TechTip: RPG IV Pointers--They're Easy!

RPG
Typography
  • Smaller Small Medium Big Bigger
  • Default Helvetica Segoe Georgia Times

I've heard comments about some RPG IV programmers being "pointer challenged." This bothers me, because this data type has been available since RPG IV's arrival in late 1994, almost 11 years. The pointer concept has been around since the invention of the computer! This tip explains the basics of RPG IV pointers and includes some examples.

The concept of a pointer (in any language) is that some variable contains an address of something else, rather than some kind of data item or procedure. If I refer to a standalone variable by its name, such as "Field5," and Field5 has been defined as a character variable with length 5, then every time I use Field5 I am referring to the storage (memory) address assigned by the compiler and loader. You could say that Field5 "points" to the five characters set up in storage. The address associated with Field5 is fixed. In contrast, a pointer variable contains an address that can change.

In RPG IV, a pointer variable is defined using the internal data type code of *. A quick check of the compile listing shows that the compiler sets aside 16 bytes for each pointer. Even though the i5 (or iSeries) uses a 64-bit address (8 bytes), 16 bytes are still used. Maybe IBM is planning ahead for 128-bit addressing--who knows? A pointer variable can contain either a procedure address or the address of a data item. If the pointer contains a procedure address, the keyword PROCPTR is used on the pointer definition. A pointer that is used to "point to" a data item is called a "basing pointer." Variables, data structures, and arrays may be used with basing pointers, and if so, the BASED keyword is specified and a pointer name is the value of the keyword.

Here's an important concept: When a data item is based, there is no storage allocated to it. It is a definition with a "yet to be determined" home in storage. I like to think of a based data item as a clear plastic ruler (or template) that's floating above the storage plane. It cannot be used until the pointer associated with it has a storage address. The value in the basing pointer determines where the template will be laid in the storage plane. Here is a simple example:

 

D Phrase S 25 Inz('Peter Piper picked a Plum')
D Field5 S 5 Based(Ptr)
D Ptr S *

/free

// You can't use Field5 yet.
// Now set the pointer.

Ptr = %addr(Phrase);

// Ptr has a value, the memory address of
// Phrase (left-most byte). Field5 can now
// be used.

// The current value of Field5 is 'Peter'
// Now move the template (address in Ptr) to the right 6 bytes

Ptr += 6;

// The value of Field5 is now 'Piper'
// Change Field5.

Field5 = 'Smith';

// The value of Phrase is now 'Peter Smith picked a Plum'

/end-free

 

Based data structures and arrays can use dynamic storage (storage outside the program, but within the activation group). With dynamic storage, the amount of storage allocated to the program can vary as the needs of the program change. In the past, if we loaded an array with a variable number of elements, we would allocate the largest number of elements (Dim value) we thought might happen, even though the maximum number was seldom needed. Much main memory was wasted, and system performance could be adversely impacted. With a basing pointer and dynamic allocation of storage, we can allocate storage for an array as the needs arise. Here is an example of using an array that is based on a pointer and using storage from a dynamic "heap."

D Array            s           10     Dim(20000) Based(Ptr)
D Size             s            6  0
     
 /free
  // Allocate storage for the first 1000 elements
  Size = 1000 * %size(Array); // Size becomes 10,000
  Ptr  = %alloc(Size); // A pointer to the allocated storage is
                       // the return value.
  // The first 1000 elements of the array can now be loaded and
  // used. The storage allocated is not pre-initialized.

  // Additional storage can be allocated using the %realloc 
  // built-in function, using the same pointer, Ptr
 /end-free

Basing pointers are often used as parameters by APIs.

The other use of pointers is to contain the address of a procedure. The PROCPTR keyword is used on the pointer definition, and the built-in function %paddr (retrieve procedure address) can be used to load the pointer. In the following code segment, an ILE condition-handling procedure (ILEPROC) is specified on a pointer definition for use by the API prototype.

D CEEHDLR         PR
D  ptrCH                    *       Procptr
D  ptrStatus                *       Const
D  Feedback             12          Options(*Omit)

D ptrCH           s         *       Procptr(Inz(%paddr('ILEPROC'))

D Status         sds
D  Progname            *Proc

 /free
  ---
  ---
  // Register condition handler
  CallP     CEEHDLR(ptrCH:%addr(Status):*Omit);
  ---
  ---
 /end-free

Pointers provide great flexibility in programming, allowing us to write generic routines that pass pointers to data and procedures.

Jim Martin, the author of Free-Format RPG IV, is a corporate technical instructor at Jack Henry & Associates in Monett, Missouri. He is a veteran of RPG programming, beginning in 1967 with a position at IBM as a systems engineer and later becoming a staff programmer at the Rochester systems programming lab. For eight years, he was at Lakeview Technology as an AS/400 and RPG instructor and was a speaker at various local midrange user group meetings and conferences. He can be reached by email at This email address is being protected from spambots. You need JavaScript enabled to view it..

BLOG COMMENTS POWERED BY DISQUS

LATEST COMMENTS

Support MC Press Online

$0.00 Raised:
$