Re: Possible intereting SETLL behavior

Yeah, I wonder a little, too - now there is the use of *LOVAL, as well. This sets each of the key fields to its respective *LOVAL, X‘99...9D’ for numeric, basically, and blanks for alpha. Not used with RRNs, methinks.

*START was, so far as I know, a pretty late addition - came out at 7.1, I believe. I don't recall anything about using a value for this, rather, that the file is positioned "...to the beginning of the file...", with no regard for actual values. Paraphrased is that, eh?

On 12/8/2015 11:02 AM, CRPence wrote:

On 08-Dec-2015 07:09 -0600, Vernon Hamberg wrote:

<<SNIP>> I tried this -

setll 0 filename;

Now %equal(filename) is, again, *OFF - there is
no record whose RRN = 0.

But %found(filename) is *OFF, which seems odd, because the following
description of SETLL is matched -

The SETLL operation positions a file at the next record
that has a key or relative record number that is greater
than or equal to the search argument

I also did this using fixed-format and checked the NR indicator,
which is *ON. Because the NR is *ON (%found() is *OFF), we are
positioned (see replies to my recent post) at end-of-file, so a READ
turns on EOF.

The doc says this - "If access is by relative record number,
search-arg must be an integer literal or a numeric field with zero
decimal positions." It doesn't say, for RRN, that the search-arg has
to be > 0.

I'm working on 7.1 of the OS.

Seems to me that there ARE records found when using 0 as the search
argument, so the NR indicator should be *OFF - is this a special
case?

Or have I missed something again?

Possibly, like with Command (CMD) Parameter (PARM) definitions, a Special Value is implemented using the same Data Type as the argument. As such, I wonder if the effect of using a zero-value matches the special-value of *START; i.e. *START effects passing a zero-value to the run-time I\O program. And if the effects match between use of zero and use of *START, then is there perhaps mention [or allusion to] in the docs that the use of *START does not effect the condition of /found=*on/ [as with actual /counting/ integer specifications]? I do not ever recall testing any effects from SETLL *START, instead, only testing the effects of the READ or READE that follows that SETLL.