Clone Detection

Abstract

Different clone detection techniques can be used to identify the known parts of a code and to avoid analyzing the same code portions again. Existing methods are found to be neither robust enough to accommodate the mutations brought by compilers nor scalable enough when querying against modern code base of high volume. To address these limitations, in this chapter we present BinSequence, a two-step clone detection engine. The proposed fine-grained fuzzy matching detection engine can perform code comparison accurately and as a result, the false correlation to irrelevant code can be avoided. The fingerprint-based detection engine can efficiently prune the search space without notably compromising the accuracy.