C language-oriented source code clone detection method

Abstract

The invention discloses a C language-oriented source code clone detection method. The technical solution of the present invention is: adopting the context facial features grammar to define the grammar of C language, for realizing the parsing of the source program, generating the parsing tree of the source program, and then transforming the whole parsing tree to obtain the transformed parsing tree, and then Then restore the source code in text form. In addition, it also includes formatting and normalizing the obtained source code in text form. Perform clone detection on the obtained C function through the LCS algorithm to obtain the clone function detection result of the current function to be detected. During clone detection, only the code sequence length of the function falls within the length range of the clone function allowed by it as the current function to be detected The clone comparison object. The invention can realize the detection of Type3 clones, and control the calculation amount of detection to a certain extent.

Description

technical field [0001] The invention belongs to the technical field of clone code detection, and in particular relates to C language-oriented source code clone detection. Background technique [0002] The current research on clone code detection generally divides clone code into three categories: [0003] (1) Type1: Except for spaces and comments, the cloned code is exactly the same; [0004] (2) Type2: The cloned code with the same syntax and modified identifiers, constants, and types; [0005] (3) Type3: On the basis of Type2, further modify the statement, such as adding a statement, removing the statement, modifying the statement statement, etc. to generate cloned code. [0006] The early idea of ​​code cloning detection is very intuitive, and the code is directly treated as pure text (string), and the similarity of the code is judged from the similarity of the text. The representative technology is Baker's Dup, which, like a general web page similarity detection tool,...

AI Technical Summary

Problems solved by technology

The Text-based method does not need to convert the source code, and its accuracy rate is high but the recall rate is low; the Token-based method has a faster detection speed and does not depend on the development language, but it is difficult to detect Type-3 code clones; AST and PDG-based methods need to convert the source code into AST or PDG and then perform corresponding comparisons. These two methods are relatively expensive; since the metric values ​​corresponding to different code segments may be the same, the metric-based The value method has a high false positive rate

Images