Add test for inserting anchor without tx

[remix7531]

Description

This adds a test case described in issue958

All Submissions:

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

[evanlinjin]

Thanks for the work so far. Remember that we also want to test recovering the appended additions into a new graph. Refer to #958 (comment)

[evanlinjin]

Good work so far. Few notes.

1. Ensure CI passes.
2. Ensure GitHub can verify your commits (Maybe you need to give your PGP or SSH public key to GitHub?).
3. Please rebase your commits into one.
4. Use loops where you can (for example if you are inserting multiple anchors or multiple txs).

[danielabrozzoni]

A few improvements can be made:

• Fix the warnings cargo clippy is throwing
• Use a more descriptive commit message. Instead of "tests for issue958", describe which test you're adding in the title, and add "Partially fixes #958" in the description. In general, it should be possible to understand what a commit is doing from the title/description without having to open github

The rest looks good to me.

[evanlinjin]

Thank you again, sorry I think it will be better to reduce the scope of this test.

[notmandatory]

Since it looks like still a bit of work today I'm moving this out of the alpha.1 milestone.

[remix7531]

I do not understand what is missing here. The workflows failed because of this
error: package log v0.4.19 cannot be built because it requires rustc 1.60.0 or newer, while the currently active rustc version is 1.57.0

Just tested the workflows locally and it seams to work now.

[danielabrozzoni]

Hey @remix7531, sorry for the late response! I think you encountered a problem with our CI, that's why it would work locally but fail here. Can you try rebasing?

[oleonardolima]

Is this something still needed for 1.0.0-beta milestone ? (cc @notmandatory @ValuedMammal)

I got a bit confused because I noticed that the #958 issue already has been closed but without the tests covered by this PR.

[ValuedMammal]

I think it wouldn't hurt to have a test similar to this

[notmandatory]

This testing could help with verifying sqlite schema question in #1712

[evanlinjin]

Tests for inserting anchor without tx already exist:

bdk/crates/chain/tests/test_tx_graph.rs

Lines 1074 to 1102 in f6c1c61

	fn insert_anchor_without_tx() {
	let mut graph = TxGraph::<BlockId>::default();
	let tx = new_tx(21);
	let txid = tx.compute_txid();
	let anchor = BlockId {
	height: 100,
	hash: hash!("A"),
	};
	// insert anchor with no corresponding tx
	let mut changeset = graph.insert_anchor(txid, anchor);
	assert!(changeset.anchors.contains(&(anchor, txid)));
	// recover from changeset
	let mut recovered = TxGraph::default();
	recovered.apply_changeset(changeset.clone());
	assert_eq!(recovered, graph);
	// now insert tx
	let tx = Arc::new(tx);
	let graph_changeset = graph.insert_tx(tx.clone());
	assert!(graph_changeset.txs.contains(&tx));
	changeset.merge(graph_changeset);
	// recover from changeset again
	let mut recovered = TxGraph::default();
	recovered.apply_changeset(changeset);
	assert_eq!(recovered, graph);
	}