Re: GTK+ 2.2.0: gtk_window_move(window, 0, 0)

[Tor Lillqvist <tml iki fi>]

Havoc Pennington writes:
 > It's quite complicated on X11; if you read the API docs to
 > gtk_window_move(), gtk_window_set_geometry_hints(),
 > gtk_window_get_position() etc. they have a lot of details.

Thanks. Didn't notice those docs. This is more complicated than I had
thought, sigh, and the Win32 backend surely gets some things
wrong. (Fortunately for most applications it probably doesn't matter
much.)

I wonder what the loop in gdkwindow-win32.c:gdk_window_get_frame_extents() 
that is commented as "find the frame window" is supposed to do. AFAIK,
on Windows the window decorations aren't a separate window, so is this
loop necessary?  Doesn't the preceding loop (identical to the one in
the X11 backend) that traverses the GDK parent chain already find the
top-level window?

And BTW, the docs for gdk_window_get_frame_extents() says "@window: a
toplevel #GdkWindow". Isn't the loop that traverses the parent chain
then unnecessary, if the window passed to this function is supposed to
always be a top-level window?

--tml