Sergey Torgashov

@mrdziuban , > My thinking was that the different behavior made sense because a Gen is always used in a test, whereas the Prop's label is only used if the...

My initial idea was to keep the existing functionality intact while add a new one as necessary (i.e. via `lazyLabel` or something). However, I might not be the best person...

I wonder if is would be possible to make `ConnectionInfo` available in an implementation-independent way. Otherwise, implementing it in one client but allowing it not to be implemented in all...

Honestly, the entire idea of having `Connection` as an optional attribute of `Request` or `Response` is not cool at all. Such an attribute may or may not be provided depending...

Perhaps, I'm missing something.. But why should `memoize` for `Resource` result to `Resource[F, F[A]]` but not `F[Resource[F, A]]`? I mean, when we call `memoize` for some `F[A]` and get `F[F[A]]`...

Hmm.. I see you point and that makes sense, but just to clarify.. Let's try to describe `memoize` in terms of abstract effect types, `F[_]` and `G[_]`. So you mean...

On the second thought, >> but not F[Resource[F, A]] > When should the resource be released in this case? I would guess (simply by looking at the signature), it should...

@SystemFw, actually I meant is that the entire Resource with its `use` block should behave as a single memoized effect. I would say, it is not the same as a...

Thank you for the PR!

> Sending "raw" JSON primitives over the wire in general seems strange (string, numbers, whatever). I'd expect them to be encapsulated as objects or arrays. I agree with that one....