[slime-devel] slime-after-change-function performance issue
On XEmacs, this function appears to cause serious thrashing for me. When I take it out of after-change-functions, the problem goes away. -russ
* Russell McManus [2005-09-20 12:37+0200] writes:
On XEmacs, this function appears to cause serious thrashing for me. When I take it out of after-change-functions, the problem goes away.
XEmacs has global after-change-functions by default, ie. the functions are run for each buffer. I guess that's a little bit too often for our purposes. They're now buffer-local in the repl buffer. I still think that the after-change-functions are quite an ugly way to achieve something and should be replaced with something less intrusive. Helmut.
Helmut Eller <heller@common-lisp.net> writes:
* Russell McManus [2005-09-20 12:37+0200] writes:
On XEmacs, this function appears to cause serious thrashing for me. When I take it out of after-change-functions, the problem goes away.
XEmacs has global after-change-functions by default, ie. the functions are run for each buffer. I guess that's a little bit too often for our purposes.
They're now buffer-local in the repl buffer.
I still think that the after-change-functions are quite an ugly way to achieve something and should be replaced with something less intrusive.
For XEmacs, a much more elegant implementation using "extents" should be possible. For Emacs, I don't see a better way than after-change-functions, however. -- Matthias Köppe -- http://www.math.uni-magdeburg.de/~mkoeppe
participants (3)
-
Helmut Eller -
Matthias Koeppe -
Russell McManus