One of the recent changes to 'rcp-wait-for-output' inserted a one second
delay to most of the remote operations.
The specific change was a call to 'accept-process-output' with a one
second timeout. Since the purpose of this call is to, as far as I can
tell, flush any pending but unprocessed input into emacs, the wait will
usually terminate with a timeout.
As such, a one second delay was noticeable (+ 3 seconds) when doing
filename completion.
The attached patch, good for all emacsen AFAIK, reduces the wait to
10ms, a much less noticeable figure.
ChangeLog entry:
1999-10-08 Daniel Pittman <[EMAIL PROTECTED]>
* rcp.el (rcp-wait-for-output): Wait 10ms rather than 1s for
output that is already waiting. Longer delays are quite noticible
where the typical case is to exit via timeout.
Daniel
--- rcp-1.155.el Fri Oct 8 12:19:26 1999
+++ rcp.el Fri Oct 8 12:50:16 1999
@@ -2332,7 +2332,11 @@
(let ((proc (get-buffer-process (current-buffer)))
(result nil))
;; In case output is already waiting, retrieve it.
- (while (accept-process-output proc 1))
+ ;;
+ ;; A one second delay made completion painfully slow, especially
+ ;; as there should never be output waiting (I think :)
+ ;; Daniel Pittamn <[EMAIL PROTECTED]>
+ (while (accept-process-output proc 0 10))
(process-send-string proc
(format "echo %s%s"
rcp-end-of-output
--
I was part of that strange race of people aptly described as spending their
lives doing things they detest to make money they don't want to buy things
they don't need to impress people they dislike.
-- Emile Henry Gauvreay
