Unit testing is essential for software reliability, yet manual test creation is time-consuming and often neglected. Although search-based software testing improves efficiency, it produces tests with poor readability and maintainability. Although LLMs show promise for test generation, existing research lacks comprehensive evaluation across execution-driven assessment, reasoning-based prompting, and real-world testing scenarios. This study presents the first large-scale empirical evaluation of LLM-generated unit tests at the full class level , systematically analyzing four state-of-the-art models (GPT-3.5, GPT-4, Mistral 7B, and Mixtral 8x7B) against EvoSuite across 216,300 generated test cases targeting Defects4J, SF110, and CMD (a dataset mitigating LLM training data leakage). We evaluate five prompting techniques–Zero-Shot Learning (ZSL), Few-Shot Learning (FSL), Chain-of-Thought (CoT), Tree-of-Thought (ToT), and Guided Tree-of-Thought (GToT)–assessing syntactic correctness, compilability, hallucination-driven failures, readability, code coverage metrics, and test smells. Reasoning-based prompting particularly GToT significantly enhances test reliability, compilability, and structural adherence in general-purpose LLMs. However, hallucination-driven failures remain a persistent challenge, manifesting as non-existent symbol references, incorrect API calls, and fabricated dependencies, resulting in high compilation failure rates (up to 86%). Moreover, test smell analysis reveals that while LLM-generated tests are generally more readable than those produced by traditional tools, they still suffer from recurring design issues such as Magic Number Tests and Assertion Roulette, which hinder maintainability. Overall, our findings indicate that LLMs can serve as effective assistive tools for generating readable and maintainable test suites, but hybrid approaches that combine LLM-based generation with automated validation and search-based refinement are required to achieve reliable and production-ready test generation.

Accurate classification of issues is essential for effective project management and timely responses, as the volume of issue reports continues to grow. Manual classification is labor-intensive and error-prone, necessitating automated solutions. While large language models (LLMs) show promise in automated issue labeling, most research focuses on broad categorization (e.g., bugs, feature requests), with limited attention to fine-grained categorization. Understanding specific bug types is crucial, as different bugs require tailored resolution strategies. This study addresses this gap by evaluating LLMs and prompt engineering strategies for fine-grained bug report categorization. We analyze 221,184 fine-grained bug report category labels generated by selected LLMs using various prompt engineering strategies for 1,024 bug reports. We examine how LLMs and prompt engineering influence output characteristics, control over outputs, and categorization performance. Our findings highlight that LLMs and prompt engineering significantly impact output consistency and classification capability, with some yielding consistent results and others introducing variability. Based on these findings, we analyze the agreements and disagreements between LLM-generated labels and human annotations to assess category correctness. Our results suggest that examining label consistency and discrepancies can serve as a complementary method for validating bug report categories, identifying unclear reports, and detecting misclassifications in human annotations.

Large Language Models (LLMs) are increasingly used as code assistants, yet their behavior when explicitly asked to generate insecure code remains poorly understood. While prior research has focused on unintended vulnerabilities, this study examines a more direct threat: open-source LLMs generating vulnerable code when prompted. We propose a dual experimental design: (1) Dynamic Prompting, which systematically varies vulnerability type, user persona, and prompt phrasing across structured templates; and (2) Reverse Prompting, which derives natural-language prompts from real vulnerable code samples. We evaluate three open-source 7B-parameter models (Qwen2, Mistral, Gemma) using static analysis to assess both the presence and correctness of generated vulnerabilities. Our results show that all models frequently generate the requested vulnerabilities, though with significant performance differences. Gemma achieves the highest correctness for memory vulnerabilities under Dynamic Prompting (e.g., 98.6 % for buffer overflows), while Qwen2 demonstrates the most balanced performance across all tasks. We find that professional personas (e.g., “DevOps Engineer”) consistently elicit higher success rates than student personas, and that the effectiveness of direct versus indirect phrasing is inverted depending on the prompting strategy. Vulnerability reproduction accuracy follows a non-linear pattern with code complexity, peaking in a moderate range. Our findings expose how LLMs' reliance on pattern recall over semantic reasoning creates significant blind spots in their safety alignments, particularly for requests framed as plausible professional tasks.

