The Analysis And Research Of Static Testing Technology Based On C/C++ Source Code

With the growing scale of modern information system and great improvement of the software functionality and performance, the frequency of software error is obviously increasing, which demonstrates the significance of software quality. To ensure the software quality in software engineering, static testing carried out in software development has become the key technology.On the basis of the classification of software defecting patterns and static testing tools, this paper puts forward the static testing framework based on source code of C/C++ programming, and builds the test platform on the domestic operating system(kylin system). The main research objects include as follows.(1)To extract the static information of the initial source in the form of abstract syntax trees, it adopts open source GCC compiler front-end, that is, the principle of lexical analysis, syntax analysis and semantic analysis in C/C++ static analyzer.(2)Safety rules are classified and assigned with options to improve the efficiency of the system test. By employing the XML model and XQuery language, the abstract syntax tree is converted to XML model, the safety rule is abstracted to XQuery expression so as to make the safety rules matching and detect the loophole.(3)Multistage application testing: Multiple inspection tools are integrated and multiple angles are adopted for multistage application testing in order to detect more defects, decreasing false negative rate.(4)Based on the above analysis, the static software source code testing platform is built and the validation of relevant methods and technology are tested by using the open source GCC compiler front-end and safety rules of static testing that can assign options as well as the multistage program testing technology.This paper studies an accurate and efficient method, by which source code detection can be realized in the process of programming. It not only guarantees subsequent software quality, but will also reduce the development costs of software.