From: Dongsheng S. <don...@gm...> - 2010-07-28 12:32:51
|
于 2010-7-28 16:02, Kai Tietz 写道: > 2010/7/28 Dongsheng Song <don...@gm...>: >> 于 2010-7-28 15:43, Kai Tietz 写道: >>> 2010/7/28 Dongsheng Song <don...@gm...>: >>>> Hi Kai, >>>> >>>> When we cross build gcc 4.5 for windows, I found we can build windows gcc binary one >>>> week ago, but now the build failed. >>>> >>>> After I do a binary search, I found the issue caused by r2945. >>>> >>>> r2950 | 2010-07-24 05:50:28 | FAILED >>>> r2945 | 2010-07-24 02:44:15 | FAILED >>>> r2944 | 2010-07-24 02:38:30 | SUCCESS >>>> r2939 | 2010-07-23 17:55:30 | SUCCESS >>>> r2928 | 2010-07-23 05:21:20 | SUCCESS >>>> r2924 | 2010-07-22 18:32:25 | SUCCESS >>>> >>>> r2945 remove some *IMPORTANT* macros from /trunk/mingw-w64-headers/crt/float.h, >>>> e.g. FLT/DBL/LDBL_MANT_DIG, FLT_EVAL_METHOD, *ALL* decimal macros (DEC32/64/128_*, ...) >>>> >>>> When I add FLT/DBL/LDBL_MANT_DIG and FLT_EVAL_METHOD back to /trunk/mingw-w64-headers/crt/float.h, >>>> then the gcc cross build success again. >>>> >>>> So I recommend you apply the attached patch at least. >>>> >>>> btw, I know FLT_EVAL_METHOD added by C99, but libgfortran/m4/nearest.m4 use it, >>>> is it mean we should use ISO C99 compiler to build gcc 4.5 or later, not ISO C90 as >>>> http://gcc.gnu.org/install/prerequisites.html ? >>>> >>>> Regards, >>>> Dongsheng >>> >>> Hello Dongsheng, >>> >>> the recent change to float.h was necessary to support the new >>> include_next patch of 4.6. So how are you exactly installing headers? >>> As usual you should just see gcc's internal float.h for older gcc's >>> then 4.6. So I am a bit puzzled. Are you removing gcc's float.h here? >>> >>> Regards, >>> Kai >>> >> >> Hi Kai, >> >> I use Debian 6.0 i686 with latest update, gcc 4.4.4 build a gcc 4.5 cross compiler for windows, >> then use the cross compiler to build a native gcc 4.5 compiler for windows. >> >> Without the patch, both i686-windows and x64-windows failed during build native >> compiler. >> >> It's strange since I can build cross compiler, it maybe a gcc bug. >> >> The related packages is: >> gcc 4.5 branch, mingw64 trunk, binutils trunk, gmp 5.0 branch, mpfr 3.0 branch, >> mpc 0.8.2, ppl 0.10.2, cloog-ppl 0.15.9. >> >> Regards, >> Dongsheng >> >> > > Well, yes it is a gcc bug in respect to native/cross toolchains. I > assume that your search path installs headers (and libraries) in > standard_include for native. This cause that the system-headers get > included before fixed-include and gcc-include. > For this I can provide a patch. See revision 2986. But indeed this > include-order of gcc is a conceptional flaw. > > Cheers, > Kai > This build error fixed now, thank your excellent work ! Thank you very much ! But new error occurred when I use cross compiler to build native compiler: ... i686-w64-mingw32-gcc -O2 -pipe -DIN_GCC -W -Wall -Wwrite-strings -Wcast-qual -Wstrict-prototypes -Wmissing-prototypes -Wmissing-format-attribute -pedantic -Wno-long-long -Wno-variadic-macros -Wno-overlength-strings -Wold-style-definition -Wc++-compat -DHAVE_CONFIG_H -s -o f951.exe \ fortran/arith.o fortran/array.o fortran/bbt.o fortran/check.o fortran/cpp.o fortran/data.o fortran/decl.o fortran/dump-parse-tree.o fortran/error.o fortran/expr.o fortran/interface.o fortran/intrinsic.o fortran/io.o fortran/iresolve.o fortran/match.o fortran/matchexp.o fortran/misc.o fortran/module.o fortran/openmp.o fortran/options.o fortran/parse.o fortran/primary.o fortran/resolve.o fortran/scanner.o fortran/simplify.o fortran/st.o fortran/symbol.o fortran/target-memory.o fortran/convert.o fortran/dependency.o fortran/f95-lang.o fortran/trans.o fortran/trans-array.o fortran/trans-common.o fortran/trans-const.o fortran/trans-decl.o fortran/trans-expr.o fortran/trans-intrinsic.o fortran/trans-io.o fortran/trans-openmp.o fortran/trans-stmt.o fortran/trans-types.o main.o libbackend.a ../libcpp/libcpp.a ../libdecnumber/libdecnumber.a ../libcpp/libcpp.a ../libiberty/libiberty.a ../libdecnumber/libdecnumber.a attribs.o -L/home/oracle/gcc-4.5-w32_i686-linux/lib -lcloog -L/home/oracle/gcc-4.5-w32_i686-linux/lib -lppl_c -lppl -lgmpxx -L/home/oracle/gcc-4.5-w32/lib -L/home/oracle/gcc-4.5-w32/lib -L/home/oracle/gcc-4.5-w32/lib -lmpc -lmpfr -lgmp -L../zlib -lz /home/oracle/gcc-4.5-w32_i686-linux/lib/libppl_c.a(ppl_c_implementation_common.o): In function `__tcf_0': ... /home/oracle/gcc-4.5-w32/lib/libgmpxx.a(isfuns.o):isfuns.cc:(.text+0x2be): undefined reference to `std::ios_base::Init::Init()' collect2: ld returned 1 exit status make[2]: *** [cc1plus-dummy.exe] Error 1 rm fsf-funding.pod gcov.pod gfdl.pod cpp.pod gcc.pod gfortran.pod make[2]: Leaving directory `/home/oracle/tmp/gcc-4.5-w32-obj/gcc/gcc' make[1]: *** [all-gcc] Error 2 make[1]: Leaving directory `/home/oracle/tmp/gcc-4.5-w32-obj/gcc' make: *** [all] Error 2 Then I run: sed -i "s/-lgmpxx/-lgmpxx -lstdc++/g" ${NATIVE_OBJ_ROOT}/gcc/gcc/Makefile make all So I can build smoothly. Is this a known bug ? Regards, Dongsheng |