spurious declaration conflict?

[Neal Becker]

Here is a very simple example.  I would think that the complaint given 
by egcs-1.0.3 is spurious.  I have not checked the standard, but I
would think that logically this should be accepted.  

The 3rd statement is rejected by g++.  But Complex *is* a class.  It
just happens to be typedefd.  This stems from the principle (which I
just made up) that the user of a typedefd class shouldn't have to know 
that it is a typedef, and not really a class.

--------------
template<class T> class complex {};

typedef complex<double> Complex;

class Complex;
---------------