Type error feedback via analytic program repair

Sakkas, Georgios K.; Endres, Madeline; Cosman, Benjamin; Weimer, Westley; Jhala, Ranjit

doi:10.1145/3385412.3386005

Cited by 18 publications

(10 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a result, CURE generates more reasonable and compilable patches and outperforms all existing techniques. Automatic Program Repair: Many APR techniques have been proposed, which use genetic programming [1], condition synthesis [2], [5], [9], state abstraction [14], heuristics [8], [10], [12], [62], human-designed fix patterns [3], [6], mined fix patterns [4], [7], [13], [15], [56], [63], bytecode mutation [64], or neural program synthesis [65]. CURE uses a new codeaware NMT approach and fixes more bugs than previous stateof-the-art approaches.…”

Section: Re L a T E D W O R Kmentioning

confidence: 99%

CURE: Code-Aware Neural Machine Translation for Automatic Program Repair

Jiang

Lutellier

Tan

2021

2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE)

176

181

View full text Add to dashboard Cite

Automatic program repair (APR) is crucial to improve software reliability. Recently, neural machine translation (NMT) techniques have been used to fix software bugs automatically. While promising, these approaches have two major limitations. Their search space often does not contain the correct fix, and their search strategy ignores software knowledge such as strict code syntax. Due to these limitations, existing NMT-based techniques underperform the best template-based approaches. We propose CURE, a new NMT-based APR technique with three major novelties. First, CURE pre-trains a programming language (PL) model on a large software codebase to learn developer-like source code before the APR task. Second, CURE designs a new code-aware search strategy that finds more correct fixes by focusing on compilable patches and patches that are close in length to the buggy code. Finally, CURE uses a subword tokenization technique to generate a smaller search space that contains more correct fixes.Our evaluation on two widely-used benchmarks shows that CURE correctly fixes 57 Defects4J bugs and 26 QuixBugs bugs, outperforming all existing APR techniques on both benchmarks.

show abstract

Section: Re L a T E D W O R Kmentioning

confidence: 99%

CURE: Code-Aware Neural Machine Translation for Automatic Program Repair

Jiang

Lutellier

Tan

2021

2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE)

176

181

View full text Add to dashboard Cite

show abstract

“…There are also studies that mostly focus on syntax errors. In particular, Gupta et al [122] used 6,975 erroneous C programs with typographic errors, Santos et al [270] used source code files with syntax errors, and Sakkas et al [268] used a corpus of 4,500 ill-typed OCaml programs that lead to compile-time errors. Bhatia et al [47] examined a corpus of syntactically correct submissions for a programming assignment.…”

Section: Data Collectionmentioning

confidence: 99%

A Survey on Machine Learning Techniques for Source Code Analysis

Sharma¹,

Kechagia²,

Georgiou³

et al. 2021

Preprint

View full text Add to dashboard Cite

Context:The advancements in machine learning techniques have encouraged researchers to apply these techniques to a myriad of software engineering tasks that use source code analysis such as testing and vulnerabilities detection. A large number of studies poses challenges to the community to understand the current landscape. Objective: We aim to summarize the current knowledge in the area of applied machine learning for source code analysis. Method: We investigate studies belonging to twelve categories of software engineering tasks and corresponding machine learning techniques, tools, and datasets that have been applied to solve them. To do so, we carried out an extensive literature search and identified 364 primary studies published between 2002 and 2021. We summarize our observations and findings with the help of the identified studies. Results: Our findings suggest that the usage of machine learning techniques for source code analysis tasks is consistently increasing. We synthesize commonly used steps and the overall workflow for each task, and summarize the employed machine learning techniques. Additionally, we collate a comprehensive list of available datasets and tools useable in this context. Finally, we summarize the perceived challenges in this area that include availability of standard datasets, reproducibility and replicability, and hardware resources. CCS Concepts: • Software and its engineering → Software libraries and repositories; Software maintenance tools; Software post-development issues; Maintaining software; • Computing methodologies → Machine learning.

show abstract

“…7) are produced using a bug injection tool. Therefore, we plan to evaluate our approach on student submissions in programming courses, similar to previous approaches [14,18,45,51].…”

Section: Limitation and Future Workmentioning

confidence: 99%

Automated Repair of Heap-Manipulating Programs Using Deductive Synthesis

Nguyen

Sergey

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

We propose a novel method to automatically repairing buggy heap-manipulating programs using constraint solving and deductive synthesis. Given an input program C and its formal specification in the form of a Hoare triple: {P} C {Q}, we use a separation-logic-based verifier to verify if program C is correct w.r.t. its specifications. If program C is found buggy, we then repair it in the following steps. First, we rely on the verification results to collect a list of suspicious statements of the buggy program. For each suspicious statement, we temporarily replace it with a template patch representing the desired statements. The template patch is also formally specified using a pair of unknown pre-and postcondition. Next, we use the verifier to analyze the temporarily patched program to collect constraints related to the pre-and postcondition of the template patch. Then, these constraints are solved by our constraint solving technique to discover the suitable specifications of the template patch. Subsequently, these specifications can be used to synthesize program statements of the template patch, consequently creating a candidate program. Finally, if the candidate program is validated, it is returned as the repaired program. We demonstrate the effectiveness of our approach by evaluating our implementation and a state-of-the-art approach on a benchmark of 231 buggy programs. The experimental results show that our tool successfully repairs 223 buggy programs and considerably outperforms the compared tool.

show abstract

Type error feedback via analytic program repair

Cited by 18 publications

References 31 publications

CURE: Code-Aware Neural Machine Translation for Automatic Program Repair

CURE: Code-Aware Neural Machine Translation for Automatic Program Repair

A Survey on Machine Learning Techniques for Source Code Analysis

Automated Repair of Heap-Manipulating Programs Using Deductive Synthesis

Contact Info

Product

Resources

About