Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utcs!mnetor!seismo!umcp-cs!chris
From: chris@umcp-cs.UUCP
Newsgroups: net.lang,net.lang.c
Subject: Re: WITH statement
Message-ID: <2366@umcp-cs.UUCP>
Date: Sat, 12-Jul-86 02:20:19 EDT
Article-I.D.: umcp-cs.2366
Posted: Sat Jul 12 02:20:19 1986
Date-Received: Sat, 12-Jul-86 07:21:43 EDT
References: <1003@zog.cs.cmu.edu>
Reply-To: chris@maryland.UUCP (Chris Torek)
Followup-To: net.lang
Organization: University of Maryland, Dept. of Computer Sci.
Lines: 58
Xref: utcs net.lang:2514 net.lang.c:9677

In article <1003@zog.cs.cmu.edu> dlc@zog.cs.cmu.edu (Daryl Clevenger) writes:
>   I dislike Pascal as much as the next C hacker, but one feature
>that could be useful to add to C is something akin to the Pascal
>"WITH" statement. ... By using such a construct, one avoids long
>statements when using deeply nested structures/unions or the kludgy
>method of using defines.

I never did like `with'.  I would like it a lot more if . . . well,
let me see if I can construct a good bad example first.

	type
		x = record ... end;
		y = record ... end;
		z = record ... end;

	var
		foo : x;
		zarog : y;
		prullo : z;

	{ several hundred pages of code }

		with foo, zarog, prullo do begin
			...
			{ next three statments use fields }
			snert := 5 * bazzi;
			bibble := 10;
			roj := klewp mod snert;
			...
		end;
		...

To which variables do those fields refer?

If `with' is intended as shorthand, I would like to see it used
in more this way:

		with a = foo, b = zarog, c = prullo do begin
			...
			b.snert = 5 * a.bazzi;
			c.bibble := 10;
			a.roj := c.klewp mod b.snert;

Note that this allows member name `collisions' to be disambiguated:

			b.snert = 5 * a.bazzi;
			c.bibble := 10;
			a.roj = c.klewp mod a.snert;

---if that is in fact what is meant.

(I have a feeling that `WITH' originally meant `Listen, compiler,
put a pointer to this here record into a machine register now,
because I am about to use it a whole bunch.')
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 1516)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris@umcp-cs		ARPA:	chris@mimsy.umd.edu