Re: When is %EOF not an %EOF

Booth Martin wrote:
> 
> The only real difference is that you'd choose ReadE and I wouldn't. Let me
> explain why I no longer use READE.
> 

Just a warning.  If you do your own key comparison, you can get false
negatives if you have any of the fancy key features in your file(ABSVAL,
ALTSEQ etc).

The way the RPG III compiler does READE actually just does the
equivalent of an HLL comparison for the keys anyway, so with RPG III you
already suffer from this problem doing a READE.  With RPG III, you could
do a CHAIN and it would find the record, but READE (or the EQ indicator
of SETLL) would not find the record.

The RPG IV compiler gets correct key comparisons for READE and
SETLL-EQ.  But when you compare your own keys, you go back to the
incorrect comparisons.