On Wed, 27 Nov 2013 13:38:31 -0800, Phil Smith wrote: >OK, this is all leading me to think this can work. Let me add a few details. > >The module source and CSECT name are RODENT; the entry points are HAMSTER, >MOUSE, and RAT. [No, these are not the actual names!] > Why not?
>Nobody ever calls RODENT: they call HAMSTER, MOUSE, or RAT. > >In batch, the RODENT module (already out there, in use, so that name can't >change) is (usually) called dynamically, though it can be statically linked as >well. > If nobody ever calls it, it shouldn't matter if it vanishes. >In IDMS, the same code (the three functions share a bunch of code) needs to >say "Ah, look, an ICM of the address of the WXTRN is non-zero, so I need to >call that function to do some IDMS magick". > >So if there's a RODENT in BATCHLIB and one in IDMSLIB (and no, of course I'm >not going to call the libraries that!), this would Just Work. > The names must be different. E.g. NAME RODENTB(R) and NAME RODENTI(R). >And it sounds like I can do that, by having two sections in the APPLY stuff, >with different SYSLMOD DDs. Correct? > I agree. You might want to clean up by ++ DELETE RODENT, else any future update is likely to re-install RODENT (Harmless, but distracting). The MOD element may continue to be called RODENT, and that name should appear in INCLUDE statements for RODENTB and RODENTI. You could get fancy/parsimonious and name one LMOD HAMSTER with aliases MOUSE and RAT; and the other LMOD MOUSE with aliases HAMSTER and RAT. Or not. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
