Fix outpoint_state index store

[zoedberg]

Found a bug calling outpoint_state. By calling the method with an empty Vec<OutPoint> it correctly returns known allocations, but by calling it with some specific outpoints it returns no allocations.

Investigating this I've found that when storing the index into the DB:

for seal in genesis.revealed_seals().unwrap_or_default() {
    debug!("Adding outpoint for seal {}", seal);
    let index_id = ChunkId::with_fixed_fragments(seal.txid, seal.vout);
    self.store.insert_into_set(db::OUTPOINTS, index_id, contract_id)?;
}

ChunkId::with_fixed_fragments receives a Option<Txid>.

While, when searching for specific outpoints:

let indexes = if outpoints.is_empty() {
    self.store.ids(db::OUTPOINTS)?
} else {
    outpoints
        .iter()
        .map(|outpoint| ChunkId::with_fixed_fragments(outpoint.txid, outpoint.vout))
        .collect()
};

ChunkId::with_fixed_fragments receives a Txid.

I've fixed the bug on the store-side by unwrapping the Option, since I believe the TXID should always be present at that point, I hope that's correct.

I manually tested it and can confirm that with this fix outpoint_state filter works as expected.

[dr-orlovsky]

Thank you for catching this!

While your fix is correct, rust stongly discourage use of unwraps in any production code; the unwrap_or, unwrap_or_default or expect should be used instead - or return an error.

I think we should do a expect("genesis with vout-based seal which passed schema validation"), to indicate the reason of the crash (genesis must validate absece of vout-based seals at the schema level, and if it is present, the line 167 and 181 above should guarantee that we would never reach this place).

Can you pls update your PR to use expect?

[zoedberg]

sure, thanks for the tip. fixed

[dr-orlovsky]

ACK e737bd6

[dr-orlovsky]

I think this is a breaking change that will invalidate the stash for existing RGB nodes. Will reflect in a release notes of the patch version...