查看文章

This report documents the classi cation, de nition, rationale, and semantics of the mutant operators designed for the (proposed) ANSI C programming language Kern88]. The design of mutant operators was carried out by the authors of this report over a six month period starting August, 1988.

Throughout the report, the following conventions are used:(a) keywords in C are emboldened,(ii) non-terminal symbols from the C grammar are italicized,(iii) mutant operators are emboldened and appear in upper case letters,(iv) mutant operator category names are emboldened and begin with an upper case letter followed by three lower case letters, and (v) P denotes the program under test and hence the program to be mutated. In general, P will consist of several functions that could be mutated individually or as a group depending on how much of P is being tested. This report assumes that the reader is familiar with the underlying theory and techniques of mutation analysis. However, the next section provides an overview of mutation analysis. DeMi79] and DeMi87] contain the in depth background material. The remainder of the report is organized into 13 sections. The next section provides an overview of mutation analysis. Section 3 enumerates the reasons responsible for the mutant operators reported here. Section 4 presents a classi cation of the mutant operators. We expect such a classi cation to be useful for the tool designer1 and for the software tester. Section 5 consolidates the naming conventions used while naming the large number of mutant operators. The conventions are designed to simplify the expansion of the four letter mnemonics for all …