New regression in gcc-3.1 (?)

[Paolo Carlini]

Michael Veksler wrote:

> The following code no longer compiles. It used to compile few days ago 
> from the branch of gcc-3.1 ....
> I don't have the ISO C++ in front of me, so gcc may be right.
>
> What do you think?

It is a regression and a severe one. It is already filed in GNATS as 
c++/6486 (high priority) and we all hope it will fixed before the release!

Ciao,
Paolo.