Technical BriefingsICSME 2026

srcML & srcDiff: Exploring, Analyzing, Manipulating, and Differencing Source Code

Presenters

• Jonathan I. Maletic, Kent State University, Ohio, USA
• Michael J. Decker, Bowling Green State University, Ohio, USA

Abstract

This technical briefing is intended for researchers and practitioners interested in constructing custom software analysis, manipulation, and differencing tools to support research or application development.

The briefing will discuss the latest releases of srcML and srcDiff, including the addition of support for Python, JavaScript/TypeScript, and CMake. It will provide an overview of both technologies and demonstrate how they can support software engineering research.

A syntactically aware query language, srcQL, has also been added to srcML. It enables users to query complex program structures through an easy-to-use query language.

srcML is an infrastructure comprising an XML representation for C/C++, C#, Java, Python, JavaScript/TypeScript, and CMake source code, together with efficient parsing technology for converting source code to and from the srcML format. Its grammars are unified across the supported programming languages, facilitating the development of multilingual tools.

srcDiff is a syntactic differencing tool built on top of srcML. It is highly scalable and produces accurate, understandable deltas in various formats.

The srcML infrastructure is freely available for download and licensed under the GNU General Public License v3.0 (GPLv3). The project has been under development for more than 20 years, with its first public release dating back to 2005. The srcML website provides comprehensive documentation about the format and supported languages, as well as downloadable executables for Microsoft Windows, macOS, and several GNU/Linux distributions.