On 13 Feb 2014 09:35:50 -0800, [email protected] (Tom Ross) wrote:
>>If other programs that handle files created by CSP (and maybe its >>successors) were using NUMPROC(MIG) because it was more efficient than >>NUMPROC(NOPFD), then those programs may have to be recompiled with >>NUMPROC(NOPFD) because there can be problems. I ran into that with a >>program that for whatever reason did an IF NUMERIC test of a field >>that had be affected by CSP and using NUMPROC(PFD) with that program >>caused the field to test not numeric. Recompiling the program in >>question with NUMPROC(MIG) solved the problem. The problem is not the >>compiler options used for the COBOL generation of CSP programs, the >>problem is the compiler options of the programs that use the output >>from CSP programs. > >That was an error, the correct solution would have been NUMPROC(NOPFD). >It is just a coincidence that NUMPROC(MIG) solved the problem. If >preferred signs are not enforced, you use NUMPROC(NOPFD). If preferred >signs are always enforced, then you can use NUMPROC(PFD). If you are >migrating from OS/VS COBOL, and mixing OS/VS COBOL with newer COBOL, >then you might have used NUMPROC(MIG). > >>Why was NUMPROC(MIG) eliminated? > >We don't support mixing OS/VS COBOL with Enterprise COBOL V5, so >it is no longer relevant. Gjiven that NUMPROC(MIG) caused more efficient code to be generated than did NUMPROC(NOPFD) and that there are sources of data other than COBOL programs, CSP with its convolutions to force F zones for positive numbers being just one of them, eliminating NUMPROC(MIG) eliminated this efficiency unless the code generation for NUMPROC(NOPFD) has improved. My comment is based on research done at the time I first encountered the problem which was with COBOL v1.4, the first 85 standard COBOL from IBM or COBOL for MVS and VM. This also will affect running programs that depend on the current NUMPROC(MIG) behavior. Will changing to NUMPROC(NOPFD) from NUMPROC(MIG) for those programs introduce other side effects such as different results for the same inputs. Clark Morris > >Cheers, >TomR >> COBOL is the Language of the Future! << > >---------------------------------------------------------------------- >For IBM-MAIN subscribe / signoff / archive access instructions, >send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
