Re: storing an integer in a double precision



gmail-unlp <ftinetti@xxxxxxxxx> wrote:

I see (I think), the problem is that the Do loop control variable
after the loop would have a "wrong" value, which could cause that the
program never starts, for example... how is this explained in the
standard?

I'm on a trip and don't have a copy here (or feel like downloading a pdf
one over the hotel net connection). So all I can do is describe it
generally. There are 3 pieces.

1. In the descrption of DO loops, it describes what the final value of
the variable shall be. In the case in question, that value is out of teh
range of values that than can fit in the variable.

2. Somewhere, it says that a program is nonconforming if it involves an
operation whose results are out of range. I'd have to paw around a litle
to find that one and its exact wording, but it is a pretty basic
requirement of the standard. Fortran isn't Ada.

3. Way up front, in the very beginning of the standard, it says that the
standard does not specify what happens to programs that do not conform
to the standard.

Put those 3 together, and you have the result mentioned.

--
Richard Maine
email: last name at domain . net
domain: summer-triangle
.