Re: Multiple RCVJRNE firing

On 26 Oct 2012 16:04, Thomas Garvey wrote:

I'm trying to determine if there is a parameter or two I might be
missing on the RCVJRNE command, or maybe just don't fully understand
it.

I'm using the RCVJRNE command to receive journal entries for a
physical file. Then set up another RCVJRNE to do the same on a
different physical file, calling a different exit program. When a
change happens to one of the files and a journal entry is placed in
the file's journal receiver, BOTH exit programs actually execute.
I've proved this using debug.

The RCVJRNE command parameters specifies the file name for which the
exit program should be receiving entries and the exit program is
indeed only getting entries for the specified file. The exit program
for the other file ALSO executes at the same time but simply gets the
parameter values passed to it that convey nothing is being passed.
Again, I've isolated this using debug. The only common denominator
is that both files are journaled to the same journal.

This is certainly non-intuitive, given that the files desired are
specified in the RCVJRNE command execution. It seems to imply that a
single RCVJRNE should be applied to a journal and the exit program
should decide what to do with the journal entry based on the file for
which the entry is. That is, check the values in the incoming parms
to determine which file the entry actually pertains, then perform the
function necessary and intended for that file. In my case, the
purposes are different.

I guess I'm saying that the RCVJRNE command merely sets up a 'chain'
of exit programs to call whenever a journal entry is written to the
named journal. It doesn't really restrict calls to the exit program
based on the file designations given in the RCVJRNE command.

Anyone else run into this?

No, I have not experienced that. Some time ago I wrote and used some code implementing a "subscription" to a named file associated with a journal, using the same program in different jobs. Each job received only entries about the one file against which the RCVJRNE was performed. That code actually ensured one specific journal name was associated with each chosen file, so the scenario seems effectively identical to that in the quoted message. As I understood, the OS journaling code would "filter" the invocations of the exit program based on the specifications of various parameters on the RCVJRNE command invocation; i.e. that the RCVJRNE establishes an effective "journal entry received" event monitor against the named journal along with identifying the OS journal program as the event handler, such that when the event occurred, the OS journal program determined if the entry that was being notified met the criteria [and restrictions] for the active RCVJRNE request and therefore invoked the exit program only when the criteria were met.