scan-rpc: ignores wallet birthday

[GuiSchet]

I've been practicing on signet and got stuck here: scan-rpc just ignores the wallet's birthday and scans from genesis.

In cli/v2/src/main.rs, scan-rpc constructs the Emitter with start_height = 0:

// cli/v2/src/main.rs
Commands::ScanRpc { rpc_args } => {
    let rpc_client = rpc_args.new_client()?;
    let mut emitter = Emitter::new(
        &rpc_client,
        wallet.chain().tip(),
        0,                         
        NO_EXPECTED_MEMPOOL_TXIDS,
    );

scan-cbf in the same file already handles this correctly — it compares the birthday against the chain tip and uses whichever is higher:

// cli/v2/src/main.rs
let sync_point = if let (Some(height), Some(hash)) = (height, hash) {
    HeaderCheckpoint::new(height, hash)
} else if wallet.birthday.height <= wallet.chain().tip().height() {
    HeaderCheckpoint::new(wallet.chain().tip().height(), wallet.chain().tip().hash())
} else {
    HeaderCheckpoint::new(wallet.birthday.height, wallet.birthday.hash)
};

scan-rpc has no equivalent logic.

Reproduction

On signet, create a wallet with a birthday near the chain tip, then run scan-rpc.
The output will show the scanner starting from a very low block height and incrementing block by block through the entire chain history rather than jumping to the birthday height.

[nymius]

Thanks for reporting it. I've checked it locally and you're right.

[GuiSchet]

I'd be happy to work on a PR for this if you think it's a worthwhile change. Let me know!

[nymius]

It's yours. A trace of execution with the before and after in the PR would be handy.

[GuiSchet]

A few related improvements surfaced while working on the fix. Sharing in case any are worth tracking:

1. Mirror scan-cbf's --height / --hash override flags on scan-rpc.
2. Log the start point on scan-rpc (scan-cbf already does it via tracing::info!("Synchronizing from block ...")) — would've made this bug obvious from the first line of output.
3. Extract a shared start-point helper between scan-cbf and scan-rpc — the drift here is exactly what duplication tends to produce.
4. cli/v2 is currently excluded from just check/just test and from default clippy (only default-members get linted), so the workspace's print_stdout/print_stderr ban is silently waived. There's also a pre-existing clippy::question_mark at scan-cbf:413 living unnoticed because of this.

Happy to open separate issues for whichever you'd like to track — none of them block the current PR.

[nymius]

1. and 4. can be tackled in a different PR. I think they don't hurt.
2. Can be done in fix: scan-rpc ignores wallet birthday #60.
3. I explicitly left out the cli crates from the checks and tests because when I was actively working on them, they were intended more as a showcase of the happy paths of the library. So it isn't a robust implementation. But improvements here are welcomed.