A Survey of Test Based Automatic Program Repair

Abstract

Abstract:Testing and debugging have always been the most time-consuming parts of the software development procedure and require large amounts of human resources. When a bug is located, manually fixing it to repair the buggy program is still a difficult and laborious task for developers. Hence automatic program repair techniques, especially the test-based approaches, have drawn great attentions in recent years. Researchers have explored and proposed various novel methods and tools, pushing the idea closer to reality. In this paper, we systematically survey the work in mainstream of test-based program repair (TBR) and discuss the properties automatically generated patches should have. We classify the state-of-the-art approaches for TBR, and evaluate their strengths and weaknesses according to their functional mechanisms. Finally, we refer to some empirical results and propose four important issues, which are supposed to be critical and constructive in this research area.

CirFix: automatically repairing defects in hardware design code

“…CirFix requires information about expected behavior for a circuit design to assign fitness values to candidate repairs. In APR for software, guidelines for correct behavior often take the form of passing and failing test cases [46]. More generally, however, such information can be induced from a previous version of the design known to be functional [4,18,22,51,53,71] or a combination of data mining and static analyses of the design [15,23,25,72], or manually provided by the human developer [3,11,24,28].…”

“…Automated Program Repair for Software. In the realm of software, significant research effort has been devoted to repairing bugs automatically over the last decade [21,46,58]. Automated program repair usually takes as input source code with a deterministic bug and a test suite with at least one failing test that reveals the bug, and aims to automatically generate fixes to the buggy code.…”

“…Furthermore, previous work does not operate on behavioral-level descriptions of hardware circuits [13,49]. On the other hand, in the realm of software, significant research effort focuses on repairing bugs automatically [21,46,58]. Automated program repair (APR) algorithms fix defects in software by producing patches that pass all test cases while retaining required functionality.…”