From b85491e9868b78d1779acb08e9b6718cc17ff141 Mon Sep 17 00:00:00 2001 From: Andreas Baumann Date: Sat, 24 Feb 2018 20:39:38 +0100 Subject: prepared everything for stage 3 --- .Attic/README.old_stage2 | 60 +++--------------------------------------------- 1 file changed, 3 insertions(+), 57 deletions(-) (limited to '.Attic') diff --git a/.Attic/README.old_stage2 b/.Attic/README.old_stage2 index 6e7fc69..fceabbe 100644 --- a/.Attic/README.old_stage2 +++ b/.Attic/README.old_stage2 @@ -1,3 +1,5 @@ +# maybe some hints to be integrated into stage3.. + # old stage 2 based on a broken stage 1, so carefull what to pick # into the next version.. @@ -39,64 +41,8 @@ sed -i '/configure.ac/ a \ sed -i "s@\\(unit_SOURCES.*\\)@#\\1@g" Makefile.am' sed -i '/configure.ac/ a \ sed -i "s@\\(check_PROGRAMS*\\)@#\\1@g" Makefile.am' dejagnu/PKGBUILD sed -i '/Makefile.am/ a \ autoreconf' dejagnu/PKGBUILD -# glibc -####### -# use 2.25 version, not 2.26. Keep the ABI intact! -# endless loop? gawk -f ../scripts/gen-as-const.awk pthread-pi-defines.sym \ -# two reasons: clock and artifacts walk backwards or forward in time -# chroot on server with chroot, hard to have an asynchronous clock? -# or glibc is already installed on the system (where not?), so how the -# hell can it be possibly built! -# => oh, this would be cruel. -# or 3rd reason, we have a glibc shim, maybe this one causes some trouble -# => why? -# or libtool fixing, whatever fixing means (https://forums.gentoo.org/viewtopic-t-496052-start-0.html) -# => nope. -# -isystem /usr/includ ein gcc spec? quite likely! -# ah: old friend: -build/glibc/src/glibc-build/libc.a(dl-sysdep.o): In function `.L18': -dl-sysdep.c:(.text+0x8c): undefined reference to `__memcmp_ia32' -/build/glibc/src/glibc-build/libc.a(dl-sysdep.o): In function `.L87': -dl-sysdep.c:(.text+0x11f): undefined reference to `__memcmp_ia32' -/build/glibc/src/glibc-build/libc.a(dl-sysdep.o): In function `.L19': -dl-sysdep.c:(.text+0x152): undefined reference to `__memcmp_ia32' -/build/glibc/src/glibc-build/libc.a(dl-sysdep.o): In function `.L85': -dl-sysdep.c:(.text+0x19c): undefined reference to `__memcmp_ia32' -dl-sysdep.c:(.text+0x1b7): undefined reference to `__memcmp_ia32' -/build/glibc/src/glibc-build/libc.a(dl-sysdep.o):dl-sysdep.c:(.text+0x202): more undefined references to `__memcmp_ia32' follow -elf/sln -=> USE_MULTIARCH -=> --disable-multi-arch -=> let's test this time, too dangerous to loose the chroot otherwise! -# older 2.25 misses a binutils 2.29 patch: -https://git.busybox.net/buildroot/diff/package/glibc/0005-fix-binutils-2-29-build.patch?id=cf821efbd0b24690b52f379d4a9934a16073762e -{standard input}: Assembler messages: -{standard input}: Error: `loc1@GLIBC_2.0' can't be versioned to common symbol 'loc1' -{standard input}: Error: `loc2@GLIBC_2.0' can't be versioned to common symbol 'loc2' -{standard input}: Error: `locs@GLIBC_2.0' can't be versioned to common symbol 'locs' -make[2]: *** [../o-iterator.mk:9: /build/glibc/src/glibc-build/misc/regexp.os] Error 1 - PACKAGES=" \ -libgpg-error \ - \ -sysfsutils libidn iputils \ - tcl expect dejagnu gcc glibc" - -# aftermatch for gcc/binutils - -#collect2: unable to find ld: this sounds bad, like an archicture mismatch somewhere -#in binutils/gcc -=> this is the cross compiled gcc for i486 inside the chroot, it expects - its platform dependend stuff in /usr/lib/gcc/i486-unknown-linux-gnu/7.2.0 -=> we can temporarily fix this: -ln -s /usr/bin/ld /usr/lib/gcc/i486-unknown-linux-gnu/7.2.0/ld -etc. - -# aftermatch for glibc - -# bacause pacman cannot set file permissions -chmod u+x /usr/bin/* /lib/*.so* - + tcl expect dejagnu" # in 'real' 486 ############### -- cgit v1.2.3-54-g00ecf