Condition handling can be quite helpful, especially when I want to report sn error to a PL/I caller.
That said, the default LE handler certainly could put out better messages. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 ________________________________________ From: IBM Mainframe Discussion List [[email protected]] on behalf of David Crayford [[email protected]] Sent: Tuesday, June 8, 2021 12:12 AM To: [email protected] Subject: Re: U 4087 abend On 8/06/2021 9:05 am, Clark Morris wrote: > What I object to is LE obscuring the original abend code with an abend > code that doesn't have a 1 to 1 relation to the original abend. A > U4xxx-S0C7 message would be adequate. LE condition handling is to number 1 reason why assembler programmres hate LE. It's also one of the reasons why it's unsuitable for systems level programming. > In regard to the application > programmer in the original posting, Has he or she looked at all of the > documentation such as CEEDUMP which as I recall has the original abend > code? There is no CEEDUMP with U4037 abends. The condition handler can't produce a stack trace because of recursive condition handling. In my experience this is usually a storage overwrite which has clobbered the save area chains. I advise the OP to turn off condition handling using the 'TRAP(OFF,NOSPIE)' run time option and then diagnose the dump using IPCS or whatever friendlier dump analysis tool such as Fault Analyzer. > Is that programmer knowledgeable enough to know that a S0C7 is > an application error 99.999 percent of the time (there may be a rare > instance where it isn't). ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
