Re: [RFC] _NET_WORKAREA and Xinerama

[Sasha Vasko <sasha aftercode net>]

Soeren Sandmann wrote:
> Havoc Pennington <hp redhat com> writes:
>
>
> > So once you have your contiguous region, possible window positions and
> > sizes are those that are inside this region. Maybe there is a
> > reasonable way to compute that.
>
>
> Note that you don't want a window to skip over a solid area during
> drag:
>
>   ###############################################################
>   -------------------------|######|------------------------------
>                            |######|
>       +--------------+     |######|        ................
>       |              |     |######|        .              .
>       |   A          |     |######|        .   B          .
>       |              |     +------+        .              .
>       |              |                     .              .
>       |              |                     .              .
>       |              |                     .              .
>       +--------------+                     ................
>
>
> In the situation above you shouldn't be able to drag a window from
> postion A to position B even though both positions are legal.

Why ?

You drag a window untill it "bumps" into forbidden area, and then let 
user move mouse further to the left without changing position of the 
window, but once its crossing threshold - you can warp window past the 
obstacle.

At least that's how user expects it.
I don't really see  whats the problem with that approach and why struts 
has to make contiguous region. The algorithms for avoiding arbitrary 
non-contiguous area are many and trivial to implement - I do all of the 
above plus edge snapping in about 100 lines of code in AfterStep.


Sasha