Re: No focus on map hint

[Rob Adams <robadams ucla edu>]

I considered that.  My thinking for doing an explicit hint for focus
behavior was that there may be application reasons other than something
like being activated not through user input for not wanting a window to
be initially focused.  Also, the other hints all seem to explicitly
indicate desired behavior rather than UI semantics, so this is more
consistent.

On Thu, 2003-03-13 at 19:44, John Harper wrote:
> Rob Adams writes:
> |Applications such as IM clients may wish to indicate that they should
> |not be focused on initial map.  Here is an attempt to express this as a
> |WM hint:
> |
> |+_NET_WM_STATE_NO_FOCUS_ON_MAP indicates that when the window is first
> |+mapped, the window manager SHOULD NOT take focus from another window
> |+in order to focus the new window.  Note that the window manager MAY
> |+still focus the new window, if no window was previously focused.
> 
> I think this hint is too procedural. Maybe I misunderstand the intent,
> but aren't you really trying to communicate to the window manager that
> the window was mapped not as a result of anything the user did, but as
> the result of an external action (i.e. receiving a message from a third
> party)
> 
> So perhaps it would be better to call it
> _NET_WM_STATE_EXTERNALLY_ACTIVATED or something?
> 
> Alternatively, perhaps the logic could be inverted and we could have a
> separate property which would be set to the timestamp of the input
> event that caused the application to map the window. This could then be
> used by the wm to decide whether to focus the window or not (I haven't
> really thought this through, so it may not work, but one algorithm the
> wm could use is to check if the timestamp is later than the last time
> it assigned focus)
> 
> 	John
>