Snapshot testing has emerged as a critical technique for UI validation in modern software development, yet it suffers from substantial maintenance overhead due to frequent UI changes causing test failures that require manual inspection to distinguish between genuine regressions and intentional design changes. This manual triage process becomes increasingly burdensome as applications evolve, creating a need for automated analysis solutions. This paper introduces LLMShot, a novel framework that leverages Vision-Language Models (VLMs) to automatically analyze snapshot test failures through semantic classification of UI changes. To evaluate LLMShot's effectiveness, we developed a comprehensive dataset using a feature-rich iOS application with configurable feature flags, creating realistic scenarios that produce authentic snapshot differences representative of real development workflows. Our evaluation using Gemma3 models demonstrates strong classification performance, with the 12 B variant achieving over 84 % recall in identifying failure root causes while the 4 B model offers practical deployment advantages with acceptable performance for continuous integration environments. However, our exploration of selective ignore mechanisms revealed significant limitations in current prompting-based approaches for controllable visual reasoning. LLMShot represents the first automated approach to semantic snapshot test analysis, offering developers structured insights that can substantially reduce manual triage effort and advance toward more intelligent UI testing paradigms.

Automatically generated unit tests-from searchbased tools like EvoSuite or LLMs-vary significantly in structure and readability. Yet most evaluations rely on metrics like Cyclomatic Complexity and Cognitive Complexity, designed for functional code rather than test code. Recent studies have shown that SonarSource's Cognitive Complexity metric assigns nearzero scores to LLM-generated tests, yet its behavior on EvoSuitegenerated tests and its applicability to test-specific code structures remain unexplored. We introduce CCTR, a Test-Aware Cognitive Complexity metric tailored for unit tests. CCTR integrates structural and semantic features like assertion density, annotation roles, and test composition patterns-dimensions ignored by traditional complexity models but critical for understanding test code. We evaluate$\mathbf{1 5, 7 5 0}$test suites generated by EvoSuite, GPT4o, and Mistral Large-1024 across 350 classes from Defects4J and SF110. Results show CCTR effectively discriminates between structured and fragmented test suites, producing interpretable scores that better reflect developer-perceived effort. By bridging structural analysis and test readability, CCTR provides a foundation for more reliable evaluation and improvement of generated tests. We publicly release all data, prompts, and evaluation scripts to support replication.

Traditional static analysis methods struggle to detect semantic design flaws, such as violations of the SOLID principles, which require a strong understanding of object-oriented design patterns and principles. Existing solutions typically focus on individual SOLID principles or specific programming languages, leaving a gap in the ability to detect violations across all five principles in multi-language codebases. This paper presents a new approach: a methodology that leverages tailored prompt engineering to assess LLMs on their ability to detect SOLID violations across multiple languages. We present a benchmark of four leading LLMs—CodeLlama:70B, DeepSeekCoder:33B, Qwen2.5 Coder:32B, and GPT-4o Mini—on their ability to detect violations of all five SOLID principles. For this evaluation, we construct a new benchmark dataset of 240 manually validated code examples. Using this dataset, we test four distinct prompt strategies inspired by established zero-shot, few-shot, and chain-of-thought techniques to systematically measure their impact on detection accuracy. Our emerging results reveal a stark hierarchy among models, with GPT-4o Mini decisively outperforming others, yet even it struggles with challenging principles like DIP. Crucially, we show that prompt strategy has a dramatic impact, but no single strategy is universally best; for instance, a deliberative ENSEMBLE prompt excels at OCP detection while a hint-based EXAMPLE prompt is superior for DIP violations. Across all experiments, detection accuracy is heavily influenced by language characteristics and degrades sharply with increasing code complexity. These initial findings demonstrate that effective, AI-driven design analysis requires not a single "best" model, but a tailored approach that matches the right model and prompt to the specific design context, highlighting the potential of LLMs to support maintainability through AI-assisted code analysis.

Unit testing, essential for identifying bugs, is often neglected due to time constraints. Automated test generation tools exist but typically lack readability and require developer intervention. Large Language Models (LLMs) like GPT and Mistral show potential in test generation, but their effectiveness remains unclear.

Organizations and companies that collect data generated by sales, transactions, client/server communications, IoT nodes, devices, engines, or any other data generating/exchanging source, need to analyze this data to reveal insights about the running activities on their systems. Since streaming data has multivariate variables bearing dependencies among each other that extend temporally (to previous time steps).Long-Short Term Memory (LSTM) is a variant of the Recurrent Neural Networks capable of learning long-term dependencies using previous timesteps of sequence-shape data. The LSTM model is a valid option to apply to our data for offline anomaly detection and help foresee future system incidents. Anything that negatively affects the system and the services provided via this system is considered an incident.Moreover, the raw input data might be noisy and improper for the model, leading to misleading predictions. A wiser choice is to use an LSTM Autoencoder (LSTM-AE) specialized for extracting meaningful features of the examined data and looking back several steps to preserve temporal dependencies.In our work, we developed two LSTM-AE models. We evaluated them in an industrial setup at Koçfinans (a finance company operating in Turkey), where they have a distributed system of several nodes running dozens of microservices. The outcome of this study shows that our trained LSTM-AE models succeeded in identifying the atypical behavior of offline data with high accuracies. Furthermore, after deploying the models, we identified the system failing at the exact times for the previous two reported failures. While after deployment, it launched cautions preceding the actual failure by a week, proving efficiency on online data. Our models achieved 99.7% accuracy and 89.1% as F1-score. Moreover, it shows potential in finding the proper LSTM-AE model architecture when time series data with temporal dependency property is fed to the model.

Much research on software engineering relies on experimental studies based on fault injection. Fault injection, however, is not often relevant to emulate real-world software faults since it “blindly” injects large numbers of faults. It remains indeed challenging to inject few but realistic faults that target a particular functionality in a program. In this work, we introduce iBiR , a fault injection tool that addresses this challenge by exploring change patterns associated to user-reported faults. To inject realistic faults, we create mutants by re-targeting a bug-report-driven automated program repair system, i.e., reversing its code transformation templates. iBiR is further appealing in practice since it requires deep knowledge of neither code nor tests, just of the program’s relevant bug reports. Thus, our approach focuses the fault injection on the feature targeted by the bug report. We assess iBiR by considering the Defects4J dataset. Experimental results show that our approach outperforms the fault injection performed by traditional mutation testing in terms of semantic similarity with the original bug, when applied at either system or class levels of granularity, and provides better, statistically significant estimations of test effectiveness (fault detection). Additionally, when injecting 100 faults, iBiR injects faults that couple with the real ones in around 36% of the cases, while mutation testing achieves less than 4%.

Fix pattern-based patch generation is a promising direction in automated program repair (APR). Notably, it has been demonstrated to produce more acceptable and correct patches than the patches obtained with mutation operators through genetic programming. The performance of pattern-based APR systems, however, depends on the fix ingredients mined from fix changes in development histories. Unfortunately, collecting a reliable set of bug fixes in repositories can be challenging. In this article, we propose investigating the possibility in an APR scenario of leveraging fix patterns inferred from code changes that address violations detected by static analysis tools. To that end, we build a fix pattern-based APR tool, Avatar , which exploits fix patterns of static analysis violations as ingredients for the patch generation of repairing semantic bugs. Evaluated on four benchmarks (i.e., Defects4J, Bugs.jar, BEARS, and QuixBugs), Avatar presents the potential feasibility of fixing semantic bugs with the fix patterns inferred from the patches for fixing static analysis violations and can correctly fix 26 semantic bugs when Avatar is implemented with the normal program repair pipeline. We also find that Avatar achieves performance metrics that are comparable to that of the closely related approaches in the literature. Compared with CoCoNut, Avatar can fix 18 new bugs in Defects4J and 3 new bugs in QuixBugs. When compared with HDRepair, JAID, and SketchFix, Avatar can newly fix 14 Defects4J bugs. In terms of the number of correctly fixed bugs, Avatar is also comparable to the program repair tools with the normal fault localization setting and presents better performance than most program repair tools. These results imply that Avatar is complementary to current program repair approaches. We further uncover that Avatar can present different bug-fixing performances when it is configured with different fault localization tools, and the stack trace information from the failed executions of test cases can be exploited to improve the bug-fixing performance of Avatar by fixing more bugs with fewer generated patch candidates. Overall, our study highlights the relevance of static bug-finding tools as indirect contributors of fix ingredients for addressing code defects identified with functional test cases (i.e., dynamic information).

