Dataflow branch review #5

[Kenneth Zadeck]

Ian Lance Taylor wrote:
>
>
> In rtlanal.c:
>
> @@ -2901,11 +2907,7 @@ commutative_operand_precedence (rtx op)
>        if (code == SUBREG && OBJECT_P (SUBREG_REG (op)))
>          return -2;
>  
> -      if (!CONSTANT_P (op))
> -        return 0;
> -      else
> -	/* As for RTX_CONST_OBJ.  */
> -	return -3;
> +      return 0;
>  
>      case RTX_OBJ:
>        /* Complex expressions should be the first, so decrease priority
>
>
> Why this change?  I don't see why, and I couldn't find a ChangeLog
> entry for it.
>
>   
Paolo,

svn blame points to you.  why did you do this?

 78824    bonzini     case RTX_EXTRA:
 78824    bonzini       /* SUBREGs of objects should come second.  */
 78824    bonzini       if (code == SUBREG && OBJECT_P (SUBREG_REG (op)))
 78824    bonzini         return -2;
 77945     geoffk
108419    bonzini       return 0;
 77945     geoffk