On 19.06.2013 00:25, Manuel Kasser wrote:
> I just wondered what a client's leader_id contains (or, to be more
> precise, in which circumstances there IS a leader_id). I couldn't
> produce a scenario where a client's leader_id is set, but maybe I just
> tried out the wrong scenarios.
Urgh. There is so much wrong in there. First, it is leader_window, not
leader_id. This rename was done here, but the luadoc was forgotten:
commit a02d026f775a45701cd6d93a5b7c5222e043c4d7
Date: Mon Aug 17 17:02:45 2009 +0200
client: port to new object system
After querying leader_window instead of leader_id, I still didn't get any useful
result, thanks to:
commit 5d0a81c8bf4881d4a0716112e7cedfc3a096a838
Date: Sun Aug 8 18:31:07 2010 +0200
fix some deprecated atom constants
client_leader_q = xcb_get_property_unchecked(globalconf.connection,
false, c->window,
- WM_CLIENT_LEADER, WINDOW, 0, 32);
+ WM_CLIENT_LEADER, XCB_ATOM_WINDOW, 1, 32);
(This accidentally(?) replaces a 0 with a 1)
Both of these are fixed now in latest git. Yay.
> What is ment by "spawned by the same
> command" as said in the API*.
This property just gives you the WM_CLIENT_LEADER property of the client. ICCCM
defines its meaning:
The client must identify one top level window as the "client leader."
[...]
All top-level, non-transient windows created by a client on the same display
as the client leader must have a WM_CLIENT_LEADER property.
http://tronche.com/gui/x/icccm/sec-5.html
So this is "spawned by the same command" (or perhaps more like "windows which
belong to the same process").
Cheers,
Uli
--
"Every once in a while, declare peace. It confuses the hell out of your enemies"
- 79th Rule of Acquisition
--
To unsubscribe, send mail to [email protected].