A large body of the literature on automated program repair develops approaches where patches are automatically generated to be validated against an oracle (e.g., a test suite). Because such an oracle can be imperfect, the generated patches, although validated by the oracle, may actually be incorrect. While the state-of-the-art explores research directions that require dynamic information or rely on manually-crafted heuristics, we study the benefit of learning code representations in order to learn deep features that may encode the properties of patch correctness. Our empirical work investigates different representation learning approaches for code changes to derive embeddings that are amenable to similarity computations of patch correctness identification, and assess the possibility of accurate classification of correct patch by combining learned embeddings with engineered features. Experimental results demonstrate the potential of learned embeddings to empower Leopard (a patch correctness predicting framework implemented in this work) with learning algorithms in reasoning about patch correctness: a machine learning predictor with BERT transformer-based learned embeddings associated with XGBoost achieves an AUC value of about 0.803 in the prediction of patch correctness on a new dataset of 2,147 labeled patches that we collected for the experiments. Our investigations show that deep learned embeddings can lead to complementary/better performance when comparing against the state-of-the-art, PATCH-SIM, which relies on dynamic information. By combining deep learned embeddings and engineered features, Panther (the upgraded version of Leopard implemented in this work) outperforms Leopard with higher scores in terms of AUC, +Recall and -Recall, and can accurately identify more (in)correct patches that cannot be predicted by the classifiers only with learned embeddings or engineered features. Finally, we use an explainable ML technique, SHAP, to empirically interpret how the learned embeddings and engineered features are contributed to the patch correctness prediction.

Test-based automated program repair has been a prolific field of research in software engineering in the last decade. Many approaches have indeed been proposed, which leverage test suites as a weak, but affordable, approximation to program specifications. Although the literature regularly sets new records on the number of benchmark bugs that can be fixed, several studies increasingly raise concerns about the limitations and biases of state-of-the-art approaches. For example, the correctness of generated patches has been questioned in a number of studies, while other researchers pointed out that evaluation schemes may be misleading with respect to the processing of fault localization results. Nevertheless, there is little work addressing the efficiency of patch generation, with regard to the practicality of program repair. In this paper, we fill this gap in the literature, by providing an extensive review on the efficiency of test suite based program repair. Our objective is to assess the number of generated patch candidates, since this information is correlated to (1) the strategy to traverse the search space efficiently in order to select sensical repair attempts, (2) the strategy to minimize the test effort for identifying a plausible patch, (3) as well as the strategy to prioritize the generation of a correct patch. To that end, we perform a large-scale empirical study on the efficiency, in terms of quantity of generated patch candidates of the 16 open-source repair tools for Java programs. The experiments are carefully conducted under the same fault localization configurations to limit biases. Eventually, among other findings, we note that: (1) many irrelevant patch candidates are generated by changing wrong code locations; (2) however, if the search space is carefully triaged, fault localization noise has little impact on patch generation efficiency; (3) yet, current template-based repair systems, which are known to be most effective in fixing a large number of bugs, are actually least efficient as they tend to generate majoritarily irrelevant patch candidates.

A large body of the literature of automated program repair develops approaches where patches are generated to be validated against an oracle (e.g., a test suite). Because such an oracle can be imperfect, the generated patches, although validated by the oracle, may actually be incorrect. While the state of the art explore research directions that require dynamic information or that rely on manually-crafted heuristics, we study the benefit of learning code representations in order to learn deep features that may encode the properties of patch correctness. Our empirical work mainly investigates different representation learning approaches for code changes to derive embeddings that are amenable to similarity computations. We report on findings based on embeddings produced by pre-trained and re-trained neural networks. Experimental results demonstrate the potential of embeddings to empower learning algorithms in reasoning about patch correctness: a machine learning predictor with BERT transformer-based embeddings associated with logistic regression yielded an AUC value of about 0.8 in the prediction of patch correctness on a deduplicated dataset of 1000 labeled patches. Our investigations show that learned representations can lead to reasonable performance when comparing against the state-of-the-art, PATCH-SIM, which relies on dynamic information. These representations may further be complementary to features that were carefully (manually) engineered in the literature.

To ensure code readability and facilitate software maintenance, program methods must be named properly. In particular, method names must be consistent with the corresponding method implementations. Debugging method names remains an important topic in the literature, where various approaches analyze commonalities among method names in a large dataset to detect inconsistent method names and suggest better ones. We note that the state-of-the-art does not analyze the implemented code itself to assess consistency. We thus propose a novel automated approach to debugging method names based on the analysis of consistency between method names and method code. The approach leverages deep feature representation techniques adapted to the nature of each artifact. Experimental results on over 2.1 million Java methods show that we can achieve up to 15 percentage points improvement over the state-of-the-art, establishing a record performance of 67.9% F1- measure in identifying inconsistent method names. We further demonstrate that our approach yields up to 25% accuracy in suggesting full names, while the state-of-the-art lags far behind at 1.1% accuracy. Finally, we report on our success in fixing 66 inconsistent method names in a live study on projects in the wild.

Properly benchmarking Automated Program Repair (APR) systems should contribute to the development and adoption of the research outputs by practitioners. To that end, the research community must ensure that it reaches significant milestones by reliably comparing state-of-the-art tools for a better understanding of their strengths and weaknesses. In this work, we identify and investigate a practical bias caused by the fault localization (FL) step in a repair pipeline. We propose to highlight the different fault localization configurations used in the literature, and their impact on APR systems when applied to the Defects4J benchmark. Then, we explore the performance variations that can be achieved by "tweaking" the FL step. Eventually, we expect to create a new momentum for (1) full disclosure of APR experimental procedures with respect to FL, (2) realistic expectations of repairing bugs in Defects4J, as well as (3) reliable performance comparison among the state-of-theart APR systems, and against the baseline performance results of our thoroughly assessed kPAR repair tool. Our main findings include: (a) only a subset of Defects4J bugs can be currently localized by commonly-used FL techniques; (b) current practice of comparing state-of-the-art APR systems (i.e., counting the number of fixed bugs) is potentially misleading due to the bias of FL configurations; and (c) APR authors do not properly qualify their performance achievement with respect to the different tuning parameters implemented in APR systems.

We revisit the performance of template-based APR to build comprehensive knowledge about the effectiveness of fix patterns, and to highlight the importance of complementary steps such as fault localization or donor code retrieval. To that end, we first investigate the literature to collect, summarize and label recurrently-used fix patterns. Based on the investigation, we build TBar, a straightforward APR tool that systematically attempts to apply these fix patterns to program bugs. We thoroughly evaluate TBar on the Defects4J benchmark. In particular, we assess the actual qualitative and quantitative diversity of fix patterns, as well as their effectiveness in yielding plausible or correct patches. Eventually, we find that, assuming a perfect fault localization, TBar correctly/plausibly fixes 74/101 bugs. Replicating a standard and practical pipeline of APR assessment, we demonstrate that TBar correctly fixes 43 bugs from Defects4J, an unprecedented performance in the literature (including all approaches, i.e., template-based, stochastic mutation-based or synthesis-based APR).

Issue tracking systems are commonly used in modern software development for collecting feedback from users and developers. An ultimate automation target of software maintenance is then the systematization of patch generation for user-reported bugs. Although this ambition is aligned with the momentum of automated program repair, the literature has, so far, mostly focused on generate-and- validate setups where fault localization and patch generation are driven by a well-defined test suite. On the one hand, however, the common (yet strong) assumption on the existence of relevant test cases does not hold in practice for most development settings: many bugs are reported without the available test suite being able to reveal them. On the other hand, for many projects, the number of bug reports generally outstrips the resources available to triage them. Towards increasing the adoption of patch generation tools by practitioners, we investigate a new repair pipeline, iFixR, driven by bug reports: (1) bug reports are fed to an IR-based fault localizer; (2) patches are generated from fix patterns and validated via regression testing; (3) a prioritized list of generated patches is proposed to developers. We evaluate iFixR on the Defects4J dataset, which we enriched (i.e., faults are linked to bug reports) and carefully-reorganized (i.e., the timeline of test-cases is naturally split). iFixR generates genuine/plausible patches for 21/44 Defects4J faults with its IR-based fault localizer. iFixR accurately places a genuine/plausible patch among its top-5 recommendation for 8/13 of these faults (without using future test cases in generation-and-validation).

