19980824 bug in 3 user-defined operators

[Reid M. Pinchback]

At 10:46 PM 9/2/98 -0300, you wrote:
>Reid M Pinchback <reidmp@MIT.EDU> writes:
>
>> **   egcs is doing promotion for three user-defined unary
>> **   operators (~,-,+)
>
>Nope, it is doing integral promotion of integral types, as required in
>[expr.unary.op]/6,7,9.  The expression `+((short)1)', for example, has
>type `int', not `short'.

Ok, let's see if I understand what is going on.  In the following code:

 [stmt1]  short x=1;
 [stmt2]  short y=+x;

we would have the following:

 * in stmt1 the literal '1' is of type int, and the assignment to 'x'
   would convert the literal to a short.
 * in stmt2 'x' is a short, but '+x' is promoted to an int, and then
   the assignment to 'y' causes the int value of '+x' to be converted
   back to a short

correct?  In other words, the integral promotion process is entirely
determined by the expression; what type the expression is being assigned
to isn't factored into the process.  That means that in general when
using unary operators it is worth considering static_cast'ing the results
to ensure that the expressions have the type you think they should.

Sum days yah jus gotta luv dis langwage.  :-)
   
 
====================================================
= Reid M. Pinchback                                =
= I/T Delivery, MIT                                =
=                                                  =
= Email:   reidmp@mit.edu                          =
= URL:     http://web.mit.edu/reidmp/www/home.html =
====================================================