Bug#644990: NEWS.Debian.gz: s/$arch/triplet
On Fri, Oct 14, 2011 at 9:44 PM, Jonathan Nieder jrnie...@gmail.com wrote: Sedat Dilek wrote: I am unsure how to interpret you might try to pass the following option to your: -B/usr/lib/triplet -I/usr/include/triplet? Normally, I would expect to do: export CFLAGS=$CFLAGS -B/usr/lib/triplet -I/usr/include/triplet But in case of gcc-trunk upstream this change impacts: [...] It means that you might want to pass those flags to the compiler. The compiler doesn't examine any environment variables, so your export CFLAGS incantation isn't going to do that, depending on the build system. You mean something like make CC='gcc ...' ? What exactly do I need to pass? When building a compiler, there is a bootstrap step, meaning there are multiple compilers to pass the flags to. I thought there was already a bug report about making this easier (#637232) which unfortunately no one seems to be working on. Unfortunately. I just built a MIPSEL toolchain w/o my workarounds, today. Worked nicely. The problem seems to affect gcc-trunk from upstream. The generated compiler needs a wrapper-script to use -B and -I options. BTW, I could compile mesa, but not use the same wrapper-script with kernel-buildsystem from Debian Kernel Team. That's probably because in debian/config.defines, we have compiler: 'gcc-4.5' So if your gcc-4.5 comes from upstream, you would need to install a /usr/local/bin/gcc-4.5 wrapper, too. With my 2 workarounds I can use my gcc-4.7 binaries OOTB. No wrapper-script etc. (so I prefer my way). But what does this have to do with the triplet text in libc6-dev's NEWS.Debian.gz? I wanted to point out that building with -B and -I options might not be enough to really use your generated new binaries. - Sedat - -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczuvbtv81hoy-_8f_o1zveyff4daaezw+sdz7wdvut3g...@mail.gmail.com
Bug#644986: i386: Compiling gcc-snapshots from upstream with multiarch-toolchain?
Some additional informations on created gcc-4.7 compiler: # which gcc-4.7 /usr/bin/gcc-4.7 # l /usr/bin/gcc-4.7 lrwxrwxrwx 1 root root 24 Jul 31 14:28 /usr/bin/gcc-4.7 - /opt/gcc-4.7/bin/gcc-4.7 # gcc-4.7 -v Using built-in specs. COLLECT_GCC=gcc-4.7 COLLECT_LTO_WRAPPER=/opt/gcc-4.7/lib/gcc/i486-linux-gnu/4.7.0/lto-wrapper Target: i486-linux-gnu Configured with: ../gcc-4.7-20111008/configure --prefix=/opt/gcc-4.7 --libdir=/opt/gcc-4.7/lib --libexecdir=/opt/gcc-4.7/lib --program-suffix=-4.7 --enable-clocale=gnu --enable-languages=c,c++ --enable-shared --enable-threads=posix --disable-bootstrap --disable-libssp --disable-multilib --disable-nls --with-system-zlib --without-cloog --without-ppl --with-arch-32=i586 --with-tune=generic --build=i486-linux-gnu --host=i486-linux-gnu --target=i486-linux-gnu Thread model: posix gcc version 4.7.0 20111008 (experimental) (GCC) -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczuxwas7jfuy3mqpjjecmuy_expujxljcl4totn4ock3...@mail.gmail.com
Bug#644986: i386: Compiling gcc-snapshots from upstream with multiarch-toolchain?
