Automatically documenting program changes

Abstract

Source code modifications are often documented with log messages. Such messages are a key component of software maintenance: they can help developers validate changes, locate and triage defects, and understand modifications. However, this documentation can be burdensome to create and can be incomplete or inaccurate. We present an automatic technique for synthesizing succinct human-readable documentation for arbitrary program differences. Our algorithm is based on a combination of symbolic execution and a novel approach to code summarization. The documentation it produces describes the effect of a change on the runtime behavior of a program, including the conditions under which program behavior changes and what the new behavior is. We compare our documentation to 250 human-written log messages from 5 popular open source projects. Employing a human study, we find that our generated documentation is suitable for supplementing or replacing 89% of existing log messages that directly describe a code change.