feat(tests): port ethereum/tests swap opcode tests

[bomanaps]

🗒️ Description

Port test for swap opcode ported from ethereum/tests.

🔗 Related Issues

Closes #1012

✅ Checklist

• All: Set appropriate labels for the changes.
• All: Considered squashing commits to improve commit history.
• All: Added an entry to CHANGELOG.md.
• All: Considered updating the online docs in the ./docs/ directory.
• Tests: All converted JSON/YML tests from ethereum/tests have been added to converted-ethereum-tests.txt.
• Tests: A PR with removal of converted JSON/YML tests from ethereum/tests have been opened.
• Tests: Included the type and version of evm t8n tool used to locally execute test cases: e.g., ref with commit hash or geth 1.13.1-stable-3f40e65.
• Tests: Ran mkdocs serve locally and verified the auto-generated docs for new tests in the Test Case Reference are correctly formatted.

[marioevz]

Thanks so much for implementing this.

There's a couple of comments I think should be addressed before merging, but overall it's a very good PR!

[marioevz]

I think we can leave this file out of the PR for the time being 👍

[danceratopz]

Looks like @bomanaps is using uv>=0.8.22 - this is due to:

• bug(tooling): uv sync with uv>=0.5.22 modifies uv.lock #1176

[danceratopz]

@bomanaps no action required regarding your uv version at the moment, I think everyone else needs to upgrade their uv.

[bomanaps]

Okay

[marioevz]

Suggested change

	from ethereum_test_tools import StateTestFiller, Storage, Transaction, Bytecode
	from ethereum_test_tools import StateTestFiller, Transaction, Bytecode

Storage is unused.

[marioevz]

We could change the approach a bit here:

• SSTORE(0,1) at the start of the execution, just to have something that will revert when the underflow happens.
• Fill the stack up to the SWAP* opcode's required minimum minus one.
• Perform the SWAP* that should abort the execution because there's not enough values in the stack for it to operate.
• Verify that the storage is empty since the SSTORE(0,1) reverts with the execution abort caused by the underflow.

[marioevz]

The test as it is right now is producing an overflow that is accomplished by the opcode PUSH1, and in that case, the SWAPn opcode is not really used.

This is the last executed opcode in the trace:

Step 1025:
{ 'depth': 1,
  'error': 'StackOverflowError',
  'gas': '0x74315',
  'gasCost': '0x3',
  'memSize': 0,
  'op': 96,
  'opName': 'PUSH1',
  'pc': 2049,
  'refund': 0,
  'returnData': '0x',
  'stack': [ '0x0',
             '0x1',
             '0x2',
...
             '0xfe',
             '0xef']}

Step 1026:
{'error': 'StackOverflowError', 'gasUsed': '0x74f18', 'output': ''}

So we don't really need to push this many items into the stack for an underflow, the key is not having enough items for the SWAPn opcode to try to fetch outside of the bounds of the stack.

For example, for SWAP1, it requires at least two elements in the stack to work properly (see the min_stack_height property of the opcode https://ethereum.github.io/execution-spec-tests/main/library/ethereum_test_vm/#ethereum_test_vm.Opcodes.SWAP1), because it swaps the first item for the second. Therefore, to create a stack underflow, we place a single item in the stack and then execute the SWAP1.

[winsvega]

@marioevz can we automate this thing? unless we specificly want to test this

[danceratopz]

One potential solution for this would be:

• feat(fill): make transaction a pytest fixture #1065
  Then the transaction (or tx fixture could made aware of the current value of the fork fixture/parameter).

[winsvega]

why uv.lock is deleted?

[marioevz]

More comments.

[marioevz]

The test as it is right now is producing an overflow that is accomplished by the opcode PUSH1, and in that case, the SWAPn opcode is not really used.

This is the last executed opcode in the trace:

Step 1025:
{ 'depth': 1,
  'error': 'StackOverflowError',
  'gas': '0x74315',
  'gasCost': '0x3',
  'memSize': 0,
  'op': 96,
  'opName': 'PUSH1',
  'pc': 2049,
  'refund': 0,
  'returnData': '0x',
  'stack': [ '0x0',
             '0x1',
             '0x2',
...
             '0xfe',
             '0xef']}

Step 1026:
{'error': 'StackOverflowError', 'gasUsed': '0x74f18', 'output': ''}

So we don't really need to push this many items into the stack for an underflow, the key is not having enough items for the SWAPn opcode to try to fetch outside of the bounds of the stack.

For example, for SWAP1, it requires at least two elements in the stack to work properly (see the min_stack_height property of the opcode https://ethereum.github.io/execution-spec-tests/main/library/ethereum_test_vm/#ethereum_test_vm.Opcodes.SWAP1), because it swaps the first item for the second. Therefore, to create a stack underflow, we place a single item in the stack and then execute the SWAP1.