[ RESEND ] I played again with CFLAGS_FOR_TARGET and CXXFLAGS_FOR_TARGET by exporting them. -I/usr/include/${HOST_SYSTEM_MULTIARCH_TYPE} ... ...catches the gnu/stub-32.h issue. -B/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} ... ...does NOT catch the problem with crt*.o files. -B/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -B/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} ... ... NOPE, too. - Sedat - --- scripts/build_gcc-snapshot_v1.sh 2011-10-11 15:36:19.879266636 +0200 +++ scripts/build_gcc-snapshot_v3.sh 2011-10-11 16:30:00.962772133 +0200 @@ -25,6 +25,8 @@ BUILD_SYSTEM_TYPE=$(dpkg-architecture -q HOST_SYSTEM_TYPE=$(dpkg-architecture -qDEB_HOST_GNU_TYPE) TARGET_SYSTEM_TYPE=i486-linux-gnu +HOST_SYSTEM_MULTIARCH_TYPE=$(dpkg-architecture -qDEB_HOST_MULTIARCH) + PREFIX=/opt/${PKG_NAME}-${PKG_VER} LIB_DIR=${PREFIX}/lib @@ -41,6 +43,9 @@ echo # CC ... $CC echo # CXX ... $CXX echo # CPP ... $CPP +export CFLAGS_FOR_TARGET=-g -O2 -B/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -B/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -I/usr/include/${HOST_SYSTEM_MULTIARCH_TYPE} +export CXXFLAGS_FOR_TARGET=-g -O2 -B/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -B/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -I/usr/include/${HOST_SYSTEM_MULTIARCH_TYPE} + ##LD_PRELOAD_FOR_BUILD=${PREFIX}/lib/libgcc_s.so.1 ##export LD_PRELOAD=${LD_PRELOAD_FOR_BUILD}
Bug#637342: Trivial: File relicts from SVN repository
On Wed, Aug 10, 2011 at 7:35 PM, Aurelien Jarno aurel...@aurel32.net wrote: tag 637342 + unreproducible thanks On Wed, Aug 10, 2011 at 04:36:24PM +0200, Sedat Dilek wrote: Package: libc-bin Version: 2.13-16 Severity: normal Hi, on d-u I noticed some SVN file relicts. # dpkg -S /etc/ld.so.conf.d/.svn/entries /etc/ld.so.conf.d/.svn/text-base/libc.conf.svn-base libc-bin: /etc/ld.so.conf.d/.svn/entries libc-bin: /etc/ld.so.conf.d/.svn/text-base/libc.conf.svn-base # dpkg -L libc-bin | grep -e .svn /etc/ld.so.conf.d/.svn/entries /etc/ld.so.conf.d/.svn/text-base/libc.conf.svn-base Have you build this package yourself? These files are not in the packages that is present on the mirror: $ wget -q http://ftp.debian.org/debian/pool/main/e/eglibc/libc-bin_2.13-16_i386.deb $ dpkg -c libc-bin_2.13-16_i386.deb | grep svn $ -- Aurelien Jarno GPG: 1024D/F1BCDB73 aurel...@aurel32.net http://www.aurel32.net As said on IRC, it was indeed an issue in my own packages which I rebuilt with debian-dir from glibc SVN trunk. Sorry for the noise and letting you wait for my answer. - Sedat - -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczux-lsq7ztrd9xcnwzragiqsjlymxu9tuopbtt_hoer...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
On Wed, Aug 10, 2011 at 5:15 AM, Jonathan Nieder jrnie...@gmail.com wrote: Hi, Sedat Dilek wrote: Do I need additional patches to gcc trunk (I am using the weekly gcc-4.7 snapshot tarballs)? No need for patches. My build of gcc trunk finally finished: $ ./configure --prefix=$HOME/opt/gcc \ CFLAGS_FOR_TARGET=-g -O2 -B/usr/lib/x86_64-linux-gnu \ CXXFLAGS_FOR_TARGET=-g -O2 -B/usr/lib/x86_64-linux-gnu $ make FLAGS_FOR_TARGET='-B$(build_tooldir)/bin/ '\ '-B$(build_tooldir)/lib/ -isystem $(build_tooldir)/include '\ '-isystem $(build_tooldir)/sys-include '\ '-B/usr/lib/x86_64-linux-gnu' \ BOOT_CFLAGS=-g -O2 -B/usr/lib/x86_64-linux-gnu $ make FLAGS_FOR_TARGET='-B$(build_tooldir)/bin/ '\ '-B$(build_tooldir)/lib/ -isystem $(build_tooldir)/include '\ '-isystem $(build_tooldir)/sys-include '\ '-B/usr/lib/x86_64-linux-gnu' \ BOOT_CFLAGS=-g -O2 -B/usr/lib/x86_64-linux-gnu \ install The resulting compiler even seems to work, as long as you pass -B/usr/lib/x86_64-linux-gnu to it. Yes, it ought to be easier. Maybe FLAGS_FOR_TARGET does everything we need. I'd be happy to review a patch that teaches configure.ac to handle something like $ ./configure --prefix=$HOME/opt/gcc \ --extra-flags-for-target=-B/usr/lib/x86_64-linux-gnu \ BOOT_FLAGS='-g -O2 -B/usr/lib/x86_64-linux-gnu' $ make $ make install [...] I heard of a gcc multiarch patchset sent to upstream from debian-gcc team - you have an URL? I'd be interested in the answer to this if there is one, too. Thanks, Sedat. Jonathan First of all, thanks for jumping into the issue and sharing your thoughts and results/testing. For i386 I needed to expicitly set -I option with -B option (see [1]), otherwise I had several breakages with gnu/stubs-32.h when playing with *_FOR_TARGET options. ### List of *_FOR_TARGET options: 1. CFLAGS_FOR_TARGET 2. CXXFLAGS_FOR_TARGET 3. LDFLAGS_FOR_TARGET 4. FLAGS_FOR_TARGET ### Like you I have explicitly set, but with -I option (always combined with below options): $ export CFLAGS_FOR_TARGET='-g -O2 -B/usr/lib/i386-linux-gnu -I/usr/include/i386-linux-gnu' $ export CXXFLAGS_FOR_TARGET='-g -O2 -B/usr/lib/i386-linux-gnu -I/usr/include/i386-linux-gnu' ### Playing with LDFLAGS_FOR_TARGET: 1st-Try: export LDFLAGS_FOR_TARGET=-L/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -Wl,-rpath-link=/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE}:/lib/${HOST_SYSTEM_MULTIARCH_TYPE}:/usr/lib 2nd-Try: export LDFLAGS_FOR_TARGET=-B/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -I/usr/include/${HOST_SYSTEM_MULTIARCH_TYPE} 3rd-Try: export LDFLAGS_FOR_TARGET=-L/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -Wl,-rpath-link=/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} NOPE: All try-outs are passed-by but did not work as expected, I got crt*.o was not found. But, isn't LDFLAGS searching for so-libs? ### Playing with FLAGS_FOR_TARGET: $ make FLAGS_FOR_TARGET=${FLAGS_FOR_TARGET} -B/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -I/usr/include/${HOST_SYSTEM_MULTIARCH_TYPE} OK: This leads to a successful build, did not test the compiler yet (flags by-passed to xgcc). ANYWAY, these workarounds should not be necessary, it should work OOTB. Not sure, what's wrong. What do the experts say? - Sedat - [1] http://anonscm.debian.org/viewvc/pkg-glibc?view=revisionrevision=4863 -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczuw__9slouodhmjnwlfa+lopcrsex6dwmjgmqag1wt1...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
On Wed, Aug 10, 2011 at 3:09 PM, Sedat Dilek sedat.di...@googlemail.com wrote: On Wed, Aug 10, 2011 at 5:15 AM, Jonathan Nieder jrnie...@gmail.com wrote: Hi, Sedat Dilek wrote: [...] ### Playing with FLAGS_FOR_TARGET: $ make FLAGS_FOR_TARGET=${FLAGS_FOR_TARGET} -B/usr/lib/${HOST_SYSTEM_MULTIARCH_TYPE} -I/usr/include/${HOST_SYSTEM_MULTIARCH_TYPE} OK: This leads to a successful build, did not test the compiler yet (flags by-passed to xgcc). When building mesa wizh new gcc-4.7: configure:3094: error: C compiler cannot create executables 15 compiler builds for nothing, grr. - Sedat - -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczuwmtt7-9wfn55zbw3qjijaw_pn9qn8pyypdpkmvgvv...@mail.gmail.com
Bug#637342: Trivial: File relicts from SVN repository
