Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!ucsd!orion.oac.uci.edu!uci-ics!rfg
From: rfg@ics.uci.edu (Ronald Guilmette)
Newsgroups: comp.lang.c++
Subject: Re: Typedef inside of a class definition.
Message-ID: <26462BED.5865@paris.ics.uci.edu>
Date: 8 May 90 02:16:13 GMT
References: <3748@elysium.sw.mcc.com>
Reply-To: rfg@ics.uci.edu (Ronald Guilmette)
Organization: UC Irvine Department of ICS
Lines: 40

In article <3748@elysium.sw.mcc.com> downing@elysium.sw.mcc.com (Glenn Downing) writes:
>
>Is it a bug in the GNU 1.37 compiler that it does not allow me to define a
>typedef within a class definition and have its scope be global?  The Sun
>compiler does define it with global scope.
>
>class foo
>  {typedef bar int;};
>
>main()
>  {bar i = 2;}  // bar is not accessible here with GNU, but is with SUN
>--
>Glenn Downing (downing@mcc.com)
>Microelectronics Computer Corporation (MCC)
>3500 West Balcones Center Drive
>Austin, TX 78759-6509

I really should let Michael answer this one but...

The C++ language rules have changed recently with respect to the scope
of things declared within classes/structs/unions.

It used to be that if you defined a type (via a typedef statement, or via
a struct, class, union, or enum type definition) within a class, that the
scope of the new typename was made "global".

Nowadays, the scopes of such names are (or will soon be) restricted to
the class itself (and its member functions of course).

Cfront 2.0 doesn't implement this fully yet.  I haven't seen what 2.1
does in such cases, so I don't know.  g++ (1.37.1) went part of the
way by making names defined via typedef statements local to the containing
class.

Whether that is a bug or a feature is (i think) in the eye of the beholder.


// Ron Guilmette (rfg@ics.uci.edu)
// C++ Entomologist
// Motto:  If it sticks, force it.  If it breaks, it needed replacing anyway.