This seems interesting, but on a cursory reading of the docs, there is something I don’t understand. What’s the real advantage of this library over other libraries we already have for both Erlang and Elixir? At a first glance it doesn’t seem to bring anything new. Maybe there are improvements under the hood, like better shrinking, or maybe it’s more convenient because it has better generators. Or maybe it makes it easier for users to create their own generators or shrinking strategies. I can’t find any documentation on these topics as of now, but I assume that must be possible right? Working with a predetermined set of generators, even if a language with as few types as Elixir, seems very restrictive.
I think these points I’ve raised should be addressed in the documentation:
- Why should I care about StreamData instead of other options?
- How do I write my own custom generators?
- Are the generated values completely random, or do we have any guarantees? For example, when I test my code on a random list, I’d like to always test it on the empty list.
