Re: xclock -borderwidth 0?

[john cs york ac uk]

On 11 Dec, Owen Taylor wrote:
> 
> john cs york ac uk writes:
> 
>> Is there any explanation in any of the Gnome documentation of why Unix
>> commands like `xclock -borderwidth 0' don't result in a command
>> window without a window border? (Is this instead a matter of the window
>> manager in use? I've tried Enlightenment, Sawfish, Twm and Window Maker.)
>> A user here has an old self-written program that he wants to appear without
>> a border, for unintelligible reasons of his own. He complains that
>> `-borderwidth 0' used to work on the SGI Indy Magic desktop he used to
>> use before being switched to Linux and Gnome. (He is still using Gnome
>> 1.4, but hopes to switch to 2.0 shortly.) I've searched the Gnome
>> documentation without success.
> 
> The core X idea of window borders is an old concept that dates
> from the era when it was too expensive for window managers to
> draw borders on windows. (Or at least, when people weren't sure
> if they would.)
> 
> So, X has a feature where windows can specify a border that
> the window manager draws in a different color from the main
> window.
> 
> 99% of all windows will have a border width of 0, since the app
> assumes the window manager will draw something. For the 1% that
> do have a border width set, modern window managers will simply
> ignore that seting.
> 
> There probably was some magic going on either in the window 
> manager or the toolkit to turn a -borderwith 0 option into
> a window without a window manager frame, but I can't really
> guess what that magic woudl be.
> 
> Regards,
>                                         Owen

Again, many thanks. (Havoc Pennington has just cleared up a confusion I
had re. `border' and `window frame'.) The user wanted to write a shell
script that marshalled various pre-existing X Windows applications
as window-frame-less pseudo-child windows of his own application
using --geometry options to position them. I pointed out that the fact
that they wouldn't be genuine child windows would spoil the effect
he desired; that the `children' wouldn't move when the `parent' was
moved. I'll suggest that he do things properly.
-- 

John A. Murdie
Experimental Officer (Software)
Department of Computer Science
University of York
England