>I would have said that the exec type Unix functions were more like the MVS >XCTL macro.
I dare to disagree. XTCL replaces the program (load module) with the new program (load module). This all happens under the same (current) TCB and RB. exec() also replaces the current program with the new one, *but* this all has to somehow flow through the initaitor. There is an SMF step-end type 30 record being cut. You will see corresponding step end messages in the joblog. There is an additional (sub) step with stepname *OMVS (or is it OMVS*; I can't verify right now). Some UNIXish resources related to UNIX and described there to survive an exec(), will be exempt from the usual step end cleanup. Every other resource acquired before the exec(), such as virtual storage areas, DD allocations, and what have you, will be freed, release, etc. I believe STEPLIB is no exception. A new STELLIB allocation will be established in the new (sub) step. I may consist of the same data sets (STEPLIB=CURRENT), it may be something else (STEPLIB=dsn1;dsn2;....), or it may not be established (STEPLIB=NONE or undefined). -- Peter Hunkeler ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
