Let's dispose of two cases:

(1) =0C'a' is an illegal literal; zero lengths are invalid.

(2) HLASM shouldn't pad the *contents* of literals, because the
program might refer to their length attribute (say, for a move)
and moving a different number of bytes than were in the original
literal could cause program failures.

Padding within the literal pool itself is a different matter,
of course.

John Ehrman

Reply via email to