From 7c0cd4f1fb4e8bec0abc2237e4088f82f69099e0 Mon Sep 17 00:00:00 2001 From: polos Date: Tue, 28 Feb 2017 22:43:02 +0100 Subject: [PATCH] update Slime docu --- doc/Slime.htm | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/doc/Slime.htm b/doc/Slime.htm index 4dde6b5..7b33bb3 100644 --- a/doc/Slime.htm +++ b/doc/Slime.htm @@ -41,13 +41,14 @@

Loading files

You should always use qload instead of load, if the file contains any EQL code.

-

It's not strictly necessary in every case, but not doing so may result in slow loading (many thread switches, one for every EQL function call), or may not even work (e.g. if the loaded code uses Qt threads internally).

+

It's not strictly necessary in every case, but not doing so may result in slow loading (many thread switches, for each EQL function call), or may not even work (e.g. if the loaded code uses Qt threads internally).


Help

For help see the qapropos, qfind-bound, qproperties and qgui functions.

To kill the swank process (Slime), use function qquit / qq (since quitting Emacs will not kill it).


Notes

+

If you notice slow execution of a function called from the REPL, try to wrap it in qrun/qrun*, as this will reduce eventual, multiple thread switches (for each EQL function call) to a single one.

All EQL functions are wrapped in qrun* (see eql5/slime/thread-safe.lisp), so it's safe to call them either directly from the REPL or using 'eval region' (or from any other ECL thread).

If the Slime REPL hangs, you can simply try to connect again (losing the old connection): no need to restart Lisp.

This Slime mode is both convenient and simple to use, but conses a little more for every EQL function call.