Re: emacsclient's option decoding code

[Eli Zaretskii]

> Date: Mon, 3 Nov 2008 15:45:48 +0100
> From: "Juanma Barranquero" <address@hidden>
> Cc: address@hidden, address@hidden
> 
> On Mon, Nov 3, 2008 at 05:14, Eli Zaretskii <address@hidden> wrote:
> 
> > I will look into it when I have time.  You can file a bug report, if
> > you want to make sure we don't forget this.
> 
> As for the -t crash, on one hand emacsclient on Windows shouldn't ever
> request a tty.

I'm not sure I agree: a native w32 emacsclient could be talking to a
Cygwin Emacs, in which case -t should work.  Also, what about the case
where a native w32 Emacs runs with -nw, and the tty name is CONOUT$ ?

> OTOH, the underlying bug is that, on Windows, calling
> make-terminal-frame with a non-nil tty-type from inside a windowing
> Emacs crashes it.

Yes, and that should be fixed.

> Shouldn't make-terminal-frame on Windows just err out unless Emacs
> is running in -nw mode?

Probably so, but make-terminal-frame should also DTRT or at least fail
gracefully under -nw as well, when called with tty that is not the
default ("CONOUT$").