Re: Varying length subfield contains blanks instead of length 0 ?

On 05-May-2014 13:41 -0500, Koester, Michael wrote:

In this particular case, I'm working with an array DS. I'm finding
that even when I initialize the DS, the "empty" subfields contain all
x'4040'. And when I test for *blanks, or for %len(...) = 0, I get
blown out by the index exceeding the range (its dim(32) is a bit
short of 4040).

Perhaps sharing the declarative might clarify. The previously recommended initialization using INZ, AFaIK, should establish a zero-length "varying" value for each element of the DS array. That has always functioned well for me; again, AFaIK.

Looks like a MONITOR block is my only [albeit not especially elegant]
way out. Gotta love it.

If there is difficulty with the RPG default initialization phase, IMO a much better resolution than monitoring for the 115 error [RNX0115 "Length of varying length variable is out of range"], would be to effect an explicit user-coded initialization phase; at least with such a small DIM().

Another option would be to test the length, without the error and thus without a monitor, having defined an OVERLAY() to the length portion of the "varying" data; i.e. comparing that variable to the decimal value 16448 [0x4040].