Build Failure on 32-bit Ubuntu 15.10

[ghost]

Building Io (current git version) fails due to its uses of SSE2 in UArray_math.c. The problem is that -msse2 needs to be specified in the compiler flags. This only applies to 32-bit builds because all 64-bit systems have this option defined implicitly. It seems to be related to the change from GCC 4 to 5, maybe earlier systems did not need this flag.

System info:

$ uname -a
Linux system 4.2.0-25-generic #30-Ubuntu SMP Mon Jan 18 12:32:18 UTC 2016 i686 i686 i686 GNU/Linux
$ gcc --version
gcc (Ubuntu 5.2.1-22ubuntu2) 5.2.1 20151010

Error message:

[ 10%] Building C object libs/basekit/CMakeFiles/basekit.dir/source/UArray_math.c.o
In file included from /home/peter/io/libs/basekit/source/simd_cph/include/simd_cp.h:129:0,
                 from /home/peter/io/libs/basekit/source/UArray_math.c:22:
/home/peter/io/libs/basekit/source/UArray_math.c: In function ‘vfloat32_add’:
/home/peter/io/libs/basekit/source/simd_cph/include/simd_cp_x86.h:172:39: warning: SSE vector return without SSE enabled changes the ABI [-Wpsabi]
    #define simd_4f_add(v1,v2,v0)  v0.v=_mm_add_ps(v1.v,v2.v)
                                       ^
/home/peter/io/libs/basekit/source/UArray_math.c:44:30: note: in expansion of macro ‘simd_4f_add’
  for(i = 0; i < max; i ++) { simd_4f_ ## VNAME (a[i], b[i], a[i]); }\
                              ^
/home/peter/io/libs/basekit/source/UArray_math.c:49:1: note: in expansion of macro ‘LLVEC_DUALARG_OP’
 LLVEC_DUALARG_OP(add,  +=, float32);
 ^
In file included from /home/peter/io/libs/basekit/source/simd_cph/include/simd_cp_x86.h:47:0,
                 from /home/peter/io/libs/basekit/source/simd_cph/include/simd_cp.h:129,
                 from /home/peter/io/libs/basekit/source/UArray_math.c:22:
/usr/lib/gcc/i686-linux-gnu/5/include/xmmintrin.h:181:1: error: inlining failed in call to always_inline ‘_mm_add_ps’: target specific option mismatch
 _mm_add_ps (__m128 __A, __m128 __B)
 ^
In file included from /home/peter/io/libs/basekit/source/simd_cph/include/simd_cp.h:129:0,
                 from /home/peter/io/libs/basekit/source/UArray_math.c:22:
/home/peter/io/libs/basekit/source/simd_cph/include/simd_cp_x86.h:172:40: error: called from here
    #define simd_4f_add(v1,v2,v0)  v0.v=_mm_add_ps(v1.v,v2.v)
                                        ^
/home/peter/io/libs/basekit/source/UArray_math.c:44:30: note: in expansion of macro ‘simd_4f_add’
  for(i = 0; i < max; i ++) { simd_4f_ ## VNAME (a[i], b[i], a[i]); }\
                              ^
/home/peter/io/libs/basekit/source/UArray_math.c:49:1: note: in expansion of macro ‘LLVEC_DUALARG_OP’
 LLVEC_DUALARG_OP(add,  +=, float32);
 ^