systray spec modifications (RFC)

[Carsten Haitzler (The Rasterman)]

On Fri, 17 Feb 2006 10:23:17 +0100 Bradley T Hughes <bhughes at trolltech.com>
babbled:

> Carsten Haitzler (The Rasterman) wrote:
> > I've attached a test file with some suggested changes to the systray 
> > spec (as per the thread that was going on). comments sought. anyone? 
> > is there a general agreement this would be good?
> 
> The details look good to me... I'm looking forward to getting to
> implement support for it :) However, I have only one suggestion:
> 
> > a systray icon conforming to the newer scheme will have either an 
> > property of the atom _NET_SYSTEM_TRAY_ICON_ID and type ATOM, and 
> > format 32 or a property of the atom _NET_SYSTEM_TRAY_ICON_NAME and 
> > type UTF8_STRING on the tray icon window.
> 
> How about instead of saying "if either of these 2 properties, then
> assume the client is using the new scheme", we could have a single,
> authoritative way of saying "this client uses the new scheme". For
> example, we could put _NET_WM_SYSTRAY_V2 in WM_PROTOCOLS, or perhaps set
> _NET_WM_SYSTRAY_VERSION (for now, to 2 (type Cardinal format 32) on the
> systray window.

i wonder if this is worth it? ok - scenario. systray gets a message for a new
window. now it looks for these 2 properties anyway one, then the other
(whichever is has a preference for it will likely stop at and not fetch any
further). if both fetches fail - well it can just drop back to the "old method"
OR abort managing the systray icon entirely. this i would guess would be as
good a check as to what to do without adding an extra version check then
proceed to do the above code anyway? just wondering if we really need the
version property - i'm not so much against it, but no need to add more than we
REALLY need?

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster at rasterman.com
裸好多
Tokyo, Japan (東京 日本)