Package: libc-bin Version: 2.13-16 Severity: normal Hi, on d-u I noticed some SVN file relicts. # dpkg -S /etc/ld.so.conf.d/.svn/entries /etc/ld.so.conf.d/.svn/text-base/libc.conf.svn-base libc-bin: /etc/ld.so.conf.d/.svn/entries libc-bin: /etc/ld.so.conf.d/.svn/text-base/libc.conf.svn-base # dpkg -L libc-bin | grep -e .svn /etc/ld.so.conf.d/.svn/entries /etc/ld.so.conf.d/.svn/text-base/libc.conf.svn-base - Sedat - -- System Information: Debian Release: wheezy/sid APT prefers unstable APT policy: (500, 'unstable'), (200, 'experimental') Architecture: i386 (i686) Kernel: Linux 3.0.1-1-rt-686-pae (SMP w/1 CPU core; PREEMPT) Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash -- no debconf information -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczuuhpqbsy8wo9pff61a8h1225v5rvfc55jrx4cdarsb...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
On Wed, Aug 10, 2011 at 4:47 PM, Jonathan Nieder jrnie...@gmail.com wrote: Sedat Dilek wrote: When building mesa wizh new gcc-4.7: configure:3094: error: C compiler cannot create executables But of course. You'll need to pass 'CC=gcc -B/usr/lib/... -I/usr/include/...' to mesa's configure script. You can experience the same thing by building a hello world program, too. NAH... the gcc with my symlinks workaround for gnu/stubs-32.h and crt*.o works w/o such hackz. This is a step back :-(. - Sedat - -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczuu5isueo+evcyfnxwooxfpulucwwgrehrpa6b8vumw...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
Package: libc6-dev-amd64 Version: 2.13-16 Severity: normal Hi, I thought this stuff is fixed? When building gcc-4.7 snapshot, I get again this gnu/stubs-32.h error. My build breaks like this: ... /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory compilation terminated. Investigating: $ LC_ALL=C ls -l /usr/include/gnu/ total 20 -rw-r--r-- 1 root root 2943 Aug 9 15:41 lib-names.h -rw-r--r-- 1 root root 1337 Aug 9 15:41 libc-version.h -rw-r--r-- 1 root root 1813 Aug 9 15:45 option-groups.h -rw-r--r-- 1 root root 604 Aug 9 15:45 stubs-64.h -rw-r--r-- 1 root root 315 Aug 9 15:41 stubs.h $ dpkg -S /usr/include/gnu/stubs.h libc6-dev-amd64: /usr/include/gnu/stubs.h ( Not sure if this a good idea that /usr/include/{bits,gnu,sys}/*.h belongs to libc6-dev-amd64 package. ) In previous eglibc packages {bits,gnu,sys} dirs were symlinked to i386-linux-gnu/{bits,gnu,sys}, now: $ l /usr/include/i386-linux-gnu/ insgesamt 72 drwxr-xr-x 5 root root 4096 9. Aug 15:51 . drwxr-xr-x 116 root root 20480 9. Aug 15:51 .. drwxr-xr-x 2 root root 12288 9. Aug 15:51 bits -rw-r--r-- 1 root root 10883 11. Jun 21:58 ffi.h -rw-r--r-- 1 root root 3496 11. Jun 21:58 ffitarget.h -rw-r--r-- 1 root root 3291 9. Aug 15:17 fpu_control.h drwxr-xr-x 2 root root 4096 9. Aug 15:51 gnu drwxr-xr-x 2 root root 12288 9. Aug 15:51 sys $ l /usr/include/i386-linux-gnu/gnu/stubs* -rw-r--r-- 1 root root 624 9. Aug 15:20 /usr/include/i386-linux-gnu/gnu/stubs-32.h -rw-r--r-- 1 root root 315 9. Aug 15:17 /usr/include/i386-linux-gnu/gnu/stubs.h XXX: Workaround # cd /usr/include/gnu # ln -sf ../i386-linux-gnu/gnu/stubs-32.h ./ # LC_ALL=C l stubs*.h lrwxrwxrwx 1 root root 32 Aug 9 17:16 stubs-32.h - ../i386-linux-gnu/gnu/stubs-32.h -rw-r--r-- 1 root root 604 Aug 9 15:45 stubs-64.h -rw-r--r-- 1 root root 315 Aug 9 15:41 stubs.h This lets the build continue, until crt*.o breakage... Regards, - Sedat - P.S.: I have rebuilt -16 with gcc-4.6 and tested against -16 from official repo. -- System