RE: Operation Precedence questions

[Dan <dbcerpg@xxxxxxxxx>]

By that logic, shouldn't and AND & OR operators occur higher in the order of 
precedence than the 
   =, <>, >, >=, <, & <= operators?

- Dan

--- Bob Cozzi <cozzi@xxxxxxxxx> wrote:
> You also have to consider that the evaluation STOPS when the condition
> cannot be met. So if I=0 the ARRAY(I)>10 comparison is never performed.
> Hence, it works.
> 
> Bob Cozzi
> 
> -----Original Message-----
> From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx]
> On Behalf Of Dan
> Sent: Thursday, March 20, 2003 11:36 AM
> To: rpg400-l@xxxxxxxxxxxx
> Subject: Operation Precedence questions
> 
> 
> I have a question about Operation Precedence in RPG-IV.  FWIW, v5r2.  I have
> found the chapter on Expressions in the Reference and, specifically, the
> sections titled "Operation Precedence" and "Short Circuit Evaluation".  Upon
> my brief foray into these topics, it appears to me that the two sections
> permit a contradiction to the rules they set forth.
> 
> "Operation Precedence" says:
> 
> The precedence of operations determines the order in which operations are
> performed within expressions. High precedence operations are performed
> before lower precedence operations.
> 
> The following list indicates the precedence of operations from highest to
> lowest:
>   1. ()
>   2. Built-in functions, user-defined functions
>   3. unary +, unary -, NOT
>   4. **
>   5. *, /
>   6. binary +, binary -
>   7. =, <>, >, >=, <, <=
>   8. AND
>   9. OR
> 
> "Short Circuit Evaluation" says:
> 
> Relational operations AND and OR are evaluated from left to right. However,
> as soon as the value is known, evaluation of the expression stops and the
> value is returned. As a result, not all operands of the expression need to
> be evaluated.
> 
> For operation AND, if the first operand is false, then the second operand is
> not evaluated. Likewise, for operation OR, if the first operand is true, the
> second operand is not evaluated.
> 
> There are two implications of this behaviour. First, an array index can be
> both tested and used within the same expression. The expression
>       I<=%ELEM(ARRAY) AND I>0 AND ARRAY(I)>10
> will never result in an array indexing exception.
> 
> <my comments/questions follow>
> 
> If AND & OR are always the lowest precedence, how can the array index
> example above work? 
> According to the rules of precedence, as I interpret them, all of the
> operands:
>       I<=%ELEM(ARRAY) 
>       I>0 
>       ARRAY(I)>10
> would be evaluated before the AND relational operators.  If 'I' is zero, I
> would expect a run-time error on the ARRAY(I)>10 because the > operator has
> a higher precedence than the AND operator.
> 
> BTW, I know from experience that this example works.  I am questioning
> whether the precedence of operations really should be re-arranged to place
> AND & OR higher in the list.
> 
> - Dan

__________________________________________________
Do you Yahoo!?
Yahoo! Platinum - Watch CBS' NCAA March Madness, live on your desktop!
http://platinum.yahoo.com