You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently the API of this project is designed with the potential for consumption from Java projects (with the exception of geojson-dsl). The libraries will always be interoperable with Java code, as all Kotlin is, but this issue looks specifically at whether it will be worth continuing to provide a dedicated Java API to match the API provided in Kotlin.
Issues with maintaining dedicated Java API
Some possible function implementations may require use of @JvmName to avoid compilation issues. For instance, if Feature generics are added (Feature generics #76), some functions like length() that can accept different types of geometry will need multiple overloads of the function taking a Feature<?> parameter where the different generic types would cause a platform conflict. With @JvmName this can be avoided, but complicates the Java API.
Without a dedicated Java API, the libraries could take advantage of inline/value classes, specifically for the LngLat class. This could wrap an Array<Double>, but would be impractical to use from Java.
With the existence of Mapbox's Java port of Turf, the value of maintaining a dedicated API is questionable.
The text was updated successfully, but these errors were encountered:
Currently the API of this project is designed with the potential for consumption from Java projects (with the exception of
geojson-dsl
). The libraries will always be interoperable with Java code, as all Kotlin is, but this issue looks specifically at whether it will be worth continuing to provide a dedicated Java API to match the API provided in Kotlin.Issues with maintaining dedicated Java API
@JvmName
to avoid compilation issues. For instance, if Feature generics are added (Feature generics #76), some functions likelength()
that can accept different types of geometry will need multiple overloads of the function taking aFeature<?>
parameter where the different generic types would cause a platform conflict. With@JvmName
this can be avoided, but complicates the Java API.LngLat
class. This could wrap anArray<Double>
, but would be impractical to use from Java.With the existence of Mapbox's Java port of Turf, the value of maintaining a dedicated API is questionable.
The text was updated successfully, but these errors were encountered: