Re: wm-spec implementation

[Iban HATCHONDO <hatchond yahoo fr>]

Havoc Pennington wrote:

> Iban <hatchond yahoo fr> writes:
> >  - _NET_CLIENT_LIST_STACKING :
> >     As a window manager I am suppose to provide this property.
> >     But the question is :  has bottom-to-top stacking order ??
> >     I don't get the point. More, it is a "long" job to retrieve this
> > order.
>
> This is used by pagers to e.g. display miniature copies of windows.

Ok. The simplest way I found to do that is by using the "old" property
_win_client_list, and to reorder it with a query-tree.

> >  - _NET_WM_ICON : I don't really understand your description. Could you
> > explain it a little more for me please ?
>
> The format is (IIRC) one 32-bit integer for width, one 32-bit integer
> for height, then each 32-bit integer contains 1 byte for alpha, 1 byte
> for red, 1 byte green, 1 byte blue.

Still not very clear for me. :'(

>
> >  - _NET_WM_WINDOW_TYPE : This one is clear, but I make some questions
> > coming up to me :
> >    when I have a _NET_WM_WINDOW_TYPE_DESKTOP : how many of this window
> > may I have at the same time ? Because if I look at nautilus, he create
> > on window of that type, and as window manager I'm not suppose to restack
> > any window below that one ?
>
> There should only be one window of this type, yes. The window manager
> shouldn't crash if there are more, but isn't expected to do anything
> sensible either.

I suspect it should be like this, but wasn't sure. That's is a good idea.
(maybe you should precise this in the spec)

Also, an other reading of your paper, gave me another question : why decide
you for a new "system" for _net_wm_state in regards of _win_state ? They seems
quite similar, except that I think the _win_state is more rich ?

Finally, I have some troubles to understand the compound_text encoding article
from the X consortium.
Is somebody familiar with this encoding, and then can show me a "regular"
compound_text encoding with his decoding result ? I guess I will have the same
problem with the UTF-8_STRING encoding for which I am looking for information.

Otherwise, I can say that, your paper, was quite sufficient, for implementing
some "gnome-extended-management-library", and for window manager behavior.
I cant wait anymore for gnome2 and compliant applications !!

Regards
Iban.


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com