Adam Wulkiewicz comments

Results 169 comments of


                                            Adam Wulkiewicz

Formatting problem with model::d2::point_xy<unsigned char>

Thanks. I think it should be fixed even though I'd discourage from using `unsigned char` as coordinate type since it probably works only in some specific cases.

Formatting problem with model::d2::point_xy<unsigned char>

After giving it more thought I'm no longer sure if type `unsigned char` should be supported here. The reason is that in order to be consistent we should treat all...

Formatting problem with model::d2::point_xy<unsigned char>

> I think, answer to 1. is Yes Can you think about some use case for that? A reason why one may want to represent coordinates like that? > What...

Formatting problem with model::d2::point_xy<unsigned char>

> However, the `point_xy` is a template (...) Right, for completeness. But how about users who store coordinates in character types because they want to print them as characters for...

Formatting problem with model::d2::point_xy<unsigned char>

> AFAIU, coordinate_type is supposed to be a number, by type and by semantic. If someone desires to overload its meaning as a splash of paint...bad luck :) Everything is...

Formatting problem with model::d2::point_xy<unsigned char>

I wonder what @barendgehrels thinks about it?

Formatting problem with model::d2::point_xy<unsigned char>

@pauljurczak what are you using `unsigned char` for?

Formatting problem with model::d2::point_xy<unsigned char>

@pauljurczak That's interesting. Were you calling some Boost.Geometry algorithms for these geometries? I'm asking because internally while performing calculation we not necessarily promote the coordinate type to something bigger so...

boost::geometry::azimuth returns values with incorrect sign convention

I also think that we should rather add more information in the documentation. Mentioning PostGIS is ok but the difference should be clearly expressed.

Intersection of geographic segment that crosses the antimeridian with its envelope

Thanks for the report! Yes, this is a bug but it's more severe than you suggested. For set operations of linear geometry vs box the use of cartesian strategy is...