🚀 go-pugleaf

In article 46ebs0$7l2@lyra.csx.cam.ac.uk, sgt20@cam.ac.uk (Simon Tatham)
writes:
>David Mccaughan <davidb@sawnlk.cs.ualberta.ca> wrote:
>>Lawrence Crowl (crowl@jade.CS.ORST.EDU) wrote:
[snip]
>>I think you'll find that he was making reference to the fact that the C
>>standard does not specify the "size" of data types (this being largely a
>>machine dependent issue) - as such it is up to the compiler writer to
>>specify this.
>
>Not quite - Kernighan & Ritchie, second edition, specifies that "int"
>and "short" should be at least 16 bits, and "long" should be at least
>32. It *is* up to the compiler writer, of course, but those are
>minimum limits.

Not quite. :-) The Standard only specifies the minimum size of a
byte (8 bits). All other minimum sizes of arithmetic types are
_implied_ by 1. the range of values that they are required to
represent and 2. the fact that they must be multiples of a byte.
For example, the Standard does not state that an int must be
at least 16 bits long, but it does state that an int must be able
to represent the values from -32767 to +32767. Surprise! A
16-bit int will do just fine in implementations where a byte is
either 8 or 16 bits long.

I think this is the point that the person to whom you were
replying was making--the Standard does not, strictly speaking,
specify the size of data types.

Paul_Long@ortel.org
plong@perf.com