Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!snorkelwacker!apple!sun-barr!newstop!sun!gzilla!jjg
From: jjg@gzilla.Sun.COM (Jonathan Gibbons)
Newsgroups: comp.lang.c++
Subject: typedef'ed class declarations
Keywords: typedef class
Message-ID: <130625@sun.Eng.Sun.COM>
Date: 24 Jan 90 01:39:35 GMT
Sender: news@sun.Eng.Sun.COM
Lines: 33

One for the language lawyers ...

I have a macro to emulate parameterized types:  param_type(C,T).
It expands into a class declaration such that I can use it as follows

  typedef param_type(List,String) Strings;

which might not be so different from
  typedef List[String] Strings; 
if ever we get parameterized types in C++.

The macro for the above example expands into something like
  typedef class tmp { ... } Strings;


That is fine and works, but it is convenient at one point in one
header to be able to go 
	class Strings;
before the above macro expansion.


cfront is happy to have the two definitions co-exist, g++ isn't.
It seems the typedef in g++ doesn't do what I expected, which was
to declare Strings as the name of the class, as though it were
declared
	typedef Strings { ... }


Am I being optimistic in expecting this to work at all, or is
g++ to be considered at fault?

-- Jon Gibbons
   jjg@sun.com