Solidity, the emergent programming language predominantly utilized for the development of smart contracts, has been gaining increased importance in the blockchain system. Ensuring the bug-free of its accompanying language compiler, which is in charge of converting the contract source codes into executables finally deployed on the blockchain, is thus of paramount importance. Therewith, this study presents DeSCDT, a differential fuzz testing approach to explore possible defects in the young and fast evolving solidity compiler. At the core of DeSCDT lies a well-behaving deep contract generator following the Transformer architecture and learnt with diverse contract code. From an initial seed pool of smart contracts carefully picked from the wild through semantic encoding and clustering, the contract generator is capable of stably producing highly syntactic-valid and functional-rich smart contracts, with three meticulously formulated generation strategies and a set of mutation operations. Subsequently, in the meantime of compiling these compiler test cases (i.e., the generated contracts) to trigger obvious compiler crashes, a differential testing environment is setup to explore the potential defects of the compiler optimizer, by observing the inconsistencies between the expectations and actual outcomes in the aspects including the gas consumption and the opcode sequence size of the optimized and non-optimized compiled outputs.
The datasets consists of the collected raw information pages of the verified smart contracts on Etherscan, the wild smart contracts deployed on Ethereum blockchain. They are accessbile and can be downloaded from: https://pan.baidu.com/s/17ijVwj0Z2XV0dEQkSonlrQ (Fetch Code: tk5s).
python ./genTrainData.py
python ./train.py
python ./generate.py