> I don't know what a larger-then-16-bit address space has to do with
> the size of an int.

Address space and int size tend to correlate, that's all.

> [...replies to my points...]

> In my book, you haven't made a convincing argument to deliberately
> create unportable code.

Portability is a matter of degree, not an absolute thing.

Every time you write a prototype declaration or definition, you are
(presumably deliberately) creating unportable code - it's not portable
to older compilers.  Every time you write a call to ioctl(), you are
creating unportable code; lots of C implementations don't have it.

Personally, I don't hesitate to write prototypes, I don't hesitate to
call ioctl()...and I don't hesitate to assume int is reasonably large
(usually meaning at least 30 or so bits).  Each of these entails some
degree of unportability; so far I have found the tradeoffs acceptable.

In my opinion, changing that short to an int is acceptable.  In that
particular case changing it to long would also be acceptable to me; in
many other, similar, cases, I would find a long unacceptable because of
the wasted space and time it would/could imply.

Certainly changing it from short to int will not make the situation
worse anywhere.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B