LCircle.getGeometry() returns point, should return polygon?

[basisbit]

Currently LCircle.getGeometry() returns a JTS point.

v-leaflet/src/main/java/org/vaadin/addon/leaflet/LCircle.java

Line 78 in 070307d

return JTSUtil.toPoint(getPoint());

If a developer uses for example a LDraw.FeatureDrawnListener, then stores the JTS geometry ( FeatureDrawnEvent.getDrawnFeature().getGeometry() ) somewhere and later on recreates the object using LFeatureGroup.addComponent(JTSUtil.toLayer(geometry)), the result will be only a single LMarker for that point on the map.

How about changing this function to returning a polygon (buffer of size radius around the point)? JTS provides the necessary functions to calculate that: https://locationtech.github.io/jts/javadoc/org/locationtech/jts/geom/Geometry.html#buffer(double)

[mstahv]

Sounds reasonable to me. Will you create a pull request?

[basisbit]

After testing some different ways to fix/change this behaviour, I am not sure any more what actually is the best solution for storing, displaying and calculating of "circles" would be. There was a huge discussion on the jts mailing list ~5 years ago and they didn't come up with a solution that works good for all use cases.

Calculating a buffer with X points per circle-quarter as done by geotools is quite CPU expensive and makes it difficult to edit these geometries later on.

Maybe extending the LCircle.getGeometry() JTSUtil.toPoint(getPoint()) to also do (jts)Geometry.setUserData(getRadius) and add a note in the code documentation about where to find the radius would be a better solution. Then, a developer would at least be able to get the radius value from a generic FeatureDrawnEvent.getDrawnFeature().getGeometry().

[TorstenBMR]

sounds reasonable - any proceedings on this issue?
Currently we have the strange result, that drawing a circle on a map is fine, but it's not possible to automatically zoom to the right level:

LCircle circle = new LCircle(50, 10, 5000);
map.addComponent(circle);
map.zoomToContent(); // does nothing

[mstahv]

I think there is a change in the latest version that takes the center point into account in zoomToContent, but still no improvements to really obey the whole geometry.

There is a lot of things in queue for me to do in this project, this would really be a good place for contribution ;-)