Information: Debian Release: wheezy/sid APT prefers unstable APT policy: (500, 'unstable'), (200, 'experimental') Architecture: i386 (i686) Kernel: Linux 3.0.1-1-rt-686-pae (SMP w/1 CPU core; PREEMPT) Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages libc6-dev-amd64 depends on: ii libc6-amd64 2.13-16 Embedded GNU C Library: 64bit Shar ii libc6-dev 2.13-16 Embedded GNU C Library: Developmen Versions of packages libc6-dev-amd64 recommends: ii gcc-multilib 4:4.6.1-2 GNU C compiler (multilib files) libc6-dev-amd64 suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CA+icZUUOCOiU18sSBbJyKYY2Swan1ir0UR-36x-Na_TybX=e...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
On Tue, Aug 9, 2011 at 6:23 PM, Jonathan Nieder jrnie...@gmail.com wrote: retitle 637218 [i386] cannot use gnu/stubs.h with multiarch-unaware toolchain (fatal error: gnu/stubs-32.h: No such file or directory) quit Jonathan Nieder wrote: retitle 637218 libc6-dev: moving headers to /usr/include/triplet breaks external multiarch unaware applications [...] This lets the build continue, until crt*.o breakage... They are one and the same problem. On second thought, I think I missed the point. I would have thought that /usr/include/gnu would be a symlink to /usr/include/i386-linux-gnu/gnu on i386. What does ls -ld /usr/include/gnu say? It's now a directory containing header-files from libc6-dev-amd64 package (see -16 changelog). root@tbox:~# ls -ld /usr/include/gnu drwxr-xr-x 2 root root 4096 9. Aug 18:20 /usr/include/gnu ( I can live with my gnu/stubs-32.h workaround ). Trying your workaround from [1], building (removed all my workarounds)... - Sedat - [1] http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=80;bug=629819 -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CA+icZUXghHNv=8V0Wj5q=-sn_3vcgdb-thsmjrmlyn_x_vz...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
On Tue, Aug 9, 2011 at 7:23 PM, Aurelien Jarno aurel...@aurel32.net wrote: merge 629819 637218 thanks On Tue, Aug 09, 2011 at 05:30:33PM +0200, Sedat Dilek wrote: Package: libc6-dev-amd64 Version: 2.13-16 Severity: normal Hi, I thought this stuff is fixed? When building gcc-4.7 snapshot, I get again this gnu/stubs-32.h error. My build breaks like this: ... /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory compilation terminated. [...] $ l /usr/include/i386-linux-gnu/gnu/stubs* -rw-r--r-- 1 root root 624 9. Aug 15:20 /usr/include/i386-linux-gnu/gnu/stubs-32.h -rw-r--r-- 1 root root 315 9. Aug 15:17 /usr/include/i386-linux-gnu/gnu/stubs.h stubs-32.h is only there because it's an i386 only header. The problem is that your compiler is not multiarch aware. Everything is correct in the package side. Just to clarify, I am building with Debian's gcc-4.6 (4.6.1-6) from i386/sid. It's basically the same bug than #629819 (crti.o, crt1.o), but for the headers. The same workaround, explained in NEWS.Debian.gz applies. I am therefore merging the bugs. [ /usr/share/doc/libc6/NEWS.Debian.gz ] eglibc (2.13-11) unstable; urgency=low Starting with the eglibc package version 2.13-5, the libraries are shipped in the multiarch directory /lib/$arch instead of the more traditional /lib. The toolchain in Debian has been updated to cope with that, and most build systems should be unaffected. If you are using a non-Debian toolchain to build your software and it is not able to cope with multiarch, you might try to pass the following option to your compiler: -B/usr/lib/$arch -- Aurelien Jarno aure...@debian.org Sat, 23 Jul 2011 23:42:46 +0200 [...] Passing -B/usr/lib/$arch did not help here (for both issue gnu/stubs-32.h and crt*.o)! What shall I do next - report a bug against gcc-4.6 package? - Sedat - -- Aurelien Jarno GPG: 1024D/F1BCDB73 aurel...@aurel32.net http://www.aurel32.net -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczuvqxy4rlk0u69rwt5m9u3_aw6z8hkdvpww3gp+n1ut...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
