What is ABI Encoding?

[ryandotsmith]

What is ABI encoding and what data in the Ethereum world is commonly ABI encoded?

[ryandotsmith]

What is ABI encoding?

ABI encoding is a specification for turning structured data into bytes. Structured data can be: static (eg int,uint,bool, etc,...), dynamic (eg bytes, string), arrays, and tuples. Arrays can be fixed length (ie int[4]) or dynamic length (ie int[]). Static means the size of the data is known by the type. Fixed length arrays and tuples are can be static if all of their members are static.

ABI decoding requires a schema since the encoded data doesn't include any information about what has been encoded. Here is an example of decoding with a schema:

abi.Decode(b, schema.Parse("(uint8, uint8[])"))

In this example, we are decoding the bytes in b as a single tuple containing a static uint8 field and a dynamic length array of uint8s.

The official spec for ABI encoding is extremely thorough and recommended for a precise understanding:

https://docs.soliditylang.org/en/v0.8.19/abi-spec.html

What is ABI encoded?

There are 2 common scenarios in which you will encounter ABI encoded data:

1. Transaction input data
2. Event logs

Transaction input data

The data contains: a 4 byte function identifier followed by ABI encoded bytes. It is not uncommon for projects to append bytes after the abi encoded data for things like attribution.

You can find transaction input data on the Etherscan transaction page by clicking: show more

The first 4 bytes are: 08dc9f42

These 4 bytes are created by:

keccak("mint(uint256,uint256,bytes)")[:4]

The remaining 448 bytes (14 32-byte words) are the ABI encoded argument inputs:

0000000000000000000000000000000000000000000000000000000063ff8c4023bbd9768518566224ac9e44be0ccf75b7cb9331b77e5ea40a8264f9345117fa00000000000000000000000000000000000000000000000000000000000000600000000000000000000000000000000000000000000000000000000000000041778b07b4743905a2fcf18ea8de85c4fd916bc3f869336f8cc915f143791758b73dc4abe33b969a632d50d1efb20dadc4bfad94edefebddecf609875ddc59b8e81c00000000000000000000000000000000000000000000000000000000000000

The following schema: (uint256,uint256,bytes) can be used to decode the data:

timestamp: 1677691968
imageHash: 16162850975775388370804851734048201322988249485151439725861064384852108843002
signature: 0x778b07b4743905a2fcf18ea8de85c4fd916bc3f869336f8cc915f143791758b73dc4abe33b969a632d50d1efb20dadc4bfad94edefebddecf609875ddc59b8e81c

Event logs

Logs are emitted during the execution of a transaction and have cryptographic commitments via receipt_root in the block header. Each log contains an address ([32]byte), topics ([4][32]byte), and data ([]byte). The address is the contract that emitted the log. The topics contain the hash of the event signature followed by 3 optional 32-byte items and finally the arbitrarily complex data field which contains ABI encoded data.

You can find a log on the Etherscan transaction page by clicking: logs

Address: 0xcd423f3ab39a11ff1d9208b7d37df56e902c932b is the contract that emitted the log.

The first topic:

0xc42079f94a6350d7e6235f29174924f928cc2ac818eb64fed8004e115fbcca67

Is created by:

keccak("Swap(address,address,int256,int256,uint160,uint128,int24)")

sender and recipient are labeled as "indexed" meaning those inputs are stored in Topics[1] and Topics[2] respectively.

The remaining inputs are ABI encoded in the Data field:

0xffffffffffffffffffffffffffffffffffffffffffffffdda44e16cb25f3bb16000000000000000000000000000000000000000000000000259bdaa44d4f0000000000000000000000000000000000000000000010a945710f7fbc894dc45d5c000000000000000000000000000000000000000000000d992347fdcaabbb184bffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff2a8b

The following schema: (int256,int256,uint160,uint128,int24) can be used to decode the data:

amount0: -633796617047659791594
amount1: 2710000000000000000
sqrtPriceX96: 5156396549542610295333346652
liquidity: 64215658400144529496139
tick: -54645