Re: Menus with more items than the TTY can display

[Eli Zaretskii]

> Date: Thu, 17 Oct 2013 23:06:07 +0300
> From: Eli Zaretskii <address@hidden>
> Cc: address@hidden
> 
> > Date: Thu, 17 Oct 2013 20:01:12 +0200
> > From: martin rudalics <address@hidden>
> > CC: address@hidden
> > 
> >  > And if you _also_ set resize-mini-windows to nil, do the artefacts
> >  > disappear?
> > 
> > Why _also_ ?  The artefacts do not appear when I disallow resizing the
> > echo area - either by setting `resize-mini-windows' to nil or
> > `max-mini-window-height' to 1.  No need to also suppress the help-echo.
> 
> But you say that disabling the help echo exacerbates the problem, so
> maybe it will now appear even when the mini-window is not resized?
> Did you try?

Also, it might be worthwhile to try this in GDB, after starting Emacs:

  (gdb) attach EMACS-PID
  (gdb) break update_frame_1
  (gdb) commands
   > bt
   > continue
   > end
  (gdb) continue

Now open the menu and do whatever you do to trigger the problem.  Then
look at the backtraces.  If the code that calls help-echo display is
ifdef'ed away, the only calls to update_frame_1 while the menu is
active should come from update_frame_with_menu; if it is not ifdef'ed
away, you will also see calls that come via Fmessage->echo_area_display
(and the Lisp backtrace should show they are from tooltip-show-help).
any other call sequences are suspects that I want to know about.