Also, as a Lisp programmer, I absolutely loved reading the section at https://koka-lang.github.io/koka/doc/book.html#sec-handlers - it's essentially like using the Common Lisp condition system, making me feel at home, except it's more generalized (e.g. masking individual condition handlers is not generally possible in CL) and also working in a strongly statically typed environment.
As far as I understand the PL theory involved, it's statically-typed delimited continuation, which I think is even more general than the CL condition system.
In CL, you can control the visibility of restarts. In the restart-bind construct, there is a :test-function which serves as a predicate that determines whether the restart is visible. (Unfortunately, this function is referred to as a condition).
While restarts can be visible or not, I believe there is no such mechanism for handlers. However, handlers can be effectively invisible/transparent by declining to handle a condition, which they can do simply by returning instead of perpetrating a non-local transfer.
Also, there doesn't seem to be an API in Common Lisp for calculating the handlers visible at a given point for a condition of a given type. So that means that the ability of a handler to decline a condition is pretty much as good as a visibility mechanism.
