peer reviewed

Large Language Models (LLMs) have demonstrated their ability to solve tasks across various domains, including software engineering. However, their extensive number of parameters makes full fine-tuning computationally prohibitive. While Parameter-Efficient Fine-Tuning (PEFT) methods, such as adapter fine-tuning, have been proposed to address this issue, yet they typically employ default configurations that use the same adapter settings across all layers. Concurrently, Automated Machine Learning (AutoML) has demonstrated success in hyperparameter optimization, while Neural Architecture Search (NAS) has proven effective in optimizing neural network architectures. Building on these successes, we introduce AutoAdapt, a novel approach that leverages NAS to automatically discover task-specific, layer-wide adapter configurations, allowing each layer to adopt distinct adapter parameters. AutoAdapt defines a search space tailored for adapter-based fine-tuning and employs an evolutionary algorithm to explore a diverse range of configurations, thereby evaluating the benefits of customizing each layer individually. We evaluate AutoAdapt on well-established software engineering tasks, including vulnerability detection, code clone detection, and code search. Our empirical results demonstrate that AutoAdapt outperforms manually engineered adapter configurations, achieving up to a 5% improvement in F1-score for clone detection and defect detection, and up to a 25% improvement in MRR for code search. Additionally, it surpasses other PEFT techniques, such as Prefix Tuning and LoRA. Furthermore, AutoAdapt is capable of identifying configurations that outperform even full fine-tuning, while training less than 2.5% of the model parameters. A comprehensive analysis reveals that factors such as selective layer adaptation, module selection (e.g., attention versus feed-forward layers), normalization, and dropout significantly influence performance across different tasks. Additionally, our findings suggest the possibility of transferring adapter configurations to similar datasets and tasks, thus simplifying the search for optimal PEFT settings. Our code and data are available for access at: https://github.com/serval-uni-lu/AutoAdapt .

The rapid evolution of machine learning model training has outpaced the development of corresponding measurement and testing tools, leading to two significant challenges. Firstly, developers of deep learning frameworks struggle to identify operators that fail to meet precision criteria, as these issues may only manifest in a few data points. Secondly, model trainers lack effective methods to estimate precision loss caused by operators during training. To address these issues, we introduce a Pythonic framework inspired by common network layer definitions in deep learning. Our framework includes two new layers: the Fuzz Layer and the Check Layer, designed to aid in measurement and testing. The Fuzz Layer introduces minor perturbations to tensor inputs for any deterministic layer under testing (LUT). The Check Layer then measures precision by analyzing the differences before and after the perturbation. This approach estimates a lower bound of the maximal relative error and alerts developers or trainers of potential bugs if the difference exceeds a pre-defined tolerance. Additionally, Check Layers can be used independently to conduct precision tests for specific layers, ensuring the precision of operators during runtime. Despite the additional memory and time requirements, this runtime testing ensures proper training of the original model. We demonstrate the utility of our framework, FuMi, through two experiments. First, we tested 21 torch operators across nine popular machine learning models using PyTorch for various tasks, finding that the conv2d and linear operators often fail to meet precision requirements. Second, to showcase the generalizability of our framework, we tested the ATTENTION operator. By comparing different implementations of state-of-the-art ATTENTION operators, we found that the maximum relative error of the ATTENTION operator is not less than 1%, which is 13 times larger than that measured by Predoo (a unit test tool). This framework provides a robust solution for identifying precision issues in deep learning models, ensuring reliable and accurate model training.

