It would be even nicer if the re where an ACB inrface that supported a concatenation of PDS, PDSE and Unix directoried. There was a Share requirement a few decades ago, but nothing happened.
I wrote the existing code to be compatible with PL/i i/O; since REXX doesn't have a concept of buffered versus unbuffered I/O, I would be free to use READ/WRITE/CHECK exclusively for a REXX version. So the question is whether there is a gain in efficiency using QSAM and, if so, whether it is enough to worry about. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 ________________________________________ From: IBM Mainframe Discussion List <[email protected]> on behalf of Kirk Wolf <[email protected]> Sent: Friday, July 10, 2020 9:51 AM To: [email protected] Subject: Re: Any interest in PSD I/O for REXX? It would be nice if there were a "supported system interface" for mixing BPAM and QSAM this way. Dual DCB approach isn't great either. Given then, isn't it best just to use BPAM/BSAM? On Thu, Jul 9, 2020 at 11:41 PM Seymour J Metz <[email protected]> wrote: > The QSAM fiddling is what makes GET/PUT work with multiple members. I'm > fairly confident that with z/OS versions of IGGIOBEX , IGGICQE and > possibly SAMB the existing code would work in z/OS, but that would only be > for PDS. > > Refitting the code for enterprise PL/I would require knowledge of how > files are passed and mappings for th DCLB and FCB, but all of the > interfaces that I would need to do it for REXX are well documented. So my > questions are whether there is enough interest to justify doing it and > whether efficiency is important enough to justify using QSAM with games or > whether to just do the whole thing with BPAM. If the latter, it might be > asier to just add a write routine to the existing read routine on the > CBTTAPE. > > > -- > Shmuel (Seymour J.) Metz > http://mason.gmu.edu/~smetz3 > > ________________________________________ > From: IBM Mainframe Discussion List [[email protected]] on behalf > of Steve Smith [[email protected]] > Sent: Thursday, July 9, 2020 10:47 PM > To: [email protected] > Subject: Re: Any interest in PSD I/O for REXX? > > I think you'd be disappointed in FAMS, I still don't think it would help. > We all wonder why it's a secret, because there ain't that much to it. > IEBCOPY and the LISTDSI command pretty much expose all there is. > > After looking at your old program, I'd give it at least 50-50 odds for > working with PDSEs. I think your worst case would be to abandon QSAM > fiddling and just provide GET/PUT functions. You could probably make that > much more seamless in REXX than in PL/I. > > sas > > > On Mon, Jul 6, 2020 at 10:40 PM Seymour J Metz <[email protected]> wrote: > > > My STOWBLDL routine at http://mason.gmu.edu/~smetz3/source/STOWBLDL.ASM > > has code to quiesce a QSAM DCB before going to a new member and reprime > > buffering afterward; this requires dealing with fields in the SAM-E IOB > > extension and the SAM-E Interrupt Control Block. I assume that I would > need > > FAMS to do the equivalent for PDSE. > > > > > > -- > > Shmuel (Seymour J.) Metz > > http://mason.gmu.edu/~smetz3 > > > > > > ---------------------------------------------------------------------- > 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 > ---------------------------------------------------------------------- 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
