Why bytes(str).length is not enough for getting the length of a string in Solidity, and understanding the strlen method from contracts of ens.

In the world of Javascript finding the length of a string is such an easy thing. Just dostr.length and that’s all 🤌

But strings are not so friendly to work with, in Solidity ❗. In solidity, the string is a group of characters stored inside an array and stores the data in bytes.

There is no length method in string type.

I was going through Buildspace’s build-polygon-ens project and found the link to StringUtils.sol. I knew to find the length of the string in Solidity we can convert the string into bytes and find its length. So it should have been as easy as doingbytes(str).length;🤌 but the method in this util file was a bit different:

// SPDX-License-Identifier: MIT
// Source:
// https://github.com/ensdomains/ens-contracts/blob/master/contracts/ethregistrar/StringUtils.sol
pragma solidity >=0.8.4;

library StringUtils {
    /**
     * @dev Returns the length of a given string
     *
     * @param s The string to measure the length of
     * @return The length of the input string
     */
    function strlen(string memory s) internal pure returns (uint256) {
        uint256 len;
        uint256 i = 0;
        uint256 bytelength = bytes(s).length;

        for (len = 0; i < bytelength; len++) {
            bytes1 b = bytes(s)[i];
            if (b < 0x80) {
                i += 1;
            } else if (b < 0xE0) {
                i += 2;
            } else if (b < 0xF0) {
                i += 3;
            } else if (b < 0xF8) {
                i += 4;
            } else if (b < 0xFC) {
                i += 5;
            } else {
                i += 6;
            }
        }
        return len;
    }
}

It had this weird ‘for’ loop in code which I couldn’t understand.

So, the developer in me googled it 🕵️‍♀️, but all the articles I came across did this to find the length of the stringbytes(str).length;I found some similar code on Stackoverflow but no one actually explained what is happening inside.

for(len = 0; i < bytelength; len++) {
            bytes1 b = bytes(s)[i];
            if(b < 0x80) {
                i += 1;
            } else if (b < 0xE0) {
                i += 2;
            } else if (b < 0xF0) {
                i += 3;
            } else if (b < 0xF8) {
                i += 4;
            } else if (b < 0xFC) {
                i += 5;
            } else {
                i += 6;
            }
  }

After 3 hours of 🐌 self-exploration I was able to figure it out myself (a little slow but I did it 🍾),

So I thought let’s write it down so it would be helpful for all the folks like me (not so experienced with bits, bytes 0️⃣1️⃣).

Let’s try to Unblock/Decode this

How bytes(str).length works

When we convert string to bytes this is what Solidity does:

// if we do bytes("xyz"), solidity converts it as 
xyz -> 78 79 7a // 78=x, 79=y, 7a=z
ABC -> 41 42 43 // 41=A, 42=B, 43=C

If you see each character generates 1 byte that’s why when we do bytes(””).length we get the length of the string. But there are some characters for which generated bytes are more than one. For example:

€ -> e2 82 ac

For the symbol of the Euro, generated bytes are 3.

So if we try to find the length of string which includes the symbol of Euro(€) 🤑 in it, the length returned bybytes(str).length will not return the correct string length for this character as € there are 3 bytes generated:

That’s when that ‘for’ loop we've seen above comes to the rescue ⛑️

Let’s iterate over this e2 82 ac

bytes array and check what’s happening inside that loop:

for(len = 0; i < bytelength; len++) {
            bytes1 b = bytes(s)[i];
                        // b = e2 for first iteration
            if(b < 0x80) {
                i += 1;
            } else if (b < 0xE0) {
                i += 2;
            } else if (b < 0xF0) {
                i += 3;
            } else if (b < 0xF8) {
                i += 4;
            } else if (b < 0xFC) {
                i += 5;
            } else {
                i += 6;
            }
  }

For the first iteration b=e2there is a condition on the following line

if(b < 0x80) {
     i += 1;
}

Let's decode this. This condition will basically compare decimal values of these hexadecimal characters:

0x80 -> 128
// our b is e2 at the moment, decimal value for e2 = 226
0xe2 -> 226

For regular characters, decimal conversion of their hex character will be < 128 , like for a it is 97.

So, if we check all conditions like this

for(len = 0; i < bytelength; len++) {
            bytes1 b = bytes(s)[i];
            if(b < 0x80) { //0x80 = 128 => 226 < 128 ❌
                i += 1;
            } else if (b < 0xE0) { //0xE0 = 224 => 226 < 224 ❌
                i += 2;
            } else if (b < 0xF0) { //0xF0 = 240 => 226 < 240 ✅
                i += 3;
            } 
                        ...
  }

So, if our i is 3 the condition in ‘for’ loop will be 3<3, which is false and the loop will break, and the value oflen will be 1 at the moment.

And that’s it, it is the correct value for the length of string “€”

If you want to try some more strings like “€”, here is a small list of characters that occupies more than 1 byte:

€ -> e2 82 ac 
à -> c3 83
¢ -> c2 a2

Create, a random string anything like abc¢Ã, for example, and try it out.

Ta-Da 🎉, and now it works

Connect with me on Twitter: @pateldeep_eth or LinkedIn

My DMs are open to any kind of improvement or suggestions

Originally published here.