Re: Test program calls puts(), creates spool file here, but not on customer box

On 02-Dec-2011 09:56 , James Lampert wrote:

I noticed that on the system in question, QGPL is not in the default
library list (which seems even more insane than not having QTEMP in
the default library list!),

Same on the system I used to test.

Nothing insane about QGPL not being in the library list. What more closely resembles insane IMO, is an assumption by code that QGPL [or QTEMP] will be in the library list. A library not-so-affectionately referred to as the "Garbage Pile Library" [versus the "General Purpose Library" as origin for its moniker] sounds like something that is best omitted from the library list ;-)

but there is a QPRINT file in QGPL.

Hmmm. I wonder if that is the origin of the "problem" with the QC2IO; i.e. a failed assumption that a QPRINT will exist in *LIBL, and if not found, reacts by CRTSRCPF QTEMP/QPRINT. A trace could show if that is the case.

I also noticed that, from a terminal session on the box in question,
if I do a host screen print, it doesn't produce a spool file,
either, and if I look at the spool files for my job, I don't see
anything, "FIN" or otherwise.

FWiW that was not a symptom for me on the system where QGPL was not in *LIBL. The spooled host print utilized *LIBL/QSYSPRT.

A trace would show if any file was opened, and probably any error on the original open. However any error should have effected "Error occurred during print operation." instead of "Print operation complete to the default printer device file."; no mention of which was received.?

Curiouser and curiouser.

The first thing I look for when data does not appear where I expect is to review overrides. And active override for QSYSPRT TOFILE(QPRINT)

Might be worth taking care to note the differences between when a *FILE named QPRINT exists in *LIBL. For example, if the "puts()" previously created a QTEMP/QPRINT, the results of the Host Print request might be different than when no QTEMP/QPRINT exists; though the "no output" effect might have been instead, the data from spool request having been deposited in the database file instead.

Regards, Chuck