To follow-up:

1. The docs were generated a few hours ago and are now live at http://docs.racket-lang.org/frog/index.html :clap:

2. However the catalog still shows only README links (not a “frog” link to docs, or “Documentation” button) at both:
3. Summary boxes like https://pkgd.racket-lang.org/pkgn/search?q=frog
4. Details page https://pkgd.racket-lang.org/pkgn/package/frog

So that’s weird. I clicked Rescan My Packages and will wait to see if that refreshes it. (Not urgent)

https://pkgs.racket-lang.org/package/frog has a link to the docs

Yes it does, now. Also it’s showing a new dependency warning, now, which I’ll fix.

@samth: are you available to answer a few questions on local type inference?

Not right now but if you ask I’ll get back to them

Alright, thanks. I have a few questions to start with, and I think they’re fairly basic. First of all, they introduce what appears to be two mechanisms of local inference: local constraint solving for type argument synthesis, and bidirectional checking rules. These mechanisms are meant to be used together, yes? I only ask because there are a few statements that seem to portray them as relatively independent, but most of them seem to say that they solve different problems.

That said, it’s the constraint generation piece that I think I least understand. The variable elimination section loses me a little bit. In a language without subtyping, is that part even necessary? Either way, what exactly is that step’s purpose? The section sort of describes the mechanical process of variable elimination, but it doesn’t really seem to motivate it.

I also start to get pretty lost by the notation in the following section on constraint generation, though I can probably figure that out if I understand at a high level what’s going on.