Ruby

rb_uint2big and rb_int2big are declared as such:

VALUE rb_uint2big(VALUE);
VALUE rb_int2big(SIGNED_VALUE);

and VALUE is one of

typedef uintptr_t VALUE;
typedef unsigned long VALUE;
typedef unsigned LONG_LONG VALUE;

and SIGNED_VALUE similar but without unsigned.

Should the signatures actually be:

VALUE rb_uint2big(unsigned long);
VALUE rb_int2big(long);

Of course, there is not much difference here.
But it seems a bit strange to make this kind of conversion functions taking C integers depend on the bitwidth of VALUE.