fix #H105 - inconsistent handling of "meditating" monsters (trunk only)

fix #H105 - inconsistent handling of "meditating" monsters (trunk only)

     From a bug report, monsters with the wait
strategy (described as "meditating" by stethoscope probing) could be
affected by music but left meditating.  Various wake up attempts shared
the same situation.  Finish waiting if the monster would have been woken
(or pacified).  I didn't search for places that diddle the msleeping bit
directly instead of calling one of the assorted wake() routines.

     A fair bit of this is making usage of DEADMONSTER() be consistent.
Sooner or later there'll be another monster movement overhaul and those
    if (DEADMONSTER(mon)) continue;
statements will all go away.  (Probably just wishful thinking.)