Revisiting reopened bugs in open source software systems

Reopened bugs can degrade the overall quality of a software system since they require unnecessary rework by developers. Moreover, reopened also lead to loss trust in end-users regarding software. Thus, predicting that might be could extremely helpful for developers avoid rework. Prior studies on bug prediction focus only three open source projects (i.e., Apache, Eclipse, and OpenOffice) generate insights. We observe one out Apache) has data leak issue – status was included as training predict bugs. In addition, prior used an outdated model pipeline with old techniques constructing model) Therefore, we revisit study large scale dataset consisting 47 tracked JIRA using modern such SMOTE, permutation importance together 7 different machine learning models. mixed methods approach both quantitative qualitative study). find that: 1) After updated pipeline, 34% give acceptable performance AUC $\geqslant $ 0.7. 2) There are four major reasons getting reopened, is, technical patch/integration issues), documentation, human due incorrect assessment), not shown reports. 3) AUC, 94% patch issues usage patch) identified before reopening. Our revisits provides new insights into developer’s reopening activities.