BUG IN GFORTRAN

[Richard E Maine]

On Dec 1, 2005, at 3:12 AM, Erik Edelmann wrote:
   [code elided]
>> Please see line marked (A). Here "fa" should be reinitialized to 
>> default
>> value 10. But this is not happening in gfortran. I think this is a 
>> Bug.
>
> I also think it's a bug.

Me too, but... I had to look multiple times in order to get past 
confusing statements earlier. All discussion of how arguments "should 
be undefined" is misleading. At first it made me think that the OP had 
fallen into the common error of thinking that there was ever a 
situation in which a compiler is required to make something undefined. 
(There isn't; if the standard says that something is undefined, that 
just places restrictions against the user doing things with it; no 
requirements on a compiler are implied).

The compiler is required to apply default initialization here. That 
part I agree with. It just has nothing to do with anything being 
undefined; that part of the preliminary discussion threw me off at 
first.

-- 
Richard Maine                |  Good judgment comes from experience;
Richard.Maine@nasa.gov       |  experience comes from bad judgment.
                             |        -- Mark Twain