Test conflicting tx situations.

[rajarshimaitra]

Description

partial fixes #958 . These tests attempt to cover all the conflicting scenarios.

The following tests are included:

• Mempool conflicts are chosen by the latest last seen.
• Conflicts in mempool and orphaned block filters out the orphaned block tx.
• Conflicts with mempool and confirmed tx, filters out mempool.
• Conflicts with two mempool tx at same seen_at.

Some helper functions in the common module.

Notes to the reviewers

As noted by @evanlinjin on the test description.

What happens if the conflicts have the same last_seen value?

A possible solution for this situation is mentioned in the comment of the last test, which covers this case. I will attempt a PR to enforce the behavior if that is a reasonable approach.

Changelog notice

Checklists

All Submissions:

• I've signed all my commits
• I followed the contribution guidelines
• I ran cargo fmt and cargo clippy before committing

[notmandatory]

Looks like this one needs a cargo fmt .. otherwise is it ready to review or dependent on #963 ?

[evanlinjin]

Looks like this one needs a cargo fmt .. otherwise is it ready to review or dependent on #963 ?

No dependencies, I saw that CI wasn't passing so didn't review.

[rajarshimaitra]

My bad. Had the fmt fix, but forgot to push. This had a dependency on #963 as the Anchor type changed. Fixed now..

[evanlinjin]

Will look at this first thing tomorrow.

[evanlinjin]

Why could we not assert all of utxo?

[rajarshimaitra]

We can, but will make the tests code too verbose. The point here is to test various txout positions. If we wanna test the whole structure of FullTxOut, could be done in a separate test.

[evanlinjin]

@rajarshimaitra as we discussed.

struct TxTemplate<'a, K> {
    pub tx_name: &'a str,
    pub spends: &'a [(&'a str, usize)], // (tx_name, vout)
    pub outputs: &'a [TxOutTemplate<K>],
    pub anchors: &'a [(u32, BlockId)],
    pub last_seen: Option<u64>,
}

struct TxOutTemplate<K> {
    pub value: u64,
    pub owned_spk: Option<(K, u32)>, // some = derive a spk from K, none = random spk
}

fn init_graph<'a, K: 'a>(
    descriptors: HashMap<K, Descriptor<DescriptorPublicKey>>,
    tx_templates: impl IntoIterator<Item = TxTemplate<'a, K>>,
) -> (
    IndexedTxGraph<ConfirmationHeightAnchor, KeychainTxOutIndex<K>>,
    HashMap<&'a str, Txid>,
) {
    todo!()
}

#[test]
fn example_test() {
    struct TestCase<'a, K> {
        tx_templates: &'a [TxTemplate<'a, K>],
        chain_tip: BlockId,
        expected_list_chain_txs: (),
    }
}

[evanlinjin]

@LagginTimes Could you take over on this?

[LagginTimes]

@LagginTimes Could you take over on this?

Sure, I can start looking into this.

[notmandatory]

@LagginTimes any updates on this? should I keep it in the alpha.1 milestone?

[evanlinjin]

Replaced by #1064