David Sheets

Could you create a type to encapsulate the percent-encoded set? I believe it is a property of the serialization of the URI only and not inherent to the identifier (but...

This is an API to consider for the fabled 2.0 release.

I thought a pretty minimal regex parser for this made it in a while ago... This will happen after it gets into ocaml-ipaddr which will take it to 2.0. Then,...

I think there is some place for a type that sums addresses and host names. End users shouldn't (necessarily) have to consider that the host string that they get back...

See 1ef3f1dfb41bdb4f33f223ffe16e62a33975661a which proposes also to escape `!`, `$`, and `,`. Are these query string conventions scheme-specific?

This is a longstanding issue which is quite simple to fix but requires breaking interface compatibility (similar to most of the other awaited changes). Unfortunately, there is no way to...

It looks like we would need some way to control the literalness of encoding particularly for the unencoded set and '/'. The '/' issue is well-known but this is the...

Are you proposing the opposite convention of the initial proposal? I don't understand your comment about giving up early. If renames clash, aliases can't preserve compat... I think the idea...

Yes, query is opaque and these functions should be separated.

Agreed. Most likely we should expose the internal structure of the library which has each of the syntactic components as nested modules. When this design is refined for 2.0, a...