Stub out facet-kdl

[fasterthanlime]

As discussed with @TheLostLambda in kdl-org/kdl-rs#83 (comment)

[fasterthanlime]

@TheLostLambda I'm sending you an invite to the org now — these aren't formalized anywhere yet but:

• Please go through PRs for most changes, just to get feedback from us folks (me for json, @tversteeg for TOML, @Veykril & @epage for pro bono consulting re: language and crate design respectively)
• Try to keep PRs relatively small if possible (although I expect the initial scaffolding to be big) — I'd rather merge a bunch of small PRs in a row than have a few big ones that are rotting away. Sometimes we need to change things in facet-reflect or facet-core for soundness reasons, better adapt the format crates heavy.
• There will be churn, we're still figuring out API surfaces, expect to have to follow the changes with facet-kdl — I'm happy to do some of it, we just have to coordinate
• Add yourself to the facet-kdl authors in the crate metadata once you've done something with it (also, @tversteeg plz do that to the TOML crate — make sure to keep me in the array if possible, I uhh have done work there)
• You can use jujutsu/jj, but make sure to run just precommit / just prepush / just ci (by increasing level of annoyingness but clairvoyance re: your future in CI) if you don't want to have a bad time when it comes to the merge queue.

That's about it! Excited to see where this goes and use KDL for my own stuff!

[TheLostLambda]

Thanks @fasterthanlime ! I should be able to throw together something in the next couple of days to give people a first look at! Even just doing something like differentiating between named and positional arguments (properties and arguments in KDL) will be farther than we've ever gotten with serde!