Generation of Semantically Correct Solidity Programs for the Automated Testing of its Compiler

Abstract

Ensuring compiler correctness is a challenging but essential task, particularly as modern compilers grow in size and complexity. Among various techniques, random program generation has proven effective at uncovering subtle compiler bugs that traditional test suites may overlook. In parallel, with the growing popularity of blockchain platforms such as Ethereum, the Solidity programming language has become a widely used tool for developing smart contracts. The Solidity compiler, solc, is still relatively new and continues to evolve, making it susceptible to bugs and regressions. This work presents SolGen, a generator of semantically correct Solidity programs which aims to test the Solidity compiler.