Re: STRPDM Option 12...Doesn't Work

On 08-Jan-2016 15:40 -0700, John Yeung wrote:

On Fri, Jan 8, 2016 at 1:26 PM,rob wrote:

A proxy command is not a copy and a rename. It's more like the
following:
ADDLNK OBJ('/QIBM/ProdData/Java400/bin') NEWLNK('/rob/ibmstuff')
LNKTYPE(*SYMBOLIC)

I never said it was actually a copy and rename, just that it felt
about as "heavy" as copying and renaming. The help for CRTPRXCMD
says that a new *CMD object will be created. But the new object will
be a "proxy command definition object" rather than a "command
definition object", so I suppose it could be smaller/lighter than an
actual command. It would still be a persistent object in the file
system, though. Is there anything like a DOS or Unix alias?

AFaIK most *CMD objects with several parameters will be twice the size of the smallest; with few or no parameters, then the smallest size. So although every proxy command likely will be, presumably is, the smallest possible *CMD object size, the perception that there is little difference between the CRTPRXCMD and use of RNMOBJ after CRTDUPOBJ is quite close to reality from the perspective of both overall size [even the result of (2*small) is still fairly small] and their tracking as /objects/.

Given the OP\subject is about STRPDM, the Program Development Manager (PDM) feature does provide an alias feature [of sorts], via the User-Defined Options (often options-file QAUOOPT). Unfortunately the options are not accepted on the command-line, nor is there a direct ability to customize a particular options-file to any particular invocation :-( i.e. there is no OPTFILE() or USRDEFOPTF() parameter available on the various WRKxxxPDM commands, though I had suggested that capability in the past.

However I have known several who had implemented synonym\alias features in their own CL command shell; per lack of any such support in the QCMD shell\request-processor. I recall trying-out one of those shells once, as an actual CL Request Processor program, but do not recall what was their implementation; presumably a database file, though perhaps a user space or user index. And I used one occasionally that I had written myself, implemented as a REXX routine, accepting input from the screen [to include interpreting a SYN request that would add a new synonym local to that shell invocation or to be loaded again the next time the REXX was invoked as the shell or as running a single command]. As I recall, my REXX shell had attempted to mimic what IIRC was called the SYNONYM support of the ¿MVS EXEC? command language on the IBM VM mainframe, though the feature may have been an add-on; I seem to have forgotten many of the terms for that environment except that the access was via 3270 vs 5250 :-)

Unrelated, but that mention of QCMD shell had me recalling also: A different request processor I recall using was QCMD with F9=Retrieve modified to search for the last invocation using the input on the command-line as a prefix search; e.g. with WRK typed on the command-line, pressing F9 resulted in the prior WRK* command\request-message that was typed-in. I do not recall if the feature retrieved the last _functional_ invocation [only returning a prior request with that prefix if that prior request message had completed without a Escape message], but I have long wanted that capability in QCMD; similarly to limit the number of F9 requests to get what I want to appear on the command-line [given the most recent failed command remains as with the current design of QCMD].