Tests on testnet are buggy when using Contract.at

[PatrickAlphaC]

Environment information

• brownie Version: 1.14.6
• ganache-cli Version: N/A
• solc Version: N/A
• Python Version: 3.9.5
• OS: OSX

What was wrong?

When running tests on a live network like kovan, brownie attempts to "tear down" instances with the close function and calls _remove_contract from a list it gathers:

On line 468

for contract in [x for v in self._containers.values() for x in v._contracts]

This can bring in duplicates, so when brownie dels them, it runs into a keyerror since it will have already deleted it.

Here is a full output:

patrick@iMac: [~/code/nft-mix -  (main)] $ brownie test --network kovan -s
Brownie v1.14.6 - Python development framework for Ethereum

================================================================================= test session starts ==================================================================================
platform darwin -- Python 3.9.5, pytest-6.2.3, py-1.10.0, pluggy-0.13.1 -- /Library/Frameworks/Python.framework/Versions/3.9/bin/python3
cachedir: .pytest_cache
hypothesis profile 'brownie-verbose' -> verbosity=2, deadline=None, max_examples=50, stateful_step_count=10, report_multiple_bugs=False, database=DirectoryBasedExampleDatabase(PosixPath('/Users/patrick/.brownie/hypothesis'))
rootdir: /Users/patrick/code/nft-mix
plugins: eth-brownie-1.14.6, web3-5.18.0, xdist-1.34.0, forked-1.3.0, hypothesis-6.10.0, flaky-3.7.0
collected 3 items                                                                                                                                                                      

tests/integration/test_advanced_int.py::test_can_create_advanced_collectible_integration PASSED

============================================================================= 1 passed, 2 skipped in 8.72s =============================================================================
  File "/Users/patrick/code/brownie/brownie/_cli/__main__.py", line 64, in main
    importlib.import_module(f"brownie._cli.{cmd}").main()
  File "/Users/patrick/code/brownie/brownie/_cli/test.py", line 61, in main
    return_code = pytest.main(pytest_args, ["pytest-brownie"])
  File "pytest-6.2.3-py3.9.egg/_pytest/config/__init__.py", line 162, in main
    ret: Union[ExitCode, int] = config.hook.pytest_cmdline_main(
  File "pluggy/hooks.py", line 286, in __call__
    return self._hookexec(self, self.get_hookimpls(), kwargs)
  File "pluggy/manager.py", line 93, in _hookexec
    return self._inner_hookexec(hook, methods, kwargs)
  File "pluggy/manager.py", line 84, in <lambda>
    self._inner_hookexec = lambda hook, methods, kwargs: hook.multicall(
  File "pluggy/callers.py", line 208, in _multicall
    return outcome.get_result()
  File "pluggy/callers.py", line 80, in get_result
    raise ex[1].with_traceback(ex[2])
  File "pluggy/callers.py", line 187, in _multicall
    res = hook_impl.function(*args)
  File "pytest-6.2.3-py3.9.egg/_pytest/main.py", line 316, in pytest_cmdline_main
    return wrap_session(config, _main)
  File "pytest-6.2.3-py3.9.egg/_pytest/main.py", line 311, in wrap_session
    config._ensure_unconfigure()
  File "pytest-6.2.3-py3.9.egg/_pytest/config/__init__.py", line 987, in _ensure_unconfigure
    self.hook.pytest_unconfigure(config=self)
  File "pluggy/hooks.py", line 286, in __call__
    return self._hookexec(self, self.get_hookimpls(), kwargs)
  File "pluggy/manager.py", line 93, in _hookexec
    return self._inner_hookexec(hook, methods, kwargs)
  File "pluggy/manager.py", line 84, in <lambda>
    self._inner_hookexec = lambda hook, methods, kwargs: hook.multicall(
  File "pluggy/callers.py", line 208, in _multicall
    return outcome.get_result()
  File "pluggy/callers.py", line 80, in get_result
    raise ex[1].with_traceback(ex[2])
  File "pluggy/callers.py", line 187, in _multicall
    res = hook_impl.function(*args)
  File "/Users/patrick/code/brownie/brownie/test/managers/base.py", line 263, in pytest_unconfigure
    project.close(raises=False)
  File "/Users/patrick/code/brownie/brownie/project/main.py", line 469, in close
    _remove_contract(contract)
  File "/Users/patrick/code/brownie/brownie/network/state.py", line 563, in _remove_contract
    del _contract_map[contract.address]
KeyError: '0xdD3782915140c8f3b190B5D67eAc6dc5760C46E9'

Please include information like:

Fix

This can be fixed easily by changing that one line to:

for contract in list(set([x for v in self._containers.values() for x in v._contracts])):

[PatrickAlphaC]

After doing some more testing, it looks like the Contract.at function is adding the duplicates.

