It is written in many places how these things are. But what not easy to decide when to use which of them, as it needs a deeper understanding to choose.
So I want to note the quick heuristic methods to choose these.
Public vs. External
TL;DR;
So if you know the function you create only allows for external calls, go for
external. It provides performance benefits and you will save on gas.
Basically, public means it can be external or internal, the compiler need additional work for the public function. With the external only, it allows arguments to be read directly from calldata, saving the copying step.
Easy read here:
- https://gus-tavo-guim.medium.com/public-vs-external-functions-in-solidity-b46bcf0ba3ac
- https://ethereum.stackexchange.com/a/19391/76945
Memory vs. Calldata vs. Storage
TL;DR;
- use
calldatawhen you only need read-only data, avoiding the cost of allocating memory or storage.- use
memoryif you want your argument to be mutable.- use
storageif your argument will already exist in storage, to prevent copying something into storage over into memory unnecessarily.
Ref:
To research more
I am still not sure why most of the OpenZeppelin libs choose to use memory over calldata, although they don't mutate them. For example: https://github.com/OpenZeppelin/openzeppelin-contracts/blob/master/contracts/token/ERC721/ERC721.sol#L179
Top comments (1)
Found this one on why OpenZeppelin team uses
memoryovercalldata.forum.openzeppelin.com/t/why-does-...
It seems
calldatais limited in functionalities when using withinternalfunction, so they go formemory.