Re: Window History Placement

[Lubos Lunak <l lunak suse cz>]

On Friday 10 of January 2003 15:03, Matthias Clasen wrote:
> >  I don't understand why such property would be needed.
> > WM_CLASS+WM_WINDOW_ROLE
> > are supposed to uniquely identify every window in an application (+PID if
> > needed). In one of the URLs above I only saw that there might be some
> > problems with role, but not what problems, and I don't see how adding
> > this
> >
> > extra property is going to fix the problem, whatever it is.
>
> For one thing, the extra property gives clients a way to turn off history
> saving, which seems valuable.

 Doesn't of of the two geometries in WM_NORMAL_HINTS do the same?

> Another clever thing with the number part of the id is this "reshuffling",
> which seems to implement something like the n best positions for a window
> with name/class/role, although the spec is not clear enough here. What
> happens if I start two instances of an app (so that the wm remembers two
> positions n/c/r/1 and n/c/r/2. If I now close 1 and move 2, will the wm
> still remember two n/c/r positions, or is
> the old n/c/r/2 position forgotten when 2 is moved ?

 Not sure what you mean. Are you saying that the WM may remember more 
positions for a window? I don't see why the extra property is really needed 
there.

>
> Here is what the ICCCM has to say about uniquely identifying windows, its
> not quite what you say,
> since the ROLE is only guaranteed to be unique withing a given
> SM_CLIENT_ID.

 Yes. But there's one SM_CLIENT_ID per application in 99,999999% cases, so in 
practive what I said should be enough (or one can use the SM property instead 
of PID, if needed).

-- 
Lubos Lunak
KDE developer
---------------------------------------------------------------------
SuSE CR, s.r.o.  e-mail: l lunak suse cz , l lunak kde org
Drahobejlova 27  tel: +420 2 9654 2373
190 00 Praha 9   fax: +420 2 9654 2374
Czech Republic   http://www.suse.cz/