Computing summaries of string loops in C for better testing and refactoring

Abstract

Analysing and comprehending C programs that use strings is hard: Using standard library functions for manipulating strings is not enforced and programs often use complex loops for the same purpose. We introduce the notion of memoryless loops that capture some of these string loops and present a counterexample-guided inductive synthesis approach to summarise memoryless string loops using C standard library functions, which has applications to testing, optimization and refactoring.We prove our summarization is correct for arbitrary input strings and evaluate it on a database of loops we gathered from a set of 13 open-source programs. Our approach can summarize over two thirds of memoryless loops in less than 5 minutes of computation time per loop. We then show that these summaries can be used to (1) enhance symbolic execution testing, where we observed median speedups of 79x when employing a string constraint solver, (2) optimize native code, where certain summarizations led to significant performance gains, and (3) refactor code, where we had several patches accepted in the codebases of popular applications such as patch and wget.

Trident: Controlling Side Effects in Automated Program Repair

“…Several existing techniques synthesize memorymanipulating transformations. Kapus et al [43] use symbolic execution to synthesize refactorings for string manipulating loops, but focus only on memoryless loops that do not carry information across iterations. TRIDENT's techniques might support more complex refactoring involving non-memoryless loops.…”