A method and apparatus for software fault location based on program invariants

Abstract

The invention discloses a software fault locating method and a device of program invariant. The method comprises the following steps that: the source code of the target software is inserted at the level of sentence, value and logical expression, and the source code after insertion is executed by a preset test case set to obtain the execution information; the source code of the target software is inserted at the level of sentence, value and logical expression; the preset failure test case set is clustered and the successful test case set is selected which is helpful to distinguish the defectivestatements for each clustering; the execution information of the preferred successful test case set is learnt, the program invariant set is obtained, including the set type, truth table type and floating point type range invariant; according to the execution information of the failed test case set and the program invariant set, the invariant violation is detected, and the suspicious statement setis obtained; dependency analysis is used to filter out invariant violation error caused by fault propagation, and invariant violation at each sentence is statistically analyzed to calculate sentencesuspicion. The invention improves the accuracy of software fault location, and overcomes the problem of missing detection of logical expression defect location.

Description

technical field [0001] The invention relates to the technical field of computer software, in particular to a software fault location method and device based on program invariants. Background technique [0002] Software systems often contain software defects, which reduce the reliability, usability and security of the software system. One of the most important tasks to fix a defect is to identify the location of the program element related to the defect, and then the programmer can modify the program according to the location of the suspicious program element and its program context to correct the defect. However, this process is very time-consuming and labor-intensive. Software debugging even accounts for 80% of the entire software overhead. Therefore, there is an urgent need for automated software fault location methods. [0003] In addition to directly supporting software developers, automated software error localization techniques are also used in automated program cor...

AI Technical Summary

Problems solved by technology

If the learned invariants are too broad, the real cause of failure may not be included in the invariant violation candidate set, resulting in missed detection of defect locations; if the learned invariants are too narrow (missing some invariants), then May lead to a large number of false detections, unable to accurately locate the root cause of software failure

Images