Radek Szymczyszyn

It seems so, indeed.

I'll close this PR since there's been no activity here for a while. Please let me know if we should reopen it.

I've rebased this on top of the current master and made sure all the tests pass - they do, together with the self-check. I'll merge it now, as otherwise it...

I accidentally pushed to this branch yesterday instead of my WIP fork. Fixing it in a second...

...and fixed.

Superseded by #450.

This is still a problem, but I'm cleaning up some of my repos, so I'll be closing this PR. If one needs to use Poison with Phoenix, Postgrex, or PhoenixSwagger,...

Thanks for raising this, @tsloughter! Hmmm, [there's code to make any record a subtype of `tuple()`](https://github.com/josefs/Gradualizer/blob/7773a04540e1d7699d59f7f1c4f8e86ddb52a3fc/src/typechecker.erl#L321-L322) in place, which leads me to believe that the problem is going in the...

In general, there are type system mechanisms to free the user from such mundane wrapping into type-safe abstractions. One of them would be bounded quantification, which we, at least partially,...

@zuiderkwast > Perhaps we can work around that by omitting the constraint: > > ```erlang > -spec keydelete(Key, N, [A]) -> [A]. > ``` I tried, but the constraint `A...