On Tue, Aug 9, 2011 at 9:22 PM, Aurelien Jarno aurel...@aurel32.net wrote: On Tue, Aug 09, 2011 at 07:33:57PM +0200, Sedat Dilek wrote: On Tue, Aug 9, 2011 at 7:23 PM, Aurelien Jarno aurel...@aurel32.net wrote: merge 629819 637218 thanks On Tue, Aug 09, 2011 at 05:30:33PM +0200, Sedat Dilek wrote: Package: libc6-dev-amd64 Version: 2.13-16 Severity: normal Hi, I thought this stuff is fixed? When building gcc-4.7 snapshot, I get again this gnu/stubs-32.h error. My build breaks like this: ... /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory compilation terminated. [...] $ l /usr/include/i386-linux-gnu/gnu/stubs* -rw-r--r-- 1 root root 624 9. Aug 15:20 /usr/include/i386-linux-gnu/gnu/stubs-32.h -rw-r--r-- 1 root root 315 9. Aug 15:17 /usr/include/i386-linux-gnu/gnu/stubs.h stubs-32.h is only there because it's an i386 only header. The problem is that your compiler is not multiarch aware. Everything is correct in the package side. Just to clarify, I am building with Debian's gcc-4.6 (4.6.1-6) from i386/sid. Without any change? gcc-4.6 from Debian is supposed to be patched, and should find gnu/stubs-32.h into /usr/include/i386-linux-gnu/ . It looks like a bug in the gcc-4.6 package. On the other hand I am not able to reproduce the bug with glibc 2.13-16 and gcc-4.6 4.6.1-6. What are you doing to reproduce the issue? Can you give an sample-script? Can you explain how to pass -B/usr/lib/$DEB_HOST_MULTIARCH option? Is this correct (which did not work here BTW)? $ export CFLAGS_FOR_TARGET=-g -O2 -B/usr/lib/$DEB_HOST_MULTIARCH Backlog from #multiarch: [22:54:17] vorlon DEB_CPPFLAGS_APPEND=-I/usr/include DEB_LDFLAGS_APPEND=-L/usr/lib/arm-linux-gnueabi -Wl,-rpath-link=/usr/lib/arm-linux-gnueabi:/lib/arm-linux-gnueabi:/usr/lib debuild -uc -us -B -aarmel Some could force to include/look-for-libs with such settings (of course should be adapted to i386-linux-gnu), but as far as I understood you this is obsolete and used automagically? I am building now GCC 4.6-20110805 Snapshot, to see if it is a problem with GCC 4.7-20110806 Snapshot [2]. - Sedat - [1] http://ftp-stud.fht-esslingen.de/Mirrors/sourceware.org/gcc/snapshots/LATEST-4.6/gcc-4.6-20110805.tar.bz2 [2] http://ftp-stud.fht-esslingen.de/Mirrors/sourceware.org/gcc/snapshots/LATEST-4.7/gcc-4.7-20110806.tar.bz2 -- Aurelien Jarno GPG: 1024D/F1BCDB73 aurel...@aurel32.net http://www.aurel32.net -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ca+iczuvlpjvmoe8tqtkzu1gqmwhh_zme1gxs6ykemvxwxif...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
On Tue, Aug 9, 2011 at 9:48 PM, Jonathan Nieder jrnie...@gmail.com wrote: Sedat Dilek wrote: [...] I've just started a new gcc build with $ git reset --hard origin/master $ git clean -fdx $ ./configure --prefix=$HOME/opt/gcc \ CFLAGS_FOR_TARGET=-g -O2 -B/usr/lib/x86_64-linux-gnu $ make BOOT_CFLAGS=-g -O2 -B/usr/lib/x86_64-linux-gnu Can you explain why you use CFLAGS_FOR_TARGET as configure-option only and BOOT_CFLAGS as make-option? Is this export not enough? $ export CFLAGS_FOR_TARGET=-g -O2 -B/usr/lib/$DEB_HOST_MULTIARCH - Sedat - Seems to be working well so far; I'm cautiously optimistic. gcc is 4.6.1-6, libc 2.13-16. If that works, next step will be to try a build on i386. I'm cautiously optimistic. -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CA+icZUUfU1C9hOid0nghz6apquE=wjgtadrh7dh-r3vnp__...@mail.gmail.com
