During the last few years, Continuous Integration (CI) has become a common practice in open-source and industrial environments to reduce the scope for errors and increase the speed to market through the automated build and test processes. However, despite this wide adoption throughout the years, little is known about the challenges developers discuss. Analyzing the discussions of developers is required to understand what researchers, educators and practitioners should focus on, and how discussion communities can be helpful to shed the light on CI challenges. In this study, we examine Stack Overflow (SO), the most popular crowd-sourced forum, to understand the challenges developers face in the CI context. We collect a corpus of 27,728 CI-related developers posts from SO and analyze those posts through a mixed method with quantitative and qualitative analyzes. To study the trends of CI discussions, we investigated the metadata of CI questions, users and tags. Then, we extract the CI main topics using Latent Dirichlet Allocation (LDA) tuned with Genetic Algorithm (GA). Finally, we investigate the most popular and difficult topics faced by developers based on unanswered questions to get further insights into CI challenges. The LDA clustering reveals that developers face challenges with six main topics namely Build, Testing, Version Control, Configuration, Deployment, and CI Culture. Particularly, we found that the build topic is the most popular among the studied topics and that version control and testing topics are the most difficult for the SO community. Our study uncovers insights about CI challenges and adds evidence to existing knowledge about CI issues related especially to software build.

Continuous Integration (CI) consists of integrating the changes introduced by different developers more frequently through the automation of build process. Nevertheless, the CI build process is seen as a major barrier that causes delays in the product release dates. One of the main reasons for such delays is that some simple changes (i.e., can be skipped) trigger the build, which represents an unnecessary overhead and particularly painful for large projects. In order to cut off the expenses of CI build time, we propose in this paper,SkipCI, a novel search-based approach to automatically detect CI Skip commits based on the adaptation of Strength-Pareto Evolutionary Algorithm (SPEA-2). Our approach aims to provide the optimal trade-off between two conflicting objectives to deal with both skipped and non-skipped commits. We evaluate our approach and investigate the performance of both within and cross-project validations on a benchmark of 14,294 CI commits from 15 projects that use Travis CI system. The statistical tests revealed that our approach shows a clear advantage over the baseline approaches with average scores of 92% and 84% in terms of AUC for cross-validation and cross-project validations respectively. Furthermore, the features analysis reveals that documentation changes, terms appearing in the commit message and the committer experience are the most prominent features in CI skip detection. When it comes to the cross-project scenario, the results reveal that besides the documentation changes, there is a strong link between current and previous commits results. Moreover, we deployed and evaluated the usefulness ofSkipCIwith our industrial partner. Qualitative results demonstrate the effectiveness ofSkipCIin providing relevant CI skip commit recommendations to developers for two large software projects from practitioner’s point of view.

Continuous Integration (CI) aims at supporting developers in inte-grating code changes quickly through automated building. How-ever, there is a consensus that CI build failure is a major barrierthat developers face, which prevents them from proceeding furtherwith development. In this paper, we introduceBF-Detector, anautomated tool to detect CI build failure. Based on the adaptationof Non-dominated Sorting Genetic Algorithm (NSGA-II), our toolaims at finding the best prediction rules based on two conflictingobjective functions to deal with both minority and majority classes.We evaluated the effectiveness of our tool on a benchmark of 56,019CI builds. The results reveal that our technique outperforms state-of-the-art approaches by providing a better balance between bothfailed and passed builds.BF-Detectortool is publicly available,with a demo video, at: https://github.com/stilab-ets/BF-Detector.

Community smells are symptoms of organizational and social issues within the software development community that often increase the project costs and impact software quality. Recent studies have identified a variety of community smells and defined them as suboptimal patterns connected to organizational-social structures in the software development community such as the lack of communication, coordination and collaboration. Recognizing the advantages of the early detection of potential community smells in a software project, we introduce a novel approach that learns from various community organizational and social practices to provide an automated support for detecting community smells. In particular, our approach learns from a set of interleaving organizational-social symptoms that characterize the existence of community smell instances in a software project. We build a multi-label learning model to detect 8 common types of community smells. We use the ensemble classifier chain (ECC) model that transforms multi-label problems into several single-label problems which are solved using genetic programming (GP) to find the optimal detection rules for each smell type. To evaluate the performance of our approach, we conducted an empirical study on a benchmark of 103 open source projects and 407 community smell instances. The statistical tests of our results show that our approach can detect the eight considered smell types with an average F-measure of 89% achieving a better performance compared to different state-of-the-art techniques. Furthermore, we found that the most influential factors that best characterize community smells include the social network density and closeness centrality as well as the standard deviation of the number of developers per time zone and per community.