Path: utzoo!attcan!uunet!husc6!mailrus!ncar!tank!mimsy!chris
From: chris@mimsy.UUCP (Chris Torek)
Newsgroups: comp.lang.c
Subject: Re: Something new for C?
Keywords: offset of vars within structures
Message-ID: <14487@mimsy.UUCP>
Date: 11 Nov 88 02:33:15 GMT
References: <73@dsoft.UUCP> <2865@ingr.UUCP> <399@babbage.acc.virginia.edu>
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
Lines: 31

>>#define  offset(type,field) ((unsigned int)&((type *)0)->field)

In article <399@babbage.acc.virginia.edu> mac3n@babbage.acc.virginia.edu
(Alex Colvin) writes:
>This works by casting a pointer to an unsigned.  That sometimes isn't a
>good idea.  To make integers out of pointers, it's best to subtract.
>
>#define offset(type,field) ((void *)&((type *)0)->field - (void *)((type *)0))

The advice above is reasonable; the implementation is not.  You may not
perform arithmetic on pointers to void.  This is considered a feature.

The dpANS provides a (compiler-dependent!) macro called `offsetof' that
might be defined as

	#define	offsetof(t, f) ((char *)&((t *)0)->f - (char *)(t *)0)

or more simply as

	#define	offsetof(t, f) ((int)&((t *)0)->f)

or as some horrible internal compiler reference:

	#define	offsetof(t, f) __C_compiler_internal_label__offset_of(t, f)

Various restrictions in the dpANS make all of these definitions
suspect-at-best, but one (whatever it might be) is to be provided
and called `offsetof'.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163)
Domain:	chris@mimsy.umd.edu	Path:	uunet!mimsy!chris