/aesopt.o /usr/lib/gcc/i686-pc-linux-gnu/3.4.5/../../../../i686-pc-linux-gnu/bin/ld: warning: creating a DT_TEXTREL in object. collect2: ld returned 1 exit status make[2]: *** [libbeecrypt.la] Error 1 make[2]: Leaving directory `/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/var/tmp/portage/beecrypt-4.1.2-r1/work/beecrypt-4.1.2' make: *** [all] Error 2
All of the beecrypt ebuilds fail on arch pentium-m with a linking error.
*** Bug 136269 has been marked as a duplicate of this bug. ***
*** Bug 136506 has been marked as a duplicate of this bug. ***
It seems that this "little" bug is causing a lot of trouble. I haven't commited the ebuild due to the problem I have on it with the AES assembler in general and amd64 in particular. Due to lack of time to do some real assembler magic I think the best solution could be disabling the assembler. However if someone on this CC list is skilled enough to take care of the assembler code... I think I can do some more testing on the ebuild disabling any assembler reference trying to solve the issues during the test phase and while compiling AES. If all works fine I'll commit the ebuild without the use flags mmx, sse and sse2 this weekend.
Created attachment 89070 [details] new ebuild without mmx, sse, sse2 use flags This ebuild uses upstream patches from cvs. Fixes python test failures on pentium3, athlon-xp, pentium-mmx; compiles fine with pentium-m but fails test phase due to lack of a symbol in libbeecrypt, fails on atholn64 so I keyword it -amd64 since amd64 team can look at it. Upstream patches follows... be patient.
Created attachment 89071 [details, diff] arch-update from upstream Replace arch-update patch, seems to work around some TEXTREL issue. Don't ask me how :-)
Created attachment 89072 [details, diff] fix import error in make check fix import error in make check adding ../.libs to PYTHONPATH. Thanks to #gentoo-python irc channel for pointing me to the right place.
Created attachment 89073 [details, diff] Fix missing symbol in _bc.so Upstream patch from cvs.
Looking for the multilib issue I noticed that the configure.ac have some ugly check on java_home like java_home=`javac conftest.java; java -classpath . conftest` and java_home="`gcj --main=conftest -o conftest conftest.java; ./conftest`" I add java team in CC in order to let them know that there's something to look at here.
for the multilib issue, just adding --libdir=/usr/$(get_libdir) to econf doesn't solve the issue. There's something wrong somewhere.
The ebuild is available in CVS and will hit the mirrors in the next hours. The bug is not fixed but at least some of the troubles this package cause are closed. I know amd64 and java teams are pretty busy, so I'll try to fix the remaining issue by myself. Feel free to help :-)
*** Bug 136325 has been marked as a duplicate of this bug. ***
On all my system pentium-m, p4, athlon-xp and amd64 beecrypt-4.1.2-r1 does not compile with Activated profiles: i686-pc-linux-gnu * i686-pc-linux-gnu-4.1.1/hardened I always get: i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I. -I../.. -I../../.. -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -DOPTIMIZE_I68 6 -DOPTIMIZE_MMX -O2 -march=athlon-xp -fomit-frame-pointer -DNDEBUG -Wa,--noexecstack -c ByteArrayInputStream.cxx -fPIC -D PIC -o .libs/ByteArrayInputStream.o In file included from ../../beecrypt/c++/io/ByteArrayInputStream.h:30, from ByteArrayInputStream.cxx:25: ../../beecrypt/c++/mutex.h:58:7: error: #error ../../beecrypt/c++/mutex.h:78:7: error: #error ../../beecrypt/c++/mutex.h:110:7: error: #error ../../beecrypt/c++/mutex.h:129:7: error: #error ../../beecrypt/c++/mutex.h:148:7: error: #error ../../beecrypt/c++/mutex.h:40: error: 'bc_mutex_t' does not name a type make[3]: *** [ByteArrayInputStream.lo] Error 1 root@cheetah:/root(5)# emerge --info Portage 2.1.1_pre1-r1 (default-linux/x86/2006.0, gcc-4.1.1/hardened, glibc-2.4-r3, 2.6.17-rc4 i686) ================================================================= System uname: 2.6.17-rc4 i686 AMD Athlon(tm) XP 1800+ Gentoo Base System version 1.12.1 distcc 2.18.3 i686-pc-linux-gnu (protocols 1 and 2) (default port 3632) [disabled] ccache version 2.4 [disabled] dev-lang/python: 2.4.3-r1 dev-python/pycrypto: 2.0.1-r5 dev-util/ccache: 2.4-r2 dev-util/confcache: [Not Present] sys-apps/sandbox: 1.2.18.1 sys-devel/autoconf: 2.13, 2.59-r7 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2 sys-devel/binutils: 2.16.1-r2 sys-devel/gcc-config: 2.0.0_rc1 sys-devel/libtool: 1.5.22 virtual/os-headers: 2.6.11-r5 ACCEPT_KEYWORDS="x86 ~x86" AUTOCLEAN="yes" CBUILD="i686-pc-linux-gnu" CFLAGS="-O2 -march=athlon-xp -fomit-frame-pointer" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/grass60/etc /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/lib/mozilla/defaults/pref /usr/share/X11/xkb /usr/share/config" CONFIG_PROTECT_MASK="/etc/env.d /etc/eselect/compiler /etc/gconf /etc/revdep-rebuild /etc/terminfo /etc/texmf/web2c" CXXFLAGS="-O2 -march=athlon-xp -fomit-frame-pointer" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig distlocks fixpackages metadata-transfer sandbox sfperms strict" GENTOO_MIRRORS="http://linux.rz.ruhr-uni-bochum.de/download/gentoo-mirror http://ftp-stud.fht-esslingen.de/pub/Mirrors/gentoo/ rsync://ftp.fi.muni.cz/pub/linux/gentoo ftp://ftp.wh2.tu-dresden.de/pub/mirrors/gentoo rsync://gd.tuwien.ac.at/opsys/linux/gentoo ftp://sunsite.informatik.rwth-aachen.de/pub/Linux/gentoo ftp://ftp.tu-clausthal.de/pub/linux/gentoo http://gentoo.oregonstate.edu" LINGUAS="de fr" MAKEOPTS="-j1" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --delete-after --stats --timeout=180 --exclude='/distfiles' --exclude='/local' --exclude='/packages'" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="x86 3dnow 3dnowext X Xaw3d a52 aac alsa ao aotuv apache2 apm arts asf atlas auctex audiofile automount avi berkdb bitmap-fonts blas bonobo bzip2 cairo cdda cddb cdf cdio cdparanoia cdr cgi cli corba crypt cups curl daap dbus devmap dga dlloader dri dvb dvd dvi dxr3 dynagraph eds effects elf emacs emboss encode epiphany esd evo evolution exif expat fam fame ffmpeg fftw fits flac fltk foomaticdb fortran fpx gdal gdbm geos gif gimp ginac glade gml gmp gnome gnuplot gphoto2 gpm graphviz grass gs gsl gsm gstreamer gtk gtk2 guile hal hardened hddtemp hdf hdf5 hlapi icq id3 imagemagick imlib innodb ipv6 isdnlog jabber jbig jp2 jpeg jpeg2 kde kexi lame lapack latex lcms ldap libg++ libgda libsamplerate libwww lirc lm_sensors lzo lzw mad math matroska matrox mikmod mmx mng mod mono motif mozcalendar mozdevelop mozilla mozsvg mozxmlterm mp3 mp4live mpeg mpeg2 mplayer musicbrainz mysql mysqli nautilus ncurses netcdf network nfs nls nntp nptl nptlonly numarray numeric ocaml octave odbc ogdi ogg opengl oss pam pcre pdf pdflib perl plotutils plugin png posix postgres pppd proj python qhull qt quicktime readline real reflection reiserfs rhythmbox rle rrdcgi rrdtool samba sdl session slang slp sndfile snmp sox spell spl sqlite sse sse2 ssl subtitles svg t1lib tcltk tcpd tetex theora tidy tiff truetype truetype-fonts type1-fonts udev unicode v4l2 video_cards_mga vorbis win32codecs wmf xemacs xext xine xml xml2 xmlreader xmlrpc xmms xorg xpm xv xvid xvmc zlib zvbi elibc_glibc input_devices_keyboard input_devices_mouse kernel_linux linguas_de linguas_fr userland_GNU video_cards_vesa video_cards_fbdev" Unset: CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG, LC_ALL, LDFLAGS, PORTAGE_RSYNC_EXTRA_OPTS
Just to note - gcc-4.1.1/hardened is in fact gcc-4.1.1/vanilla; >=sys-devel/gcc-4 is masked on the hardened profile as it's not yet hardened.
