Re: Re: changing terminal size

[Mike Miller]

On Wed, Jun 24, 2015 at 15:04:30 -0400, John W. Eaton wrote:
> On 06/23/2015 02:02 PM, Francesco Potortì wrote:
>
>> Hm.  Disabling line editing is conceptually different from assuming that
>> you cannot ask for terminal size, as far as I know.  Am I wrong?
>
>
> It is a separate issue, but I think that we have used readline functions to
> query terminal size and the current implementation assumes that those
> functions are not available if you disable line editing.  I guess it should
> really use them if they are available, even if line editing is not enabled.

Yeah, we also use readline for its ability to process events in the
background while waiting for input, so that FLTK plotting doesn't work
when line editing is disabled. Those concepts are even further removed
but currently bound together. *shrug*

I looked at the "comint" package in Emacs, which is what I think is
used to run Octave, and it does already handle setting the COLUMNS
environment variable correctly when a subprocess is started, so Octave
should just need to be taught to override the terminal size from
readline if the COLUMNS and/or LINES variables are present in the
environment. I think this is the easiest solution and worth doing, and
is not just for Emacs.

Now if you also need the number of lines to be set correctly, you
should probably file a bug to add that to the Emacs comint library.

-- 
mike