Re: twm? How to test [was: Re: Various comments, mostly on ImplementationNotes]

[Nathan Clemons <nathan windsofstorm net>]

On 18 Aug 2000, Owen Taylor wrote:
> 
> Matthias Ettrich <ettrich@trolltech.com> writes:
> 
> > 
> > now, what _is_ again the compliant behaviour? I thought we agreed that
> > x,y configure requests means (at least in NorthWestGravity) the top left corner
> > of the window frame.
> > 
> > Oven, please clearify.
> 
> [ Actually, that's "Owen". ]
> 
> According to the ICCCM, for NorthWestGravity, the x, y referse to
> the top-left corner of the window manager frame.
> 
> xv is broken in rather strange ways. That is, it works fine with
> twm, which is ICCCM compliant, and also works fine with fvwm,
> which is not ICCCM compliant. But it doesn't work with old versions
> of sawfish, which were not ICCCM compliant.
> 
> My best guess is that it expects ICCCM behavior, but it's
> algorithm for finding out with the top-left corner of the window
> manager decoration is is broken in such a way that it works
> with some non-ICCCM compliant window managers. 
> 
> Regards,
>                                         Owen
> 

In reference to this (above post) and the following post:

On Fri, 18 Aug 2000, Julian Adams wrote:                                                            
>                                                                                                   
> It would be a disaster if afterstep, kwin, sawfish, blackbox etc.                                 
> implemented differing versions of the spec, but all posted the same                               
> _NET_SUPPORTED hints.                                                                             
>                                                                                                   

Something to consider would be for someone to sit down and write a testing
application (or several). It would walk through switching gravities,
handling window movement/resizing, etc., and making sure that everything
is compliant according to the ICCCM and this spec. 

I think it would have to be written in raw X-based code, to be portable,
unbiased, and non-reliant upon any toolkit (Qt or Gtk, etc.). This would
gain the benefit of making sure that all "released as supporting the new
spec" WMs match in behavior. It would also gain the benefit of allowing
new WMs (or DEs, for that matter) to be able to implement the spec and
then verify themselves as compliant.

Normally, I'd offer to do this, but my time and X-based knowledge limits
to a prohibitive degree. This is by far an in-depth "grokking" of the
workings of X.

Side note: does anyone happen to have the current spec in SGML form? I
keep getting emails every now and then about the section on
developer.gnome.org being broken (the way SGML parses into HTML files was
never too friendly with CVS when I was working with it), and I'd like to
fix it. I can't seem to find my older copy, and I'm thinking it was lost
in a restructing of my home network and a failing drive with home
directories (summer in an apartment building didn't play nicely with an
older 500M Maxtor SCSI drive)...

-- 
Nathan P. Clemons                       "Peace favor your code."
nathan@windsofstorm.net                 ICQ: 2810688
(v) 401.725.6061			(f) 603.372.9737