Windows port

[calctopian]

Please note that the configure script may have some issues depending on your system:
- x86_64-w64-mingw32-gcc.exe (64 bits): detected size_t is "unsigned long long" but the configure doesn't accept it and cil.ml doesn't recognize it. Edit the configure script, search for "real_type=" and explicitly declare it (e.g., "real_type='unsigned long'"). Then run: "./configure CC=x86_64-w64-mingw32-gcc.exe" and "make RELEASE=1"
- i686-w64-mingw32 (32 bits): if you get the following error message "/usr/i686-w64-mingw32/sys-root/mingw/include/stdlib.h:336: Error: Undefined function", define out the function _abs64:
#ifdef __MINGW_INTRIN_INLINE
__MINGW_INTRIN_INLINE __int64 __cdecl _abs64(__int64 x) { return __builtin_llabs(x);}
#endif

Steps to generate oblivc.dll and oblivc.lib:

x86_64-w64-mingw32-gcc.exe -c -g atomic_queue.c -o atomic_queue.o -I .
x86_64-w64-mingw32-gcc.exe -c -g bcrandom.c -o bcrandom.o -I .
x86_64-w64-mingw32-gcc.exe -c -g commitReveal.c -o commitReveal.o -I .
x86_64-w64-mingw32-gcc.exe -c -g dualex.c -o dualex.o -I .
x86_64-w64-mingw32-gcc.exe -c -g nnob.c -o nnob.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_bits.c -o obliv_bits.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_network_utils.c -o obliv_network_utils.o -I .
x86_64-w64-mingw32-gcc.exe -c -g ot.c -o ot.o -I .
x86_64-w64-mingw32-gcc.exe -c -g privacy-free.c -o privacy-free.o -I .
x86_64-w64-mingw32-gcc.exe -c -g psi.c -o psi.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_float_add.c -o obliv_float_add.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_float_div.c -o obliv_float_div.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_float_eq.c -o obliv_float_eq.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_float_le.c -o obliv_float_le.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_float_lt.c -o obliv_float_lt.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_float_mult.c -o obliv_float_mult.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_float_neg.c -o obliv_float_neg.o -I .
x86_64-w64-mingw32-gcc.exe -c -g obliv_float_sub.c -o obliv_float_sub.o -I .

x86_64-w64-mingw32-gcc.exe -shared -g -o oblivc.dll *.o -Wl,--output-def,oblivc.def,--out-implib,oblivc.a -lgcrypt -pthread -lws2_32

lib /machine:x64 /def:oblivc.def

Compiling with Obliv-C:
Please note that the CIL compiler underlying Obliv-C doesn't support intrinsics: you may have to comment/define out all the intrinsics definitions from system's headers (e.g., avx512fintrin.h, avxintrin.h, emmintrin.h, f16cintrin.h, mm3dnow.h, xmmintrin.h)