feat(test-benchmark): validate post-state for stateful filler (#3148)

* feat: add post-state verification to stateful filler

* fix(test-plugins): Fix types, simplify

---------

Co-authored-by: Mario Vega <marioevz@gmail.com>

Author: LouisTsai-Csie

packages/testing/src/execution_testing/cli/pytest_commands/plugins/execute/execute.py

@@ -10,13 +10,13 @@
 from pytest_metadata.plugin import metadata_key
 
 from execution_testing.base_types import Account
-from execution_testing.base_types import Alloc as BaseAlloc
 from execution_testing.base_types.base_types import HexNumber
 from execution_testing.execution import BaseExecute
 from execution_testing.forks import Fork, TransitionFork
 from execution_testing.logging import get_logger
 from execution_testing.rpc import EngineRPC, EthRPC
 from execution_testing.specs import BaseTest
+from execution_testing.test_types import Alloc as BaseAlloc
 from execution_testing.test_types import (
     Environment,
     EnvironmentDefaults,

packages/testing/src/execution_testing/cli/pytest_commands/plugins/execute/pre_alloc.py

@@ -20,9 +20,6 @@
     Storage,
     StorageRootType,
 )
-from execution_testing.base_types import (
-    Alloc as BaseAlloc,
-)
 from execution_testing.base_types.conversions import (
     BytesConvertible,
     NumberConvertible,
@@ -42,6 +39,7 @@
     TransactionTestMetadata,
     compute_deterministic_create2_address,
 )
+from execution_testing.test_types import Alloc as BaseAlloc
 from execution_testing.tools import Initcode
 from execution_testing.vm import Bytecode, Op
 

packages/testing/src/execution_testing/cli/pytest_commands/plugins/shared/execute_fill.py

@@ -9,11 +9,7 @@
 from pytest import StashKey
 
 from execution_testing.base_types import Account, Number
-from execution_testing.base_types import Alloc as BaseAlloc
-from execution_testing.execution import (
-    BaseExecute,
-    LabeledExecuteFormat,
-)
+from execution_testing.execution import BaseExecute, LabeledExecuteFormat
 from execution_testing.fixtures import BaseFixture, LabeledFixtureFormat
 from execution_testing.logging import get_logger
 from execution_testing.rpc import EthRPC
@@ -65,7 +61,7 @@ def _validate_and_cache_address_stubs(
     eth_rpc = EthRPC(rpc_endpoint)
     labels = list(address_stubs.root.keys())
     addresses = [address_stubs.root[k].addr for k in labels]
-    query = BaseAlloc(root={addr: Account() for addr in addresses})
+    query = Alloc(root={addr: Account() for addr in addresses})
     alloc = eth_rpc.get_alloc(query)
     empty: list[str] = []
     accounts: Dict[str, Account] = {}

packages/testing/src/execution_testing/client_clis/client_backend.py

@@ -20,6 +20,7 @@
 from execution_testing.forks import Fork, TransitionFork
 from execution_testing.logging import get_logger
 from execution_testing.rpc import (
+    BlockNumberType,
     DebugRPC,
     EngineRPC,
     EthRPC,
@@ -33,7 +34,12 @@
     PayloadAttributes,
     PayloadStatusEnum,
 )
-from execution_testing.test_types import Requests, Transaction, Withdrawal
+from execution_testing.test_types import (
+    Alloc,
+    Requests,
+    Transaction,
+    Withdrawal,
+)
 from execution_testing.test_types.block_access_list import BlockAccessList
 from execution_testing.test_types.receipt_types import TransactionReceipt
 
@@ -306,6 +312,12 @@ def evaluate(
             ),
         )
 
+    def get_post_state_alloc(
+        self, expected: Alloc, *, block_number: BlockNumberType = "latest"
+    ) -> Alloc:
+        """Fetch the post-state that ``expected`` constrains from client."""
+        return self.eth_rpc.get_alloc(expected, block_number=block_number)
+
     def extract_block_opcode_count(
         self, block_hash: Hash
     ) -> OpcodeCount | None:

packages/testing/src/execution_testing/execution/transaction_post.py

@@ -5,7 +5,7 @@
 import pytest
 from pytest import FixtureRequest
 
-from execution_testing.base_types import Address, Alloc, Hash
+from execution_testing.base_types import Address, Hash
 from execution_testing.forks import Fork
 from execution_testing.logging import get_logger
 from execution_testing.rpc import (
@@ -14,6 +14,7 @@
     SendTransactionExceptionError,
 )
 from execution_testing.test_types import (
+    Alloc,
     Environment,
     NetworkWrappedTransaction,
     TestPhase,

packages/testing/src/execution_testing/rpc/rpc.py

@@ -37,14 +37,14 @@
 from execution_testing.base_types import (
     Account,
     Address,
-    Alloc,
     Bytes,
     Hash,
     to_json,
 )
 from execution_testing.logging import (
     get_logger,
 )
+from execution_testing.test_types import Alloc
 
 from .rpc_types import (
     EthConfigResponse,

packages/testing/src/execution_testing/specs/blockchain.py

@@ -1384,7 +1384,8 @@ def make_stateful_fixture(
         - Payloads are partitioned by ``FixtureEngineNewPayload.phase``
           into ``setup_payloads`` (setup-phase txs) and ``payloads``
           (execution-phase txs).
-        - ``verify_post_state`` is skipped: the client is the oracle.
+        - ``post`` is verified against the live client (the oracle) via
+          ``get_post_state_alloc``; there is no t8n post alloc to diff.
         """
         if not isinstance(t8n, ClientBackend):
             raise RuntimeError(
@@ -1557,6 +1558,10 @@ def make_stateful_fixture(
             )
             head_hash = client_hash
 
+        if self.post.root:
+            got_alloc = t8n.get_post_state_alloc(self.post)
+            self.post.verify_post_alloc(got_alloc)
+
         fixture = BlockchainEngineStatefulFixture(
             fork=self.fork,
             last_block_hash=head_hash,

packages/testing/src/execution_testing/test_types/account_types.py

@@ -324,6 +324,13 @@ def __contains__(
             address = Address(address)
         return address in self.root
 
+    def get(self, address: Address) -> Account | None:
+        """Get an account if it's present in the allocation, otherwise None."""
+        account = self.root.get(address)
+        if not account:
+            return None
+        return account
+
     def empty_accounts(self) -> List[Address]:
         """Return list of addresses of empty accounts."""
         return [
@@ -372,12 +379,10 @@ def verify_post_alloc(self, got_alloc: "Alloc") -> None:
         for address, account in self.root.items():
             if account is None:
                 # Account must not exist
-                if (
-                    address in got_alloc.root
-                    and got_alloc.root[address] is not None
-                ):
+                got_account = got_alloc.get(address)
+                if got_account:
                     raise Alloc.UnexpectedAccountError(
-                        address=address, account=got_alloc.root[address]
+                        address=address, account=got_account
                     )
             else:
                 if address in got_alloc.root: