On Sat, 5 Jan 2008 02:19:44 +0000, Bill Wilkie wrote: >Tom: > >No, only the one that is coded. If you have PARM=" " coded, it works as >always. If you code NEWPARM="" they Getmain an area for something larger like >256 bytes and everything else works the same. That way you could have a >controlled implementation without mucking up what's there. Once the support >for newparm is in the system, you can add it where needed. > In fact, there's the seed of a compromise here. Provided that NEWPARM doesn't introduce any unexpected restrictions; i.e.:
o It supports symbolic substitution in the same way as PARM. o It can be used in PROCs and open code alike, even as PARM. There could be a bridge load module; preferably from IBM; as an alternative from CBT which would reformat the NEWPARM into classic PARM format, verifying that the length does not exceed 65,535 (no need to limit it to 32,767) and XCTL to the program named in the first 8 characters of NEWPARM. Advantages: o The number of job steps would not be inflated. o (as before) symbolic substitution would work o (as before) it could be used either in PROCs or in open code. o The installation could choose to install it with either AC=1 or AC=0; in the latter case thwarting the particular hazard of invoking a program in authorized state with the hazard of buffer overflow. -- gil ---------------------------------------------------------------------- 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
