[RP] window in wrong frame?, defframesels

[Nicolas Rachinsky]

Hallo,

I discovered the following strange behaviour, with a dual head setup
(no xinerama).

+---------++---------+
|Screen .0||Screen .1|
+---------++---------+

I start with just one frame per screen, and only two xterms in screen0.

Now I start an xclock in Screen1. 

Now switch back to one of the xterms and kill xclock.

After that ratpoison thinks that the other xterm is in the frame
previously occupied by xclock. It's particularly annoying if the
second screen is smaller or contains small frames, since the window
will be scaled to the frames size.


I think the test "cur->win->scr == s" in group.c:336 is true, since
current_screen() is screen0 which contains the active xterm (and the
second xterm), and not screen1, which contains the frame which will is
emptied.


diff -r -u ratpoison-1.3.0.orig/src/events.c ratpoison-1.3.0/src/events.c
--- ratpoison-1.3.0.orig/src/events.c   Sat Aug 21 01:51:24 2004
+++ ratpoison-1.3.0/src/events.c        Sat Aug 21 01:51:43 2004
@@ -108,6 +108,7 @@
       frame = find_windows_frame (win);
       if (frame)
        {
+         rp_current_screen=win->scr->xine_screen_num;
          cleanup_frame (frame);
          if (frame->number == win->scr->current_frame) set_active_frame 
(frame);
        }

This patch seems to fix the problem, but I'm not sure if it's correct.



defframesels seems to be missing in the manual.

Thanks,
Nicolas