Remek Zajac

A a note on this. It appears https://github.com/boostorg/geometry/pull/479 fixes the original problem reported, however, we continue observing two related, albeit slightly different classes of problems: 1. geometries defined over points...

Hi @barendgehrels .. my time to finish my migration is coming to an end, so I must ask whether there is any progress on this?

Thanks @barendgehrels. Is there a branch i could test against?

I will just keep filling this comment up with my findings as i diagnose this. I see that cartesian and spherical scenarios start to differ [here](https://github.com/boostorg/geometry/blob/develop/include/boost/geometry/algorithms/detail/overlay/overlay.hpp#L306) where the cartesian finds...

@awulkiew are there any news on this? I could have another go at diagnosing this, but the code is quite cryptic. Is it described somewhere? Does it implement a specific,...

Thanks for the reply @awulkiew and for diving into this. > Is this the only case where you get wrong result in spherical CS? Nope. Attaching more samples. I can't...

Found another possible problem. ``` namespace cartesian { using point = boost::geometry::model::d2::point_xy; using linestring = boost::geometry::model::linestring; using polygon = boost::geometry::model::polygon; using multipolygon = boost::geometry::model::multi_polygon; using mpoint = boost::geometry::model::multi_point; } namespace...

Sure.. the vertical edges of this tile are of course non-parallel in geometric sense, but they run along their respective meridians so i can't imagine the picture changing post-densify and...

Very interesting - i guess i assumed the winding order will be a hint stronger than edge lengths. So all i have to do (if i understand correctly) is to...

Thanks @awulkiew ! I confirm that the fix solves the original problem, but i continue seeing [empty unions](https://github.com/boostorg/geometry/files/1975615/out.txt). Fewer of them now (119 down to 91) as i re-run my...