Path: utzoo!mnetor!uunet!steinmetz!ge-dab!codas!pdn!alan
From: alan@pdn.UUCP (Alan Lovejoy)
Newsgroups: comp.lang.modula2
Subject: Re: union types
Message-ID: <2641@pdn.UUCP>
Date: 28 Mar 88 15:06:22 GMT
References: <8803221947.AA00248@nrl-iws6.ARPA> <8803242203.AA28027@ndsuvax.UUCP>
Reply-To: alan@pdn.UUCP (0000-Alan Lovejoy)
Organization: Paradyne Corporation, Largo, Florida
Lines: 37

In article <8803242203.AA28027@ndsuvax.UUCP> Info-Modula2 Distribution List <INFO-M2%UCF1VM.bitnet@jade.berkeley.edu> writes:
>Mark, I believe the closest you could come to simulating the C union
>statement would be using a Variant record.  I know this involves an
>extra identifier, but I don't know any way around this.  I am currently
>trying to write a program that translates C procedures to Modula2
>procedures and that is how I chose to translate the C union type.

Why does a variant record require an extra identifier?  The only extra
identifier I know of is OPTIONAL:

   TYPE

     TaggedVariant = 
       RECORD
	 CASE type: CARDINAL OF
	   0: foo: Foo;
	   |
	   1: bar: Bar;
         END;
       END;

     UntaggedVariant =
       RECORD
	 CASE CARDINAL OF
	   0: foo: Foo;
	   |
	   1: bar: Bar;
         END;
       END;

Variant records do not have to have a tag field.  What other extra
identifier could there be?  Did you mean the identifier that specifies
the type of the case labels (CARDINAL in the examples above)?  If so,
that identifier is a compile-time creature only--it has no existence 
at run time.

--alan@pdn