MVSFORUMS.com

[pcmoonbeam]

I have a production COBOL II program in which a subscript reaches a value of zero yet does not cause the program to abend. I have put displays in the code to verify this, because I couldn't believe that the program could continue to execute with a subscript pointing to a zero occurrance. This zero subscript is computed as shown here:

[kolusu]

pcmoonbeam,

Add the following line as the first line in your program(before ID Division) and then try again.

[pcmoonbeam]

Thanks Kolusu,
I checked the compile options that we use and wouldn't you know it, NOSSRAGNE is in effect.
_________________
Thanks,
PCMOONBEAM

[kolusu]

pcmoonbeam,

Now are you clear as to why your program did not abend?

Kolusu
_________________
Kolusu
www.linkedin.com/in/kolusu

[pcmoonbeam]

Yes, crystal clear, thanks!
We are about to embark on a decision in our shop as to whether or not we should change this option to be in effect. If we do, it will have a huge impact when we recompile existing programs.
_________________
Thanks,
PCMOONBEAM

[semigeezer]

Just as an aside, there are times when not checking subscript ranges can be useful: reading prefixes of control blocks for example.

Array index use is just a storage address calculation: address of structure + (index -1)*length of element

Using that, you can define a simple array of pointers containing OCCURS n TIMES based at address 0 where n is any positive number and access any location in storage by simply manipulating the index. Not good programming practice because it is hard to maintain, but in a quick-and-dirty programming situation, it can be very useful.

Also, note that the logic of the program should avoid out of range conditions. If you need SSRANGE to catch the error, you are adding runtime overhead to everything inorder to catch a design flaw in the program. (Yes, that is a philosophical point, but that is my opinion on it)

[pcmoonbeam]

Thank you for your interest. However, my biggest problem is that I'm taking on a position from an individual who didn't test thoroughly, so now I'm having to do damage repair. Having SSRANGE in place on compiles (at least for testing purposes) points directly to the problems that were created.
_________________
Thanks,
PCMOONBEAM

[semigeezer]

Hmmm I thought all programming was an excersise in damage repair. I know it is when I write code

[edkir98]

I was browsing the forum on SSRANGE and got hold of this topic.

In our system, we dont specify the compiler options. The cobol compiler does not work. We do it thru endevor. If we would want to compile a code then we would add it to ENDEVOR(Version Control Tool). There they have the default compiler options listed and it compiles like that.

I'm able to see from the compiler listing that SSRANGE parameter is set while compiling.

I had used an two dimensional array (Internal Table) and in a particular paragraph, the array index was set as (1,0). This caused memory overlap and hence a wrong value had moved to another variable and hence it caused a S0C7 abend.

Why is that, although SSRANGE parameter was set why dint the program abend when it referenced a invalid starting position? Why dint it do the "range-checking"?

Am I clear with my question? Please advise if there are any additional options that would override this paramer?

[mf_user]

Hi, semigeezer.

[CraigG]

SSRANGE also produces extra code in the program, which is why we con't use it for production compiles.

[edkir98]

Craig.. You are correct!! I see NOSSRANGE is defined in the PROD environment. But why is that?
_________________
Thanks

[jsharon1248]

A limited sample to be sure, but the shops I've visited in my career enable SSRANGE for the test environments and then NOSSRANGE for production. The thought being that adequate testing will reveal out of range conditions prior to rolling out the code to production.

edkir
Please be careful when you make a statement like

[acevedo]