Sergei Winitzki
Sergei Winitzki
I see no reason to drop 2.11 support. The failing tests depend on internal details of Scala standard library (e.g., the name of accessors in Either). Tests should not depend...
Right now tests pass in 2.12 and 2.13. Some inessential tests fail in 2.11. Is this situation okay or should it be fixed asap?
In my job, we are still using 2.10, 2.11, and 2.12, while upgrading to 2.13 even for a few projects has been dragging on because we don't have enough scala...
This simplification seems to be problematic; it prevents correct typing of case objects for the example `Option[Option[A]] => Option[Option[A]]`.
`bitLength` is used to estimate the number of iterations required for a `fold` to achieve convergence for certain algorithms. In a language like `dhall`, all functions are required to terminate...
A validation-based approach will look like this, as far as I understand: ```haskell let x: Natural = ... let validate_x = assert : Prelude.Natural.isZero x === False let y: Natural...
I think we would like to keep the number of built-in functions to a minimum. However, `foldWhile` and `div` with `rem` seem to be important. Other numerical functions can be...
Here is an implementation of “safe division” `divide x y _` where the third argument must be of type `Unit` and represents a proof that the second argument is nonzero....
@Gabriel439 Would it be good to add a built-in `div` and `mod` functions that work only with `Natural` constants (or alternatively return `Optional Natural`)? And what do you think about...
@Gabriel439 Do you think `foldWhile` is not worth adding as a built-in (for `Natural` and `List`)?