Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!rutgers!iuvax!pur-ee!j.cc.purdue.edu!k.cc.purdue.edu!l.cc.purdue.edu!cik
From: cik@l.cc.purdue.edu (Herman Rubin)
Newsgroups: comp.lang.c
Subject: Re: Question on structures
Message-ID: <608@l.cc.purdue.edu>
Date: Sat, 7-Nov-87 07:03:42 EST
Article-I.D.: l.608
Posted: Sat Nov  7 07:03:42 1987
Date-Received: Wed, 11-Nov-87 06:33:25 EST
References: <1025@phoenix.Princeton.EDU> <8888@utzoo.UUCP>
Organization: Purdue University Statistics Department
Lines: 55
Keywords: Sturctures records
Summary: I find the "." a nuisance too

In article <8888@utzoo.UUCP>, henry@utzoo.UUCP (Henry Spencer) writes:
> > I have gotten really tired of typing things like a.b.c.d etc  to refer to
> > elements of a structure. ....
......
> Nothing quite the same, but you can come close.  If all you want is the
> abbreviation, just
> 
> 	#define	IT	a.b.c
> 
> and then refer to IT.d or whatever.  If you also want the efficiency boost
> that some Pascal compilers give you this way, then
> 
> 	struct whatever *it = &a.b.c;
> 
> and then refer to it->d or whatever.  ....
 
> Those who do not understand Unix     |  Henry Spencer @ U of Toronto Zoology
> are condemned to reinvent it.        | {allegra,ihnp4,decvax,utai}!utzoo!henry

This is another example of the use of unnecessarily complicated syntax.  
The use of unions in C is far more messy than that of equivalence in FORTRAN.
What is needed is a massive extension of the common block idea in FORTRAN
without forcing common together with extension of equivalence.  In addition,
if the block is external, only the first part need be given, with an indication
that this is the case.  As an example, I use a subroutine linkage call in which
I would like my declaration to read

#extern array  block{*routine r, *TYPE abc[3]};

and my call to read

	r(array);

and referring the the I-th pointer in abc as abc[I],

instead of what I am now doing, which is using

union point array[];

and then making the call

	array[0].r(array);

and having to use array[I+1].T instead of abc[I].

Since "a pointer is a pointer is a pointer", someone like myself who does not
believe in requiring strong typing would like to have a type "pointer" which
can be any type of pointer.  This would help since I could use pointer array.

We should let the compiler read through the chain of structures instead of
requiring the programmer to use ->'s and .'s.
-- 
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907
Phone: (317)494-6054
hrubin@l.cc.purdue.edu (ARPA or UUCP) or hrubin@purccvm.bitnet