Use StakeTableV2 and new events in the GCL #3289

sveitser · 2025-05-19T15:24:11Z

Verify signatures on new events in GCL.
Deploy StakeTableV2
Allow CLI to register and update keys with both versions of the contract for backwards compat.
Run all stake table CLI tests against the V1 and V2 contract.

What's not done

There's currently no E2E test where we start on the original stake table and move to V2 and generate all the events for that. I think we can consider adding it but we are missing good testing infra for doing all of this, so I think it's better to work on this as a separate PR since this one is already much bigger than anticipated.

To see the specific tasks where the Asana app for GitHub is being used, see below:
- https://app.asana.com/0/0/1210174253354699
- https://app.asana.com/0/0/1210174253354700

I mainly want to test if the asana linking from github works https://app.asana.com/1/1208976916964769/project/1210098463466349/task/1210174253354698

Using current jellyfish APIs it's not possible to construct the type currently used. Since the contract doesn't need it we can send a bytes type instead.

contracts/rust/adapter/src/sol_types.rs

…ures

This isn't complete yet, but it compiles.

sveitser · 2025-05-22T12:30:25Z

types/src/v0/impls/stake_table.rs

+
+                let stake_table_key: BLSPubKey = blsVK.clone().into();
+                let state_ver_key: SchnorrPubKey = schnorrVK.clone().into();
+                // TODO(MA): The stake table contract currently enforces that each bls key is only used once. We will


Need to update this comment

contracts/rust/deployer/src/lib.rs

staking-cli/src/registration.rs

staking-cli/src/deploy.rs

alxiong · 2025-05-22T13:55:56Z

contracts/rust/adapter/src/sol_types.rs

+impl Serialize for ValidatorRegisteredV2 {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>


why are we adding on this? so .abi_encode() is not sufficient?

We store the events locally in the DB for that I think they are wrapped into an enum.

espresso-network/types/src/v0/v0_3/stake_table.rs

Lines 96 to 105 in 7617c22

#[derive(Clone, derive_more::From, PartialEq, serde::Serialize, serde::Deserialize)]

pub enum StakeTableEvent {

Register(ValidatorRegistered),

RegisterV2(ValidatorRegisteredV2),

Deregister(ValidatorExit),

Delegate(Delegated),

Undelegate(Undelegated),

KeyUpdate(ConsensusKeysUpdated),

KeyUpdateV2(ConsensusKeysUpdatedV2),

}

It would probably also work with abi_encode / decode but to change now we could have to migrate storage of any nodes that have this deployed already. We would still have to implement serde I think to work with our storage code.

Correct me if I'm wrong @imabdulbasit .

Foundry has a workaround for the lack of function overloading in rust that leads to confusing function names in the bindings. Avoiding the function name collison makes the workaround unnecessary.

…ures

- Add tests with V2 register and key update events. - Fix bug where update keys does not check for key-reuse, with regression test. - Make types a bit more convenient with From and Copy impls.

sveitser · 2025-05-27T14:06:46Z

types/src/v0/impls/stake_table.rs

                if bls_keys.contains(&stake_table_key) {
-                    bail!("bls key {} already used", stake_table_key.to_string());
+                    bail!("bls key already used: {}", stake_table_key.to_string());


The diff is a bit misleading here. For the V1 there shouldn't be any changes other than doing the formatting for errors differently and removing some clones because some types are Copy now. I guess since the two branches for v1 and v2 are so similar the diff got cut in half.

alysiahuggins

lgtm with a few clarifying qs

alysiahuggins · 2025-05-27T14:24:10Z

contracts/src/StakeTableV2.sol

-        BN254.G2Point blsVk,
-        EdOnBN254.EdOnBN254Point schnorrVk,
+        BN254.G2Point blsVK,
+        EdOnBN254.EdOnBN254Point schnorrVK,
        uint16 commission,
        BN254.G1Point blsSig,


since we made schnorrSig bytes, any chance it makes sense to do so for blsSig too? cc @alxiong

I actually don't like to use Bytes (anymore) if we have a type of known size because it means we can't make the type Copy in rust.

But IIRC we are missing some code so if we don't use bytes things would have gotten more complicated.

personally I like bytes because it's more future proof. but whenever we want to do something with these objects in the contract, we should consider giving them a concrete types.

In this case, we do verify the BLS signature, thus giving it a type is better IMO.

I hear Mathis's complain about no Copy for bytes, but my personal tolerance for Clone is higher and I can live with lots of .clone() everywhere, (clippy will inform us about unnecessary clone once alloy bindings decide to add Copy on them)

alysiahuggins · 2025-05-27T14:26:53Z

contracts/rust/adapter/src/copy.rs

+impl Copy for crate::sol_types::G1PointSol {}
+impl Copy for crate::sol_types::G2PointSol {}
+impl Copy for crate::sol_types::EdOnBN254PointSol {}
+impl Copy for crate::sol_types::StakeTableV2::ValidatorRegistered {}


not totally sure but would this also need to be called ValidatorRegisteredV2?

