Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!caen!spool.mu.edu!munnari.oz.au!goanna!ok
From: ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe)
Newsgroups: comp.lang.c
Subject: Re: Help needed with: sizeof (struct a)
Keywords: sizeof,struct,help,cc
Message-ID: <6398@goanna.cs.rmit.oz.au>
Date: 20 Jun 91 07:11:47 GMT
Article-I.D.: goanna.6398
References: <10569@aspect.UUCP>
Organization: Comp Sci, RMIT, Melbourne, Australia
Lines: 29

In article <10569@aspect.UUCP>, dave@aspect.UUCP (Dave Corcoran) writes:
> My Sun cc prints 4 for both structs, prints 3 for the array
> struct a {char x;char y;char z};
> struct b {char x[3]};
> char c[3];

Syntax error:  that should be
	struct a {char x; char y; char z;};
					^
and the same for struct b.  Omitting the last ; before the } of a struct
was a Berkeley extension.  Many compilers do not support it, and it is
not standard.

> Is there any way to cause cc to force sizeof return the actual size of the
> structs without rounding up to the next highest sizeof (short)?

Four bytes *IS* the actual size of the structs in that implementation of C.
Remember, "sizeof (T)" means "if you had an array of N elements each of
type (T), the number of bytes in the array would be N times <what>?".
If you have an array of (struct a)s, they will be 4 bytes apart, not 3.

There isn't anything you can portably do that would profit from trying
to outsmart the compiler in this way, and very little that is _safe_
even for a program that is never going to see another compiler or machine.

(You could try taking offsetof the last field + sizeof the last field.
But of course the last field itself may contain padding.)
-- 
I agree with Jim Giles about many of the deficiencies of present UNIX.