Re: Weird issue with a display file

On 12-Jan-2012 10:52 , James Lampert wrote:

RPG is library-qualifying the reference to the display file, when
qualification is neither necessary nor desirable. And since the
program in question was compiled on our V4R4 box (with dummy service
programs), using the V4R2 compiler, it's doing it at a level where I
couldn't (without an OVRDBF) qualify a database file if I wanted to.

It makes no sense.

While the RPG compiler may indeed be library-qualifying the "combined" file name for the data provided for the "Where Used" information as presented by the DSPPGMREF [at least as seen for device WORKSTN, from a quick test], the RPG Programmer's Gd suggests very clearly that the run-time is supposed to be searching *LIBL. See topic "Naming Files" in the following links:
IBM i 6.1 Information Center -> Programming -> Programming languages -> RPG -> ILE RPG Programmer's Guide -> Working with Files and Devices -> Defining Files -> Naming Files
http://publib.boulder.ibm.com/infocenter/iseries/v6r1m0/topic/rzasc/sc092507367.htm
http://publib.boulder.ibm.com/iseries/v5r2/ic2924/books/c092507422.htm#Header_313

There is no mention of the existence of the library from the compile as a requirement for that run-time processing. Thus I infer that a failure by the run-time for that reason alone, would be a defect. I can not test [because I can not delete a library], but that should be easy enough to do on a newer release, using none of the newer features like EXTFILE and EXTDESDC, to verify that is the origin for the problem.

IMO the description of the expected effect at run-time also implies that the Where-Used information should probably indicate *LIBL, just as occurs for the typical Input or Output file and what I recall of most device independent file specifications; i.e. to imply, perhaps the 'Device' specification is origin for the behavior.?

Regards, Chuck