Re: Work in Progress: draft 1.9a

[Dominik Vogt <dominik vogt fvwm org>]

On Tue, Jul 06, 1999 at 12:32:30AM +0100, pdw@ferret.lmh.ox.ac.uk wrote:
> On Mon, 5 Jul 1999, Dominik Vogt wrote:
> > On Mon, Jul 05, 1999 at 08:05:26PM +0100, M.Rogers@cs.ucl.ac.uk wrote:
> > > >This modal thing reminds me of something else: Currently
> > > >application-modal windows are not handled very nicely (in Qt anyway).
> > > >When you click on one of the main windows, it will be raised and then Qt
> > > >raised the app-modal dialog again. This is annoying. 
> > > 
> > > Would it make sense for window managers to keep all transients stacked above 
> > > their parents?
> > 
> > At least several fvwm users keep complaining about it. It's
> > definitely a matter of personal taste. I see no reason why one
> > should force netscape's 'find' window fully above the browser
> > window (just an example).
> 
> Take xcdroast, which has application-modal dialogs [1] which you can
> very easily lose behind the main window.  Confuses the hell out of new
> users, and I definitely consider not keeping it on top a bug.   As a point
> of reference, icewm does keep TRANSIENTS_FOR on top.
> 
> > > (I'm not suggesting this should go in the spec.)
> > 
> > Um, if you don't suggest it goes in the spec, what are you suggesting
> > then? At what other place should it be stated?
> 
> I asked Michael privately if he could clarify this: "would there be any
> case in which it would be wrong to keep a WM_TRANSIENT_FOR window on top
> of its parent?"

Hm, there is a noticeable number of people who are quite capable
of tracking even lowered transients who do complain about that.
And I don't think windows staying above others is easy to understand.
Windows don't bear a label 'I am a transient window' and often it
is not clear why it should be transient. Currently I'm working on a
'more usable' solution in fvwm. For once you can turn off the special
treatment of transients completely. But if you do activate it you
are still able to lower transients below their parent if you want:
Usually when raising the window it is just raised along with all its
transients, but if it is already on top (but below all its transients)
it is brought on top of them all. Has some unwanted effects when moving
the window though (because moving usually raises the window too).

In other words: I wouldn't like it if a spec demanded this.

> The ICCCM is ambiguous as to whether this is what is
> intended.  If this is not what is intended by WM_TRANSIENT_FOR, then we
> need some further hint.  Is this a modal/non-modal question perhaps
> (modals should stay on top).
> 
> I personally think that WM_TRANSIENT_FOR should force the window on top.
> IMHO Netscape's find window should be forced on top of the main window -
> it is useful, otherwise you keep losing the window behind the main window.
> If you don't want the find window there, you kill it.

Bye

Dominik ^_^

-- 
Dominik Vogt, Hewlett-Packard GmbH, Dept. BVS
Herrenberger Str.130, 71034 Boeblingen, Germany
phone: 07031/14-4596, fax: 07031/14-3883, dominik_vogt@hp.com