Re: core dump when window is too short; infinite loop when too narrow

[Algorithm Petrofsky]

> From: Eli Zaretskii <eliz@is.elta.co.il>
> On Wed, 16 Jan 2002, Algorithm Petrofsky wrote:
> 
> > -- Create a window with height two, but zero usable lines because
> >    mode-line-format and header-line-format are both non-nil:
> > 
> >      emacs -q
> >      C-u C-x 2 C-h i C-x o M-2 C-x ^ M-C-v
> > 
> >    Symptom: Fatal error (6).Aborted (core dumped)
> 
> Could you please try the pretest of Emacs 21.2, ... 21.1.80

I get the same symptom.  Is there are any version that passes this
test on a tty?

By the way, I would like to request that one-line and two-line windows
be supported, with the semantics that mode-lines and header-lines are
not displayed on one-line windows, and header-lines are not displayed
on two-line windows if there is a mode-line.  These rules would still
ensure that there is at least one line of buffer display.  I think
these rules are preferable to disallowing windows shorter than three
even when they are not trying to use a mode-line or header-line.

An application I have in mind is an emulator of an Infocom Z-machine's
display (see malyon.el), for which the natural implementation is two
windows with no mode-line or header-line between them.  The size of
the upper window is often only one or two lines.

-al

P.S.  I'm talking about absolute minimums here.  I'm not suggesting
changing the default values of window-min-height and window-min-width
(which are four and ten).