On my system (Windows XP SP3, 32 Bit, JRE 1.6.0_18, JDK 1.6.0_12) it
can be reproduced as follows:
1. Install Clojure Box 1.1.0 (January 5, 2010)
2. Start the Slime REPL and enter these lines:
(use 'clojure.contrib.repl-utils)
(javadoc "snafu")
Result: The Slime REPL stalls and doesn't return a result.
4. Visit the *inferior-lisp* REPL buffer and place your cursor at the
command line. Press 'return'.
Result: The Slime thread continues and opens a Javadoc browser window.
The Slime REPL returns.
When the above lines are entered directly into the *inferior-lisp*
buffer they get executed promptly and without errors.
I haven't noticed any such problems on my Mac and Linux boxes.
On Windows, I keep some Elisp around to swiftly revive a stuck Slime
session:
(defun slime-poke-clojure ()
(interactive)
(slime-send-to-process "\n"))
(defun slime-send-to-process (command)
(interactive)
(let ((slime (slime-inferior-process)))
(if slime (comint-redirect-send-command-to-process
command (process-buffer slime) slime nil t)
(message "No Slime process running."))))
Sorry, I don't have a Windows machine available to debug this problem.
Could you report it on the swank-clojure issue tracker so we can have
a record of it?
http://github.com/technomancy/swank-clojure/issues
Thanks,
Phil
Thanks,
Rickard