Automatically Fixing Syntax Errors Using the Levenshtein Distance

Abstract

Abstract:

To ensure high quality software, much emphasis is laid on software testing. While a number of techniques and tools already exist to identify and locate syntax errors, it is still the duty of programmers to manually fix each of these uncovered syntax errors. In this paper we propose an approach to automate the task of fixing syntax errors by using existing compilers and the levenshtein distance between the identified bug and the possible fixes. The levenshtein distance is a measure of the similarity between two strings. A prototype, called ASBF, has also been built and a number of tests carried out which show that the technique works well in most cases. ASBF is able to automatically fix syntax errors in any erroneous source file and can also process several erroneous files in a source folder. The tests carried out also show that the technique can also be applied to multiple programming languages. Currently ASBF can automatically fix software bugs in the Java and the Python programming languages. The tool also has auto-learning capabilities where it can automatically learn from corrections made manually by a user. It can thereafter couple this learning process with the levenshtein distance to improve its software bug
correction capabilities.

Keywords: Automatically fixing syntax errors, bug fixing, auto-learn, levenshtein distance, Java, Python

(Article history: Received 16 September 2016 and accepted 9 December 2016)