On Wed, Jun 12, 2013 at 4:29 PM, Justin Deoliveira <jdeol...@opengeo.org>wrote:
> Its important to note that on the encoding side the type/element qname
> declared by the binding is only used if there is type in the context which
> is often not there for encoding (think encoding a straight feature
> collection without specifying the feature schema), so it falls back to the
> java class declared. If there is type information in the context then it
> can choose explicitly the binding.
>
> So let's consider both LineStringTypeBinding and
> CircleByCenterPointBinding both bound to LineString. If we have schema that
> tells us that a property of a feature is one or the other, than its no
> problem, the right one will get chosen. If we don't though, then we have a
> problem. The easy way around this would be to have our data structure for
> curves actually extend LineString. Then we have a unique type mapping.
> Which I think is the way we are going correct?
>
> If we don't have a unique type mapping the other thing we can do is have
> either binding check properties of the object to determine which to use.
> For instance, if LineStringTypeBinding gets called but the LineString
> object has some property that indicates the CircleByCenterPointBinding
> should be used, it could delegate. And vice versa for
> CircleByCenterPointBinding. A bit of a hack.
>
Ok, so, if I use a hierarchy like this one, then it should be good?
LineString
+---LinearRing
+----Ring (composed of CircularArc and LineString that form a closed
Ring)
+----Circle (composed of two CircularArc of the same circle, closing
on each other, again, this might not be needed)
MultiLineString
+---Curve (composed of CircularArc only, not sure if this class is actually
necessary yet)
+---CompoundCurve (composed of CircularArc and LineString internally, but
only visible if you cast to Curve)
There would not be a specific CurvePolygon class, I guess the code
interested in circular stuff will have to check
shell and holes to see if they are curved classes? GML wise at least the
specific class does not seem to be needed.
With the above we'd make specific bindings for those classes to the
respective elements in GML and this is it?
Cheers
Andrea
--
==
Our support, Your Success! Visit http://opensdi.geo-solutions.it for more
information.
==
Ing. Andrea Aime
@geowolf
Technical Lead
GeoSolutions S.A.S.
Via Poggio alle Viti 1187
55054 Massarosa (LU)
Italy
phone: +39 0584 962313
fax: +39 0584 1660272
mob: +39 339 8844549
http://www.geo-solutions.it
http://twitter.com/geosolutions_it
-------------------------------------------------------
------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:
Build for Windows Store.
http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
GeoTools-Devel mailing list
GeoTools-Devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel
