add: support for getblocktemplate in template mode

[0xB10C]

Partly addresses #133

This adds basic support for the getblocktemplate RCP to the client. The default 'template' mode, in which Bitcoin Core builds and returns a block template to the RPC client, is implemented. The 'proposal' mode, where the client submits a block to Bitcoin Core for validation (before it's mined; proof-of-work is not checked) as well as longpolling for changes to a given block template are not implemented yet. However, the basic implementation proposed here can be extended to support both of these features without changing the API.

I've focused on the subset of features (also called "capabilities") that Bitcoin Core actually implements from BIP 22 and BIP 23. I've tested this against mainnet, testnet, regtest and signet with the Rust code provided below.

Reviewers might find the following resources useful:

• getblocktemplate() in bitcoin/bitcoin/src/rpc/mining.cpp @ master
• BIP-22: getblocktemplate - Fundamentals: Non Optional Sections
• BIP-23: getblocktemplate - Pooled Mining: Mutations

Code for playing around with this

use bitcoincore_rpc::json::{GetBlockTemplateOptionsModes, GetBlockTemplateOptionsRules};
use bitcoincore_rpc::{Auth, Client, RpcApi};

fn main() {
    // mainnet rpc port: 8332
    // testnet rpc port: 18332
    // regtest rpc port: 18444
    // signet rpc port: 38332
    let rpc = Client::new(
        "http://127.0.0.1:8332".to_string(),
        Auth::UserPass("your user".to_string(), "your pass".to_string()),
    )
    .unwrap();

    let get_block_template = rpc
        .get_block_template(
            GetBlockTemplateOptionsModes::Template,
            vec![
                GetBlockTemplateOptionsRules::SegWit,
                // GetBlockTemplateOptionsRules::Signet, for signet
            ],
            vec![],
        )
        .unwrap();

    println!("template: {:#?}", get_block_template);
}

[0xB10C]

The one CI job failing seems to be unrelated. Happens on master as well.

[stevenroose]

Left some comments, looks quite good!

[stevenroose]

Could you also rebase to make CI happy?

[0xB10C]

Thanks! Addressed your review and rebased.

[stevenroose]

LGTM

[stevenroose]

Perhaps @shesek or @sgeisler could review too. One small nit that could be improved would be changing the Vec function arguments with &[ ] slices.

[shesek]

The PR looks great! Thanks for including doc comments for all the fields, the resource links in the PR's description and the example code.

My only nit is what @stevenroose already brought up regarding slice arguments.

[0xB10C]

Pushed f36e4e5 which has slice arguments for the get_block_template function. Wasn't to sure about to_vec(), but happy to squash if that's what you meant.

edit: forgot to adapt the tests in previously pushed commit

[stevenroose]

One more nit :)

[0xB10C]

One more nit :)

Done!

[stevenroose]

ACK, let's await CI

[0xB10C]

The failing job seems to be unrelated

[stevenroose]

thanks!