Flaky tests, which exhibit non-deterministic behavior and fail without changes to the codebase, pose significant challenges to the reliability and efficiency of software testing processes. Despite extensive research on flaky tests in traditional unit and integration testing, their impact and prevalence within web user interface (UI) testing remains relatively unexplored, especially concerning Document Object Model (DOM) events. In web applications, DOM-related flakiness, resulting from unstable interactions between DOM and events, is particularly prevalent. This study conducts an empirical analysis of 123 flaky tests in 49 open-source web projects, focusing on the correlation between DOM event interactions and test flakiness. Our findings indicate that DOM events, and their associated interactions with the application, can introduce flakiness in web UI tests; these events are frequently associated with Event-DOM interactions (32.5 %), Event operations (22.8 %), and Response evaluations (16.3 %). The analysis of DOM consistency and event interaction levels reveals that element-level interactions across multiple DOMs are more likely to cause flakiness than interactions confined to a single DOM or occurring at the page level. Furthermore, the primary strategies used by developers to handle these issues involve synchronizing DOM interactions (50.4%), managing conditional event completion (38.2%), and ensuring consistent DOM state transitions (11.4%). We discovered that the Event-DOM category has the highest fixed frequency (2.6 times), while the DOM category on sole takes the longest time to resolve (153.4 days). This study provides practical insights into improving web application testing practices by highlighting the importance of understanding and managing DOM event interactions.

As DRAM scales to smaller nodes, commercial off-the-shelf (COTS) memory achieves higher density and lower cost but faces increased inherent single-event upset (SEU) susceptibility, complicating the adoption of energyefficient Low-Power Double Data Rate (LPDDR) SDRAM, ideal for power-constrained satellites like CubeSats. Radiationhardened memory, while more SEU-resistant, is increasingly becoming impractical for modern space systems due to lower bit density, higher cost, and infeasibility with computeintensive tasks like AI/ML applications. Additionally, limited public SEU radiation test data on LPDDR devices, particularly for emerging failure modes, constrains reliability assessment of COTS ECC adequacy in masking radiation-induced memory faults. This also hinders the development of fault-tolerance measures that could enable wider LPDDR adoption in space systems through improved reliability. In this paper, we evaluate the SEU susceptibility of COTS LPDDR memory through proton irradiation tests using 20-58 MeV proton beams on two architecturally distinct SoC platforms: the Raspberry Pi Zero 2 W (512 MB LPDDR2 SDRAM) and the NXP i.MX 8M Plus (4096 MB LPDDR4 SDRAM). Constant memory stress testing through Linux during irradiation exposed several previously unreported failure modes, including SBUs, MBUs, row-hammer faults, and stress-ng-detected modulo-x, gray code, rand-set, and modulo X faults. The resulting dataset provides key insights for COTS ECC optimization and informs the suitability of LPDDR for modern space systems.

Regularly testing deep learning-powered systems on newly collected data is critical to ensure their reliability, robustness, and efficacy in real-world applications. This process is demanding due to the significant time and human effort required for labeling new data. While test selection methods alleviate manual labor by labeling and evaluating only a subset of data while meeting testing criteria, we observe that such methods with reported promising results are simply evaluated, e.g., testing on original test data. The question arises: are they always reliable? In this paper, we explore when and to what extent test selection methods fail. First, we identify potential pitfalls of 11 selection methods based on their construction. Second, we conduct a study to empirically confirm the existence of these pitfalls. Furthermore, we demonstrate how pitfalls can break the reliability of these methods. Concretely, methods for fault detection suffer from data that are: 1) correctly classified but uncertain, or 2) misclassified but confident. Remarkably, the test relative coverage achieved by such methods drops by up to 86.85%. Besides, methods for performance estimation are sensitive to the choice of intermediate-layer output. The effectiveness of such methods can be even worse than random selection when using an inappropriate layer.

Asynchronous waits are a common root cause of flaky tests and a major time-influential factor of Web application testing. We build a dataset of 49 reproducible asynchronous wait flaky tests and their fixes from 26 open source projects to study their characteristics in Web testing. Our study reveals that developers adjusted wait time to address asynchronous wait flakiness in about 63% of cases (31 out of 49), even when the underlying causes lie elsewhere. From this, we introduce TRaf , an automated time-based repair for asynchronous wait flakiness in Web applications. TRaf determines appropriate wait times for asynchronous calls in Web applications by analyzing code similarity and past change history. Its key insight is that efficient wait times can be inferred from the current or past codebase since developers tend to repeat similar mistakes. Our analysis shows that TRaf can statically suggest a shorter wait time to alleviate async wait flakiness immediately upon the detection, reducing test execution time by 11.1% compared to the timeout values initially chosen by developers. With optional dynamic tuning, TRaf can reduce the execution time by 16.8% in its initial refinement compared to developer-written patches and by 6.2% compared to the post-refinements of these original patches. Overall, we sent 16 pull requests from our dataset, each fixing one test, to the developers. So far, three have been accepted by the developers.