Bug#637218: /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory
On Tue, Aug 9, 2011 at 9:32 PM, Jonathan Nieder jrnie...@gmail.com wrote: Aurelien Jarno wrote: Without any change? gcc-4.6 from Debian is supposed to be patched, and should find gnu/stubs-32.h into /usr/include/i386-linux-gnu/ . It looks like a bug in the gcc-4.6 package. On the other hand I am not able to reproduce the bug with glibc 2.13-16 and gcc-4.6 4.6.1-6. I am guessing Sedat is building gcc trunk. The failing commands are using xgcc, so the relevant compiler is (unpatched) upstream gcc, which does not look at multiarch paths unless you tell it to. Do I need additional patches to gcc trunk (I am using the weekly gcc-4.7 snapshot tarballs)? From Debian gcc-snapshot (aka gcc-4.7) package - which ones? I heard of a gcc multiarch patchset sent to upstream from debian-gcc team - you have an URL? - Sedat - -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CA+icZUV=mojh2ou7c8mm4nm6siyju12msxsornjbazzuqbd...@mail.gmail.com
Bug#636116: Overwriting identical include header files fails from libc6-dev and libc6-dev-amd64
Package: libc6-dev Version: 2.13-13 Severity: normal Hi, I upgraded to eglibc (2.13-13) and see the following error-messages: dpkg: error processing /var/cache/apt/archives/libc6-dev-amd64_2.13-13_i386.deb (--unpack): trying to overwrite '/usr/include/bits', which is also in package libc6-dev 2.13-13 When asking on #multiarch and looking into the libc6 changelog, the problem might occur from: eglibc (2.13-12) unstable; urgency=low ... * Fix installation of biarch headers (Closes: #635685): - Use a symlink for bits/ and gnu/ directories ... I have de-installed all multilib stuff for now. Regards, - Sedat - -- System Information: Debian Release: wheezy/sid APT prefers unstable APT policy: (500, 'unstable'), (200, 'experimental') Architecture: i386 (i686) Kernel: Linux 3.0.0-1-rt-686-pae (SMP w/1 CPU core; PREEMPT) Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages libc6-dev depends on: ii libc-dev-bin 2.13-13Embedded GNU C Library: Developmen ii libc6 2.13-13Embedded GNU C Library: Shared lib ii linux-libc-dev3.0.0-1Linux support headers for userspac Versions of packages libc6-dev recommends: ii gcc [c-compiler] 4:4.6.1-2 GNU C compiler ii gcc-4.4 [c-compiler] 4.4.6-7GNU C compiler ii gcc-4.5 [c-compiler] 4.5.3-3The GNU C compiler ii gcc-4.6 [c-compiler] 4.6.1-5GNU C compiler Versions of packages libc6-dev suggests: pn glibc-doc none (no description available) pn manpages-dev none (no description available) -- no debconf information -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CA+icZUVJKEmKJPjmJYyWDwZGKs5PJN8LZyO=xfs4gdrtevw...@mail.gmail.com
Bug#622116: eglibc: patch: unrecognized option '--unified-reject-files' (wrong QUILT_PATCH_OPTS)
