Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site watmath.UUCP
Path: utzoo!watmath!kpmartin
From: kpmartin@watmath.UUCP (Kevin Martin)
Newsgroups: net.lang.c
Subject: A comment on aggregate constants
Message-ID: <9649@watmath.UUCP>
Date: Wed, 31-Oct-84 13:51:18 EST
Article-I.D.: watmath.9649
Posted: Wed Oct 31 13:51:18 1984
Date-Received: Thu, 1-Nov-84 03:29:26 EST
Organization: U of Waterloo, Ontario
Lines: 35

>From mark (mwm@ea.UUCP)
>	struct gort {
>		int	type ;
>		union	info {
>				struct	gfloat	{ float	gfx, gfy } ;
>				struct	gint	{ long	gix, giy } ;
>				string	*name ;
>				} ;
>		} sample[] = {
>			{T_FLOAT, (struct glfoat) { 3.4, 5.7 } } ,
>			{T_INT, (struct gint) { 243, 56 } } ,
>			{T_STRING, (char *) "this is a test" }
>			} ;
>Comments?

Although I think that aggregate constants are a nice thing in themselves,
I still don't like the name-the-type approach to initualizing unions. I
prefer:
	struct gort {
		int	type ;
		union	info {
				struct	gfloat	{ float	gfx, gfy } gfl;
				struct	gint	{ long	gix, giy } gin;
				string	*name ;
				} ;
		} sample[] = {
			{T_FLOAT, gfl = { 3.4, 5.7 } } ,
			{T_INT, gin = { 243, 56 } } ,
			{T_STRING, name = "this is a test" }
			} ;
I believe that most C compilers now require you to give names to s/u elements
even if they are themselves s/u elements, so the first two union elements
are improperly declared in the first example.
               Kevin Martin, UofW Software Development Group