On Mon, 23 Nov 2015 07:42:13 -0500, Peter Relson wrote: >Whether it is a good idea or not to free the ddname, each API that intends >to be called with user-allocated DD's (rather than with system-allocated >DD's) should document its behavior in this regard. > FWIW, my case was the reverse: I invoked the SDSF API and it allocated the DD for me to use.
I need to RTFM; if there is a restriction, it might appear in the SDSF reference, or in more general JES documentation, in which case I wouldn't know where to look. What are the restrictions on spool data sets? I've successfully concatenated SYSIN in JCL both before and after other other data sets. I might expect spool data sets allocated by SDSF to have similar capabilities. >I differentiate the two cases because if the only expected environment is >"EXEC PGM=" (and yes some utilities do expect such an environment, even if >you might like to "call" from your program) I don't think that the >behavior is of interest. > Where is the "expected environment" documented? I once had IBM support give me WAD on an SR (against HLASM. IIRC) that I had invoked with ATTCHMVS, arguing that the *examples* in the manual showed only invocation by "EXEC PGM=". I feel it's "common knowledge" that any program that can be invoked by "EXEC PGM=" can, absent an explicit restriction, be invoked alike by ATTACH. (I am aware of AC=1 restrictions; I'll accept those as "common knowledge".) Thanks, gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
