This works similarly to the previous command, except that it performs matching
with extension for those theories that need it (currently just those including
the assoc attribute). If the subject (after theory-normalization)
has a symbol f from an extension theory on top, only a piece of the top
theory layer with f on top need be matched. This kind of matching is used
by the engine for applying equations and rules in order to accurately simulate
congruence class rewriting. The result is a list of all matches. If only part
of the subject was matched, that part is given. Example:
xmatch +(*(P, Q), *(X, Y)) <=? +(*(a, b), *(c, d), *(a, e)) .