Mutation testing aims to simulate real-world defects, but existing tools often struggle to replicate method invocation defects accurately. To address this, we propose MIN (Method INvocation mutator), which uses a mapping strategy to pair method names with corresponding values, ensuring that methods share argument and return types. This method enhances the feasibility and realism of mutants by considering factors such as library methods, access control, inheritance, and static methods. Experimental results show that integrating MIN into Major (a popular mutation tool) improves semantic similarity to real defects by 11%, increases mutant set diversity to 97.5%, and reduces undetected faults by 38.5%. Furthermore, MIN’s performance rivals that of state-of-the-art machine learning-based mutators like CodeBERT, with a 10x speed advantage over CodeBERT and 4x over DeepMutation in generating compilable mutants. These findings demonstrate that MIN can significantly enhance defect simulation and improve the efficiency of mutation testing.

LTL solvers check the satisfiability of Linear-time Temporal Logic (LTL) formulas and are widely used for verifying and testing critical software systems. Thus, potential bugs in the solvers' implementations can have a significant impact. We present SpecBCFuzz, a fuzzing method for finding bugs in LTL solvers, that is guided by boundary conditions (BCs), corner cases whose (un)satisfiability depends on rare traces. SpecBCFuzz implements a search-based algorithm that fuzzes LTL formulas giving relevance to BCs. It integrates syntactic and semantic similarity metrics to explore the vicinity of the seeded formulas with BCs. We evaluate SpecBCFuzz on 21 different configurations (including the latest and past releases) of four mature and state-of-the-art LTL solvers (NuSMV, Black, Aalta, and PLTL) that implement a diverse set of satisfiability algorithms. SpecBCFuzz produces 368,716 bug-triggering formulas, detecting bugs in 18 out of the 21 solvers' configurations we study. Overall, SpecBCFuzz reveals: soundness issues (wrong answers given by a solver) in Aalta and PLTL; crashes, e.g., segmentation faults, in NuSMV, Black and Aalta; flaky behaviors (different responses across re-runs of the solver on the same formula) in NuSMV and Aalta; performance bugs (large time performance degradation between successive versions of the solver on the same formula) in Black, Aalta and PLTL; and no bug in NuSMV BDD (all versions), suggesting that the latter is currently the most robust solver.

With the release of powerful language models trained on large code corpus (e.g., CodeBERT, trained on 6.4 million programs), a new family of mutation testing tools has arisen that promises to generate more “natural” mutants, where the mutated code aims at following the implicit rules and coding conventions produced by programmers. In this paper, we empirically study the observable behavior of CodeBERT-generated mutants and to what extent are these coupled with software vulnerabilities. To do so, we carefully analyze 45 reproducible vulnerabilities from the Vul4J dataset to determine whether the mutants and vulnerabilities fail the same tests and whether the failures are for the same reasons or not. Hence, we define different degrees of vulnerability-coupling classes. Strongly coupled mutants fail the same tests for the same reasons as the vulnerabilities, while test coupled mutants fail the same tests but for some different reason as the vulnerabilities. Partial coupling classes are also considered. Overall, CodeBERT-generated mutants strongly coupled with 32 out of these 45 vulnerabilities (i.e. The mutants fail on the same tests for the same reasons), while another 7 vulnerabilities are test-coupled by CodeBERT mutants (i.e. The mutants fail on the same tests but not for the same reasons). Interestingly, CodeBERT mutants are diverse enough to couple vulnerabilities from 14 out of the 15 types of vulnerabilities explored, i.e., CWEs (Common Weakness Enumeration). Finally, we observe that strongly coupled mutants are scarce (1.17 % of the killable mutants), test coupled mutants represent 7.2 %, and 64.9 % of the killable mutants are not coupled with the vulnerabilities.

