Re: emacsclient's option decoding code

[Eli Zaretskii]

> Date: Wed, 12 Nov 2008 22:05:11 +0100
> From: "Juanma Barranquero" <address@hidden>
> Cc: address@hidden, address@hidden
> 
>       (t
>        (unless (string-match-p "\\`[^:]*:[0-9]+\\(\\.[0-9]+\\)?\\'" display)
>          (error "Invalid display, not HOST:SERVER or HOST:SERVER.SCREEN"))
>        (when (and (boundp 'x-initialized) (not x-initialized))
>          (setq x-display-name display)
>          (x-initialize-window-system))
>        (make-frame `((window-system . x)
>                      (display . ,display) . ,parameters)))))
> 
> and it would fail for Windows. At some point or other, you've gonna
> add a check for Windows, because, really, ":0.0" has no meaning right
> now.

I meant for it to be checked inside a w32 specific code of make-frame.

> I agree.  But the things we've been discussing until now aren't more
> uniform IMO; they just move the Windows specific code around or
> introduce ad-hoc hacks to force Windows frames to simulate being what
> they are not...

I suggested a different approach, but it sounds like it's being
rejected.