> On Dec 20, 2017, at 2:25 PM, John McKown <[email protected]> wrote: > > It would need to front in LOAD in addition to LINK. As I understand it, > the first time a COBOL program does a dynamic CALL, it goes to a subroutine > which maintains a table. If the name is in the table, then the program has > already been LOADed and control is given to it. If this is the first time, > then the routine will do a LOAD and update the table with the name and EPA. > I don't know that for certain, but I'm fairly sure that's how it works. > This is why there are special LE routines for CICS because the CICS version > does an EXEC CICS LOAD instead of a z/OS LOAD. Again, as best as I know. > Damn OCO!John,
IIRC it does that as well. This was at a time before LE (if memory serves me). My memory says that our boss was trying to find “something” and forgot to turn the trace off and all of a sudden SMF grew really large and fast. If memory serves me three times a day and climbing. I ran a utility that summarized the number of smf record by number. When we saw that it was from the product then we started to back track. In the morning status meeting I asked if anyone had turned on the tracking as SMF had gotten a lot larger all of a sudden and before we got into exactly what was being traced it would be nice to turn if off. My boss was not amused and he said he had forgotten to turn off the trace. We turned it off and SMF went back to normal. Ed > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
