Consider implementing a check on msg.data.length

[GriffGreen]

Consider implementing a check on msg.data.length

[koeppelmann]

please be careful doing this. There is some low level solidity padding happening that might cause the payload to be bigger than expected. So while a check against a too small payload should be fine a check for "equal" is dangerous.
It might lead to a situation that those tokens could get stuck in e.g. a multisig wallet.

Read more here: https://blog.coinfabrik.com/smart-contract-short-address-attack-mitigation-failure/
or here:
http://vessenes.com/notice-about-fun-token-support-for-multisignature-wallets-2/

[jbaylina]

PR #15 implements it.

But I'm not convinced on this, because this functions might be called internally from a derived contract and what's checking are not the parameters but the data field.

[acedev0816]

assert(msg.data.length == numwords * 32 + 4);
why +4?