Re: Still need a hint for undecorated windows

[Carsten Haitzler (The Rasterman) <raster rasterman com>]

On Mon, 20 Jun 2005 20:08:38 -0400 Adam Jackson <ajax nwnk net> babbled:

> On Monday 20 June 2005 19:30, Owen Taylor wrote:
> > On Mon, 2005-06-20 at 18:32 -0400, Adam Jackson wrote:
> > > I've seen a few comments about this in the archives, but no motion on it.
> > > Apps still need a way to draw themselves undecorated, and right now there
> > > are two ways to do it while still being managed.  The first is mwm hints,
> > > which have no formal blessing by netwm; and the other is setting yourself
> > > to TYPE_DOCK, which works on every wm I've tested but is semantically
> > > bogus.
> > >
> > > So I guess the question is whether to make this its own atom, or a new
> > > value for _NET_WM_WINDOW_TYPE.
> >
> > What's the *semantic* reason? (If "dock" is wrong, then you must
> > have an idea :-)
> 
> Well _NET_WM_WINDOW_TYPE_DOCK is wrong because xmms is not a dock, and because
> 
> in most wms it has a side effect on stacking order.  Likewise floating pagers 
> or Dashboard-style applets are not docks.  Docks are the things that stick to 
> the edge of your screen undecorated, motionless, and usually in a higher 
> layer.  The way the spec is written now, the TYPE is a description of the 
> function of the window.  So I don't think it's valid to add TYPE_FRAMELESS, 
> because that's not a class of functionality, that's a decoration style.
> 
> The spec does mention the mwm hints, but does not say the wm should honor 
> them, and in fact makes the claim that WINDOW_TYPE is the intended 
> replacement for the motif hints.  So a client that strictly conforms to netwm 
> cannot be drawn undecorated without side effects, and a wm that strictly 
> conforms has no standard way of providing undecorated windows.
> 
> The alternative, of course, is just documenting that conforming wms must honor
> 
> at least some subset of the motif decoration hints.

i would say this is a case for "if the type and mwm decorations are not mutually
exclusive in their hints, then use mwm for decoration hinting and type to
indicate everything else, otherwise window type takes precedence". as for being
mutually exclusive - this is up to the wm to decide based on its idea of what to
do with that window type (ie if it believes all NORMAL type windows MUST have a
titlebar - then so be it, or if all DOCK windows must have no border - and mwm
hints say "give me a border" then the DOCK prefs take precedence.)

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster rasterman com
裸好多                              raster deephackmode org
Tokyo, Japan (東京 日本)