WRT "WTO then ABEND:
That's one possibility.
My preference is to use dead traps,
sometimes containing a documented message,
most times containing only a number that is unique throughout the source.
Both methods will usually result in a support call,
so it's a coin flip on that score.
My method will also produce summary dump messages in SYSLOG,
so there's that.
Dave
At 6/20/2025 12:07 PM, Tom Brennan wrote:
I figured that about executing text - I was joking around :)
More seriously though, I hope I'd never come
across any of these in production code, since
they are all for testing. If we really need to
die in production I'd expect something more like this:
WTO 'XYZ1234E THIS CANNOT BE HAPPENING'
ABEND 1234
On 6/20/2025 2:13 AM, David Cole wrote:
Hi Tom,
There is a problem with using DC C'AB...' to trigger an intentional 0C1...
Somewhere in the Principles, there is a
statement, a promise if you will, that...
- X'00' will never be used as a valid opcode.
- No such promise is made for X'C1'.
As it happens, as of the latest edition of
Principles (edition -14 for the Z17), X'C1'
still has not been used as an opcode. But there
is no guarantee that will remain true in the future.
In other news. I, being the created of the
X'00DEAD00' "instruction", have a couple of comments about it...
The "instruction's" "opcode" (if you will) is
X'00DEAD'. The trailing X'00' is an operand.
When non-zero, it is the length of message text that follows the "opcode".
The macro that generates the dead trap is named
#DIE. It has an operand you can use to provide
the message text. Absent that text, the macro
will generate a four-digit number that will be
unique in the assembly. Example: X'00DEAD04',C'3208'
z/XDC has logic that recognizes dead traps and
displays its message if it has one. But as
someone else mentioned, even without z/XDC,
when have an 0C1 failure at a dead trap, you
can instantly see that it is a dead trap, even
if all you have is summary dump messages in SYSLOG.
Finally, you don't need to have z/XDC to use the #DIE macro.
IHTH,
Dave Cole, Developer
[email protected] (personal)
[email protected] (business)
540-456-6518 (cell)
At 6/19/2025 04:08 PM, Tom Brennan wrote:
Nice! These seem to work too:
DC C'ABEND 0C1'
DC C'S0C6'
On 6/19/2025 11:55 AM, Leonard D Woren wrote:
DC X'00C1' abend S0C1
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
