feat: expose ElectrumClient::block_headers_subscribe method

[thunderbiscuit]

This is a bit of a cheat: the method is not intended for a 1-time use, but rather is meant to open a subscription. In any case the Electrum client does return what I think our user wants (LDK-node please confirm!).

I'm not ready to merge this one just yet because I actually didn't expose the Header type fully (which would require 4 new types: Version, BlockHash, TxMerkleNode, and CompactTarget). If we do want to expose the block_header() method, then it might actually be worth taking the time exposing the whole Header type from rust-bitcoin.

Let me know if you have an opinion either way, because this would be a breaking change if we expose a simplified version and eventually want to return the full Header.

Changelog

Added:
    - Expose ElectrumClient::block_headers_subscribe method [#664]

[#664]: https://github.com/bitcoindevkit/bdk-ffi/pull/664

[reez]

I'm not ready to merge this one just yet because I actually didn't expose the Header type fully (which would require 4 new types: Version, BlockHash, TxMerkleNode, and CompactTarget). If we do want to expose the block_header() method, then it might actually be worth taking the time exposing the whole Header type from rust-bitcoin.

Let me know if you have an opinion either way, because this would be a breaking change if we expose a simplified version and eventually want to return the full Header.

Thanks for getting this PR going!

Yeah this solves the specific use case by providing the information needed/asked. It would be great to expose Header fully, but would also definitely be nice to have other specific use cases to help clarify how important it is to expose Header fully, I'm personally not using Electrum nearly as much as Esplora so I don't personally have a strong feeling (balancing this is what known user needs vs default of opening up types fully + not having future breaking changes)

[thunderbiscuit]

We should also note that the Electrum client is not part of our "semantic 1.0", so if we do end up with small breaking changes there it's ok. Doesn't mean we shouldn't try to get it right on the first shot, but I just thought I'd mention it.