Flaky tests are tests that pass and fail on different executions of the same version of a program under test. They waste valuable developer time by making developers investigate false alerts (flaky test failures). To deal with this issue, many prediction methods have been proposed. However, the utility of these methods remains unclear since they are typically evaluated based on single-release data, ignoring that in many cases tests that fail flakily in one release also correctly fail (indicating the presence of bugs) in some other, meaning that it is possible for subsequent correctly-failing cases to pass unnoticed. In this paper, we show that this situation is prevalent and can raise significant concerns for both researchers and practitioners. In particular, we show that flaky tests, tests that exhibit flaky behaviour at some point in time, have a strong fault-revealing capability, i.e., they reveal more than 1/3 of all encountered regression faults. We also show that 76.2%, of all test executions that reveal faults in the codebase under test are made by tests that are classified as flaky by existing prediction methods. Overall, our findings motivate the need for future research to focus on predicting flaky test executions instead of flaky tests.

Abstract Graph Neural Networks (GNNs) have gained prominence in various domains, such as social network analysis, recommendation systems, and drug discovery, due to their ability to model complex relationships in graph-structured data. GNNs can exhibit incorrect behavior, resulting in severe consequences. Therefore, testing is necessary and pivotal. However, labeling all test inputs for GNNs can be prohibitively costly and time-consuming, especially when dealing with large and complex graphs. In response to these challenges, test selection has emerged as a strategic approach to alleviate labeling expenses. The objective of test selection is to select a subset of tests from the complete test set. While various test selection techniques have been proposed for traditional deep neural networks (DNNs), their adaptation to GNNs presents unique challenges due to the distinctions between DNN and GNN test data. Specifically, DNN test inputs are independent of each other, whereas GNN test inputs (nodes) exhibit intricate interdependencies. Therefore, it remains unclear whether DNN test selection approaches can perform effectively on GNNs. To fill the gap, we conduct an empirical study that systematically evaluates the effectiveness of various test selection methods in the context of GNNs, focusing on three critical aspects: 1) Misclassification detection : selecting test inputs that are more likely to be misclassified; 2) Accuracy estimation : selecting a small set of tests to precisely estimate the accuracy of the whole testing set; 3) Performance enhancement : selecting retraining inputs to improve the GNN accuracy. Our empirical study encompasses 7 graph datasets and 8 GNN models, evaluating 22 test selection approaches. Our study includes not only node classification datasets but also graph classification datasets. Our findings reveal that: 1) In GNN misclassification detection, confidence-based test selection methods, which perform well in DNNs, do not demonstrate the same level of effectiveness; 2) In terms of GNN accuracy estimation, clustering-based methods, while consistently performing better than random selection, provide only slight improvements; 3) Regarding selecting inputs for GNN performance improvement, test selection methods, such as confidence-based and clustering-based test selection methods, demonstrate only slight effectiveness; 4) Concerning performance enhancement, node importance-based test selection methods are not suitable, and in many cases, they even perform worse than random selection.

Graph Neural Networks (GNNs) have achieved promising performance in a variety of practical applications. Similar to traditional DNNs, GNNs could exhibit incorrect behavior that may lead to severe consequences, and thus testing is necessary and crucial. However, labeling all the test inputs for GNNs can be costly and time-consuming, especially when dealing with large and complex graphs, which seriously affects the efficiency of GNN testing. Existing studies have focused on test prioritization for DNNs, which aims to identify and prioritize fault-revealing tests (i.e., test inputs that are more likely to be misclassified) to detect system bugs earlier in a limited time. Although some DNN prioritization approaches have been demonstrated effective, there is a significant problem when applying them to GNNs: They do not take into account the connections (edges) between GNN test inputs (nodes), which play a significant role in GNN inference. In general, DNN test inputs are independent of each other, while GNN test inputs are usually represented as a graph with complex relationships between each test. In this article, we propose GraphPrior ( GNN -oriented Test Prior itization), a set of approaches to prioritize test inputs specifically for GNNs via mutation analysis. Inspired by mutation testing in traditional software engineering, in which test suites are evaluated based on the mutants they kill, GraphPrior generates mutated models for GNNs and regards test inputs that kill many mutated models as more likely to be misclassified. Then, GraphPrior leverages the mutation results in two ways, killing-based and feature-based methods. When scoring a test input, the killing-based method considers each mutated model equally important, while feature-based methods learn different importance for each mutated model through ranking models. Finally, GraphPrior ranks all the test inputs based on their scores. We conducted an extensive study based on 604 subjects to evaluate GraphPrior on both natural and adversarial test inputs. The results demonstrate that KMGP, the killing-based GraphPrior approach, outperforms the compared approaches in a majority of cases, with an average improvement of 4.76% ~49.60% in terms of APFD. Furthermore, the feature-based GraphPrior approach, RFGP, performs the best among all the GraphPrior approaches. On adversarial test inputs, RFGP outperforms the compared approaches across different adversarial attacks, with the average improvement of 2.95% ~46.69%.

