Josh DuBois <[EMAIL PROTECTED]> writes:
> While it does seem better in principle to me to have to use 'signed
> char' in situations where you care about sign, I don't like the fact
> that, unless we make all chars signed by default on all platforms, I
> don't think we can preserve char* & LPSTR as the same types. It seems
> to me prima facie reasonable that they should be.
That's a good point yes. CHAR should probably remain as 'char' then,
and we should change the few functions that really need signed chars
to take a 'signed char' instead of a CHAR argument. This means they
won't have the exact same prototype as under Windows, but on i386 it
doesn't matter and on other platforms it's the right thing to do IMO.
--
Alexandre Julliard
[EMAIL PROTECTED]
