window-system-version on Windows

window-system-version on Windows

[Juanma Barranquero]

Why is Fwindow_system_version initialized to 1 on dispnew.c

  #ifdef HAVE_NTGUI
    if (!inhibit_window_system)
      {
        Vinitial_window_system = intern ("w32");
        Vwindow_system_version = make_number (1);
        adjust_frame_glyphs_initially ();
        return;
      }
  #endif /* HAVE_NTGUI */

and then unconditionally set to 3

  XSETFASTINT (Vwindow_system_version, 3);

at two places of w32fns.c (x_display_info_for_name and Fx_open_connection)?

Also, what is its meaning on Windows? What does the 3 stands for?

    Juanma

Re: window-system-version on Windows

[Eli Zaretskii]

> Date: Thu, 4 Dec 2008 12:21:41 +0100
> From: "Juanma Barranquero" <lekktu@gmail.com>
> 
> Why is Fwindow_system_version initialized to 1 on dispnew.c
> 
>   #ifdef HAVE_NTGUI
>     if (!inhibit_window_system)
>       {
>         Vinitial_window_system = intern ("w32");
>         Vwindow_system_version = make_number (1);
>         adjust_frame_glyphs_initially ();
>         return;
>       }
>   #endif /* HAVE_NTGUI */
> 
> and then unconditionally set to 3
> 
>   XSETFASTINT (Vwindow_system_version, 3);
> 
> at two places of w32fns.c (x_display_info_for_name and Fx_open_connection)?

Probably a result of patchwork of some kind.

> Also, what is its meaning on Windows? What does the 3 stands for?

The 3rd version of the native Windows GUI for Emacs?  You can see that
it's 11 for the X11 GUI.

IOW, I doubt that there's some deep meaning to this number.

Re: window-system-version on Windows

[Juanma Barranquero]

On Thu, Dec 4, 2008 at 20:51, Eli Zaretskii <eliz@gnu.org> wrote:

> Probably a result of patchwork of some kind.

Would it be OK to initialize it to 3 and remove the XSETFASTINT macros?

> The 3rd version of the native Windows GUI for Emacs?  You can see that
> it's 11 for the X11 GUI.

Yeah, but for example mule-diags concatenates window-system and
window-system-version to form "x11". On Windows it outputs
"Window-system: w323", which is quite meaningless.

    Juanma

Re: window-system-version on Windows

[Eli Zaretskii]

> Date: Thu, 4 Dec 2008 20:58:31 +0100
> From: "Juanma Barranquero" <lekktu@gmail.com>
> Cc: emacs-devel@gnu.org
> 
> On Thu, Dec 4, 2008 at 20:51, Eli Zaretskii <eliz@gnu.org> wrote:
> 
> > Probably a result of patchwork of some kind.
> 
> Would it be OK to initialize it to 3 and remove the XSETFASTINT macros?

Don't see why not.

> > The 3rd version of the native Windows GUI for Emacs?  You can see that
> > it's 11 for the X11 GUI.
> 
> Yeah, but for example mule-diags concatenates window-system and
> window-system-version to form "x11". On Windows it outputs
> "Window-system: w323", which is quite meaningless.

That'd be a bug, I'd say.  How about "x, version 11" instead?

Re: window-system-version on Windows

[Stephen J. Turnbull]

Eli Zaretskii writes:

 > > > The 3rd version of the native Windows GUI for Emacs?  You can
 > > > see that it's 11 for the X11 GUI.
 > > 
 > > Yeah, but for example mule-diags concatenates window-system and
 > > window-system-version to form "x11". On Windows it outputs
 > > "Window-system: w323", which is quite meaningless.
 > 
 > That'd be a bug, I'd say.  How about "x, version 11" instead?

"X11" is the standard way to refer to "The X Window System, version
11".  "x11" case-mangled but interpretable; "x, version 11" causes me
to double-take.

There are political reasons for not using Win32 like everybody else
does, I know, but why let that screw up a perfectly good custom for a
free windowing system?

Also, you phrase "3rd version ... for Emacs" is somewhat ambiguous,
but I'd take that as "the third try at a Windows GUI by the Emacs
project", not as "Microsoft's third version".  OTOH, "X11" most
definitely means "version 11 as defined by the MIT/X Consortium".