Code search is a common yet important activity of software developers. An efficient code search model can largely facilitate the development process and improve the programming quality. Given the superb performance of learning the contextual representations, deep learning models, especially pre-trained language models, have been widely explored for the code search task. However, studies mainly focus on proposing new architectures for ever-better performance on designed test sets but ignore the performance on unseen test data where only natural language queries are available. The same problem in other domains, e.g., CV and NLP, is usually solved by test input selection that uses a subset of the unseen set to reduce the labeling effort. However, approaches from other domains are not directly applicable and still require labeling effort. In this article, we propose the k NN-b a sed p erformance t e sting ( KAPE ) to efficiently solve the problem without manually matching code snippets to test queries. The main idea is to use semantically similar training data to perform the evaluation. Extensive experiments on six programming language datasets, three state-of-the-art pre-trained models, and seven baseline methods demonstrate that KAPE can effectively assess the model performance (e.g., CodeBERT achieves MRR 0.5795 on JavaScript) with a slight difference (e.g., 0.0261).

This article presents a comprehensive survey on test optimization in deep neural network (DNN) testing. Here, test optimization refers to testing with low data labeling effort. We analyzed 90 papers, including 43 from the software engineering (SE) community, 32 from the machine learning (ML) community, and 15 from other communities. Our study: (i) unifies the problems as well as terminologies associated with low-labeling cost testing, (ii) compares the distinct focal points of SE and ML communities, and (iii) reveals the pitfalls in existing literature. Furthermore, we highlight the research opportunities in this domain.

Applying deep learning (DL) to science is a new trend in recent years, which leads DL engineering to become an important problem. Although training data preparation, model architecture design, and model training are the normal processes to build DL models, all of them are complex and costly. Therefore, reusing the open-sourced pre-trained model is a practical way to bypass this hurdle for developers. Given a specific task, developers can collect massive pre-trained deep neural networks from public sources for reusing. However, testing the performance (e.g., accuracy and robustness) of multiple deep neural networks (DNNs) and recommending which model should be used is challenging regarding the scarcity of labeled data and the demand for domain expertise. In this article, we propose a labeling-free (LaF) model selection approach to overcome the limitations of labeling efforts for automated model reusing. The main idea is to statistically learn a Bayesian model to infer the models’ specialty only based on predicted labels. We evaluate LaF using nine benchmark datasets, including image, text, and source code, and 165 DNNs, considering both the accuracy and robustness of models. The experimental results demonstrate that LaF outperforms the baseline methods by up to 0.74 and 0.53 on Spearman’s correlation and Kendall’s τ, respectively.

