>David McPaul wrote: > > renamed their float.h float.c to nasmfloat.h nasmfloat.c and fix the > makefiles etc to cope. This was done because their version of float.h > was being included instead of gcc's version. Ah, that's the part I was getting stuck on. Still wondering where the type "efunc" comes from that is used in their float.h > Added the following defines to compiler.h These are c99 constructs > that we don't have in our gcc headers. > > #define PRIu32 "u" > #define PRIX32 "X" > #define PRId32 "d" > #define PRIX64 "X" > #define PRId64 "d" > #define PRIu64 "u" > #define PRIx64 "x" > #define PRIx32 "x" > #define PRIx16 "x" > #define PRIx8 "x" > My understanding is that the 64bit values need to have the length specified to output proper values. For 32-bit this would be: #define PRIX64 "llX" #define PRId64 "lld" #define PRIu64 "llu" #define PRIx64 "llx" > #define INT32_MAX 2147483647 > #define UINT64_C(val) val##ULL What it should be for 32-bit. Code style wise I prefer using value instead of val, but that's nitpicking on my part. ;) > If anyone wants a copy to try let me know. If you don't upload it somewhere (i.e. haikuware.com or osdrawer.net), I would like a copy. Curtis