> A watch guard on the full height parameter would notice the effect ctwm
 > restart has taken place and the value has changed but maybe emacs can't
 > tell the difference between human or environment invoked change on the
 > height parameter.  A heuristic could suggest to emacs ctwm restarted
 > when the bottom edge is raised less than 10 pixels.

I attached a function you can add to 'window-size-change-functions'.  It
seems to work here but make sure to use it only with a Lucid build.

With a GTK build I was able to crash CTWM as

ctwm:  restarting:  ctwm
sh: 1: /usr/X11R7/libexec/ctwm_font_size: not found
sh: 1: /usr/X11R7/libexec/ctwm_app_menu: not found
/usr/bin/m4: syscmd subprocess failed
xterm: fatal IO error 11 (Die Ressource ist zur Zeit nicht verfügbar) or KillClient on X server ":0.0"
xterm: fatal IO error 11 (Die Ressource ist zur Zeit nicht verfügbar) or KillClient on X server ":0.0"
X connection to :0 broken (explicit kill or server shutdown).

 > Why is it that F11 fullscreen maximizes the frame without gap at bottom?
 > When frame-resize-pixelwise is nil, shouldn't the frame's line height
 > effect leave a gap at bottom?

Apparently F11 is not affected by size hint increments.  The X11 manuals are
not clear about this.

martin