No this is the the V1 event defined in the V2 abi. The rust code is now written to always use the V2 ABI/bindings (except for when deploying the original StakeTable contract) because the V2 ABI it's a superset of the V1 ABI. I think I left some comments about this in the stake table contract.

alysiahuggins · 2025-05-27T14:29:19Z

contracts/rust/adapter/src/stake_table.rs

+            sigma: sigma_affine.into_group(),
+        }
+    };
+    if !bls_vk.validate(&sig, &msg) {


I might have missed it but do we have a test for bls vk validation against its sigs?

You mean unittests? I didn't add them because @alxiong wrote it, so it should be correct 😂 added some unittests here: 985ce8d

contracts/rust/adapter/src/stake_table.rs

contracts/rust/adapter/src/sol_types.rs

alxiong · 2025-05-28T08:32:45Z

contracts/rust/deployer/Cargo.toml

@@ -14,6 +14,7 @@ derive_more = { workspace = true }
 espresso-types = { path = "../../../types" }
 hotshot-contract-adapter = { workspace = true }
 hotshot-types = { workspace = true }
+sequencer-utils = { version = "0.1.0", path = "../../../utils" }


Only test-utils? So dev-dependency?

contracts/rust/deployer/src/builder.rs

…ures

staking-cli/src/registration.rs

alxiong · 2025-05-28T10:11:23Z

staking-cli/tests/cli.rs

+#[case::v1(StakeTableContractVersion::V1)]
+#[case::v2(StakeTableContractVersion::V1)]


Are they supposed to be the same?

uff, no, thanks

I haven't found a way to sort of "inherit" the test name from the struct. The test function name is more explicit if #[values] are used such as here:

espresso-network/staking-cli/tests/cli.rs

Lines 361 to 372 in ebe7026

#[rstest::rstest]

#[tokio::test]

async fn stake_for_demo_delegation_config_helper(

#[values(StakeTableContractVersion::V1, StakeTableContractVersion::V2)]

version: StakeTableContractVersion,

#[values(

DelegationConfig::EqualAmounts,

DelegationConfig::VariableAmounts,

DelegationConfig::MultipleDelegators

)]

config: DelegationConfig,

) -> Result<()> {

test stake_for_demo_delegation_config_helper::version_2_StakeTableContractVersion__V2::config_1_DelegationConfig__EqualAmounts ... ok test stake_for_demo_delegation_config_helper::version_2_StakeTableContractVersion__V2::config_2_DelegationConfig__VariableAmounts ... ok test stake_for_demo_delegation_config_helper::version_1_StakeTableContractVersion__V1::config_1_DelegationConfig__EqualAmounts ... ok test stake_for_demo_delegation_config_helper::version_1_StakeTableContractVersion__V1::config_3_DelegationConfig__MultipleDelegators ... ok test stake_for_demo_delegation_config_helper::version_2_StakeTableContractVersion__V2::config_3_DelegationConfig__MultipleDelegators ... ok

But for the applied template it just uses case_1 and case_2 by default.

Base automatically changed from 3267-stake-table-v2 to main May 19, 2025 16:34

WIP: use StakeTableV2

db1ed76

I mainly want to test if the asana linking from github works https://app.asana.com/1/1208976916964769/project/1210098463466349/task/1210174253354698

sveitser mentioned this pull request May 20, 2025

Functions overwritten in StakeTableV2 should revert if the V1 functions are called #3293

Closed

Use bytes for schnorr sig in stake table contract

c789fc2

Using current jellyfish APIs it's not possible to construct the type currently used. Since the contract doesn't need it we can send a bytes type instead.

sveitser force-pushed the ma/staking-cli-signatures branch from 7a67542 to c789fc2 Compare May 20, 2025 14:38

sveitser commented May 21, 2025

View reviewed changes

contracts/rust/adapter/src/sol_types.rs Show resolved Hide resolved

alxiong and others added 5 commits May 22, 2025 15:51

add bls key verify function

0166bb0

Merge remote-tracking branch 'origin/main' into ma/staking-cli-signat…

fb2ca09

…ures

stake table: verify validator event signatures

a056f32

This isn't complete yet, but it compiles.

