Re: [PATCH] Fix PR middle-end/20030, fold_indirect_ref and non zero lower bounds

[Andrew Pinski <pinskia at physics dot uc dot edu>]

On Feb 17, 2005, at 1:26 PM, Richard Henderson wrote:

> On Thu, Feb 17, 2005 at 12:44:37PM -0500, Andrew Pinski wrote:
> > -	return build4 (ARRAY_REF, type, op, size_zero_node, NULL_TREE, 
> > NULL_TREE);
> > +	{
> > +	  tree min_val = TYPE_MIN_VALUE (TYPE_DOMAIN (optype));
> > +	  return build4 (ARRAY_REF, type, op, min_val, NULL_TREE, 
> > NULL_TREE);
> > +	}
>
> TYPE_DOMAIN isn't always set.  Neither is TYPE_MIN_VALUE.
> You need to check for nulls and use zero.

Good point.  I should have checked the documentation.

Bootstrapping and testing in process, ok if it passes?

Thanks,
Andrew Pinski

	* fold-const.c (fold_indirect_ref_1): Use the correct index for zero 
access,
	the lower bound of the array type if it exists.

Attachment: temp.diff.txt
Description: Text document

Testcase (modified from the pervious one to be able to commit the 
testsuite,
I was reminded that abort exists in gfortran):
      program foo
      character*1 a1, a2, b
      a1='A'
      a2='A'
      b='B'
      x = LSAME(a1,a2)
      if ( x.ne.1 ) then
        call abort  ();
      endif
      end

      logical function LSAME( CA, CB )
      character CA, CB
      integer   INTA, INTB
      INTA = ICHAR( CA )
      INTB = ICHAR( CB )
      LSAME = INTA.EQ.INTB
      end