In D2444#50409, @nienhueser wrote:

This is already pushed, right?

This revision solves the issue mentioned with the bathymetry polygons missing. The problem was with the orientation of the inner rings: they are counterclockwise polygons, but the clipping algorith works ONLY on clockwise polygons, so implementing an efficient way to reverse the order of the nodes solved the problem.

The clipping algorithm works perfectly now, there are no major issues, except for the case when the same polygon appears twice on the viewport(ie. when zoomed out enough to see the world map twice or three times), but that case will not apply for the osm-simplify tool's tile cutting algorithm anyway, so I think we can overlook that.

Added const qualifiers.

Generated the diff to the latest version of Marble and some minor changes.

Improved the algorithm to be faster and more readable. Still need to write comments for it. There is one little issue which I worked out with a simple break, so it wouldn't crash. The cause is somehow related to the intersections added at the cornerpoints, because the Weilner-Atherton algorithm is not prepared for that(in a perfect, mathematical world, it can't even occur). Besides that little issue(which is not even visible most of the times), the clipping works fine.

Updated to the latest changes. Still some way to go, but it's even better now.
There is only one main issue remaining(besides some smaller ones), to find a way to reimplement the clipMultiple() algorithm for polygons in order to draw corner points when there is no crossing point with the viewport. This was the main cause of the tile borderline issue btw, calling the clipMultiple() method for closed polygons.

Updated the diff to the latest version, one-by-one the cases are getting solved, but there are still some issues left, like how to handle when there is no crossing point(ie. zoomed on the middle of the polygon).

Update the diff.

I corrected most of the issues and little changes pointed out by Dennis.

The previous diff was wrong.

Reaction for Dennis' comments.