That explains why the output is created. It probably fails when it tries to reopen it.
El vie, 12 sept 2025 a las 2:44, Frank Swarbrick (< [email protected]>) escribió: > If I am recalling correctly (I may not be), I believe SYSMDECK has to be > assigned to an actual dataset because it is used for input and not just > output. > > ________________________________ > From: IBM Mainframe Discussion List <[email protected]> on behalf > of Paul Gilmartin <[email protected]> > Sent: Thursday, September 11, 2025 10:05 AM > To: [email protected] <[email protected]> > Subject: Re: Would you consider this a compiler bug? > > On Thu, 11 Sep 2025 00:44:06 -0400, Phil Smith III wrote: > > ... > >After much tinkering, I found that it didn't like > >//SYSMDECK DD SYSOUT=* > > > >Changing that to > >//SYSMDECK DD UNIT=SYSDA,SPACE=(CYL,(1,1)) > >made it happy. > > > What else did you try? "SYSOUT=*" refers to MSGCLASS, which > may not be meaningful within compilation. "SYSOUT=A"? > > What would DYNALLOC or BPXWDYN accept? > > >Seems like I shouldn't be able to cause an internal compiler error that > trivially, no? Alas, we don't have a "real" IBM customer number, so I can't > report this normally. Just wondered if I was missing something obvious > about this. > > ... > When I was enhancing a FOSS Pascal compiler, any abnormal > termination of the compiler automatically was reportable. > > -- > gil > > ---------------------------------------------------------------------- > 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 > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
