Re: Weird compiler message and failure

On 03-Dec-2014 09:13 -0600, Thomas Garvey wrote:

On 12/2/2014 3:58 PM, James H. H. Lampert wrote:

On 12/2/14 1:06 PM, Thomas Garvey wrote:

<<SNIP>>
A program generates RPGLE source and compiles the source into a
program (using CRTBNDRPG).
The compile is failing with message ID RNS9328, 'The INFOSTMF
parameter must be specified when PGMINFO is not *NO.'

However, the PGMINFO parameter _is_ *NO and the INFOSTMF
parameter is blank. The source is being compiled from a *SRC
source file member, NOT a stream file.
<<SNIP>>

Is there (and I haven't bothered to check whether there is such a
thing) an H-spec overriding the PGMINFO parameter?

FWiW, there is:

<http://www.ibm.com/support/knowledgecenter/api/content/ssw_ibm_i_71/rzasd/sc092508389.htm>
"PGMINFO(*PCML | *NO { : *MODULE } )

The PGMINFO keyword specifies how program-interface information is to be generated for the module or program.
..."

PGMINFO is something I've only used in connection with web
services and the like.

Nothing in H spec. I think the failure must be happening before the
source member is even opened, right? It would seem to be a failure
at the command processing level, no?

Yet the subject does say "compiler message and failure". And the explanation says "compile fails" vs 'command invocation fails'.

Seems from those comments then, that the msg RNS9328 is *not* being logged in the joblog; thus not being issued from the Command Analyzer and followed by a CPF0001 [if not a prompted command request]? If not logged as a message for the [un]prompted command, then probably the compiler issues the message and logs that in the compiler listing vs the joblog?

If the message is issued to the joblog, then best to provide the full context for the failure; provide the spooled joblog taken with both LOG(4 0 *SECLVL) LOGCLPGM(*YES) so as to include the Request message for the failing request [the CRTBNDRPG with all specified parameters] and all messages up to the next Request message. Presumably if the command parameter-checking failed, then the diagnostic error was the msg RNS9382 issued F/QCAIFLD followed by the Escape msg CPF0001 issued F/QCADRV2 to indicate the command invocation failed. If the command request was made without being library-qualified and a value for the PGMINFO() was not made explicitly, then checking for alternate\duplicate commands that have been customized to 'PGMINFO(*PCML)' might assist.