Re: window-system-version on Windows

[jasonr]

Quoting "Stephen J. Turnbull" <stephen@xemacs.org>:

> There are political reasons for not using Win32 like everybody else
> does, I know, but why let that screw up a perfectly good custom for a
> free windowing system?

Renaming from w32 to Win32 would not solve this, so I don't know why you even
bring this up.

> Also, you phrase "3rd version ... for Emacs" is somewhat ambiguous,
> but I'd take that as "the third try at a Windows GUI by the Emacs
> project", not as "Microsoft's third version".

I think the 3 refers to the major version of NT 3.1 which is what this port of
Emacs was originally written for. All these hardcoded constants to satisfy old
X specific window system informational functions date from the original port by
Microsoft interns.

Re: window-system-version on Windows

[Juanma Barranquero]

On Fri, Dec 5, 2008 at 04:56, Stephen J. Turnbull <stephen@xemacs.org> wrote:

> "X11" is the standard way to refer to "The X Window System, version
> 11".  "x11" case-mangled but interpretable; "x, version 11" causes me
> to double-take.
>
> There are political reasons for not using Win32 like everybody else
> does, I know, but why let that screw up a perfectly good custom for a
> free windowing system?

A mountain out of a molehill, IMHO. window-system-version is almost
unused. The concatenation of window-system with window-system-version
is only used (in the Emacs sources) in the output of the quite
uncommon mule-diag command. If the user perusing its output does not
understand what "Window system: x, version 11" means, he will have a
hard time grasping the much more obscure info that follows about
fonsets, charsets, etc. On the other hand, the previous "Window
system: w323" was quite unhelpful for Windows users. I though it was
some kind of error and had to go looking at the source to see where it
came from.

One way or the other, this will affect the... what, 1% of users who
will ever do M-x mule-diag? ;-)

    Juanma

Re: window-system-version on Windows

[Juanma Barranquero]

On Fri, Dec 5, 2008 at 05:25,  <jasonr@f2s.com> wrote:

> I think the 3 refers to the major version of NT 3.1 which is what this port of
> Emacs was originally written for.

We should bump it up to 4 now, then. I don't think Emacs on Windows
would run on NT 3.1 now, would it? ;-)

    Juanma

Re: window-system-version on Windows

[Stephen J. Turnbull]

Juanma Barranquero writes:
 > On Fri, Dec 5, 2008 at 04:56, Stephen J. Turnbull <stephen@xemacs.org> wrote:
 > 
 > > "X11" is the standard way to refer to "The X Window System, version
 > > 11".  "x11" case-mangled but interpretable; "x, version 11" causes me
 > > to double-take.
 > 
 > A mountain out of a molehill, IMHO.

Precisely.  Why fix what ain't broke?

Re: window-system-version on Windows

[Juanma Barranquero]

On Fri, Dec 5, 2008 at 11:23, Stephen J. Turnbull <stephen@xemacs.org> wrote:

> Precisely.  Why fix what ain't broke?

I use Windows. "Window system: w323" is broke.

    Juanma

Re: window-system-version on Windows

[Stephen J. Turnbull]

Juanma Barranquero writes:
 > On Fri, Dec 5, 2008 at 11:23, Stephen J. Turnbull <stephen@xemacs.org> wrote:
 > 
 > > Precisely.  Why fix what ain't broke?
 > 
 > I use Windows. "Window system: w323" is broke.

I have no clue where you got the idea that anybody thinks a broken
Windows doesn't need fixing.  Just don't chop a hole in the floor
while you're replacing the pain. ;-)

Re: window-system-version on Windows

[Juanma Barranquero]

On Sat, Dec 6, 2008 at 04:52, Stephen J. Turnbull <stephen@xemacs.org> wrote:

> I have no clue where you got the idea that anybody thinks a broken
> Windows doesn't need fixing.  Just don't chop a hole in the floor
> while you're replacing the pain. ;-)

I'll do, the day I find a system which doesn't cause equivalent paint.
GNU/Linux isn't that system, alas.

But let's not turn this into a political discussion. I fixed a short
message in a little used diagnostic function. There's nothing more to
it.

    Juanma