Thanks, Peter - that's interesting; I was unaware of that change despite working on IBM software development for many years. In my area (HLASM) we only talked about "intended programming interfaces", but our manuals still referred to "general-use programming interface" information. HLASM last had a new release in 2008, so perhaps those considerations only applied to new releases. For HLASM, the listing and message texts were not considered an intended programming interface, but the message ids were. However, we tried to minimize changes to the listing because we were aware of post-processing tools that used it. The optional ADATA output is an intended programming interface which contains similar information to the listing, but can change from one release to the next, and can be extended (compatibly) within a release.
Jonathan Scott -----Original Message----- From: IBM Mainframe Assembler List <ASSEMBLER-LIST@LISTSERV.UGA.EDU> On Behalf Of Peter Relson Sent: 03 September 2025 13:17 To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Is HLASM efficient WAS: Telum and SpyreWAS: Vector instruction performance IBM stopped differentiating between GUPI and PSPI many years ago. There is now only PI, although you will see many macros that still show both (and the differentiated cases remain interesting information). <snip>>Generally, you can add new fill-in choices Let me qualify this because there is a risk of a PE'd PTF because a new fill-in field is inserted before an existing fill-in field.</snip> Adding a new fill-in field would not generally fall into what I was thinking of as adding a new fill-in choice.But, sure, make mine "can add new fill-in choices to an existing fill-in" for clarity. Peter Relson
