Overloading vs. string functions, etc.

[Benjamin Kosnik]

>     Stephen> It looks like they, or working shadow headers, might be
>     Stephen> desireable for the 3.0 release if the c_std stuff just
>     Stephen> won't work. 
> 
> It'll work good enough.  I don't think we want a wholesale revision
> right now; I was just tickling around the edges.  (I removed a
> misfeature in the compiler that showed up some problems that had been
> hidden.)

Phew. I was contemplating trying to help Stephen with this. I don't 
imagine it would be easy, and would rather work on some of the other stuff.
Plus I don't think it could be really solidified before the release. 
(Although trying might be fun, in some peverse way.)

Obviously, it would be nice to have the "C" header solution in better 
shape. Conformant code, and judicious use of explicit qualifications, does 
work. However, it would be nice to have a real solution that got rid of the ambiguities 
of the current approach without wholesale defining away of "C" functions. 

Anyway. 

-benjamin