Created attachment 92299 [details, diff] patch to fix name qualifications in friend declarations OK, this here has nothing to do with asm or sse or amd, but it is an issue with gcc4 and as other gcc4 beecrypt bugs have been duped here, I submit this patch here as well. After applying it, beecrypt emerged fine for me on prescott. http://gcc.gnu.org/gcc-4.0/changes.html: "When declaring a friend class using an unqualified name, classes outside the innermost non-class scope are not searched" But the "using" declarations are in the global level scope, causing the friend declarations to fail, leading to access errors. The error message solved by this patch: /bin/sh ../../../libtool --tag=CXX --mode=compile i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../.. -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -march=prescott -O2 -DNDEBUG -Wa,--noexecstack -c -o CertificateFactory.lo CertificateFactory.cxx ../../../beecrypt/c++/security/Security.h: In static member function 'static beecrypt::security::cert::CertificateFactory* beecrypt::security::cert::CertificateFactory::getInstance(const beecrypt::lang::String&)': ../../../beecrypt/c++/security/Security.h:81: error: 'struct beecrypt::security::Security::spi' is private CertificateFactory.cxx:48: error: within this context ../../../beecrypt/c++/security/Security.h:89: error: 'static beecrypt::security::Security::spi* beecrypt::security::Security::getSpi(const beecrypt::lang::String&, const beecrypt::lang::String&)' is private CertificateFactory.cxx:48: error: within this context ../../../beecrypt/c++/security/Security.h: In static member function 'static beecrypt::security::cert::CertificateFactory* beecrypt::security::cert::CertificateFactory::getInstance(const beecrypt::lang::String&, const beecrypt::lang::String&)': ../../../beecrypt/c++/security/Security.h:81: error: 'struct beecrypt::security::Security::spi' is private CertificateFactory.cxx:63: error: within this context ../../../beecrypt/c++/security/Security.h:90: error: 'static beecrypt::security::Security::spi* beecrypt::security::Security::getSpi(const beecrypt::lang::String&, const beecrypt::lang::String&, const beecrypt::lang::String&)' is private CertificateFactory.cxx:63: error: within this context ../../../beecrypt/c++/security/Security.h: In static member function 'static beecrypt::security::cert::CertificateFactory* beecrypt::security::cert::CertificateFactory::getInstance(const beecrypt::lang::String&, const beecrypt::security::Provider&)': ../../../beecrypt/c++/security/Security.h:81: error: 'struct beecrypt::security::Security::spi' is private CertificateFactory.cxx:78: error: within this context ../../../beecrypt/c++/security/Security.h:91: error: 'static beecrypt::security::Security::spi* beecrypt::security::Security::getSpi(const beecrypt::lang::String&, const beecrypt::lang::String&, const beecrypt::security::Provider&)' is private CertificateFactory.cxx:78: error: within this context make[4]: *** [CertificateFactory.lo] Error 1
*** Bug 136716 has been marked as a duplicate of this bug. ***
(In reply to comment #27) > Looking for the multilib issue I noticed that the configure.ac have some ugly > check on java_home like > java_home=`javac conftest.java; java -classpath . conftest` > > and > > java_home="`gcj --main=conftest -o conftest conftest.java; ./conftest`" > > I add java team in CC in order to let them know that there's something to look > at here. The best way to handle this probably is to patch configure and company to respect the JAVA_HOME env variable if its defined. This should always be defined properly at merge time. Regretfully, my autotools-fu is weak, otherwise I'd pitch in with a patch.
