The Tricks and Techniques of Contract Development

Recently, while learning about the development of decentralized exchanges, I discovered some interesting contract development techniques. These techniques are derived from studying the code of a well-known DEX and should be very helpful for beginners who want to get started with smart contract development.

Predictable Contract Address

The addresses generated by deploying contracts usually appear random and are difficult to predict. However, in certain scenarios, we need to derive the contract address from the information of the trading pairs, which is very useful for determining trading permissions or obtaining pool addresses.

You can create contracts using the CREATE2 method by adding a salt parameter, making the generated address predictable. The calculation logic for the new address is: hash("0xFF", creator address, salt, initcode).

Clever Use of Callback Functions

In certain scenarios, it is useful for Contract A to call a method of Contract B, and then B to callback a method of A. For example, during a trade, the pool contract will call swapCallback, passing in the actual amount of Tokens needed, and the caller transfers the Tokens in the callback. This ensures the integrity and security of the entire trading logic.

Use Exceptions to Pass Information

When estimating a trade, the execution of the swap method can be wrapped in a try-catch block. Since the estimate will not actually produce a token swap, it will throw an error. A special error can be thrown in the callback, which can then be caught to parse the required data from the error message. This way, there is no need to specifically modify the swap method for estimation purposes, making the logic simpler.

Solving Precision Problems with Big Numbers

When it comes to price and liquidity calculations, to avoid precision loss from division operations, you can first left shift 96 bits (, which is equivalent to multiplying by 2^96), and then perform the calculations. This ensures precision without overflowing. Although there will still be a theoretical minimum unit precision loss, it is acceptable in practical applications.

Share Model Calculating Earnings

When recording LP fee earnings, it is not necessary to record for each LP with every transaction, as this would consume a large amount of Gas. Instead, you can only record the total fees and the fees that should be earned per unit of liquidity, and then calculate the withdrawable amount based on the liquidity held when the LP withdraws. This is similar to the principle of stock dividends.

Off-chain Data Storage

Not all information needs to be on-chain or retrieved from the chain. The transaction pool list, pool information, etc. can be stored in traditional databases and synchronized from the chain periodically. This can improve access efficiency and reduce costs. Of course, key transactions still need to be conducted on-chain.

Contract Splitting and Reuse

A project can be split into multiple actual deployed contracts, or the code can be split into multiple contracts for maintenance through inheritance. At the same time, it is also important to make good use of existing standard contracts, such as ERC721, to improve development efficiency.

No amount of theory can compare to hands-on practice. Attempting to implement a simplified version of a DEX will allow you to gain a deeper understanding of the various techniques involved in contract development. I hope these tips will be helpful for your journey in smart contract development.