As someone who usually (almost always) only tells people to use "documented and supported" interfaces, I almost hesitate to mention this. HOWEVER,
There is a currently TOTALLY unsupported and undocumented (it was documented in VS COBOL II, R1.0 - but never since then) "feature" that MIGHT give you some information that would be helpful for things like this. Back when VS COBOL II was "current" and some sites were weighing whether to go to it or not and to give up their old friend "CAPEX Optimizer" - I created a GUIDE (yes GUIDE, not SHARE) requirement to make some of this a "generally available programming interface". I can't remember if this died or was rejected by IBM. Any way, I scanned the OLD documentation for "RNDMP" and it is available on line at: http://home.comcast.net/~wmklein/IBM/rndmp_1.pdf http://home.comcast.net/~wmklein/IBM/rndmp_2.pdf http://home.comcast.net/~wmklein/IBM/rndmp_3.pdf http://home.comcast.net/~wmklein/IBM/rndmp_4.pdf http://home.comcast.net/~wmklein/IBM/rndmp_5.pdf The last time I checked, the support was still in the current compilers, but I don't know for sure that it still is. What you would want from all of this is (with OPT and LIST specified) to put : $RNDMP FLAGS=1,3 shown in "page 4" and described in page 5. If there is part of this that you think WOULD be "generally useful", then you might (or might not <G>) want to try a SHARE requirement to see what IBM thinks of it now. (The "$" goes in column 8 of the first line of your source code. "McKown, John" <[EMAIL PROTECTED]> wrote in message news:<[EMAIL PROTECTED]>... > I was just looking at some COBOL code because it just crashed one of our > production CICS regions. I found the problem. Basically, an infinite > loop where each iteration of the loop added one to a subscript value and > put an "error message" into a table. The infinite loop is because the > PERFORM of the paragraph used a WHEN clause which tested a variable > which was not even referenced within the PERFORM'ed paragraphs. > > So now I wondering if there is any automated system which can detect > things like this. I do know that in general, it would be difficult > because flow of control can be very complicated. I also noticed that the > "counter" for the array was PIC 9(3) USAGE DISPLAY. Another thing that I > wish I could detect. > > Is there anything on the market to detect errors or poor coding which is > relatively OBVIOUS, if only somebody had the time to examine it? > > -- > John McKown > Senior Systems Programmer > HealthMarkets > Keeping the Promise of Affordable Coverage > Administrative Services Group > Information Technology ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html