Re: [PATCH] exit function

Hi. First on the code. The patch should look like:
+  (define (exit action)
+    (when (eq action 'quit)
+       (quit))
+    (call-hook 'before-exit-hook)
+    (map-windows delete-window)
+    (case ...
Let me explain. 

before-exit-hook shouldn't be called for quit, or it will be called
twice. (The call is done by rep, and not coded in Sawfish.)

(window-history-save) shouldn't be called unless the user do
(require sawfish.wm.ext.window-history). And if they do require,
it is already hooked into before-exit-hook, so there's no need for its
call from (exit action).

It means (exit action) does not add anything for quit, so it can be
removed? (See also my vote at the end of this message.)

But I'm not sure if it's correct altogether. Look at main(argc, argv):
 rc = rep_top_level_exit ();

  /* call all exit funcs... */
  server_kill ();
  functions_kill ();
  windows_kill ();
  frames_kill ();
  cursors_kill ();
  fonts_kill ();
  colors_kill ();
  images_kill ();
  events_kill ();
  session_kill ();
(exit action) skips all of them, but is it ok? My guess is yes;
they're needed if X server remains to run, but I haven't read the
code, so can't say anything for sure.

On Sat, 28 Nov 2009 14:52:26 +0100, Christopher Roy Bratusek wrote:
>> Sorry, I wanted to say (defgroup exit "Exit Sawfish" :group misc).
> Nope. It uses "External Applications" from sawfish.wm.commands.launcher

Commands for term and browser are set in "external applications", since
there's no other way to name them. But reboot/shutdown/suspend can
be called 'exit' (or exit-session) , so it's natural.

>> >> * Func & command 'exit' is so confusing with 'quit'. But what'll be
>> >>   good? 
>> [...]
> [] exit
> [] safe-quit
> [] exit-session

If quit is to be dropped, then exit-session.

Teika (Teika kazura)

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]