Re: This is weird: RNX0105 with a statement number that's a comment

On 10/12/15 11:56 AM, James H. H. Lampert wrote:

I'm looking at an RNX0105 ("A character representation of a numeric
value is in error.") that gives a statement number (8218) that appears
to be a comment in some /FREE code.

. . .

As if that weren't weird enough, I'm also seeing multiple RNX0103s ("The
target for a numeric operation is too small to hold the result.") giving
a statement number that shows as an OPEN statement (also in /FREE code)
on the compilation listing.

I've been asked to provide some more details about the environment. It's V7R1 across the board, and there doesn't appear to be any SQL preprocessing involved.

And I'm definitely looking at the compilation listing spool file. Not from the same compilation as the live program, but the source dates are the same, and the only discrepancy I see in DSPPGM is a 16-byte difference in the static storage size (for which I can find no reason).

Also from the DSPPGM, I see:

Maximum optimization level . . . . . . . . . . : *FULL
Debug data . . . . . . . . . . . . . . . . . . : *YES

and they look the same if I STRDBG them (but the statement numbers in STRDBG end long before the statement numbers cited above, so they're no help).

--
JHHL