Abstract Large software projects follow a continuous development process with regular releases during which bugs are handled. In recent years, many software projects shifted to rapid releases that reduce time-to-market and claim a faster delivery of fixed issues, but also have a shorter period to address bugs. To better understand the impact of rapid releases on bug handling activity, we empirically analyse successive releases of the Eclipse Core projects, focusing on the bug handling rates and durations as well as the feature freeze period. We study the impact of Eclipse’s transition from a yearly to quarterly release cycle. We confirm our findings through feedback received from five Eclipse Core maintainers. Among others, our results reveal that Eclipse’s bug handling process is becoming more stable over time, with a decreasing number of reported bugs before releases, an increasing bug fixing rate and an increasingly balanced bug handling workload before and after releases. The transition to a quarterly release cycle continued to improve bug handling. In addition, more effort is spent on bug fixing during the feature freeze period, while the bug handling rates do not differ between both periods.

中文翻译：

---

关于发布政策对错误处理活动的影响：Eclipse 的案例研究

摘要 大型软件项目遵循持续的开发过程，定期发布，在此期间会处理错误。近年来，许多软件项目转向快速发布，以缩短上市时间并声称修复问题的交付速度更快，但解决错误的时间也更短。为了更好地理解快速发布对错误处理活动的影响，我们对 Eclipse Core 项目的连续发布进行了实证分析，重点关注错误处理率和持续时间以及功能冻结期。我们研究了 Eclipse 从每年发布到每季度发布周期的影响。我们通过从五位 Eclipse Core 维护者那里收到的反馈确认了我们的发现。其中，我们的结果表明 Eclipse 的错误处理过程随着时间的推移变得更加稳定，随着发布前报告的错误数量的减少，错误修复率的提高以及发布前后错误处理工作量的日益平衡。向季度发布周期的过渡继续改进错误处理。此外，在功能冻结期间，更多的精力用于修复错误，而两个时期的错误处理率没有差异。