On 2016-10-26, at 09:22, Pew, Curtis G wrote: > On Oct 26, 2016, at 9:17 AM, Charles Mills <charl...@mcn.org> wrote: >> >> Sounds like this is slightly bleeding edge. I think I will try to sync my >> "code management" processes (C versus HLASM) a little more closely together >> but continue to keep HLASM source and object in a PDSE. > A mischievous programmer wishing to invalidate depending on PDS(E) might craft #include with such as: o member names with multiple unconventional filename extensions. o long filenames. o case sensitive filenames. o multi-evel directory paths.
In fact, how does C/C++ in the JCL environment deal with such as "#include <sys.wombat.h>"? > I’ll just throw out what I do. I keep the source in HFS (but not object; > that’s still in a PDSE) and use OCOPY to put it in a temporary dataset for > the assembler. ... > What's the benefit in keeping the object in a PDSE? Binder is competent with UNIX files. I dynalloc various UNIX directories with generated DDNAMES then filter SYSLIN to such as: CASE MIXED INCLUDE SYS00011(foobar.o) INCLUDE SYS00012(wombat.o) (The unfiltered SYSLIN becomes SMP/E JCLIN.) It remains a PITA that Binder doesn't support UNIX directories in a SYSLIB concatenation, but a boon that Binder (unlike HLASM) does not support explicit DDNAME specification on copybook members. What's the benefit in OCOPY rather than just allocating UNIX files/directories to SYSIN/SYSLIB? (But I understand that ISPF does something similar to create temporary SYSLIBs.) We use an IBM product. CM Synergy, that creates volatile directories of symbolic links to selected member-versions rather than copying the member bodies. On 2016-10-26, at 00:59, John McKown wrote: > > ... One thing which made this much easier > was FLOWASM ( ftp://ftp.phoenixsoftware.com/pub/demo/asmmods.xmi ), ref: > http://planetmvs.com/hlasm/tips.html . FLOWASM > > is an HLASM user exit program for reading the SYSIN & SYSLIB. It makes > HLASM "free format". Quite nice! This was a case of "eating your own dog > food" because I was developing some UNIX utilities. > > If you're curious, these are available on the CBTTape.org site, in file > #864 http://cbttape.org/cbtdowns.htm > This doesn't fit our current cross-assembler process well, but thank you for doing this. Formerly, we used those exits (but not FLOWASM). How do you deal with nested COPY instructions? With persent-day resources opening multiple file handles seems more sensible than replacing the NOTE-POINT shell game with similar ftell(); fseek(). And I mentioned earler that I had to get PTF fixes for problems (not using the exits) with nested COPY instructions in a SYSLIB of mixed UNIX and PDS directories. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
