Anil Madhavapeddy

Because then dune will unambiguously be able to know which subdirectories are vendored code and which are direct local code. This could, e.g., affect default profiles so that we filter...

> As I've said in slack, this idea is outside of this feature's scope. Once duniverse creates a lock file for dune, that should be the source of truth for...

Given that a Host header is essential in HTTP/1.1+, and that we have a dedicated Http.Header module, it makes more sense to specialise the internals of Http.Header to put "Host"...

I'm still not a huge fan of this approach, which doesn't seem to take advantage of what we _do_ know about the HTTP request structure. How about modifying Cohttp.Header to...

There are a few more constraints beyond just the Host header, aren't there? E.g. you must have a content-length or be chunked encoding. Or you can have an Upgrade field,...

While there's no callback for "post binding" in cohttp, you can avoid launching the binary as root on Linux by running: ``` setcap 'cap_net_bind_service=+ep' ``` This will then allow it...

Hannes, how important is it for you to use Fedora with these variants? We can promote an RPM distro to Tier1, but there's a significant ongoing cost in terms of...

Introducing a submodule sounds fine, and we can move it out to a toplevel Absolute_uri in the future. I'm not sure that `Absolute` is the right prefix though; perhaps `Uri_with_host`...

I see that Absolute_uri is a term from RFC9110, so we could just call this Uri.Absolute to indicate the mandatory presence of a non-empty path.

I think it might be worth separating the errors out for a deprecated user_info (which clients could validly choose to allow to pass through, after stripping it out) from the...