🚀 go-pugleaf

Problem: backward-sentence in the scratch buffer misbehaves,
apparently infinitely recursing.  This problem has only been observed
in the scratch buffer.

Affected versions:  This has been observed in both Emacs releases I
have access to - 18.55.3 on bu-pub (.bu.edu), 18.55.9 on 
bucsf (.bu.edu).

Manifestation:
Two different behaviors have been seen, dependent on whether text is
in the buffer or not.

1)  If there is text in the buffer, with blank lines above:

* TOP OF BUFFER *


(A)
text text text.  text text text.
(B)

* END OF BUFFER *

A backward-sentence with point following the first non-blank line (ie,
point B) will work properly.  However, backward-sentence from point A
causes Emacs to "disappear" for a short time, then return the message
"Lisp nesting exceeds max-lisp-eval-depth".  Emacs remains functional,
with no loss of data.

2) if there is no text in the buffer, only blank lines:

* TOP OF BUFFER *





* END OF BUFFER *

A backward-sentence anywhere in the buffer will cause Emacs to
"disappear" indefinitely, and, on a Sun 4, raise the system load by 2.
(It has been timed to 10 minutes.)  Control-G will cause a return,
with the normal "Quit" response.

Resolution:
Unfortunately, I lack the programming expertise to find or rectify the
problem.  The problem can be avoided by not using the *scratch* buffer
for processing, but it managed to confuse me quite badly when I
started up Emacs with an improperly set terminal, and ended up running
backward-sentence in the condition mentioned above.


Bruce Howells,  beh@bu-pub.bu.edu