Currently, the library uses unsigned integers for indexing, since indices may not be negative and to match the standard library. However, for some use-cases indices may be computed using negative integers. Casting back-and-forth between signed and unsigned creates messy code.
What should we support? Both?
Currently, the library uses unsigned integers for indexing, since indices may not be negative and to match the standard library. However, for some use-cases indices may be computed using negative integers. Casting back-and-forth between signed and unsigned creates messy code.
What should we support? Both?