(In reply to comment #27) > Looking for the multilib issue I noticed that the configure.ac have some ugly > check on java_home (In reply to comment #35) > The best way to handle this probably is to patch configure and company to > respect the JAVA_HOME env variable if its defined. This should always be > defined properly at merge time. Regretfully, my autotools-fu is weak, > otherwise I'd pitch in with a patch. We only do all this stuff to find a dir that contains jni.h. :-) Simply exporting ac_cv_java_include=-I$(dirname /path/to/jni.h) might be enough. I don't know if there are portability issues with JNI, so I'm not sure how important it is /what/ jni header file is used by the build. Currently we have neither a DEPEND on jdk that I can see, nor do we inherit any javaish eclasses. Probably we should inherit java-pkg-opt-2.eclass. Establishing the java home directory through the java.home system variable inside a java program seems to work well enough. Especially it is flexible enough so that VM switching should be no problem at all, if need be. That is the "ugly check" currently performed in configure. The configure script handles gcj specially and prefers it to other java VMs. Simply exporting ac_cv_have_gcj=no might work, but I'm not too sure how that interacts with caching, especially if the confcache FEATURE is enabled. Same problem arises of course for setting ac_cv_java_include mentioned above. Sadly configure just checks for the existence of the include dir, and does not explicitly look for the jni.h file. Furthermore, it looks for $java_home/lib/gcc-lib/$gcj_machine/$gcj_version/include but not for $java_home/lib/gcc/$gcj_machine/$gcj_version/include so in my case, it just uses /usr/include (from gcj java.home /usr) which does not contain a jni.h at all. That it still does compile is something of a surprise to me, probably gcc automaically looks in the /usr/lib/gcc/... dir by itself. I'd try writing a patch for this, the relevant sections seem easy AC code and lot of verbatim bash. But I'd need to know which direction to take. Does it mater what jni.h we use? Is there some easier way to find the jni.h of the current VM? How should we treat gcj and its jni.h in relation to other VMs? Does someone know the extent of interaction with the confcache FEATURE?
Should this bug here block a gcc-4 Tracker, bug 117482 or bug 140707?
(In reply to comment #33) I was receiving the error messages in comment #33. Martin's patch has allowed successful compilation. Thanks, Martin!
I'm sorry if in these weeks I haven't worked much on this bug. I had no time to work on it. I'll be away for nearly all the next month, so if someone have the skills and the time to solve this bug in the next month he/she will be wellcome.
*** Bug 144115 has been marked as a duplicate of this bug. ***
Same problem here with -march=pentium-m and gcc-3.4.6. Works with march=pentium3.
I can confirm that gcc-3.4.6 and -march=pentium-m is a problem, see my (duplicate?) bug http://bugs.gentoo.org/show_bug.cgi?id=144115. I don't know wether it is a good idea to mix this with the gcc-4 bug...
*** Bug 144218 has been marked as a duplicate of this bug. ***
*** Bug 144294 has been marked as a duplicate of this bug. ***
adding a block on [Tracker] gcc-4.x porting.
My failure: i686-pc-linux-gnu-gcc -shared aes.lo base64.lo beecrypt.lo blockmode.lo blockpad.lo blowfish.lo dhaes.lo dldp.lo dlkp.lo dlpk.lo dlsvdp-dh.lo dsa.lo elgamal.lo endianness.lo entropy.lo fips186.lo hmac.lo hmacmd5.lo hmacsha1.lo hmacsha256.lo md5.lo memchunk.lo mp.lo mpbarrett.lo mpnumber.lo mpprime.lo mtprng.lo rsa.lo rsakp.lo rsapk.lo sha1.lo sha256.lo timestamp.lo aesopt.lo blowfishopt.lo mpopt.lo sha1opt.lo -lrt -lpthread -Wl,-O1 -Wl,--sort-common -Wl,-soname -Wl,libbeecrypt.so.6 -o .libs/libbeecrypt.so.6.2.0 blowfishopt.lo: In function `blowfishEncrypt': : multiple definition of `blowfishEncrypt' blowfish.lo:blowfish.c:(.text+0x520): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `blowfishEncrypt' changed from 1116 in blowfish.lo to 815 in blowfishopt.lo blowfishopt.lo: In function `blowfishDecrypt': : multiple definition of `blowfishDecrypt' blowfish.lo:blowfish.c:(.text+0xc0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `blowfishDecrypt' changed from 1117 in blowfish.lo to 815 in blowfishopt.lo mpopt.lo: In function `mpzero': : multiple definition of `mpzero' mp.lo:mp.c:(.text+0x0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpzero' changed from 31 in mp.lo to 15 in mpopt.lo mpopt.lo: In function `mpfill': : multiple definition of `mpfill' mp.lo:mp.c:(.text+0x20): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpfill' changed from 44 in mp.lo to 17 in mpopt.lo mpopt.lo: In function `mpeven': : multiple definition of `mpeven' mp.lo:mp.c:(.text+0x60): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpeven' changed from 19 in mp.lo to 18 in mpopt.lo mpopt.lo: In function `mpodd': : multiple definition of `mpodd' mp.lo:mp.c:(.text+0x50): first defined here mpopt.lo: In function `mpaddw': : multiple definition of `mpaddw' mp.lo:mp.c:(.text+0xb00): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpaddw' changed from 84 in mp.lo to 42 in mpopt.lo mpopt.lo: In function `mpsubw': : multiple definition of `mpsubw' mp.lo:mp.c:(.text+0xc50): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpsubw' changed from 84 in mp.lo to 42 in mpopt.lo mpopt.lo: In function `mpadd': : multiple definition of `mpadd' mp.lo:mp.c:(.text+0xb60): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpadd' changed from 93 in mp.lo to 41 in mpopt.lo mpopt.lo: In function `mpsub': : multiple definition of `mpsub' mp.lo:mp.c:(.text+0xcb0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpsub' changed from 98 in mp.lo to 41 in mpopt.lo mpopt.lo: In function `mpdivtwo': : multiple definition of `mpdivtwo' mp.lo:mp.c:(.text+0x1140): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpdivtwo' changed from 59 in mp.lo to 24 in mpopt.lo mpopt.lo: In function `mpmultwo': : multiple definition of `mpmultwo' mp.lo:mp.c:(.text+0x11f0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpmultwo' changed from 62 in mp.lo to 29 in mpopt.lo mpopt.lo: In function `mpsetmul': : multiple definition of `mpsetmul' mp.lo:mp.c:(.text+0xe10): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpsetmul' changed from 159 in mp.lo to 49 in mpopt.lo mpopt.lo: In function `mpaddmul': : multiple definition of `mpaddmul' mp.lo:mp.c:(.text+0xeb0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpaddmul' changed from 175 in mp.lo to 55 in mpopt.lo mpopt.lo: In function `mpaddsqrtrc': : multiple definition of `mpaddsqrtrc' mp.lo:mp.c:(.text+0x1080): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mpaddsqrtrc' changed from 133 in mp.lo to 50 in mpopt.lo mpopt.lo: In function `mppndiv': : multiple definition of `mppndiv' mp.lo:mp.c:(.text+0x19b0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `mppndiv' changed from 99 in mp.lo to 13 in mpopt.lo sha1opt.lo: In function `sha1Process': : multiple definition of `sha1Process' sha1.lo:sha1.c:(.text+0x0): first defined here /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: Warning: size of symbol `sha1Process' changed from 4168 in sha1.lo to 3290 in sha1opt.lo collect2: ld returned 1 exit status make[2]: *** [libbeecrypt.la] Error 1 make[2]: Leaving directory `/var/tmp/portage/beecrypt-3.1.0-r2/work/beecrypt-3.1.0' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/var/tmp/portage/beecrypt-3.1.0-r2/work/beecrypt-3.1.0' make: *** [all] Error 2 !!! ERROR: dev-libs/beecrypt-3.1.0-r2 failed. Call stack: ebuild.sh, line 1539: Called dyn_compile ebuild.sh, line 939: Called src_compile beecrypt-3.1.0-r2.ebuild, line 53: Called die
*** Bug 147286 has been marked as a duplicate of this bug. ***
*** Bug 150021 has been marked as a duplicate of this bug. ***
there is a whole suite of unrelated bugs here ... i dont really think this bug is useful anymore sync up and if the latest beecrypt-4.1.2-r1 has failures, file new bugs
*** Bug 151606 has been marked as a duplicate of this bug. ***
*** Bug 156055 has been marked as a duplicate of this bug. ***
