WIP: nanoarrow refactor

[kylebarron]

Closes #41, closes #42

Change list:

• Parse C FFI structs into plain JS objects. Then these intermediate objects can be stored on their own or converted to Arrow JS
• Add writing support for writing the JS representation of these structs into Wasm memory directly, using a malloc function exported from wasm.

Todo:

• Support converting Arrow JS objects into these structs.
• Am I going to hit an issue with setting the release callback to the null pointer? It works with arrow2, but that might be implementation specific.

[brettsaunders21]

Do you have plans to continue working on this feature and merge it eventually? Or was it just an experiment?

[kylebarron]

Which part? The nanoarrow refactor or writing to wasm memory? This PR is kinda intertwined among the two. The nanoarrow refactor is less likely to get merged; the main goal of that is to remove an Arrow JS dependency, but I don't think I have the development bandwidth to maintain a separate (even extremely minimal) Arrow implementation, so I'll probably shelve that. The writing part is useful even with Arrow JS and I'd like to have that merged, though I don't have a use case for it myself yet.

[kylebarron]

If there's any of this that you'd like to work on, feel free to discuss more and/or make a PR

[kylebarron]

I'm going to close this because I don't have the maintenance bandwidth to try and implement data structures for Arrow outside of Arrow JS, and I don't have a use case at this point where Arrow JS's bundle size is a deal-breaker.