There are certainly cases where a base-displacement operand on
USING would appear at first glance to make sense, as in the
suggested case:
USING SOMELABEL,24(R3)
or with a symbolic offset expression
USING SOMELABEL,TWENTY_FOUR(R3)
However, even in concept some of the implications are not
obvious. This would obviously be expected to work like a
dependent USING, but it does not have a base USING giving the
list of registers to be used, so it is not the same. One might
expect it to act as if there was an implicit base USING for R3
pointing 24 bytes earlier, but without a label, which would be a
new concept. But does that mean one could issue a DROP R3 and
that R3 would be currently considered in use as a base so that a
new USING for R3 would drop this use? At least since the
language enhancement for PH42188 in 2021 to allow DROP for a
relocatable symbol, one would be able to use DROP SOMELABEL.
These are examples of the sorts of issues that arise when we
investigate possible language enhancements. If there is a
strong justification for some enhancement, it may be worth
working through such issues and finding solutions. However, as
far as I can see, the current suggestion would appear to require
new concepts and raise complications which appear to be out of
proportion to the potential benefits.
Jonathan Scott, HLASM
IBM Hursley, UK
