[gfortran] Fix NULL reference types.

[Mark Mitchell]

Richard Henderson wrote:

>On Wed, Aug 18, 2004 at 02:10:29PM +0100, Paul Brook wrote:
>  
>
>>As part of his int_cst sharing work Nathan spotted that gfortran was creating 
>>null pointers with REFERENCE_TYPE. Apparently the optimizers [could] assume 
>>that a reference type argument is never NULL.
>>    
>>
>
>So?  We can assume anything, since it would be incorrect to actually
>touch such an argument, would it not?
>  
>
I'm not sure what you're asking.

In C++, it's never valid to have a reference to NULL, and so it is safe 
for an optimizer to assume that any expression with REFERENCE_TYPE is 
non-NULL.

Kenner and I discussed this back in 1998:

http://gcc.gnu.org/ml/gcc-bugs/1998-07/msg00040.html

RTH, do you object to Nathan's documentation patch?  It would give us an 
easy way to give the optimizers a little more information.

-- 
Mark Mitchell
CodeSourcery, LLC
(916) 791-8304
mark@codesourcery.com