> 1) Aligned data > If I am right there is no way to portably cause the alignment of > variables in a function declaration.
Hugh? In a function declaration? I don't get you here. > 2) Pointers as integers > As above, the SWAR code often needs to do math on the integer representation > of a pointer. Unfortunately the size of a pointer may vary, so providing > a "uintp" internal declaration would be handy. Hmm - Pointer Arithmetics won't work? Usually (unsigned char *)ptr+bla will do what you want. It is _not_ guaranteed in any way by C standards, that you can do arithmetics on pointers yourself by casting them to integer types. AFAIK there even existed some perverse architectures, that had very strange representations for pointers. But if you really feel it cannot be avoided, something like an uintp might be a good idea, yes. > 3) Interesting bug Yeah. Seen it. Integer overflow ... nasty stuff. CU, Andy -- = Andreas Beck | Email : <[EMAIL PROTECTED]> =
