feat(ethexe): introduce ScaleCodec library for Solidity #4237
Conversation
osipov-mit
changed the title
| bytes memory result = new bytes(end - start); | ||
|
|
||
| for (uint256 i = 0; i < end - start; i++) { | ||
| result[i] = data[i + start]; |
There was a problem hiding this comment.
why do you need to fill each slice/bytes etc?
| return result; | ||
| } | ||
|
|
||
| function concatBytes(bytes[] memory value) public pure returns (bytes memory) { |
There was a problem hiding this comment.
gpt suggests to pre-allocate resulting bytes memory here:
// SPDX-License-Identifier: MIT
pragma solidity ^0.8.0;
contract ByteConcatenation {
function concat(bytes[] memory data) public pure returns (bytes memory) {
// Calculate total length
uint totalLength = 0;
for (uint i = 0; i < data.length; i++) {
totalLength += data[i].length;
}
// Allocate memory for the result
bytes memory result = new bytes(totalLength);
uint offset = 0;
// Copy each bytes element into the result
for (uint i = 0; i < data.length; i++) {
bytes memory currentBytes = data[i];
uint currentLength = currentBytes.length;
// Use inline assembly to copy memory
assembly {
// Copy currentBytes to result at the offset
// The `mload` is used to get the length of currentBytes
let src := add(currentBytes, 0x20) // skip the length
let dest := add(result, add(0x20, offset)) // skip the length of result
mcopy(dest, src, currentLength)
}
offset += currentLength;
}
return result;
}
// Helper function for memory copy
function mcopy(uint dest, uint src, uint length) internal pure {
for (uint i = 0; i < length; i++) {
assembly {
mstore8(add(dest, i), mload8(add(src, i)))
}
}
}
}| } | ||
|
|
||
| function decodeBool(bytes memory _bytes, uint256 offset) public pure returns (bool) { | ||
| return _bytes[offset] == 0x01; |
There was a problem hiding this comment.
you should return result here maybe
for optimised code impl decodeBoolUnchecked
There was a problem hiding this comment.
Since there is no native Result type in Solidity I think it's not the best idea.
Maybe calling revert is better if the bytes are wrong?
| return int16(decodeUint16(_bytes, offset)); | ||
| } | ||
|
|
||
| function encodeUint32(uint32 value) public pure returns (bytes memory) { |
There was a problem hiding this comment.
do you always need bytes?
seems that uint32 value has zero cast to bytes4 - so it could be also useful for future usage?
There was a problem hiding this comment.
Actually encodeUint32To function is supposed to be used most of the time (at least in contracts generated by sails). And here I belief bytes is better cuz you probably need to combine those bytes with some others and you can't do it if you have type bytes4. And even if you don't need to combine them you need to pass those bytes to the sendMessage function which takes bytes type
No description provided.