Machine learning has achieved remarkable success across diverse domains. Nevertheless, concerns about interpretability in black-box models, especially within Deep Neural Networks (DNNs), have become pronounced in safety-critical fields like healthcare and finance. Classical machine learning (ML) classifiers, known for their higher interpretability, are preferred in these domains. Similar to DNNs, classical ML classifiers can exhibit bugs that could lead to severe consequences in practice. Test input prioritization has emerged as a promising approach to ensure the quality of an ML system, which prioritizes potentially misclassified tests so that such tests can be identified earlier with limited manual labeling costs. However, when applying to classical ML classifiers, existing DNN test prioritization methods are constrained from three perspectives: 1) Coverage-based methods are inefficient and time-consuming; 2) Mutation-based methods cannot be adapted to classical ML models due to mismatched model mutation rules; 3) Confidence-based methods are restricted to a single dimension when applying to binary ML classifiers, solely depending on the model’s prediction probability for one class. To overcome the challenges, we propose MLPrior, a test prioritization approach specifically tailored for classical ML models. MLPrior leverages the characteristics of classical ML classifiers (i.e., interpretable models and carefully engineered attribute features) to prioritize test inputs. The foundational principles are: 1) tests more sensitive to mutations are more likely to be misclassified, and 2) tests closer to the model’s decision boundary are more likely to be misclassified. Building on the first concept, we design mutation rules to generate two types of mutation features (i.e.,model mutation featuresandinput mutation features) for each test. Drawing from the second notion, MLPrior generatesattribute featuresof each test based on its attribute values, which can indirectly reveal the proximity between the test and the decision boundary. For each test, MLPrior combines all three types of features of it into a final vector. Subsequently, MLPrior employs a pre-trained ranking model to predict the misclassification probability of each test based on its final vector and ranks tests accordingly. We conducted an extensive study to evaluate MLPrior based on 185 subjects, encompassing natural datasets, mixed noisy datasets, and fairness datasets. The results demonstrate that MLPrior outperforms all the compared test prioritization approaches, with an average improvement of 14.74%∼66.93% on natural datasets, 18.55%∼67.73% on mixed noisy datasets, and 15.34%∼62.72% on fairness datasets.

The costly human effort required to prepare the training data of machine learning (ML) models hinders their practical development and usage in software engineering (ML4Code), especially for those with limited budgets. Therefore, efficiently training models of code with less human effort has become an emergent problem. Active learning is such a technique to address this issue that allows developers to train a model with reduced data while producing models with desired performance, which has been well studied in computer vision and natural language processing domains. Unfortunately, there is no such work that explores the effectiveness of active learning for code models. In this paper, we bridge this gap by building the first benchmark to study this critical problem - active code learning. Specifically, we collect 11 acquisition functions (which are used for data selection in active learning) from existing works and adapt them for code-related tasks. Then, we conduct an empirical study to check whether these acquisition functions maintain performance for code data. The results demonstrate that feature selection highly affects active learning and using output vectors to select data is the best choice. For the code summarization task, active code learning is ineffective which produces models with over a 29.64% gap compared to the expected performance. Furthermore, we explore future directions of active code learning with an exploratory study. We propose to replace distance calculation methods with evaluation metrics and find a correlation between these evaluation-based distance methods and the performance of code models.

Mutation testing drives the creation and improvement of test cases by evaluating their ability to identify synthetic faults. Over the past decades, the technique has gained popularity in academic circles. In practice, however, little is known about its adoption and use. While there are some pilot studies applying mutation testing in industry, the overall usage of mutation testing among developers remains largely unexplored. To fill this gap, this paper presents the results of a qualitative study among open-source developers on the use of mutation testing. Specifically, we report the results of a survey of 104 contributors to open-source projects using a variety of mutation testing tools. The findings of our study provide helpful insights into the use of mutation testing in practice, including its main benefits and limitations. Overall, we observe a high degree of satisfaction with mutation testing across different programming languages and mutation testing tools. Developers find the technique helpful for improving the quality of test suites, detecting bugs, and improving code maintainability. Popularity, usability, and configurability emerge as key factors for the adoption of mutation tools, whereas performance stands overwhelmingly as their main limitation. These results lay the groundwork for new research contributions and tools that meet the needs of developers and boost the widespread adoption of mutation testing.

