Sure, it's a continuum. The logical limit of "you didn't follow the
requirements" is a system whose error messages are all just a big red "NO".
You're not suggesting that, of course, but since this is a behavior change from
the previous release AND there is an assertion being tested, it feels like
someone just was being lazy and didn't handle an error case that they did
realize could happen. That's not great.
"Applications people worry about how it's going to work.
Systems people worry about how it's going to fail."
-- me
-----Original Message-----
From: IBM Mainframe Discussion List <[email protected]> On Behalf Of
Peter Relson
Sent: Friday, September 12, 2025 12:12 PM
To: [email protected]
Subject: Re: Would you consider this a compiler bug?
It would rarely be considered an APARable defect if something unexpected
happened when you did not follow the documented requirements (this is certainly
not unique to the COBOL compiler; the same is true for almost any documented
interface). Handling the situation "nicer" sometimes is necessary but often
would be an "enhancement". I'm fully in favor of such enhancements, but they do
need to be prioritized and paid for (which could be at the expense of something
else that you might prefer having).
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