Fix pattern-based patch generation is a promising direction in Automated Program Repair (APR). Notably, it has been demonstrated to produce more acceptable and correct patches than the patches obtained with mutation operators through genetic programming. The performance of pattern-based APR systems, however, depends on the fix ingredients mined from fix changes in development histories. Unfortunately, collecting a reliable set of bug fixes in repositories can be challenging. In this paper, we propose to investigate the possibility in an APR scenario of leveraging code changes that address violations by static bug detection tools. To that end, we build the AVATAR APR system, which exploits fix patterns of static analysis violations as ingredients for patch generation. Evaluated on the Defects4J benchmark, we show that, assuming a perfect localization of faults, AVATAR can generate correct patches to fix 34/39 bugs. We further find that AVATAR yields performance metrics that are comparable to that of the closely-related approaches in the literature. While AVATAR outperforms many of the state-of-the-art pattern-based APR systems, it is mostly complementary to current approaches. Overall, our study highlights the relevance of static bug finding tools as indirect contributors of fix ingredients for addressing code defects identified with functional test cases.

Automated program repair (APR) has extensively been developed by leveraging search-based techniques, in which fix ingredients are explored and identified in different granularities from a specific search space. State-of-the approaches often find fix ingredients by using mutation operators or leveraging manually-crafted templates. We argue that the fix ingredients can be searched in an online mode, leveraging code search techniques to find potentially-fixed versions of buggy code fragments from which repair actions can be extracted. In this study, we present an APR tool, LSRepair, that automatically explores code repositories to search for fix ingredients at the method-level granularity with three strategies of similar code search. Our preliminary evaluation shows that code search can drive a faster fix process (some bugs are fixed in a few seconds). LSRepair helps repair 19 bugs from the Defects4J benchmark successfully. We expect our approach to open new directions for fixing multiple-lines bugs.

Bug fixing is a time-consuming and tedious task. To reduce the manual efforts in bug fixing, researchers have presented automated approaches to software repair. Unfortunately, recent studies have shown that the state-of-the-art techniques in automated repair tend to generate patches only for a small number of bugs even with quality issues (e.g., incorrect behavior and nonsensical changes). To improve automated program repair (APR) techniques, the community should deepen its knowledge on repair actions from real-world patches since most of the techniques rely on patches written by human developers. Previous investigations on real-world patches are limited to statement level that is not sufficiently fine-grained to build this knowledge. In this work, we contribute to building this knowledge via a systematic and fine-grained study of 16,450 bug fix commits from seven Java open-source projects. We find that there are opportunities for APR techniques to improve their effectiveness by looking at code elements that have not yet been investigated. We also discuss nine insights into tuning automated repair tools. For example, a small number of statement and expression types are recurrently impacted by real-world patches, and expression-level granularity could reduce search space of finding fix ingredients, where previous studies never explored.

In this work, we investigate the practice of patch construction in the Linux kernel development, focusing on the differences between three patching processes: (1) patches crafted entirely manually to fix bugs, (2) those that are derived from warnings of bug detection tools, and (3) those that are automatically generated based on fix patterns. With this study, we provide to the research community concrete insights on the practice of patching as well as how the development community is currently embracing research and commercial patching tools to improve productivity in repair. The result of our study shows that tool-supported patches are increasingly adopted by the developer community while manually-written patches are accepted more quickly. Patch application tools enable developers to remain committed to contributing patches to the code base. Our findings also include that, in actual development processes, patches generally implement several change operations spread over the code, even for patches fixing warnings by bug detection tools. Finally, this study has shown that there is an opportunity to directly leverage the output of bug detection tools to readily generate patches that are appropriate for fixing the problem, and that are consistent with manually-written patches.