On Fri, 12 Apr 2019 09:23:40 -0400, Steve Smith <sasd...@gmail.com> wrote:
>I think the grumpiness and sarcasm might be a bit over the top. The error
>does have roots going back to the middle ages,
>
IBM should operate a booth at Renaissance Faire. Or Comic Con.
> ... that when combined with
>modern functionality, causes some odd results. Anyway it's something that
>doesn't happen very often. If IBM thinks it's worthwhile, they'll improve it.
>
That would be an RFE. As new facilities such as zFS appear it happens
increasingly often.
> ... I doubt there's any reason to worry about error message compatibility.
>
I'm staying with my assumption that the message is formatted by SYNADAF.
SYNADAF is admirable; it's reusable code. Not every facility needs to write
its own diagnostic formatter. But wide use engenders wide exposure to
incompatibility. How many callers make assumptions about SYNADAF message
format? Probably nonzero. In:
https://www.ibm.com/support/knowledgecenter/en/SSLTBW_2.3.0/com.ibm.zos.v2r3.idad500/syndf.htm
The SYNADAF macro uses register 1 to return the address of an area
containing
a message. The message describes the error, and can be printed by a later
PUT,
WRITE, or WTO macro. The message consists mainly of EBCDIC information and
is in variable-length record format. The format of the area is shown
following the
descriptions of the SYNADAF parameters.
I assume that's Y(count),X'0000',up to 32,752 bytes of message text.
What callers would suffer overruns copying an unexpectdly long message into a
static buffer? Would some simply attempt to PUT it, as returned, to SYSTERM
and trigger a cascade of similar errors?
Regardless, it's lousy, and should be fixed. As reusable code, not only for
HLASM.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
