2

I'm looking at some legacy programs, where code like the following is found:

C/free
    Dow 1=1;
        SubRoutine();
    EndDo;
    *INLR = *On;
/end-free

The program in question is a server-like program, where the sub routine handles incoming requests over the network. To me, it seems that the *INLR = *On statement will have no effect, but then again, my knowledge of RPG is severely limited, and this pattern repeats itself in several programs I'm looking at.

My question is thus: is the last line before /end-free unreachable and thus redundant?

WarrenT
  • 4,502
  • 19
  • 27
Eyvind
  • 5,221
  • 5
  • 40
  • 59
  • I'd tend to agree (though i also have little knowledge of rpg). It looks like it is only ever going to be reached when 1 != 1. – mcalex Jan 03 '13 at 07:41

1 Answers1

4

You are correct, that in the example, as shown, there is no normal way to exit this loop. It appears tat this program will only end from an error or job cancellation. It is sometimes better to check %shutdown() in the main DOW, to allow a never ending job to close gracefully, but that was either not chosen, or not thought of, in this case.

I might speculate that the program was written by copying another, and that the *INLR line may have been left in the code as a conventional way of marking the end. So someone might argue that the line is harmless and could make it easier to understand that is the end.

WarrenT
  • 4,502
  • 19
  • 27
  • 4
    RPG programs 'end' by RETURN or LR being on when the detail calculations end (either through end of file or SETON). If the compiler doesn't detect one of these, the compiler will terminate with an RNF7023 error - 'The Compiler cannot determine how the program can end.' – Buck Calabro Jan 04 '13 at 16:21
  • 1
    I agree with Buck. That line was probably put there to appease the compiler. – Tracy Probst Jan 05 '13 at 13:24