Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!uwvax!tank!mimsy!chris
From: chris@mimsy.UUCP (Chris Torek)
Newsgroups: comp.lang.c
Subject: Re: mutual reference in structures
Message-ID: <17441@mimsy.UUCP>
Date: 11 May 89 11:20:54 GMT
References: <6712@medusa.cs.purdue.edu> <1822@ubu.warwick.UUCP> <2346@wheat-chex.ai.mit.edu>
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
Lines: 51

In article <2346@wheat-chex.ai.mit.edu> jym@wheaties.ai.mit.edu
(Jym Dyer) writes:
>VAX C accepts this (which seems wrong to me):
>
>	typedef struct
>	{
>	  struct NODE_T *  flink_p;
>	  struct NODE_T *  blink_p;
>	  char  data[512];
>	}  NODE_T;

This is legal; it is merely stupid.  It is the moral equivalent of

	typedef struct {
		struct never_defined *forw;
		struct never_defined *back;
		char data[512];
	} NODE_T;

Here, you cannot use the pointers `forw' and `back' without first
declaring a structure tag `struct never_defined'.  In the quoted
article, you cannot use the pointers `flink_p' and `blink_p' because
there is no `struct NODE_T' (there is only a NODE_T; the typedef
is in a different namespace from structure tags).

>It would be nice if C typedefs knew about themselves ... :
>
>	typedef
>	{
>	  NODE_T *  flink_p;
>	  NODE_T *  blink_p;
>	  char  data[512];
>	}  NODE_T;

This would require the ability to `back up', or at least to defer
`thinking about' the declaration.  There is nothing in C that now
requires this, so it would be in effect a radical change.

What is so wrong with


	typedef struct node_t {
		struct node_t *flink_p;
		struct node_t *blink_p;
		char data[512];
	} NODE_T;

anyway?
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163)
Domain:	chris@mimsy.umd.edu	Path:	uunet!mimsy!chris