[Gvsig_english] clean polygon? Issue with layers imported fromOracle Spatial

[Francisco José Peñarrubia]

Hi Wolfgang.

As Jukka said, polygon orientation is different from Oracle and 
Shapefiles. And some other formats, may come with different orientation 
even inside the same file (DXF, for example). So, gvSIG uses a rule from 
Java:

<H3>WIND_NON_ZERO</H3><PRE>static final int <B>WIND_NON_ZERO</B></PRE> 
<DL> <DD>The winding rule constant for specifying a non-zero rule for 
determining the interior of a path. The non-zero rule specifies that a 
point lies inside the path if a ray drawn in any direction from that 
point to infinity is crossed by path segments a different number of 
times in the counter-clockwise direction than the clockwise direction. 
</DD></DL>

Is very common (at least, it was some years ago) that even shape files 
come with some polygons in one order and others the reverse.

That said, I agree about shape file writer in gvSIG. We should test the 
polygon before writing to the shape file, and change the orientation if 
needed. Even in edition mode could be a good idea to always use JTS 
orientation, and check every polygon with Geometry.isValid method.
To testers: We should point this as a bug, and try to fix them in 1.9 or 
in 2.0.

Best regards.

Fran Peñarrubia
www.scolab.es

Wolfgang Qual escribió:
> Dear list,
> in fact, exporting a shapefile to Oracle Spatial will result in change of
> the orientation of the vertices. When re-importing that Oracle spatial layer
> to shapefile (export Oracle Spatial layer to shapefile), this "oracle
> spatial" orientation of vertices will remain! Is it a bug? 
>
> Best,
> Wolfgang
>
>