🚀 go-pugleaf

Hardware: Sun-4c/25
OS:       SunOS 4.1.1
epoch:    4.0p1
X lib:	  X11R5

The following behavior is probably not desirable

	epoch -nm	# attached minibuffer

In the interactive lisp *scratch* buffer evaluate the following:

	(progn
	  (duplicate-screen)
	  (epoch::unmap-screen (current-screen))
	  (remove-screen)
	)

now epoch is still running but with no way for the user to send input
to it.

In screen.c, in the funciton delete_screen(), a check is made to see
if the user is attempting to delete the only edit screen. However, the
check should probably only count screens that are mapped OR it should
force one of the unmapped screens to be mapped.

Failing this, it might be reasonable that a kill -HUP signal sent to
an epoch process with no mapped screens should cause one (or more) of
the unmapped screens to be mapped.

	-- Mark

PS: Is there any easy way to determine if a screen is an icon rather
than being unmapped?