feat!: remove send wait

[Thunkar]

Closes: https://linear.app/aztec-labs/issue/F-112/reevaluate-basecontractinteractionsend
Closes: #13895

This PR removes the infamous SendTx class and its derivations, avoiding the confusing syntax described in #13895 of

const tx = contract.methods.fn().send();
// tx is not actually sent here
console.log(await tx.getTxHash());
// tx is sent now? I just wanted the txhash!
await tx.wait();
// tx is confirmed 

sendTx now allows the consumer to specify wait options. Furthermore, we leverage typescript's type system to dynamically change the return type depending on the input options:

const receipt = await contract.methods.fn.send({ from: address }); // Result is TxReceipt
const txHash = await contract.methods.fn.send({ from: address, wait: false }); // Result is TxHash

A utility called waitForTx has been introduced so things like these are possible:

const txHash = await contract.methods.fn.send({ from: address, wait: false });
await waitForTx(aztecNode, txHash, { timeout: 300 });

All of this allows us to make the Wallet interface leaner too, so getTxReceipt has been removed from it.

[Thunkar]

These are ported from the old sent_tx.test.ts

[mverzilli]

I finally understand what deployed is 😆 ! Suggest to rename:

Suggested change

	public async deployed(options: DeployOptions): Promise<TContract> {
	public async deployedInstance(options: DeployOptions): Promise<TContract> {

Ack it's a bit more verbose but this doesn't seem something that people need to write all the time so I think the declarativeness is worth the mild extra typing

[Thunkar]

I'm leaning towards just killing it. It does not seem like a lot of gain when you can do:

const { contract } = await MyContract.deploy(wallet).send({ from: address });

[mverzilli]

oh I wholeheartedly agree 💯

[Thunkar]

Found an alternative approach:

const contract = await MyContract.deploy(wallet).send({ from: address });
const txHash = await MyContract.deploy(wallet).send({ from: address, wait: NO_WAIT });
const receipt = await MyContract.deploy(wallet).send({ from: address, wait: { returnReceipt: true }});

WDYT?

[mverzilli]

this type is what I would consider rewriting as a tagged union

[Thunkar]

WDYT about replacing the false here with a const expression?

type NO_WAIT = "NO_WAIT";

That way we don't need the builder pattern and the thing wouldn't be falsy

[mverzilli]

better than false definitely!

[mverzilli]

another issue with false is it's very natural for someone approaching this API to go like "wait, sure yeah:" => wait: true

[Thunkar]

This has been implemented as a symbol (which turns to a boolean during serialization and then back)
branded string

[mverzilli]

Colossal work! Sorry for the spam of API opinions, none of those should block this and in any case they can feed into future conversations

[mverzilli]

Very nice!

[AztecBot]

Flakey Tests

🤖 says: This CI run detected 2 tests that failed, but were tolerated due to a .test_patterns.yml entry.

\033FLAKED\033 (8;;http://ci.aztec-labs.com/c940ed46abcfdb19�c940ed46abcfdb198;;�):  yarn-project/end-to-end/scripts/run_test.sh simple src/e2e_epochs/epochs_invalidate_block.parallel.test.ts "proposer invalidates multiple blocks" (126s) (code: 1) group:e2e-p2p-epoch-flakes (\033Aztec Bot\033: feat!: remove send wait (#19778))
\033FLAKED\033 (8;;http://ci.aztec-labs.com/059ef1534c86d246�059ef1534c86d2468;;�):  yarn-project/end-to-end/scripts/run_test.sh simple src/e2e_p2p/valid_epoch_pruned_slash.test.ts (394s) (code: 1) group:e2e-p2p-epoch-flakes (\033Aztec Bot\033: feat!: remove send wait (#19778))