Effective layering error locating method

Abstract

The invention discloses an effective layering error locating method. The method comprises the following steps: initializing an abnormal value set V by use of a value of a fault assert sentence in an error path; then selecting an abnormal value v from the set V and calculating the unsatisfied weakest precondition of the abnormal value v; calculating the generation reason c of the abnormal value v; judging whether c contains an error sentence or not, if yes, stopping the algorithm, otherwise, starting from c, identifying possible other abnormal value set V' by virtue of data and control denpendency, and adding abnormal values in the set V' into the set V; iterating the above steps till finding the error sentence containing reason; finally outputting the error sentence containing minimal reason and transferring the abnormal value generated by the reason to a reason set through which a fault assertion sentence passes. The effective layering error locating method disclosed by the invention can be used for assisting debugging personnel in executing error locating, and further providing effective information to assist the debugging personnel in understanding the error formation reason and further to assist the debugging personnel in repairing the error.

Description

technical field [0001] The present invention relates to data flow analysis technology and error location technology. Aiming at the problem that the error location method based on single error execution path information is difficult to provide effective information to help debuggers understand the cause of errors, an effective semi-automatic (semi-automatic) layering is proposed. Error location method. Background technique [0002] Bug localization in program debugging is one of the most expensive and time-consuming activities in software development. By analyzing the source program, test results and other various information, the error location technology gives the possible location of the source code statement that causes the software defect, and helps the programmer to fix the program defect. [0003] According to the principles based on and the characteristics possessed, the existing automated error localization techniques can be roughly divided into three categories: co...

AI Technical Summary

Problems solved by technology

[0004] The error location method based on comparison can provide a more accurate ordering of error statements, but it often needs to obtain a large number of failures and successful program execution path information, and the ordering results do not use the debugger's understanding of the error, and it is difficult to correct the program. mistake

The state-based error location method can give certain information to assist the debugger to understand the cause of the error. However, this type of method needs to search a large amount of state space, which limits its effectiveness and usability.

Finally, error location methods based on dependency analysis can help programmers understand the formation of errors very well, but they often mark too many possible error statements, making it difficult for developers to directly use the location results provided by them

Images