Flaky tests are tests that yield different outcomes when run on the same version of a program. This non-deterministic behaviour plagues continuous integration with false signals, wasting developers’ time and reducing their trust in test suites. Studies highlighted the importance of keeping tests flakiness-free. Recently, the research community has been pushing towards the detection of flaky tests by suggesting many static and dynamic approaches. While promising, those approaches mainly focus on classifying tests as flaky or not and, even when high performances are reported, it remains challenging to understand the cause of flakiness. This part is crucial for researchers and developers that aim to fix it. To help with the comprehension of a given flaky test, we propose FlakyCat, the first approach to classify flaky tests based on their root cause category. FlakyCat relies on CodeBERT for code representation and leverages Siamese networks to train a multi-class classifier. We train and evaluate FlakyCat on a set of 451 flaky tests collected from open-source Java projects. Our evaluation shows that FlakyCat categorises flaky tests accurately, with an F1 score of 73%. Furthermore, we investigate the performance of our approach for each category, revealing that Async waits, Unordered collections and Time-related flaky tests are accurately classified, while Concurrency-related flaky tests are more challenging to predict. Finally, to facilitate the comprehension of FlakyCat’s predictions, we present a new technique for CodeBERT-based model interpretability that highlights code statements influencing the categorization.

Recently many mutation testing tools have been proposed that rely on bug-fix patterns and natural language models trained on large code corpus. As these tools operate fundamentally differently from the grammar-based traditional approaches, a question arises of how these tools compare in terms of 1) fault detection and 2) cost-effectiveness. Simultaneously, mutation testing research proposes mutant selection approaches based on machine learning to mitigate its application cost. This raises another question: How do the existing mutation testing tools compare when guided by mutant selection approaches? To answer these questions, we compare four existing tools – μBERT (uses pre-trained language model for fault seeding), IBIR (relies on inverted fix-patterns), DeepMutation (generates mutants by employing Neural Machine Translation) and PIT (applies standard grammar-based rules) in terms of fault detection capability and cost-effectiveness, in conjunction with standard and deep learning based mutant selection strategies. Our results show that IBIR has the highest fault detection capability among the four tools; however, it is not the most cost-effective when considering different selection strategies. On the other hand, μBERT having a relatively lower fault detection capability, is the most cost-effective among the four tools. Our results also indicate that comparing mutation testing tools when using deep learning-based mutant selection strategies can lead to different conclusions than the standard mutant selection. For instance, our results demonstrate that combining μBERT with deep learning-based mutant selection yields 12% higher fault detection than the considered tools.

Deep Neural Networks (DNNs) have gained considerable attention in the past decades due to their astounding performance in different applications, such as natural language modeling, self-driving assistance, and source code understanding. With rapid exploration, more and more complex DNN architectures have been proposed along with huge pre-trained model parameters. A common way to use such DNN models in user-friendly devices (e.g., mobile phones) is to perform model compression before deployment. However, recent research has demonstrated that model compression, e.g., model quantization, yields accuracy degradation as well as output disagreements when tested on unseen data. Since the unseen data always include distribution shifts and often appear in the wild, the quality and reliability of models after quantization are not ensured. In this paper, we conduct a comprehensive study to characterize and help users understand the behaviors of quantization models. Our study considers four datasets spanning from image to text, eight DNN architectures including both feed-forward neural networks and recurrent neural networks, and 42 shifted sets with both synthetic and natural distribution shifts. The results reveal that 1) data with distribution shifts lead to more disagreements than without. 2) Quantization-aware training can produce more stable models than standard, adversarial, and Mixup training. 3) Disagreements often have closer top-1 and top-2 output probabilities, and Margin is a better indicator than other uncertainty metrics to distinguish disagreements. 4) Retraining the model with disagreements has limited efficiency in removing disagreements. We release our code and models as a new benchmark for further study of model quantization.

