Signal on error

[Joseph Reichman]

Hi

I trying to capture B37 space or directory space errors in my rexx routine
The documentation says that special variable RC in the case of a system abend negative of the decimal value of the abend code

Wonder if anybody would anything about this

Thanks

----------------------------------------------------------------------
For TSO-REXX subscribe / signoff / archive access instructions,
send email to LIST...@VM.MARIST.EDU with the message: INFO TSO-REXX

[Paul Gilmartin]

On 2021-04-15, at 12:45:33, Joseph Reichman wrote:
>
> I trying to capture B37 space or directory space errors in my rexx routine
> The documentation says that special variable RC in the case of a system abend negative of the decimal value of the abend code
>
> Wonder if anybody would anything about this
>

I do.

You have beeen advised many times on various lists to show your
code and trace output. It's the best way to get help from a list.

-- gil

[Willy Jensen]

Quick test shows rc=20 for EXECIO D37 abend. But you can trap the message by
the TRAPMSG function like in this sample:

signal on error name er

zz=outtrap('lst.')

mt=trapmsg('ON')

do n=1 to 999999

"execio" rec.0 "diskw dd1 (stem rec.)"

end

zz=outtrap('off')

zz=trapmsg(mt)

say 'end..'

exit

er:

zz=outtrap('off')

zz=trapmsg(mt)

say 'rc:' rc 'n:' n

do n=1 to lst.0

say '->' lst.n

end

exit

Shows:

-> IRX0250E System abend code D37, reason code 00000004.


-> IRX0255E Abend in host command execio or address environment routine TSO.


-> IRX0670E EXECIO error while trying to GET or PUT a record.






-----Oprindelig meddelelse-----
Fra: TSO REXX Discussion List <TSO-...@VM.MARIST.EDU> På vegne af Joseph
Reichman
Sendt: 15. april 2021 20:46
Til: TSO-...@VM.MARIST.EDU
Emne: [TSO-REXX] Signal on error

Hi



I trying to capture B37 space or directory space errors in my rexx routine
The documentation says that special variable RC in the case of a system
abend negative of the decimal value of the abend code



Wonder if anybody would anything about this



Thanks



----------------------------------------------------------------------

For TSO-REXX subscribe / signoff / archive access instructions, send email

to <mailto:LIST...@VM.MARIST.EDU> LIST...@VM.MARIST.EDU with the message:
INFO TSO-REXX



--
Denne mail er kontrolleret for vira af AVG.
http://www.avg.com

[Joseph Reichman]

Thanks

I appreciate it somewhat new to Rexx



> On Apr 15, 2021, at 3:24 PM, Willy Jensen <wi...@harders-jensen.com> wrote:
>
> Quick test shows rc=20 for EXECIO D37 abend. But you can trap the message by

[Joseph Reichman]

Paul

Thanks for some reason even tough I post from my personal email reichmanjoe@gmail
I don’t get messages back

I’m able to receive messages on my work email Joseph....@IRS.gov but because of irs security I can send them
Out any way

Thank you very much

[Seymour J Metz]

Maybe you set NOMAIL for that address?


--
Shmuel (Seymour J.) Metz
http://mason.gmu.edu/~smetz3

________________________________________
From: TSO REXX Discussion List [TSO-...@VM.MARIST.EDU] on behalf of Joseph Reichman [reich...@GMAIL.COM]
Sent: Thursday, April 15, 2021 5:18 PM
To: TSO-...@VM.MARIST.EDU
Subject: [TSO-REXX] Signal on error

[Seymour J Metz]

ABEND uses a system completion code and a user completion code in the same word; normally one of those fields is zero. Does REXX return both, negated? If not, what is in RC for a system code versus a user cod?

--
Shmuel (Seymour J.) Metz
http://mason.gmu.edu/~smetz3

________________________________________
From: TSO REXX Discussion List [TSO-...@VM.MARIST.EDU] on behalf of Joseph Reichman [reich...@GMAIL.COM]

Sent: Thursday, April 15, 2021 2:45 PM

To: TSO-...@VM.MARIST.EDU
Subject: [TSO-REXX] Signal on error

[Robin]

What possible recovery do you think your REXX program can do.

Suppose a D37 abend occurs. That happens not because of any error in your
program. It is not even any error in the program performing the I/O. Such an
error is in the design of the whole application wherein the number of
members being created exceeds what was anticipated by the designer. Recovery
from such is not the responsibility of the program but the application
designer. I suggest this is the case in any Abend.

-----Original Message-----
From: TSO REXX Discussion List [mailto:TSO-...@VM.MARIST.EDU] On Behalf Of
Joseph Reichman
Sent: April 15, 2021 2:46 PM
To: TSO-...@VM.MARIST.EDU
Subject: [TSO-REXX] Signal on error

[Paul Gilmartin]

On 2021-04-16, at 08:59:26, Robin wrote:
>
> What possible recovery do you think your REXX program can do.
>
> Suppose a D37 abend occurs. That happens not because of any error in your
> program. It is not even any error in the program performing the I/O. Such an
> error is in the design of the whole application wherein the number of
> members being created exceeds what was anticipated by the designer. Recovery
>

Reallocate as PDSE; IEBCOPY; deletee; rename; retry.

> from such is not the responsibility of the program but the application
> designer. I suggest this is the case in any Abend.
>

And yet SMP/E performs similar recovery from some x37 ABENDs.

-- gil

[Phil Smith III]

Robin wrote:

> What possible recovery do you think your REXX program can do.

> Suppose a D37 abend occurs. That happens not because of any error in your

> program. It is not even any error in the program performing the I/O. Such
an

> error is in the design of the whole application wherein the number of

> members being created exceeds what was anticipated by the designer.
Recovery

> from such is not the responsibility of the program but the application

> designer. I suggest this is the case in any Abend.

That's silly. First, a friendly error is more useful than "Something ran out
of space". Second, depending on what the Rexx program was doing, it might
need to unwind stuff, or at least tell the user how far it got, or that it
didn't do any damage yet.



Not to pick on you, Robin, but z/OS user-hostile behavior is not a good
thing. There are far too many places where diagnosing a simple and common
problem requires trolling through job or other logs, rather than receiving a
coherent and meaningful message somewhere trivially accessible.



I assume this behavior all goes back to OS/360, where these errors were the
best it could produce on tiny, slow machines; this ain't 1964 any more. Some
of these problems may in fact be impractical to fix because there are four
different commercial products (two from Broadcom, one from BMC, and one from
IBM) dedicated to helping users with them, but that's no excuse for anyone
writing new code not to think about failure modes and make them as easy to
diagnose and fix as possible.



z/OS isn't the only offender here, of course-cf. the infamous "The operation
failed" in Outlook (with antecedent for "the"), or (15 years ago, hopefully
since fixed) Postgres startup, which provided NO data if it failed: you had
to hack the startup script to get it to emit errors.



I could rant on this for hours but I'll stop now, ending with my
gross-but-not-inaccurate generalization:

Applications programmers think about how the code will work.

Systems programmers think about how the code will fail.

Feel free to change the A/S words to other, less kind delineations.



...phsiii (it must be Monday)

[Joseph Reichman]

The Rexx program assembles and links
A program

Blowing of the directory is possible on assemble sysprint and linker syslmod

In which can compress and reassemble or relink



> On Apr 19, 2021, at 4:08 PM, Phil Smith III <li...@akphs.com> wrote:

[zMan]

Is that a haiku? Seriously, please think, use complete sentences, like
that. Your post is gibberish.

On Mon, Apr 19, 2021 at 4:13 PM Joseph Reichman <reich...@gmail.com>
wrote:

--
zMan -- "I've got a mainframe and I'm not afraid to use it"

[Paul Gilmartin]

On 2021-04-19, at 14:07:55, Phil Smith III wrote:
>
> Robin wrote:
>> [... "such" could be anything]
>> Recoveryfrom such is not the responsibility of the program but the application

>> designer. I suggest this is the case in any Abend.
>
> That's silly. First, a friendly error is more useful than "Something ran out
> of space". Second, depending on what the Rexx program was doing, it might
> need to unwind stuff, or at least tell the user how far it got, or that it
> didn't do any damage yet.
>

I sense a clash of cultures. If the C RTL traps S0C4 and reports it
as SIGSEGV, programmers conversant with the former will be dumbfounded
by the latter. Programmers with a broad C experience would be mystified
by S0C4. Should the C RTL display both forms?

-- gil

[Frank Clarke]

That's pretty unfriendly.   They say "never criticise someone who can operate in more than one language".   I would add "especially if you're an American".
The underlying beauty of the English language is that it has built-in auto-correction.  Even if someone is not "native-speaker fluent", they can still be understood.  I'm disappointed that you don't have that capability.   When you develop it, please come back.   Until then, take a hike, asshole.

[Bob Bridges]

I don't disagree about the "gibberish" part, zMan, but be gentle; the name "Joseph Reichman" is a clue that English may not be his first language. He may, I submit, be ~trying~ to construct English sentences, but botching it through mere inability. If so, we might cut him a little slack.

---
Bob Bridges, robhb...@gmail.com, cell 336 382-7313

/* Never make fun of someone who speaks broken English. It means they know another language. -H Jackson Brown Jr */

-----Original Message-----
From: TSO REXX Discussion List <TSO-...@VM.MARIST.EDU> On Behalf Of zMan
Sent: Monday, April 19, 2021 16:21

Is that a haiku? Seriously, please think, use complete sentences, like that. Your post is gibberish.

--- On Mon, Apr 19, 2021 at 4:13 PM Joseph Reichman <reich...@gmail.com> wrote:
> The Rexx program assembles and links
> A program
>
> Blowing of the directory is possible on assemble sysprint and linker
> syslmod
>
> In which can compress and reassemble or relink

[David Spiegel]

"...  the name "Joseph Reichman" is a clue that English may not be his
first language ..."
I wouldn't say that it IS a clue. (It might be a clue.)

> .

[Don Leahy]

With some effort, you may be able to alter your Rexx program to reallocate
and resize your output file on the fly. “All you have to do” is to obtain
the allocation attributes of the file that failed and allocate a new one
under a different name with more space/directory blocks allocated. Then
copy the from the failed file to the new one. And finally, delete the
failed file and rename the new one to the original name. In other words,
you can automate what you would do manually.

[zMan]

Sure, but there's also no punctuation. Pretty sure Germanic languages use
sentences. Just sayin'.

--
zMan -- "I've got a mainframe and I'm not afraid to use it"