
Geometry : 
Subject: Re: [geometry] difference algorithm produces invalid polygon?
From: Barend Gehrels (barend_at_[hidden])
Date: 20120305 18:25:37
Hi Volker,
On 2922012 15:04, Volker Schöch wrote:
>> Can you send a (WKT) sample of such a polygon? I think I understand but don't see how the xconstraint fit into this.
> MULTIPOLYGON(((1716 1554,2076 2250,2436 2352,2796 1248,3156 2484,3516 2688,3516 2688,3156 2484,2796 1248,2436 2352,2076 2250, 1716 1554)))
>
> This seems to me like a more general form of the "spike only"type polygon that I originally submitted. I merely mentioned the xconstraint to rule out any real selfintersections (and this constraint incidentially holds in my application, allowing to ensure correct order of points).
>
In the meantime I tried your polygon. The problem our library currently
has with this polygon, actually with every spike (spikeonly or a spikey
feature) is that it does not know which way to go. As soon as the
intersection point (two at a spot) is encountered, it decides going left
or write, but because of the two it can make a wrong decision.
The good news is I think this can be helped. Will study this later in
more detail.
Regards, Barend
Geometry list run by mateusz at loskot.net