Add submitpackage method to docs

[f321x]

Adds the submitpackage method to the protocol documentation.
See e-x implementation in spesmilo/electrumx#288
Note that the function signature contains a semicolon after [raw_txs]; as sphinx otherwise shows it incorrectly as [raw_txs,] in the generated html.

[t-bast]

Thanks for adding this 🙏

[SomberNight]

note: the bitcoind response lists both txid and wtxid - although it keys results based on wtxid. In this use case either one is sufficient. There can be no collision of txids within a package.
The electrum protocol uses txids generally, I think we can keep to that and do the same here.

[t-bast]

That sounds good, I don't see any issue with using txid either.

[SomberNight]

I expect most clients to want the default, simple, verbose=false option.

However, AFAICT it does not cost a server implementation ~anything to offer the rich bitcoind response. The server has to call the submitpackage bitcoind RPC anyway. (Contrast this with e.g. blockchain.transaction.get, where the verbose=false option is implemented by e.g. esplora without calling bitcoind, while the verbose=true option must call bitcoind)

Having access to the full bitcoind result can be useful e.g. if you trust the server, or for quick scripts. And maybe also for some unforeseen use-cases where the verbose=false response is not sufficient, the rich result can be used a stopgap.

So I think we can also add this verbose=true variant. (with this explicit caveat that its structure is not guaranteed)

[t-bast]

Fair enough!

[t-bast]

LGTM 👍

[tnull]

Any update on this?

[SomberNight]

I cherry-picked and squashed this into #6, which is now a concrete minimal proposal for a new protocol version that I intend to release quickly (unless a valid objection comes up).