Jerry James wrote:
I can reproduce it with XEmacs 21.2.32. It has all the earmarks of a
timing or synchronization error. If I slow the process down at all
(e.g., by making some of the functions involved interpreted, rather than
byte-compiled), it works properly. When the process is "fast" (i.e.,
everything is byte-compiled and there are no delay loops), the bad
behavior appears.
I know nothing about the XEmacs event code, so I will have to let more
experienced hands debug this the rest of the way.
And, hey, in spite of all the traffic about psgml, so far we have
uncovered only 1 actual bug, and it had a 1-line fix. Not bad, eh?
--
Jerry James, trying not to break his arm while patting himself on the
back
Jerry,
I am relieved that you have been able to reproduce this. I have in the
meantime compiled the latest 21.1 and there is, as you note, no problem.
No, it's not bad at all.
Peter
--
__ /__ Peter B. West pbwest(a)netscape.net
/
http://www.powerup.com.au/~pbwest
/ "Lord, to whom shall we go?"