With the fi, or workaround, for the narrowing problem, in place, the
issue I reported regarding the emacs buffer being messed up
(apologies for my lack of technical precision there! The buffer was
being truncated to almost half the length in a way quite hard to
figure out from where to where, and the point placed at the end, not
on the object in question. The file had to be reloaded.) appears to
have gone away. Time will tell, but it was happening every six or
seven edits, and today in a heavy score session it has not occurred.