On 02/28/2013 06:52 AM, Peter Relson wrote:
As I understand the example, it was dependent on SYSPARM having 7
characters. For what it's worth, that happens to be the typical SYSPARM
that IBM products are built with (the "FMID" is 7 characters long, as is
an APAR number) when SYSPARM is used at all. Nevertheless, if a sample is
truly wrong, that is APARable. If documentation is truly wrong, that is
APARable. Emphasis on truly wrong. Whether a fix will be provided in
service or in a subsequent release is a different question.

Clearly no sample ought to be using undocumented (hence unsupported)
parameters. And since this cost someone time to figure out, it might cost
someone else time later so should be brought to the attention of the owner
to address.

As to the change activity comment of "fix J", SAVE was "jumpified"
incorrectly (by me) in z/OS 1.3 (based on the user's specification of
SYSSTATE ARCHLVL). The J instruction I had used was incorrect. The APAR
corrected that.
As to the "botched" handling -- IBM will certainly not change the SAVE
macro to do anything about this. You may not like the behavior, but it is
not really wrong. It may well be not what you happen to prefer. Since it
has been this way likely for longer than you (or I) have been someone who
might use it, perhaps it is the macro that is right and "we users" who are
wrong <grin>?

Peter Relson
z/OS Core Technology Design

...
I would contend that the length of time a behavior has been in place doesn't really count if the bad consequences of that behavior have only recently been revealed.

If IBM insists on retaining a SAVE macro definition that behaves badly when a character string parameter contains quotes, then at the very least the manual documenting the macro should document that as a restriction and those writing sample code for other IBM manuals should be made aware of and honor that restriction. If it is felt that a terse and accurate description of what constitutes a valid form for this SAVE parameter cannot be written, then perhaps that is the best argument for fixing the macro.

SYSPARM is a documented system variable that is set based on parameter options specified when the Assembler is invoked. Installations may use this parameter for their own purposes, which makes any assumptions of there being a conventional length for this parameter invalid.

--
Joel C. Ewing,    Bentonville, AR       [email protected] 

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to