Package: libc6 Version: 2.11.2-11 Severity: normal Tags: patch Hi, while discussing Debian BR #619963 with Edwin Török I tried to apply two glibc/upstream patches he proposed. By unpacking the eglibc source, I fell over: ... Applying patches... Applying patch locale/check-unknown-symbols.diff patch: unrecognized option '--unified-reject-files' patch: Try `patch --help' for more information. Patch locale/check-unknown-symbols.diff does not apply (enforce with -f) make: *** [/home/sd/src/eglibc/eglibc-2.11.2/stamp-dir/patch] Error 1 dpkg-buildpackage: error: debian/rules build gave error exit status 2 debuild: fatal error at line 1329: dpkg-buildpackage -rfakeroot -d -us -uc failed What's the help of patch saying? $ patch --help | grep reject-format --reject-format=FORMAT Create 'context' or 'unified' rejects. BTW, patch has (2.6.1-1) version here. So, I started my investigations and checked debian/quiltrc and changed: -QUILT_PATCH_OPTS=--unified-reject-files +QUILT_PATCH_OPTS=--reject-format=unified This lets debuild apply patches again. Regards, - Sedat - P.S.: Patch quiltrc.diff attached -- System Information: Debian Release: wheezy/sid APT prefers unstable APT policy: (500, 'unstable'), (200, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.39-rc2-next20110408.4-686-small (SMP w/1 CPU core) Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages libc6 depends on: ii libc-bin 2.11.2-11 Embedded GNU C Library: Binaries ii libgcc1 1:4.6.0-2 GCC support library Versions of packages libc6 recommends: ii libc6-i6862.11.2-11 Embedded GNU C Library: Shared lib Versions of packages libc6 suggests: ii debconf [debconf-2.0] 1.5.38 Debian configuration management sy pn glibc-doc none (no description available) ii locales 2.11.2-11 Embedded GNU C Library: National L -- debconf information: * glibc/upgrade: true glibc/disable-screensaver: glibc/restart-failed: * glibc/restart-services: cups cron --- debian/quiltrc.orig 2011-04-10 11:48:36.0 +0200 +++ debian/quiltrc 2011-04-10 12:08:43.696587720 +0200 @@ -1,4 +1,4 @@ QUILT_PATCHES=debian/patches -QUILT_PATCH_OPTS=--unified-reject-files +QUILT_PATCH_OPTS=--reject-format=unified QUILT_DIFF_ARGS=--no-timestamps --no-index QUILT_REFRESH_ARGS=-pab --no-timestamps --no-index --diffstat
Bug#622116: eglibc: patch: unrecognized option '--unified-reject-files' (wrong QUILT_PATCH_OPTS)
On Sun, Apr 10, 2011 at 1:04 PM, Jonathan Nieder jrnie...@gmail.com wrote: unarchive 612540 forcemerge 612540 622116 quit Hi, Sedat Dilek wrote: -QUILT_PATCH_OPTS=--unified-reject-files +QUILT_PATCH_OPTS=--reject-format=unified Thanks for reporting. Should be fixed in 2.11.2-12. Thanks for the rocket-fast reply. Has eglibc (2.11.2-12) ever been released or it is going to be released? I see a Closes #612540 in eglibc/experimental (2.13-0exp5): ... [ Clint Adams ] * Patch from Nobuhiro Iwamatsu to cope with the removal of patch --unified-reject-files. closes: #612540. Can you enlighten what will be next version of eglibc in sid? 2.13? Can someone switch/jump to it w/o (big, any) concerns? - Sedat - [1] eglibc (2.11.2-12) -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/banlktikoo69q3wf7i-aczsowtlyz1mk...@mail.gmail.com
Bug#526823: [glibc] DNS problems with version = (2.9-8) - Can you provide i386 packages for testing?
Hi Aurelien, since upgrading to glibc (2.9-8) and higher I have serious DNS problems. Using OpenDNS dns-servers [1] and adding single-request to /etc/resolv.conf did *not* help. For amd64 arch you provided deb-files, can you offer packages for i386? Is it possible to re-add glibc (2.9-7) back to the mirrors or give a web-link where I can download it? Thanks in advance. Kind Regards, Sedat [1] Use OpenDNS https://www.opendns.com/start/ -- To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org