Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!samsung!uunet!taumet!steve
From: steve@taumet.com (Stephen Clamage)
Newsgroups: comp.lang.c++
Subject: Re: Local typedef possible?
Message-ID: <780@taumet.com>
Date: 23 Jun 91 15:02:00 GMT
References: <1991Jun22.182615.16885@murdoch.acc.Virginia.EDU>
Organization: Taumetric Corporation, San Diego
Lines: 32

ecl2v@opal.cs.Virginia.EDU (Edmond C. Loyot) writes:

>The ARM says that what you want to do is legal.  Consider this example from
>page 189 of the ARM:  [ example of typedef in a class ]

>However when I run it through my 2.0 C++ compiler, it does not flag the error.
>I think this is a throwback to the days of 1.2 when type names
>(including classes) declared in a class where in the global scope...

Nested types were introduced with version AT&T C++ 2.1.  Up through 2.0,
a type defined in a class was exported to the outer scope.  In fact,
AT&T 2.1 is a compromise between local and exported interpretation of
nested types.

AT&T 2.1 example:
	struct A { ...
		struct B { ... };	// exported to global
	};
	struct B { ... };		// ERROR -- B already defined
	struct E { ... };
	struct D { ...
		struct E { ... };	// local to D
	};

I assume this compromise is to allow existing code to continue to work,
while allowing new code with nested types to be written.  What I show
in this example violates both the ARM and earlier language definitions!

The ARM is for the post-2.1 brave new world.
-- 

Steve Clamage, TauMetric Corp, steve@taumet.com