This is fixed, see
http://mid.gmane.org/20223.7965.212018.49660@parhasard.net and friends.

I’ve also seen something related to this, I think. See the
comments in #'easy-menu-add, 

http://thread.gmane.org/gmane.emacs.xemacs.beta/21654/focus=21736

and the thread starting with

http://mid.gmane.org/17337.50619.323602.622218@parhasard.net

Happily it only seems to be relevant in code run immediately
after startup.

In the process of producing the patch for the isearch crash recently
discussed, I produced a test of the form

(with-temp-buffer
  (execute-kbd-macro vector-of-keystrokes))

I saved it to a file, and ran xemacs -vanilla, then M-x load-file RET
search-crash.el RET, and XEmacs didn't crash.  The reason was that
instead of being executed in the temp buffer, it was executed in
*scratch*, which contained the usual leading comment, so the buffer
underrun didn't happen.

Is this expected behavior?

I'll try to remember to produce a real test case tomorrow, but I need
some sleep now....

_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta@xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta