Embracing modern C++ features: An empirical assessment on the KDE community

Abstract

Similar to software systems, programming languages evolve substantially over time. Indeed, the community has more recently seen the release of new versions of mainstream languages in shorter and shorter time frames. For instance, the C++ working group has begun to release a new version of the language every 3 years, which now has a greater number of modern C++ features and improvements in modern standards (C++11, C++14, C++17, and C++ 20). Nonetheless, there is little empirical evidence on how developers are transitioning to use modern C++ constructs in legacy systems, and not understanding the trends and reasons for adopting these new modern C++ features might hinder software developers in conducting rejuvenation efforts. In this paper, we conduct an in‐depth study to understand the development practices of KDE contributors to evolve their projects toward the use of modern C++ features. Our results show a trend in the widespread adoption of some modern C++ features (lambda expressions, auto‐typed variables, and range‐based for) in KDE community projects. We also found that developers in the KDE community are making large efforts to modernize their programs using automated tools, and we present some modernization scenarios and the benefits of adopting modern C++ features of the C++ programming language. Our results might help C++ software developers, in general, to evolve C++ legacy systems and tools builders to implement more effective tools that could help in rejuvenation efforts.

Understanding the Motivations, Challenges, and Practices of Software Rejuvenation

“…Time constraints arise from the need to modern language features need some time for being adopted 28 (P2, P3, P9, P10, P12-P15, P18-P20, P12, and P23), and It is hard to keep updated with language evolution 14 (P1, P3, P6, P7, P10, P15, P16, P18, P20, P21 and P23), which require developers to allocate time for understanding and incorporating these features effectively. The findings of our mining study [9] indicate that the prompt and extensive integration of modern language features is not instantaneous. The findings of our study indicate that the widespread adoption of modern features occurred approximately five years after the release of the C++11 specification.…”

“…In the second study our goal was to mine the source code repository of C++ and JavaScript open source projects to find evidence that developers replace legacy programming language constructs and idioms by new ones and even conduct large rejuvenation efforts. [9]. The KDE Open Community is a globally recognized open-source community that is actively engaged in the development of a wide range of applications across various domains, such as games, education, and development frameworks.…”

“…The early integration of specific features revealed that JavaScript developers have a forward-looking mindset and eagerness to infuse quality improvements into their projects, thereby rejuvenating their codebases. Although we have already published the results of the C++ study [9], our plan is to submit the results of the JavaScript study in October 2023. We detail the second study in Section V.…”

“…• Second Study aims to identify (a) how C++ and JavaScript developers replaces legacy constructs by modern language features and (b) whether or not they conduct large rejuvenation efforts. The second study relies on mining software repository efforts [1], [9]. • Third Study aims to improve and validate our theory on software rejuvenation efforts with the feedback from industry.…”

“…We analyzed C++ repositories that we checked out from the GitHub of The KDE Open Community [9]. We filtered out projects with a small percentage of C++ code (below 50%) and projects that started after 2010 or that did not have recent updates-that is, we only consider projects that have at least one commit in 2022.…”