> Term::ReadLine allows defining explicit IN and OUT handles which may be redirected to a separate xterm window?

tried this but wasn't sufficiently reliable, (but this couldhave been be caused by gnome-terminal weirdness)

The following hack works with xterm , put it at the very beginning of your program.

It restarts the program within a separate xterm-window but redirects STDERR to the parent process.

This allows running a terminal script from within emacs and should be easily adapted to other environments using l/unix.

All STDERR warnings show within emacs and are clickable to navigate to source file/line.

my $ptty=$ARGV[0];
if ($ENV{INSIDE_EMACS}) {
    my $tty=`tty`;
    delete $ENV{INSIDE_EMACS};
    warn "restart in xterm $0 $tty\n";
    my $x=`xterm -e '/usr/bin/perl $0 $tty'`;
    warn "xterm closed\n";
    exit;
} elsif ($ptty) {
    close STDERR;
    open STDERR,">",$ptty;
    local $\="\n";
    warn "redirecting STDERR to $ptty\n";
    $|=1;
}
[download]

This approach can also be used by the perldebugger when debugging another Term::ReadLine application

When sufficiently tested this logic can be put into a separate module RunXTerm , no adjustments of editor needed.

disclaimer

It's a proof of concept hack!!!

I'm not overly familiar with tty and forking and open for cleaner suggestions. :)