If I have in my map.json a contract already, and then I call Contract.at it appears to add a duplicate contract into brownie, but only 1 is in themap.json

Commenting out this line also appears to do the trick of fixing the error

brownie/brownie/network/contract.py

Line 231 in 4a399b1

self._contracts.append(contract)

        contract._save_deployment()
        _add_contract(contract)
        # self._contracts.append(contract)
        if CONFIG.network_type == "live":
            _add_deployment(contract)

[PatrickAlphaC]

Ok, I think I got it.

It's adding a Contract the first time, and then a ProjectContract the second time:

> /Users/patrick/code/brownie/brownie/project/main.py(469)close()
    468         breakpoint()
--> 469         for contract in [x for v in self._containers.values() for x in v._contracts]:
    470             _remove_contract(contract)

ipdb> stuff = [x for v in self._containers.values() for x in v._contracts]
ipdb> stuff
[<VRFCoordinatorMock Contract '0xdD3782915140c8f3b190B5D67eAc6dc5760C46E9'>, <VRFCoordinatorMock Contract '0xdD3782915140c8f3b190B5D67eAc6dc5760C46E9'>, <AdvancedCollectible Contract '0x31989dd1D2F718561Fbe1dCaa8Cf00933C0800c3'>, <AdvancedCollectible Contract '0x39ebC36a317f4537829Cd67aC65E85497886a2bD'>, <LinkToken Contract '0xa36085F69e2889c224210F603D836748e7dC0088'>, <LinkToken Contract '0xa36085F69e2889c224210F603D836748e7dC0088'>]
ipdb> type(stuff[0])
<class 'brownie.network.contract.Contract'>
ipdb> type(stuff[1])
<class 'brownie.network.contract.ProjectContract'>

[PatrickAlphaC]

contract = _find_contract(address) in contract.py will then return a Contract the first time.

[PatrickAlphaC]

Aha.... I should use from_abi instead of at....

[PatrickAlphaC]

Looks like we could do a few things:

1. Stop recording contracts in the _contract_map
2. Check to see if the address is already full in the _contract_map
3. Disallow people to use at with contracts deployed outside of brownie

Happy to help here if there are any thoughts

[PatrickAlphaC]

I've gone deep down the brownie hole on this, and I think I see a few easy wins. Let me know what you think.

[zmeghji]

@PatrickAlphaC How did you even get the tests running on Kovan? Whenever I try running tests in Brownie on a test network, I get the following error. They work fine on the mainnet-fork network though.

    def _request(method: str, args: List) -> int:
        try:
            response = web3.provider.make_request(method, args)  # type: ignore
            if "result" in response:
                return response["result"]
        except (AttributeError, RequestsConnectionError):
            raise RPCRequestError("Web3 is not connected.")
>       raise RPCRequestError(response["error"]["message"])
E       brownie.exceptions.RPCRequestError: The method evm_snapshot does not exist/is not available

../../.local/pipx/venvs/eth-brownie/lib/python3.8/site-packages/brownie/network/rpc/ganache.py:129: RPCRequestError

[xrave110]

@PatrickAlphaC How did you even get the tests running on Kovan? Whenever I try running tests in Brownie on a test network, I get the following error. They work fine on the mainnet-fork network though.

    def _request(method: str, args: List) -> int:
        try:
            response = web3.provider.make_request(method, args)  # type: ignore
            if "result" in response:
                return response["result"]
        except (AttributeError, RequestsConnectionError):
            raise RPCRequestError("Web3 is not connected.")
>       raise RPCRequestError(response["error"]["message"])
E       brownie.exceptions.RPCRequestError: The method evm_snapshot does not exist/is not available

../../.local/pipx/venvs/eth-brownie/lib/python3.8/site-packages/brownie/network/rpc/ganache.py:129: RPCRequestError

I have the same issue, kovan works when I write brownie run but not working for brownie test.

Does anybody know how to fix it ?

UPDATE 04-08-2022:
Removing this function helped (it is default function after baking the aave/flashloans).

[PatrickAlphaC]

I just used from_abi instead of at