Comments on 1.0pre3, part 1

["Matthias Clasen" <Matthias Clasen poet de>]

Here is a list of comments on 1.0pre3.

Matthias


2. Root Window Properties (+ Related Messages)

Add the following:

Whenever this spec speaks about "sending a message to the root window",
it is understood that the client is supposed to create a ClientMessage
event with the specified contents and send it by using a SendEvent
request with the following arguments:

destination root
propagate False
event-mask (SubstructureNotify|SubstructureRedirect)
event the specified ClientMessage


2.2 _NET_CLIENT_LIST 

Remove XA_ prefix from type names.
Instead of "mapping order" it may be clearer to say "initial mapping 
order" or "time the WM started to manage the window".  Also, it is 
unclear if the youngest windows are at the beginning or at the end of
the list. Likewise, it is unclear if "stacking order" is "bottom-to-top
stacking order" (as in QueryTree) or "top-to-bottom stacking order".


2.4 _NET_DESKTOP_GEOMETRY

Should read

_NET_DESKTOP_GEOMETRY width, height, CARDINAL[][2]/32


2.5 _NET_DESKTOP_VIEWPORT

Replace "current view in pixel" by "viewport".


2.6 _NET_CURRENT_DESKTOP

Should probably read

_NET_CURRENT_DESKTOP desktop CARDINAL/32


2.8 _NET_ACTIVE_WINDOW

It would be more in line with X terminology to speak about window IDs
instead of window handles. It should be specified that _NET_ACTIVE_WINDOW
is set to None if no window has the focus.


2.9 _NET_WORKAREA

Should probably read

_NET_WORKAREA left, right, top, bottom, CARDINAL[][4]/32


2.10 _NET_SUPPORTING_WM_CHECK

Remove XA_ prefix from type names.


2.11 _NET_VIRTUAL_ROOTS

Remove XA_ prefix from type names.
I think the motivation for this property is misleading, since 
conforming pagers shoudln't have to listen for substructure notifies
anymore, or should they ? I was under the impression that they get
all they need (except for window geometries) via the _NET_DESKTOP_...
properties.  A better motivation for _NET_VIRTUAL_ROOTS is that is
allows background drawing and figuring out WM frame windows.


3.2 _NET_WM_MOVERESIZE

Add the following:

The client MUST release all grabs on Pointer events, prior to sending
such message. 


4.5 _NET_WM_STATE

Replace "state window" by "window state" in the first sentence.
Just out of curiosity: Why not _NET_SKIP_PAGER as well ?


4.6 _NET_WM_STRUT

In other cases where the value of the property is a fixed-size struct,
the names of the struct members have been listed, so this should probably
read

_NET_WM_STRUT left, right, top, bottom, CARDINAL[4]/32


4.7 _NET_WM_ICON_GEOMETRY

Should probably read 

_NET_WM_ICON_GEOMETRY x, y, width, height, CARDINAL[4]/32