Nelson Vides

@zuiderkwast I moved that test to `test/known_problems/should_fail/`, I tried having a look at fixing it, but I honestly didn't know how for now. I also changed the matching on the...

@josefs > It's nice to see binary comprehensions getting some love because they're almost completely forgotten right now. In order to type check them properly we need a new function...

@zuiderkwast It's all rewritten now, following the style of `expect_` functions. I was aware of those from the very beginning, but I had the feeling they were a bit of...

About the moving stuff to different modules, I sketched this quick refactor yesterday: https://github.com/NelsonVides/Gradualizer/tree/typechecker_refactor Not saying it's good, but it kinda follows your suggestion: error messages to one module, subtyping...

Here. That's how typechecker looks like on a graph. When @erszcz showed me that diagram without telling me what was it, for a moment I thought that his screen was...

@zuiderkwast reading the source code is getting increasingly complicated, specially for this `typechecker.erl` module, at >4k LOC, it's quite a cognitive overhead. I think that module should get some simple...

I have a question, in line with the last things said last time I posted here. I have some prototype for a better `check_guards` that refines the types from the...

Indeed, this is giving me trouble now. When some variable was given the type `any()`, then whatever refinement I do when checking the guards is ignored when doing `union_var_binds` or...

I have another question for a function I cannot find. `union_var_binds` does a glb over all the variables given, but what if I want a lup (least upper bound, i.e.,...

@zuiderkwast when it comes to > This is what I want to solve next, whenever I get enough time. Also other very simple guards such as `Var == atom1; Var...