Add CofreeT

[eviefp]

Description of the change

I'm looking for early feedback.

Here's my attempt at implementing CofreeT. I'm not very sure it's correct, so I would love some feedback. I've mostly followed the existing FreeT code and looked at the Haskell CofreeT signatures.

Fixes #14

I will definitely add documentation, and follow the rest of the checklist as appropriate. I'm currently looking for early feedback on the implementation. If it would make it easier for you to provide feedback on a "complete" PR (i.e. including docs, tests, etc.), I would be happy to oblige.

---

Checklist:

• Added the change to the changelog's "Unreleased" section with a link to this PR and your username
• Linked any existing issues or proposals that this pull request should close
• Updated or added relevant documentation in the README and/or documentation directory
• Added a test for the contribution (if applicable)

[JordanMartinez]

I'm not familiar enough with CofreeT to comment on this. Overall, it looks like the initial PR is good. My guess is that we're not looking for the most performant implementation of this, or even the most complete implementation; merely, an implementation in the first place, correct?

[eviefp]

Alright, thank you for the feedback. I'll proceed to add docs, tests, etc., as to make it ready for review/merge!

[eviefp]

I have no idea how to write unit tests for CofreeT since all the functions seem to only have one implementation which type-checks. I think the only "interesting" choice in the module is actually the type itself.

Note that, despite me having added a file under test/, it's not actually a test: it's just an example.

I think this is now ready for review!

[JordanMartinez]

LGTM. The docs helped me understand why I would use this, so thanks!

[thomashoneyman]

I'll have time to review this certainly by the end of the week. Thanks for the PR!

[eviefp]

I realised I might want to do one more thing: consider what other instances I could add to CofreeT. I think there's quite a few. I'll have them ready before the weekend :)

EDIT: nevermind, they are done 👇

[thomashoneyman]

I'm not particularly familiar with Cofree, so it would be nice to have another look from @garyb (if you have a moment or suggestions for anyone else who should verify this). That said, it looks pretty good to me! Just a couple of suggestions.

[thomashoneyman]

There is also this implementation which relies on Lazy (though yours via Unit -> ... accomplishes the same thing), which you may find interesting to compare:

https://github.com/arthurxavierx/purescript-comonad-ui-todos/blob/master/src/Control/Comonad/Cofree/Trans.purs

[eviefp]

I addressed all review comments.

About the Lazy vs Unit -> bit, the only reason I would keep it as-is is because we're not already depending on lazy. But if you think it's better to add that dependency, it's a trivial change. I haven't used lazy much (I even forgot it exists) -- that's the only reason I didn't implement it like that from the beginning.

[thomashoneyman]

No need to introduce it, although since we’re already bringing in new dependencies it wouldn’t be that big of a deal.

I’d like to let this sit for a week or so for others to weigh in, but if no one does by then I will go ahead and merge because it appears correct to me (with the caveat that I’m not particularly familiar with this construct).

[paf31]

This won't be guaranteed to terminate in most actual use cases. Is there a reason for adding it?

[eviefp]

Is it because we expect CofreeTs to generally be infinite? The main reason I added it is because this instance exists in Haskell as well, but I see your point. I'm not sure this would get a lot of (legitimate) use. Maybe it's better to remove and consider re-adding if we get requests/issues for real use cases.

[garyb]

Would it be possible to use Eq1 m instead of this constraint, if we keep it?

[eviefp]

I removed both Eq and Ord for now. It might be possible to use Eq1 though, so I'll keep that in mind if we end up implementing it eventually.

[eviefp]

I think I addressed all review comments. Thank you for taking your time to review my PR!