Simon Cruanes

> I think that the current design of moonpool is okay in this respect, but I wish there was a clear mention in the documentation of the parameter `num_threads`, or...

Is this going to kill `ppx_deriving_yojson` ? I think it targets the safe fragment so it might expect Tuple and Variant as inputs…

lists might be faster for small sizes, but what about large sizes? does it get worse?

Sounds like a good idea. I'd say, by default, small `bytes` should be printed in hex, and large bytes should be like "deadbeef[...2415151 more bytes]". This should probably be configured...

Right, it's never been there. It is in the `iter` package. Implementing reasonably efficiently requires a dynamic array and I didn't have that at the time.

An alternative is to port the [unrolled list](https://github.com/c-cube/iter/blob/main/src/Iter.ml#L191-L309) from iter, and add it to containers as a separate module, and then use that in `Array.of_iter`.

It's just a bit hard to know where to stop :-). In my projects I tend to have a `common_.ml` module with the functor instantiations I need, opened in the...

I think that means that `CCParse.bind` is the wrong one, tbh. In general I use the infix operators anyway.

I think I started this way because of `Lwt.bind` a long time ago. As I say, it's better to use `let*` these days :-).

Yes, that's' a good idea.