Xref: utzoo gnu.emacs.help:807 comp.emacs:9879
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!UUNET.UU.NET!kyle
From: kyle@UUNET.UU.NET (Kyle Jones)
Newsgroups: gnu.emacs.help,comp.emacs
Subject: bug in Freplace_buffer_in_windows (really window_loop)
Message-ID: <9101171848.AA13535@wendy-fate.UU.NET>
Date: 17 Jan 91 18:48:22 GMT
Sender: daemon@tut.cis.ohio-state.edu
Followup-To: gnu.emacs.help
Organization: Gatewayed from the GNU Project mailing list help-gnu-emacs@prep.ai.mit.edu
Lines: 34

GNU Emacs version 18.55:

In src/window.c, inside the definition of window_loop, at the
part of the main switch statement that Freplace_buffer_in_windows
uses, we see

	  case 6:
	    if (EQ (p->buffer, obj))
	      {
		tem = Fother_buffer (obj);
		if (NULL (tem))
		  tem = Fget_buffer_create (build_string ("*scratch*"));
		Fset_window_buffer (w, tem);
		Fset_buffer (p->buffer);
	      }
	    break;

I think the call to Fset_buffer is a bug.  Why does the current
buffer need to be changed here?

To illustrate the buggy behavior, consider this code:

(delete-other-windows)
(switch-to-buffer "aaa")
(split-window)
(switch-to-buffer "bbb")
(kill-buffer "aaa")
(message "%s" (current-buffer))

At this point the current buffer should still be "bbb", but it is
not.  If you put this code into a file and run "emacs -q -l <file>",
you'll see #<buffer *scratch*> in the echo area.  The bug nails
Fkill_buffer because it calls Freplace_buffer_in_windows which in
turn calls window_loop.