10:05 <m811> Hey! Are there any good talks on the internals of how OCaml's type checking is implemented? Haskell has more than a few on its constraint checking etc, but OCaml's talks don't show up that easily.

10:17 <d_bot> <colin> http://okmij.org/ftp/ML/generalization.html this is often talked about

10:37 <m811> colin: Thanks. Though if I'm not mistaken, that's a description of speeding up your run of the mill type unification algorithm. OCaml does a lot more these days I think.

10:38 <m811> E.g. subtyping, GADTs(?), polymorphic variants etc.

10:43 <d_bot> <colin> Internals are always glossed over, I doubt you'll find much of substance. Just allusions to a "first-order unification algorithm" and "union-find data structures" etc. You may find http://gallium.inria.fr/~fpottier/publis/emlti-final.pdf interesting (it's the chapter from ATAPL) - it doesn't have much in the way of implementation specifics but is a good view of the landscape.

10:46 <d_bot> <colin> If you search for "objective caml" in that PDF, you'll find brief mentions of things related to what you said. Example: equirecursive types w/ OCaml objects and polymorphic variants at the bottom of page 71 of the PDF.

10:58 <m811> Ah, ATTAPL, you say. I've got that book. I'll go over the 10th chapter too. Thank you!

11:01 <m811> Ah, I've actually read the 10th chapter already. I found "OutsideIn(X) -- Modular Type Inference with Local Assumptions (2011) - Dimitrios Vytiniotis, Simon Peyton, Jones Tom Schrijvers, Martin Sulzmann" to be far more accessible.

11:04 <m811> Would it be fair to assume OCaml does its checking/inferring much like Haskell and the constraint system described in HM(X)?

11:05 <m811> I'm starting to skim OCaml's source too for answers, starting with the parser AST.

11:05 <m811> Fortunately found a HACKING.adoc in the latter's directory.

11:29 <d_bot> <ggole> The Remy paper mentioned in colin's first link has some concrete code in the appendix that might be helpful

11:31 <m811> ggole: You mean http://okmij.org/ftp/ML/generalization.html?

11:35 <d_bot> <ggole> http://gallium.inria.fr/~remy/ftp/eq-theory-on-types.pdf

11:36 <m811> Oh, interesting. I don't see him pasting that link.

11:36 <m811> I'll take a look at that PDF, too. Thank you!

11:37 <d_bot> <ggole> There's a (rotted) link at the bottom of one of the sections

11:37 <m811> Ah, you meant a link at the end of generalizations.html. Got it!

11:38 <d_bot> <ggole> Oh yeah, should have been a bit clearer about that.

11:38 <m811> I'm curious, are you two chatting via some other mechanism that both of your messages come via <d_bot>?

11:40 <d_bot> <ggole> We're on discord, which is bridge via that bot

11:43 <m811> Ah, okay. Thanks!

14:49 <d_bot> <mbacarella> You are invited to the next OCaml Cafe on Wed, Aug 4th @ 7pm (US, Central).

14:49 <d_bot> <mbacarella>

14:49 <d_bot> <mbacarella> This month, OCaml Café will consist of two parts. First, Rudi Grinberg of OCaml Labs 2 will present an informal introduction to Dune, the OCaml build system. Learn about Dune from one the people developing it. Following Rudi’s presentation, we will open the discussion to all things OCaml-related.

14:49 <d_bot> <mbacarella>

14:49 <d_bot> <mbacarella> Zoom meeting details with add-to-calendar link here https://hfpug.org/event/ocaml-cafe-introduction-to-dune-and-open-forum/

14:49 <d_bot> <mbacarella>

14:49 <d_bot> <mbacarella> OCaml Cafe is a friendly, low stakes opportunity to ask questions about the OCaml language and ecosystem, work through programming problems that you’re stuck on, and get feedback on your code. Especially geared toward new and intermediate users, experienced OCaml developers will be available to answer your questions. Bring your code and we’ll be happy to review it, assist with debugging, and provide recommendations for imp

14:51 <Corbin> Just to double-check my understanding of what's recommended and supported: If I want to have a foreign language (a DSL) compiled to bytecode, even if the foreign language has no names, I should compile it to OCaml source and generate names as needed. Right?

20:56 <companion_cube> That's a way to do it, but bytecode is pretty stable, you could directly produce it I think.