This is the mail archive of the crossgcc@sourceware.org mailing list for the crossgcc project.
See the CrossGCC FAQ for lots more information.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
Hi, On Wed, Sep 1, 2010 at 12:43 PM, Doug Crawford <dcraw101@yahoo.com> wrote: > I ran the build again with "Build shared companion libraries" checked, but now > it complains about not finding GMP version 4.1.3 duing the ppl configure stage. > After looking through the mail archives I see that others have this problem > also, but I did not see a definitive solution.? Just for grins I decided to try > a canadian cross compile from my Ubuntu 10.04 system where build=x86-linux, > host=mingw and target=powerpc-linux-uclibc.? During the first toolchain build > iteration where build/host=x86-linux, target=mingw I encountered the same ppl > configure error as my cygwin host?build: "configure: error: Cannot find GMP > version 4.1.3 or higher."? In one of the previous emails someone mentioned that > it was related to the version of libstcc++.?? On my Ubuntu system my > /usr/lib/libstdc++ is pointing to libstdc++.so.6.0.13. > > config for my build/host=cygwin target=powerpc-linux-uclibc > $ ct-ng show-config > powerpc-860-linux-uclibc? [l? ] > ??? OS???????????? : linux-2.6.34.1 > ??? Companion libs : gmp-4.3.2 mpfr-3.0.0 ppl-0.10.2 cloog-ppl-0.15.9 > ??? binutils?????? : binutils-2.20.1 > ??? C compiler???? : gcc-4.4.3 (C,C++) > ??? C library????? : uClibc-0.9.30.2 > ??? Tools????????? : gdb-6.8 strace-4.5.19 > > config for my build/host=x86-linux target=mingw > i386-unknown-mingw32? [l X] > ??? OS???????????? : mingw32 > ??? Companion libs : gmp-5.0.1 mpfr-3.0.0 ppl-0.10.2 cloog-ppl-0.15.9 > ??? binutils?????? : binutils-2.20.1 > ??? C compiler???? : gcc-4.4.3 (C,C++) > ??? C library????? : mingw > ??? Tools????????? : > there is a known issue and a patch for gmp-5.0.1/ppl-0.10.2, I'm not sure being aware of anything between prior gmp and ppl. Concerning your original link trouble I'm not sure ld(1) behave the same on *nix and cygwin. I found an old post concerning a similar behavior: http://osdir.com/ml/sysutils.autoconf.general/2004-06/msg00104.html I confirmed this with a trivial test program; On cygwin: % gcc -lz main.c # FAILS: undefined refs to `gzopen' % gcc main.c -lz # WORKS both works on *nix. This might be related to PE/COFF import handling (say some IRC guy on #cygwin). Anyhow, `-lstdc++' must definitively appear after, as it is done in 4.5. We are just lucky (and plainly wrong) to hijack LDFLAGS to specify -lstdc++ before. The short answer might be "static linking of companion library prior to gcc 4.5 is not reliable", beside hack like: diff --git a/gcc/Makefile.in b/gcc/Makefile.in index 88c40b6..9881ac4 100644 --- a/gcc/Makefile.in +++ b/gcc/Makefile.in @@ -1642,8 +1642,8 @@ gcc-cross$(exeext): xgcc$(exeext) dummy-checksum.o : dummy-checksum.c cc1-dummy$(exeext): $(C_OBJS) dummy-checksum.o $(BACKEND) $(LIBDEPS) - $(CC) $(ALL_CFLAGS) $(LDFLAGS) -o $@ $(C_OBJS) dummy-checksum.o \ - $(BACKEND) $(LIBS) $(BACKENDLIBS) + $(CC) $(ALL_CFLAGS) -o $@ $(C_OBJS) dummy-checksum.o \ + $(BACKEND) $(LIBS) $(BACKENDLIBS) $(LDFLAGS) cc1-checksum.c : cc1-dummy$(exeext) build/genchecksum$(build_exeext) build/genchecksum$(build_exeext) cc1-dummy$(exeext) > $@ @@ -1651,8 +1651,8 @@ cc1-checksum.c : cc1-dummy$(exeext) build/genchecksum$(build_exeext) cc1-checksum.o : cc1-checksum.c cc1$(exeext): $(C_OBJS) cc1-checksum.o $(BACKEND) $(LIBDEPS) - $(CC) $(ALL_CFLAGS) $(LDFLAGS) -o $@ $(C_OBJS) cc1-checksum.o \ - $(BACKEND) $(LIBS) $(BACKENDLIBS) + $(CC) $(ALL_CFLAGS) -o $@ $(C_OBJS) cc1-checksum.o \ + $(BACKEND) $(LIBS) $(BACKENDLIBS) $(LDFLAGS) which should work. - Arnaud -- For unsubscribe information see http://sourceware.org/lists.html#faq
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |