Automated program repair recently received considerable attentions, and many techniques on this research area have been proposed. Among them, two genetic-programming-based techniques, GenProg Par, shown the promising results. In particular, has used as baseline technique to check effectiveness of new in much literature. Although Par their strong ability fixing real-life bugs nontrivial programs, what extent can benefit from genetic programming, by them guide patch search process, is still unknown.

Code comment generation aims at generating natural language descriptions for a code snippet to facilitate developers' program comprehension activities. Despite being studied long time, bottleneck existing approaches is that given snippet, they can only generate one while developers usually need know information from diverse perspectives such as what the functionality of this and how use it. To tackle limitation, study empirically investigates feasibility utilizing large models (LLMs)...

Test-based automated program repair has been a prolific field of research in software engineering the last decade. Many approaches have indeed proposed, which leverage test suites as weak, but affordable, approximation to specifications. Although literature regularly sets new records on number benchmark bugs that can be fixed, several studies increasingly raise concerns about limitations and biases state-of-the-art approaches. For example, correctness generated patches questioned studies,...

Fault localization aims at identifying suspicious statements potentially responsible for failures. The recent rapid progress on deep learning shows the promising potential of many neural network architectures in making sense data, and more importantly, this offers a new prospective probably benefiting fault localization. Thus, paper proposes CNN-FL: an approach localizing faults based convolutional networks to explore Specifically, CNN-FL constructs customized localization, then trains with...

Smart contracts with natural economic attributes have been widely and rapidly developed in various fields. However, the bugs vulnerabilities smart brought huge losses, which has strengthened people's attention to security issues of contracts. The immutability makes people more willing conduct checks before deploying Nonetheless, existing contract vulnerability detection techniques are far away from enough: static analysis approaches rely heavily on manually crafted heuristics is difficult...

Despite the capability in successfully fixing more and real-world bugs, existing Automated Program Repair (APR) techniques are still challenged by long-standing overfitting problem (i.e., a generated patch that passes all tests is actually incorrect). Plenty of approaches have been proposed for automated correctness assessment (APCA ). Nonetheless, dynamic ones those needed to execute tests) time-consuming while static built on top code features) less precise. Therefore, embedding recently,...

Data is the fuel to models, and it still applicable in fault localization (FL). Many existing elaborate FL techniques take code coverage matrix failure vector as inputs, expecting could find correlation between program entities failures. However, input data high-dimensional extremely unbalanced since real-world programs are large size number of failing test cases much less than that passing cases, which posing severe threats effectiveness techniques.

Most techniques for automated program repair use test cases to validate the effectiveness of produced patches. The validation process can be time-consuming especially when object programs ship with either lots or some long-running cases. To alleviate cost testing, we first introduce regression prioritization insight into area repair, and present a novel technique called FRTP goal reducing number case executions in process. Unlike most existing frequently requiring additional gathering...

Many techniques on automated fault localization (AFL) have been introduced to assist developers in debugging. Prior studies evaluate the technique from viewpoint of developers: measuring how many benefits that can obtain used when However, these evaluation approaches are not always suitable, because it is difficult quantify precisely due complex debugging behaviors developers. In addition, recent user presented working with AFL do correct defects more efficiently than ones only traditional...

The cost of LTL model checking is highly sensitive to the length formula under verification. We observe that, some specific conditions, input can be reduced an easier-to-handle one before checking. In such reduction, these two formulae need not logically equivalent, but they share same counterexample set w.r.t model. case that symbolically represented, condition enabling reduction detected with a lightweight effort (e.g., SAT-solving). this paper, we tentatively name technique...

Test-based automated program repair (APR) has attracted huge attention from both industry and academia. Despite the significant progress made in recent studies, overfitting problem (i.e., generated patch is plausible but overfitting) still a major long-standing challenge. Therefore, plenty of techniques have been proposed to assess correctness patches either generation phase or evaluation APR techniques. However, effectiveness existing not systematically compared little known their...

Software is constantly changing, requiring developers to perform several derived tasks in a timely manner, such as writing description for the intention of code change, or identifying defect-prone changes. Considering that cost dealing with these can account large proportion (typically around 70 percent) total development expenditure, automating processes will significantly lighten burdens developers. To achieve target, existing approaches mainly rely on training deep learning models from...

Fault localization is essential for solving the issue of software faults. Aiming at improving fault localization, this paper proposes a deep learning-based with contextual information. Specifically, our approach uses neural network to construct suspiciousness evaluation model evaluate statement being faulty, and then leverages dynamic backward slicing extract The empirical results show that significantly outperforms state-of-the-art technique Dstar.

Smart contracts have been widely and rapidly used to automate financial business transactions together with blockchains, helping people make agreements while minimizing trusts. With millions of smart deployed on blockchain, various bugs vulnerabilities in emerged. Following the rapid development deep learning, many recent studies learning for vulnerability detection conduct security checks before deploying contracts. These approaches show effective results detecting whether a contract is...

Floating point computation is by nature inexact, and numerical libraries that intensively involve floating-point computations may encounter high errors. Due to the wide use of libraries, it highly desired reduce errors in them. Using higher precision will degrade performance also introduce extra for certain precision-specific operations libraries. mathematical rewriting mostly focuses on rearranging expressions or taking Taylor expansions not fit reducing evoked ill-conditioned problems are...

Abstract Many fault localization approaches recently utilize deep learning to learn an effective model showing a fresh perspective with promising results. However, models are generally learned from class imbalance datasets; that is, the number of failing test cases is much fewer than passing cases. It may be highly susceptible affect accuracy models. Thus, in this paper, we explore using data resampling reduce negative effect imbalanced problem and improve deep‐learning‐based localization....

The quality of method names is critical for the readability and maintainability source code. However, it often challenging to construct concise names. To alleviate this problem, a number approaches have been proposed automatically recommend high-quality methods. Despite being effective, existing meet their bottlenecks mainly in two aspects: (1) leveraged information restricted target itself; (2) lack distinctions towards contributions tokens extracted from different program contexts. Through...

Just-in-time comment update is a promising way to reduce the burden of developers during software maintenance and evolution. Existing approaches can be divided into two categories: heuristic-based approach deep-learning-based approach. The restricted specific type updates (i.e., code-indicative updates), but performs well on such type. effectiveness limited it handle diverse updates. Considering complementary advantages existing approaches, an intuitive idea combine them for better...

The integration of Large Language Models (LLMs) into software development has revolutionized the field, particularly through use Retrieval-Augmented Code Generation (RACG) systems that enhance code generation with information from external knowledge bases. However, security implications RACG systems, risks posed by vulnerable examples in base, remain largely unexplored. This risk is concerning given public repositories, which often serve as sources for base collection are usually accessible...