Re: [PATCH 0/4] Native dockapp support

[Teika Kazura <teika lavabit com>]

Isn't it enough to introduce a new variable "detect-dockapp", and
check its value at dockapp detection in C? Like this:
------------------------------------------------------------------------
  w->wmhints = XGetWMHints (dpy, w->id);
  if (w->wmhints && w->wmhints->flags & StateHint
      && w->wmhints->initial_state == WithdrawnState
      && w->wmhints->flags & IconWindowHint
      && w->wmhints->icon_window != 0
      // ** look here**
      && ( Fsymbol_value(Qdetect_dockapp, Qt) != Qnil)) { 
      /* window in withdrawn state with IconWindow? */
      /* This looks like dockapp... */
      ...
------------------------------------------------------------------------
This is far from the final solution, but at least acceptable.

On Fri, 28 May 2010 23:05:53 +0400, "Alexey I. Froloff" wrote:
> My implementation sucks, not because it makes wrong assumptions, it
> just sucks.

Don't curse, it's the first attempt, and you're boldly beating a new,
hard path. If it doesn't harm, and so serves as to make Sawfish
acceptable for you, why not we take it?

> Implementing the whole "appicon" and "miniwindows" stuff (as a
> separate module, yes) will be better, but I guess I'll need some
> help.

Are there any documentation on how to code an dockapp? I can't find.

Dockapps are used in wm and gnustep. (What about other window
managers?) I think WM and GS people should push the specfication to
ewmh.

Teika (Teika kazura)