A Method to Sufficiently Detect Null Pointer Reference Defects

Abstract

The invention discloses a method of fully detecting null pointer reference defects. The method comprises recognizing all addressable expressions of an application to be detected based on an abstract syntax tree; carrying out conservative interval arithmetic and pointer analysis on the application to be detected according to a control flow diagram, and generating a procedural summary according to results of the interval arithmetic and the pointer analysis; recognizing all pointer reference and referred pointers according to the procedural summary and the abstract syntax tree, and creating a null pointer reference defect state machine example on each referred pointer; and running the null pointer reference defect state machine examples based on the control flow diagram, carrying out state transition on each defect state machine example on each node of the control flow diagram according to the results of the interval arithmetic and the pointer analysis, and carrying out null pointer reference detection. By adopting the method of fully detecting the null pointer reference defects, the problem of failing to report the null pointer reference defects can be effectively solved, and zero omission and low misinformation of the null pointer reference defect detection are achieved.

Description

technical field [0001] The invention relates to a null pointer detection technology in software static testing technology, in particular to a method for fully detecting null pointer reference defects. Background technique [0002] Software testing is a process of ensuring software quality. Its fundamental purpose is to find various defects in the software with as little time and manpower as possible through some economical and effective methods, so as to ensure the quality of the software. For software testing, it is classified based on whether the software under test needs to be run. Software testing is divided into dynamic testing and static testing, wherein static testing is also called static analysis. Static testing does not actually run the software under test, but scans the source application to find structural anomalies, control flow anomalies, and data flow anomalies that may cause errors. Compared with dynamic testing, static testing has the advantages of low cost...

AI Technical Summary

Problems solved by technology

Compared with dynamic testing, static testing has the advantages of low cost, easy implementation, coverage of all paths, and does not depend on a specific operating environment; its disadvantage is that the problems found are often not real problems, and manual confirmation and troubleshooting are required

[0003] Existing static testing methods have a large number of false positives or false negatives

The existence of false positives requires manual confirmation to eliminate them; the existence of false negatives will cause the illusion that the application is of high quality, but once the defects of false negatives are triggered when the software is running, it may lead to unpredictable adverse consequences

[0005] However, by using the above test tools to test null pointer references, there will be varying degrees of false positives and false negatives

Images