How a smart contract can call it's own function if it's not in the code

Question

Please tell me how can smart contract call it's own function if this is not in smart contract code? Here is an example: https://github.com/gnosis/MultiSigWallet/blob/master/contracts/MultiSigWallet.sol. It has onlyWallet modifier applied to addOwner function, but addOwner is not called in the code. Thanks.

Answer 1

The line require(msg.sender == address(this)) in onlyWallet can be a bit of a head scratcher, so let's first look at some snippets from the MultiSig wallet tests to see how you'd interact with the contract:

var owners = [accounts[0], accounts[1], accounts[2]]
var multisigInstance = await MultiSigWallet.new(owners, confirmations)
var addOwnerData = multisigInstance.contract.addOwner.getData(accounts[3])
await multisigInstance.submitTransaction(multisigInstance.address, 0, addOwnerData, {from: accounts[0]})

As you can see, one of the contract owners can invoke submitTransaction passing along the encoded data that makes up the addOwner method call.

Inside the submitTransaction smart contract method there is a call to the addTransaction method which stores the transaction data (addOwners call data) into a mapping.

transactions[transactionId] = Transaction({
    destination: destination,
    value: value,
    data: data,
    executed: false
});

Then an owner calls the confirmTransaction method to confirm the transaction which then invokes executeTransaction(transactionId):

Finally, inside executeTransaction you can see that it's executing the encoded data (addOwner call data) which makes the contract address the sender:

txn.destination.call.value(txn.value)(txn.data)

So essentially, the MultiSig wallet allows owners to submit a transaction which gets stored until another owner approves of it and executes it.