Re: proposal for a smarter behavior for raising windows on mouse click

[John Harper <jsh unfactored org>]

sawfish implements something similar, it has a _SAWFISH_WM_RAISE_WINDOW 
wm protocol. When windows advertise they support it, user-initiated 
window raises are redirected to the client window, similar to how 
WM_TAKE_FOCUS works

The motivation for adding this was to do drag and drop correctly in 
nautilus (the version based on gtk 1.x). However, getting things to 
work correctly was fraught with race conditions, so something more 
under wm-control may be better..

	John


On Jan 27, 2004, at 11:34 AM, Rob Adams wrote:

> This is very similar to my _IS_DRAG_START protocol proposal a while
> back, although this proposal has race conditions.
>
> I think that the ideal behavior here is actually how Windows does it:
> windows are raised on mouse down if there is no possibility that the
> mouse click will be the start of a drag.
> if it could be a drag start, such as clicking on an icon in the windows
> file manager, the raise is delayed until mouse-up.
>
> My drag-start protocol was an attempt to allow the toolkit to tell the
> window manager whether it was safe to raise on mouse down, or if it
> should wait until mouse up in response to a given mouse click.
>
> I think that the whatever protocol we come up with will have to look
> something like that, though the details are of course important.
>
> -Rob
>
> On Tue, 2004-01-27 at 10:02, Martin Grimme wrote:
> > Hi,
> >
> > I have seen there has been an endless discussion going on about 
> > whether
> > mouse clicks into the window area should raise a window in metacity or
> > not. Havoc's point of view was that we should not ask for yes or no, 
> > but
> > think of a smarter window manager behavior.
> >
> > > From what I have heard, users don't want windows to be raised during:
> > - drag and drop
> > - selecting text
> > - using the scrollbars
> >
> > This led me to the idea that certain widgets could e.g. set a
> > _NET_WM_DONT_RAISE flag on their toplevel window which the window
> > manager checks before actually raising the window.
> >
> > Imagine the following scenario:
> > - user clicks on scrollbar arrow
> > - scrollbar knows that it should not raise the window
> > - scrollbar sets _NET_WM_DONT_RAISE on its toplevel window
> > - window manager detects the mouse click and checks for the
> >   flag
> > - window manager takes action depending on whether the flag is set
> >   or not and unsets the flag
> >
> > (In case that mouse clicks reach the window manager before they reach
> > the widget, the window manager may check for raising in an
> > idle handler.)
> >
> > That way, widgets can determine from the context whether the window is
> > to be raised or not. All this would of course involve code changes in
> > the toolkit, maybe applications, and the window manager.
> >
> > Any comments on this idea?
> >
> >
> > Bye, Martin
> >
> >
> > _______________________________________________
> > wm-spec-list mailing list
> > wm-spec-list gnome org
> > http://mail.gnome.org/mailman/listinfo/wm-spec-list
>
> _______________________________________________
> wm-spec-list mailing list
> wm-spec-list gnome org
> http://mail.gnome.org/mailman/listinfo/wm-spec-list