feat(chain): Introduce tx_graph::Update WIP

[LagginTimes]

Description

Introduces the tx_graph::Update struct, which will be used to update a TxGraph. This will replace the old method of updating a TxGraph with a TxGraph.

Notes to the reviewers

With the current implementation, there is some goofiness in the esplora/electrum wallet examples and test_electrum::sync_with_electrum(), where an Update is converted into a TxGraph to execute update_last_seen_unconfirmed(), before reverting the TxGraph back into an Update again. Implementing #1550 would prevent the need for this.

Changelog notice

WIP

Checklists

All Submissions:

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

[evanlinjin]

@LagginTimes forgot to mention that this is step 1 for #1543

[notmandatory]

I would like to avoid unnecessary refactoring like this for the 1.0 beta milestone, see my comment in #1543.

[evanlinjin]

@notmandatory I suggested this change because I thought it was necessary for 1.0. Please refer to @LLFourn's comment here: #1543 (comment)

[evanlinjin]

Remember that we will be moving Update into a separate crate bdk_core so we won't have access to private members.

Additionally, we don't want to provide floating outputs to our update logic when we already have the full tx in the same update struct.

With these two requirements in mind, let's see various solutions.

• Publicize all members. This is a bad idea, since we can't enforce our invariances.
• Enforce invariance when caller inserts data. When inserting full tx, rm all associated floating txouts of the tx's txid. Don't insert floating txout when there is a full tx of same txid.
• Enforce invariance when extracting data. When extracting full txs, also rm floating txouts with the same txid. When extracting floating txouts, ignore elements which have an associated full tx.

I think enforcing invariance when caller inserts data is easier to reason with.

[notmandatory]

This is replaced by #1568 right?

[evanlinjin]

Replaced by #1568