A framework for studying and predicting bug propagation through code cloning for mobile apps

通过移动应用程序代码克隆来研究和预测错误传播的框架

• 批准号: 543819-2019
• 负责人: Roy, Chanchal
• 金额: $ 1.82万
• 依托单位: University of Saskatchewan
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2019
• 资助国家: 加拿大
• 起止时间: 2019-01-01 至 2020-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=680218
• 关键词: framework; studying; predicting; bug; propagation

Code cloning (i.e., copy/pasting) is a common yet controversial software engineering practice, which is often practised by the programmers during development and maintenance for repeating common functionalities. Such a practice results the existence of identical or nearly similar code fragments, also known as code clones, in a code-base. Two code fragments that are similar to each other form a clone-pair. While a number of studies have identified some positive impacts of code clones, other studies have shown strong empirical evidence of negative impacts too. Code clones can be directly related to bugs and inconsistencies in the code-base. If a particular code fragment in a code-base contains a temporarily hidden bug, and a programmer copies that fragment to several other places being unaware of the presence of the bug, the bug in the original fragment gets propagated. Given that software bugs are of primary concerns of any software, one must need to have tools and frameworks to deal with software bugs. However, there is no study on how frequently bug-propagation occurs during code cloning. Without studying the intensity of bug-propagation we cannot properly realize the impacts of code clones on software evolution and maintenance. Furthermore, there are no studies that investigated the propagated bugs for code clones for mobile applications. Given that Push Interactions mainly develop mobile applications, it is important for Push to make sure that their software applications do not have many of these propagated bugs. However, until now they do not have any such frameworks that can study the bug propagation through code cloning (propagation of bugs through copy/pasting of existing buggy code fragments). Furthermore, it is important to have a prediction model so that the model could predict which clone fragments might experience such propagated bugs in future during the evolution. With such toolset, Push Interactions would be able to have a clear idea about the presence of code clones in their mobile applications, their relationships to bugs, will find which clones are likely involve with bugs, and can predict the buggy clones. These will ensure proper maintenance and evolution of their mobile software applications.

代码克隆（即，复制/粘贴）是一种常见但有争议的软件工程实践，程序员在开发和维护过程中经常采用这种做法，以重复通用功能。这种做法导致在代码库中存在相同或几乎相似的代码片段，也称为代码克隆。两个彼此相似的代码片段形成克隆对。虽然一些研究已经确定了代码克隆的一些积极影响，但其他研究也显示了负面影响的有力经验证据。代码克隆可能与代码库中的错误和不一致直接相关。如果代码库中的特定代码片段包含临时隐藏的bug，并且程序员将该片段复制到其他几个地方而不知道bug的存在，则原始片段中的bug会传播。考虑到软件错误是任何软件的主要关注点，我们必须拥有处理软件错误的工具和框架。然而，没有关于代码克隆期间错误传播发生频率的研究。如果不研究缺陷传播的强度，我们就不能正确认识代码克隆对软件演化和维护的影响。此外，没有研究调查移动的应用程序的代码克隆的传播错误。由于Push Interactions主要开发移动的应用程序，因此Push确保其软件应用程序没有许多此类传播的错误是很重要的。然而，到目前为止，他们还没有任何这样的框架，可以研究通过代码克隆（通过复制/粘贴现有的错误代码片段的错误传播）的错误传播。此外，重要的是要有一个预测模型，以便该模型可以预测哪些克隆片段可能会在未来的进化过程中遇到这样的传播错误。有了这样的工具集，推送交互将能够清楚地了解移动的应用程序中代码克隆的存在，它们与错误的关系，将发现哪些克隆可能涉及错误，并可以预测有错误的克隆。这将确保其移动的软件应用程序的适当维护和发展。