I think adopting ssize_t, thus improving parity with C would be more
reasonable than a size_s that can be easily mistaken for size_t when
reviewing long code.
Either way, sizediff_t isn't used that often anyway, so I don't know if
this minor issue should take too much priority.
On 10-05-2011 07:28, KennyTM~ wrote:
On May 10, 11 11:57, Andrej Mitrovic wrote:
I've noticed this thread a while ago:
http://thread.gmane.org/gmane.comp.lang.d.phobos/3881/focus=3885
Maybe it's way too late for any changes, but I think sizediff_t is a
mouthful. Doesn't size_s look nicer? It would be size_t's evil twin
brother, look:
size_s imSoNegative;
size_t neverADowner;
-sizediff_t.max
The meaning of _t and _s suffix on types in C mean "type" and "struct"
respectively, the latter never mean "signed". Using a C convention for a
totally different meaning is going to confuse people. Further more, the
difference between 'size_t' and 'size_s' is too small visually.
BTW, the signed version of size_t in C (POSIX) is called 'ssize_t'.
</bikeshed>
