A Survey on Co-occurrences of Code and Test Smells

The co-occurrence of code and test smells can have detrimental effects on software systems, causing developers more difficulty in understanding and changing code in which the smells appear. Researchers have explored code smell co-occurrences and analyzed which smells occur together, the languages and domains of the software systems in which bad smells occurred, and the effects of co-occurring code smells. This paper presents the results of a systematic mapping study we conducted to summarize the research on the subject of code smell co-occurrences. We reviewed 14 papers, identifying various aspects, including the code smells that researchers investigated, the programming languages of the software systems containing the smells, the code and test smells that often occurred together, the tools used to detect smells, the techniques employed to determine co-occurrences, as well as the similarities and differences between studies on the topic. Our results present an overview of the code smells phenomenon, including information such as which code and test smells occur with others the most, as well as how frequently code and test smells are mentioned across the studies. The following implications can be derived from our study: identifying programming languages and code smell categories that have not yet been fully explored, determining standards for code and test smells based on attributes and descriptions of smells, and prioritizing certain code and test smell co-occurrences based on their frequency across the studies we surveyed.