Péter Gömöri

thanks for the example, I have a look into it

great work, Radek I agree with the error formatting. I formulated a task expanding a bit the idea. https://github.com/josefs/Gradualizer/issues/49

:D haha, ... maybe I won't sleep till new year, it's less then 24 hours, types are too exciting

I think intersection with `any()` could be temporary solution for handling conditional expressions before refinement types are fully implemented. for exampe ``` case get_int() of PosInt when PosInt > 0...

I don't have broad knowledge of history and syntax in other languages. Could you explain me why `;` is a good fit for intersection or why it is used in...

(it was actually Victor who mentioned return-no-type) I see, so type information must always stem from specs (of functions and also records, if I see correctly) and type should be...

there is an interesting consequence of the above (I dont know if we can call this inconsistency or just surprising for some) This snippet fails type checking as a list...

Just a note, without expecting any answer As a "real user" :) (running Gradualizer on a smaller open-source and big proprietary project) I can say that not only Erlang/OTP is...

would it be an option to extend the alias format from a 2-tuple to also support a 3-tuple eg ``` {alias, [ {test_company_a, company_a, [test]} ]}. ``` and the alias...

I imagined it would work the same way as `rebar3 as company_b eunit` would run in a combined `company_b+test` profile. You mentioned that > Commands themselves can define their own...