fix(execute): add gas validation for benchmark tests in execute mode (#2219)

* fix(execute): add gas validation for benchmark tests in execute mode

Previously, execute mode was not validating that transactions consumed
the expected amount of gas when expected_benchmark_gas_used was set.
This could cause benchmark tests to incorrectly pass even when consuming
significantly less gas than expected (e.g., due to missing factory contracts).
This feature is needed by benchmark tests like the ones in #2186 in
order to make sure that the benchmarks are indeed consuming all gas
available or causing a failure otherwise when the flag is set.

Changes:
- Add expected_benchmark_gas_used and skip_gas_used_validation fields to TransactionPost
- Implement gas validation logic in TransactionPost.execute() using transaction receipts
- Pass gas validation parameters from StateTest and BlockchainTest to TransactionPost
- Add eth_getTransactionReceipt RPC method to fetch gas used from receipts

This ensures benchmark tests fail appropriately when gas consumption doesn't match
expectations, preventing false positives in performance testing.

* refactor(execute): simplify gas validation implementation

Addresses review comment to make execute mode gas validation cleaner:
- Set expected_benchmark_gas_used to gas_benchmark_value as default in execute parametrizer
- Remove gas_benchmark_value parameter from TransactionPost, StateTest, BlockchainTest, and BaseTest
- Simplify gas validation logic in TransactionPost

This ensures consistent gas validation behavior between fill and execute modes
with a cleaner implementation that sets defaults at the parametrizer level.

Author: CPerezz

src/ethereum_test_execution/transaction_post.py

@@ -20,6 +20,9 @@ class TransactionPost(BaseExecute):
 
     blocks: List[List[Transaction]]
     post: Alloc
+    # Gas validation fields for benchmark tests
+    expected_benchmark_gas_used: int | None = None  # Expected total gas to be consumed
+    skip_gas_used_validation: bool = False  # Skip gas validation even if expected is set
 
     format_name: ClassVar[str] = "transaction_post_test"
     description: ClassVar[str] = (
@@ -33,6 +36,10 @@ def execute(
         assert not any(tx.ty == 3 for block in self.blocks for tx in block), (
             "Transaction type 3 is not supported in execute mode."
         )
+
+        # Track transaction hashes for gas validation (benchmarking)
+        all_tx_hashes = []
+
         for block in self.blocks:
             signed_txs = []
             for tx_index, tx in enumerate(block):
@@ -51,11 +58,31 @@ def execute(
                 for transaction in signed_txs:
                     if transaction.error is None:
                         eth_rpc.send_wait_transaction(transaction)
+                        all_tx_hashes.append(transaction.hash)
                     else:
                         with pytest.raises(SendTransactionExceptionError):
                             eth_rpc.send_transaction(transaction)
             else:
                 eth_rpc.send_wait_transactions(signed_txs)
+                all_tx_hashes.extend([tx.hash for tx in signed_txs])
+
+        # Perform gas validation if required for benchmarking
+        # Ensures benchmark tests consume exactly the expected gas
+        if not self.skip_gas_used_validation and self.expected_benchmark_gas_used is not None:
+            total_gas_used = 0
+            # Fetch transaction receipts to get actual gas used
+            for tx_hash in all_tx_hashes:
+                receipt = eth_rpc.get_transaction_receipt(tx_hash)
+                assert receipt is not None, f"Failed to get receipt for transaction {tx_hash}"
+                gas_used = int(receipt["gasUsed"], 16)
+                total_gas_used += gas_used
+
+            # Verify that the total gas consumed matches expectations
+            assert total_gas_used == self.expected_benchmark_gas_used, (
+                f"Total gas used ({total_gas_used}) does not match "
+                f"expected benchmark gas ({self.expected_benchmark_gas_used}), "
+                f"difference: {total_gas_used - self.expected_benchmark_gas_used}"
+            )
 
         for address, account in self.post.root.items():
             balance = eth_rpc.get_balance(address)

src/ethereum_test_rpc/rpc.py

@@ -292,6 +292,18 @@ def get_transaction_by_hash(self, transaction_hash: Hash) -> TransactionByHashRe
             pprint(e.errors())
             raise e
 
+    def get_transaction_receipt(self, transaction_hash: Hash) -> dict | None:
+        """
+        `eth_getTransactionReceipt`: Returns transaction receipt.
+
+        Used to get the actual gas used by a transaction for gas validation
+        in benchmark tests.
+        """
+        response = self.post_request(
+            method="getTransactionReceipt", params=[f"{transaction_hash}"]
+        )
+        return response
+
     def get_storage_at(
         self, address: Address, position: Hash, block_number: BlockNumberType = "latest"
     ) -> Hash:

src/ethereum_test_specs/blockchain.py

@@ -898,9 +898,12 @@ def execute(
             blocks: List[List[Transaction]] = []
             for block in self.blocks:
                 blocks += [block.txs]
+            # Pass gas validation params for benchmark tests
             return TransactionPost(
                 blocks=blocks,
                 post=self.post,
+                expected_benchmark_gas_used=self.expected_benchmark_gas_used,
+                skip_gas_used_validation=self.skip_gas_used_validation,
             )
         raise Exception(f"Unsupported execute format: {execute_format}")
 

src/ethereum_test_specs/state.py

@@ -444,9 +444,12 @@ def execute(
     ) -> BaseExecute:
         """Generate the list of test fixtures."""
         if execute_format == TransactionPost:
+            # Pass gas validation params for benchmark tests
             return TransactionPost(
                 blocks=[[self.tx]],
                 post=self.post,
+                expected_benchmark_gas_used=self.expected_benchmark_gas_used,
+                skip_gas_used_validation=self.skip_gas_used_validation,
             )
         raise Exception(f"Unsupported execute format: {execute_format}")
 

src/pytest_plugins/execute/execute.py

@@ -342,6 +342,11 @@ def __init__(self, *args, **kwargs):
                     kwargs["pre"] = pre
                 elif kwargs["pre"] != pre:
                     raise ValueError("The pre-alloc object was modified by the test.")
+                # Set default for expected_benchmark_gas_used
+                if "expected_benchmark_gas_used" not in kwargs:
+                    kwargs["expected_benchmark_gas_used"] = request.getfixturevalue(
+                        "gas_benchmark_value"
+                    )
                 kwargs |= {
                     p: request.getfixturevalue(p)
                     for p in cls_fixture_parameters