Re: Fix libgomp Fortran tests

[Brooks Moses <bmoses at stanford dot edu>]

Jerry DeLisle wrote:
> To avoid the runtime penalty in computations of checking for that absurd case, I 
> think its better to just not allow it and stick with the standard which defines 
> symmetrical integers.
>
> One reason for standards and standards committees is to hash out all these things.

Note, however, that the Fortran standard only defines a symmetric 
"integer model".  The language in the standard is very careful to avoid 
the idea that the integer model exactly corresponds to the integers 
supported by the compiler/processor.

In particular, it was -- intentionally, unless I'm very much mistaken -- 
crafted in a way that allows all of the necessary loopholes to permit a 
compiler to support a slightly asymmetrical range like the usual 
two's-complement integers.

Personally, I would consider it a poor quality of implementation that 
wouldn't permit integers to take on all possible values permitted by the 
hardware, given the frequent usage of integers for storing bit-patterns.

- Brooks