Re: _NET_WM_USER_TIME reconsidered

[Matthias Clasen <mclasen redhat com>]

On Wed, 2006-09-06 at 15:00 +0200, Lubos Lunak wrote:

>  But generally no have strict objections to the change. I'd just like to make 
> it a bit more generic in case there'll more similar cases in the future 
> (i.e., as I said, _NET_WM_USER_TIME_WINDOW feels overly specific).

The value of being this specific is that we can check for
_NET_WM_USER_TIME_WINDOW being supported, and if not, fall back to the
old behavior. If you add something generic like: "If a property is not
set on the toplevel, look for it on the group window", there is no nice
way to opt out. One generic mechanism one could think of is to make the
type of the property relevant, ie something like:

IF _NET_WM_USER_TIME is of type Window, it points to a window on which
the actual user time property is stored.

This could be generalized to all (non-window) properties, but it sounds
a bit error-prone to me, there are certainly window managers out there
which just assume the correct type, and don't check it.

Matthias