There are several reported bugs that can be fixed by settting
*complexsign* to true; one such bug is SF bug 2358696 "is equal bug."
Alternatively, bug 2358696 could be fixed by setting complexsign to
true, but I think complexsign is deprecated?
Would it be OK to set *complexsign* to true in functions such as
meqp-by-csign and mgrp? The downside is that we might uncover bugs
with the new *complexsign* scheme; the upside is that we might fix
newly discovered bugs with the *complexsign* scheme.
Barton