Ulrich Drepper <drepper(a)cygnus.com&gt; writes: I was afraid so. My patch makes the race condition less rare, but it doesn't make it go away. This patch will (or should, because I didn't test it) make the crash go away while I ponder on how fix x_any_window_to_frame. If you get the message could please lookup the window id in the output of 'xwininfo -tree' Jan P.S. Greg, your lost selection bug could be related to this. Could you by any change find out the window id of the lost Selection message? Index: event-Xt.c =================================================================== RCS file: /usr/CVSroot/XEmacs/xemacs-20/src/event-Xt.c,v retrieving revision 1.43 diff -u -u -r1.43 event-Xt.c --- event-Xt.c 1999/06/16 20:45:57 1.43 +++ event-Xt.c 1999/08/20 18:59:21 ＠＠ -777,9 +777,22 ＠＠ int bufsiz = sizeof (buffer); Status status; #ifdef XIM_XLIB - XIC xic = FRAME_X_XIC (x_any_window_to_frame - (get_device_from_display (event->display), - event->window)); + XIC xic; + struct frame *f; + + /* Maybe we should convert to a widget and then also try its parent + #### */ + f = x_any_window_to_frame (get_device_from_display (event->display), + event->window); + + if (f) + xic = FRAME_X_XIC (f); + else + { + xic = NULL; + fprintf (STDERR,"Failed to find frame for window %x\n", event->window); + } + #endif /* XIM_XLIB */ #endif /* HAVE_XIM */