A Method of Fault Analysis in Software Evolution Process

Abstract

The present invention proposes a fault analysis method in the software evolution process, which is mainly used to solve the fault understanding and analysis problems in the software evolution process, including the following steps: Step 1. Using AST analyzer to construct an abstraction of the intermediate representation form of the evolution modification program Syntax tree EFAGM_AST; step 2, based on the evolutionary modification, establish the key modification node set of EFAGM, and sequentially construct the forward and backward influence nodes and paths of the modification node set; step 3, generate the static state of EFAGM according to the domain of the calculated EFAGM graph nodes Influence factors and dynamic influence factors complete the construction of the evolutionary fault analysis graph model EFAGM; when a fault occurs, use EFAGM to realize the fault analysis caused by the modification of the program evolution. The invention can not only represent the elements affected by evolution modification and the relationship among them, but also can effectively improve the efficiency of evolution fault analysis.

Description

A Method of Fault Analysis in Software Evolution Process technical field The invention provides a method for performing fault analysis on software evolution process by adopting an evolution fault analysis model, which belongs to the technical field of software debugging. Background technique With the development of technology and the improvement of requirements, the software is becoming more and more complex and large-scale. During the development and maintenance of such software products, software program evolution and modification behaviors such as iterative development, requirement changes, and performance optimization are often accompanied, and the progress of evolution increases the risk of introducing new software program failures. Therefore, in the process of software evolution, continuous modeling analysis is required to find potential faults as early as possible and minimize software development and maintenance costs. When a fault occurs in a software program, t...

AI Technical Summary

Problems solved by technology

This is based on the fact that troubleshooting personnel are extremely familiar with the program, there is a lot of subjectivity, and it is not suitable for troubleshooting of modern large-scale and complex software

(2) With the help of debugging tools, troubleshooting personnel usually use setting breakpoints to troubleshoot, which is also too subjective and inefficient, resulting in a great waste of personnel

(3) Finding faults in software based on program spectrum is a popular fault location scheme at present. The disadvantage is that during the process of program evolution and modification, the historical program spectrum may be destroyed, and there is no guarantee that Test oracle will still be valid after modification.

Although the system dependency graph provides detailed dependencies among all program elements, the time cost and space cost of continuously building the system dependency graph during the continuous evolution of the program are incalculable, and it does not have the ability to analyze evolutionary faults

Images