Paul Colomiets

> I don't personally feel too strongly one way or another which strategy is used by default (wheel or heap) I prefer heap to be used by default. There are...

> I'm wondering if promoting this upstream in pnetlink would be a better idea since I've already borrowed a decent amount of code from pnetlink. Well, this will help too,...

> It looks like there are multiple mistakes with visiting recursively. Or was non-recursive visitor intentional?

Hi! Thanks for the PR. And sorry for late reply. There are two issues with the PR. 1. `unexpected_format` should probably be replaced by `unexpected_static_message` instead where message in static...

Well, i definitely want some kind of visitor in the library. But I'm not sure which one is good. Usually, I like iterator-style visitors better than what's linked (obviously it's...

> Walking the AST once to gather data needed to answer questions like "is the type Foo a scalar or a type? Isn't it just walk over top-level definitions? I...

Follow up: I think simple visitors that map enum variant to method call are just legacy from the languages that don't have pattern-matching. I.e. this is a way to emulate...

> I don’t quite know what you mean by “iterator-style” visitor however. Do you have a link to an example? I mean a visitor where you can iterate over nodes...

I've made a sketch of "iterative visitor" in #31, which potentially allows iterating over different levels of granularity: ```rust let mut field_names = Vec::new(); for f in doc.visit::() { fields...

Oh, sorry, I've mentioned wrong PR in the previous comment. #31 is the correct one.