contract inline assembly - return array parts by using assembly

Question

I've a couple of questions.

1. mload function returns the data that resides on the address we pass to mload. Is this correct ?

2. add function is used to give us the address from 0 to whatever we pass. add(arr, 64) would return the address from arr + 64 bytes. correct. ?

3. I want to achieve same things with storage variables (x_slot and x_offset).. I have the following code.

uint[] storageArr;
    function assembly2() public view returns (uint a, uint b, uint c) {
        storageArr.push(111);
        storageArr.push(222);
        storageArr.push(333);
        
       
        assembly {
            a := sload(0)
            b := sload(32)
            c := sload(64)
        }
    }

This returns {3, 0, 0,}. The first one(slot) is 3, since it stores the length of the array. why is the second and third one 0 ?

Answer 1

Due to their unpredictable size, dynamically-sized array types use a Keccak-256 hash computation to find the starting position of the array data.

So in your example:

• The position of storageArr[0] is at keccak256(uint256(0)) + 0
• The position of storageArr[1] is at keccak256(uint256(0)) + 1
• The position of storageArr[2] is at keccak256(uint256(0)) + 2

Here is how you can put it into code:

function assembly2() public view returns (uint a, uint b, uint c) {
    uint index = uint(keccak256(abi.encode(uint(0))));
       
    assembly {
        a := sload(add(index, 0))
        b := sload(add(index, 1))
        c := sload(add(index, 2))
    }
}