test: deploy stake table upgrade

d8d7c1f

stake table: verify event signatures in GCL

6c46aea

sveitser commented May 22, 2025

View reviewed changes

clippy

20a17ad

alysiahuggins reviewed May 22, 2025

View reviewed changes

contracts/rust/deployer/src/lib.rs Show resolved Hide resolved

alysiahuggins reviewed May 22, 2025

View reviewed changes

staking-cli/src/registration.rs Show resolved Hide resolved

alysiahuggins reviewed May 22, 2025

View reviewed changes

staking-cli/src/registration.rs Outdated Show resolved Hide resolved

alysiahuggins reviewed May 22, 2025

View reviewed changes

staking-cli/src/deploy.rs Outdated Show resolved Hide resolved

alxiong reviewed May 22, 2025

View reviewed changes

sveitser added 12 commits May 23, 2025 10:51

use stake-table v2 for local demo

8a0e31e

update comment

7617c22

fix: append v2 to new functions in v2 contract

afcb93c

Foundry has a workaround for the lack of function overloading in rust that leads to confusing function names in the bindings. Avoiding the function name collison makes the workaround unnecessary.

fix: assert version is 2 after upgrade

05e7511

use deployer code in staking-cli test system

6e5b45e

staking-cli: allow using stake table V1 in tests

fa60b98

test: run staking-cli against contract v1 and v2

b027d6f

test: use rstest to parametrize tests

a659eee

Merge remote-tracking branch 'origin/main' into ma/staking-cli-signat…

fd2889e

…ures

fix: contract upgrade test

8fc8a52

cleanup and documentation

3f86bdf

espresso-types: stake table v2 unit tests

ad7c6b9

- Add tests with V2 register and key update events. - Fix bug where update keys does not check for key-reuse, with regression test. - Make types a bit more convenient with From and Copy impls.

test: filter unauthenticated stake table events

294d222

sveitser changed the title ~~WIP: use StakeTableV2~~ Use StakeTableV2 and new events in the GCL May 27, 2025

update comments

0584e25

sveitser marked this pull request as ready for review May 27, 2025 13:56

sveitser requested review from jbearer, tbro, imabdulbasit, ss-es, pls148, bfish713, rob-maron and lukaszrzasik as code owners May 27, 2025 13:56

fix TODO comment

a756313

sveitser commented May 27, 2025

View reviewed changes

alysiahuggins approved these changes May 27, 2025

View reviewed changes

sveitser added 2 commits May 27, 2025 16:44

fix clippy

d84e850

add unittests for bls sig verification

985ce8d

sveitser commented May 27, 2025

View reviewed changes

contracts/rust/adapter/src/stake_table.rs Outdated Show resolved Hide resolved

remove unused error variants

62f1923

alxiong reviewed May 28, 2025

View reviewed changes

sveitser added 3 commits May 28, 2025 11:45

move copy impl to copy.rs

521a764

fix: maybe better function, plus doc comment

e43169e

Merge remote-tracking branch 'origin/main' into ma/staking-cli-signat…

5011b37

…ures

alxiong reviewed May 28, 2025

View reviewed changes

sveitser added 4 commits May 28, 2025 12:24

fix: staking-cli, actually use V2

06144cc

staking-cli: be more explicit about ABI compat

ebe7026

fix: move dev dependency to dev-dependencies

8bf2101

fix: sqlite lock file

2a4749f

sveitser merged commit 21ff7c0 into main May 28, 2025
102 checks passed

sveitser deleted the ma/staking-cli-signatures branch May 28, 2025 12:52

		impl Serialize for ValidatorRegisteredV2 {
		fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>

	#[derive(Clone, derive_more::From, PartialEq, serde::Serialize, serde::Deserialize)]
	pub enum StakeTableEvent {
	Register(ValidatorRegistered),
	RegisterV2(ValidatorRegisteredV2),
	Deregister(ValidatorExit),
	Delegate(Delegated),
	Undelegate(Undelegated),
	KeyUpdate(ConsensusKeysUpdated),
	KeyUpdateV2(ConsensusKeysUpdatedV2),
	}

		#[case::v1(StakeTableContractVersion::V1)]
		#[case::v2(StakeTableContractVersion::V1)]

	#[rstest::rstest]
	#[tokio::test]
	async fn stake_for_demo_delegation_config_helper(
	#[values(StakeTableContractVersion::V1, StakeTableContractVersion::V2)]
	version: StakeTableContractVersion,
	#[values(
	DelegationConfig::EqualAmounts,
	DelegationConfig::VariableAmounts,
	DelegationConfig::MultipleDelegators
	)]
	config: DelegationConfig,
	) -> Result<()> {

Use StakeTableV2 and new events in the GCL #3289

Use StakeTableV2 and new events in the GCL #3289

Uh oh!

Conversation

sveitser commented May 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What's not done

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alysiahuggins left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sveitser May 27, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

sveitser commented May 19, 2025 •

edited

Loading

sveitser May 27, 2025 •

edited

Loading