CODEWITHSUNDEEP
ibm-midrangerpglerpg
Tarquila
Tarquila

Reputation: 1207

Is there an numeric type in ILE RPG which will overflow without crashing my program?

I'm looking for a numeric type in ILE RPG which will "wrap round" when it overflows, in a similar way to how a C int would. Is there such a thing?

Upvotes: 0

Views: 756

Answers (3)

Steve
Steve

Reputation: 31

You can use the fixed format mathematical operations (add,sub,mult, and div). They will truncate when overflow is reached. Cumbersome but works.

0001.00 D Fld1            s              3  0                                         
0001.01 D                                                                             
0002.00 C     999           add       3             Fld1                              
0003.00  /free                                                                        
0004.00    dsply ('The current value '+%editc(Fld1:'X'));                             
0005.00    *inlr=*on;                                                                 
0006.00    return;

Display Program Messages

Job 912834/SPRICE/DP88LT started on 01/11/11 at 15:39:15 in subsystem QINTER Message queue SPRICE is allocated to another job.
DSPLY The current value 002

Upvotes: 3

GuyB
GuyB

Reputation: 21

Or you could monitor for the error code when an overflow happens:

 D Counter         S             10I 0

  /FREE
   Monitor;
      Counter += 1;
   On-Error 103;
      Clear Counter;
   EndMon;
  /END-FREE

Upvotes: 2

Tracy Probst
Tracy Probst

Reputation: 1859

RPG won't let you do that. The best solution I can suggest would be to create a procedure that does the math for you and handles the overflow. While RPG does have the TRUNCNBR compile option and control spec keyword, it's only applicable in certain scenarios.

If you're doing a simple counter, you can create a data structure with overlapping numeric fields like this:

DCounterDS        DS                        
D Counter                 5      8  0       
D CountOverflow           1      4  0       
D WholeCounter            1      8  0 INZ(0)

Then you add to WholeCounter and then zero-out CountOverflow immediately afterwards. In this example, Counter is a 4-digit number. You can do the same thing with integer fields, but I recommend you keep them unsigned:

DCounterDS        DS                        
D Counter                 5      8U 0       
D CountOverflow           1      4U 0       
D WholeCounter            1      8U 0 INZ(0)

Again, this is best done in a procedure.

Upvotes: 3

Related Questions