It would be a good experiment to get ballpark performance numbers for the generated code from the Kotlin backend (especially when used with a library like ICU4X).
JNI/JNA can be slow, and having a rough idea of what is going on helps clients understand what they're looking for.
It would be a good experiment to get ballpark performance numbers for the generated code from the Kotlin backend (especially when used with a library like ICU4X).
JNI/JNA can be slow, and having a rough idea of what is going on helps clients understand what they're looking for.
cc @jcrist1