https://doi.org/10.1021/acs.energyfuels.9b00813

Alharbi, H. A., Morandi, G. D., Jones, P. D., Wiseman, S. B., & Giesy, J. P. (2019). Comparison of the effects of extraction techniques on mass spectrometry profiles of dissolved organic compounds in oil sand process-affected water. Energy & Fuels, 33(8), 7001-7008. https://doi.org/10.1021/acs.energyfuels.9b00813

Copying and pasting source code during software development is known as code cloning. Clone fragments with a minimum size of 5 LOC were usually considered in previous studies. In recent studies, clone fragments which are less than 5 LOC are referred as micro-clones. It has been established by the literature that code clones are closely related with software bugs as well as bug replication. None of the previous studies have been conducted on bug-replication of micro-clones. In this paper we investigate and compare bug-replication in between regular and micro-clones. For the purpose of our investigation, we analyze the evolutionary history of our subject systems and identify occurrences of similarity preserving co-changes (SPCOs) in both regular and micro-clones where they experienced bug-fixes. From our experiment on thousands of revisions of six diverse subject systems written in three different programming languages, C, C# and Java we find that the percentage of clone fragments that take part in bug-replication is often higher in micro-clones than in regular code clones. The percentage of bugs that get replicated in micro-clones is almost the same as the percentage in regular clones. Finally, both regular and micro-clones have similar tendencies of replicating severe bugs according to our experiment. Thus, micro-clones in a code-base should not be ignored. We should rather consider these equally important as of the regular clones when making clone management decisions.

https://doi.org/10.1021/acs.energyfuels.9b00813