Old redisplay bug in xdisp.c, question about change

["Jan D." <jan.h.d@swipnet.se>]

Hello.

I was looking for a redisplay bug I thought was in the GTK parts.  It turns
out that the bug is present in both the Motif and Lucid ports also.

To reproduce.
emacs -q --no-site-file
C-h i
Select a node with the mouse.

Sometimes the bug can be seen right here.  The tool bar arrows for up
and next are still disabled.  If they are disabled, do a Ctrl-l to fix.

Now press 'u' to go up.  The next and up tool bar arrows are now enabled,
they shall be disabled.  Again, Ctrl-l fixes it.

There seems to be a bit of race condition here, the bug does not always
show up.

I traced the origin to this change:

2002-03-26  Richard M. Stallman  <rms@gnu.org>

        * xdisp.c (update_menu_bar): Test only update_mode_lines;

The diff:

@@ -7435,8 +7438,14 @@
   window = FRAME_SELECTED_WINDOW (f);
   w = XWINDOW (window);
   
+#if 0 /* The if statement below this if statement used to include the
+         condition !NILP (w->update_mode_line), rather than using
+         update_mode_lines directly, and this if statement may have
+         been added to make that condition work.  Now the if
+         statement below matches its comment, this isn't needed.  */  
   if (update_mode_lines)
     w->update_mode_line = Qt;
+#endif
 
   if (FRAME_WINDOW_P (f)
       ?
@@ -7455,7 +7464,9 @@
         the rest of the redisplay algorithm is about the same as
         windows_or_buffers_changed anyway.  */
       if (windows_or_buffers_changed
-         || !NILP (w->update_mode_line)
+         /* This used to test w->update_mode_line, but we believe
+            there is no need to recompute the menu in that case.  */
+         || update_mode_lines
          || ((BUF_SAVE_MODIFF (XBUFFER (w->buffer))
               < BUF_MODIFF (XBUFFER (w->buffer)))
              != !NILP (w->last_had_star))

The comments indicate that the menu does not need to be updated, but
apparently the tool bar is not updated either, because w->update_mode_line
is not set to Qt.

There seems to be no ill effects from reverting this change, but since I
don't know why it was made in the first place, I can not be sure.
Does anybody remember the reason for this change?

Thanks,

	Jan D.