Abstract System goals are the statements that, in the context of software requirements specification, capture how the software should behave. Many times, the understanding of stakeholders on what the system should do, as captured in the goals, can lead to different problems, from clearly contradicting goals, to more subtle situations in which the satisfaction of some goals inhibits the satisfaction of others. These latter issues, calledgoal divergences, are the subject ofgoal conflict analysis, which consists of identifying, assessing, and resolving divergences, as part of a more general activity known as goal refinement. While there exist techniques that, when requirements are expressed formally, can automatically identify and assess goal conflicts, there is currently no automated approach to support engineers inresolvingidentified divergences. In this paper, we presentACoRe, the first approach that automatically proposes potential resolutions to goal conflicts, in requirements specifications formally captured using linear-time temporal logic.ACoResystematically explores syntactic modifications of the conflicting specifications, aiming at obtaining resolutions that disable previously identified conflicts, while preserving specification consistency.ACoReintegrates modern multi-objective search algorithms (in particular, NSGA-III, WBGA, and AMOSA) to produce resolutions that maintain coherence with the original conflicting specification, by searching for specifications that are eithersyntacticallyorsemanticallysimilar to the original specification. We assessACoReon 25 requirements specifications taken from the literature. We show thatACoRecan successfully produce various conflict resolutions for each of the analyzed case studies, including resolutions that resemble specification repairs manually provided as part of conflict analyses.

The reactive synthesis problem consists of automatically producing correct-by-construction operational models of systems from high-level formal specifications of their behaviours. However, specifications are often unrealisable, meaning that no system can be synthesised from the specification. To deal with this problem, we present AuRUS, a search-based approach to repair unrealisable Linear-Time Temporal Logic (LTL) specifications. AuRUS aims at generating solutions that are similar to the original specifications by using the notions of syntactic and semantic similarities. Intuitively, the syntactic similarity measures the text similarity between the specifications, while the semantic similarity measures the number of behaviours preserved/removed by the candidate repair. We propose a new heuristic based on model counting to approximate semantic similarity. We empirically assess AuRUS on many unrealisable specifications taken from different benchmarks and show that it can successfully repair all of them. Also, compared to related techniques, AuRUS can produce many unique solutions while showing more scalability.

While leveraging additional training data is well established to improve adversarial robustness, it incurs the unavoidable cost of data collection and the heavy computation to train models. To mitigate the costs, we propose *Guided Adversarial Training * (GAT), a novel adversarial training technique that exploits auxiliary tasks under a limited set of training data. Our approach extends single-task models into multi-task models during the min-max optimization of adversarial training, and drives the loss optimization with a regularization of the gradient curvature across multiple tasks. GAT leverages two types of auxiliary tasks: self-supervised tasks, where the labels are generated automatically, and domain-knowledge tasks, where human experts provide additional labels. Experimentally, under limited data, GAT increases the robust accuracy on CIFAR-10 up to four times (from 11% to 42% robust accuracy) and the robust AUC of CheXpert medical imaging dataset from 50% to 83%. On the full CIFAR-10 dataset, GAT outperforms eight state-of-the-art adversarial training strategies. Our large study across five datasets and six tasks demonstrates that task augmentation is an efficient alternative to data augmentation, and can be key to achieving both clean and robust performances.

Deep learning (DL) plays a more and more important role in our daily life due to its competitive performance in industrial application domains. As the core of DL-enabled systems, deep neural networks (DNNs) need to be carefully evaluated to ensure the produced models match the expected requirements. In practice, the de facto standard to assess the quality of DNNs in the industry is to check their performance (accuracy) on a collected set of labeled test data. However, preparing such labeled data is often not easy partly because of the huge labeling effort, i.e., data labeling is labor-intensive, especially with the massive new incoming unlabeled data every day. Recent studies show that test selection for DNN is a promising direction that tackles this issue by selecting minimal representative data to label and using these data to assess the model. However, it still requires human effort and cannot be automatic. In this paper, we propose a novel technique, named Aries, that can estimate the performance of DNNs on new unlabeled data using only the information obtained from the original test data. The key insight behind our technique is that the model should have similar prediction accuracy on the data which have similar distances to the decision boundary. We performed a large-scale evaluation of our technique on two famous datasets, CIFAR-10 and Tiny-ImageNet, four widely studied DNN models including ResNetl0l and DenseNetl21, and 13 types of data transformation methods. Results show that the estimated accuracy by Aries is only 0.03% - 2.60% off the true accuracy. Besides, Aries also outperforms the state-of-the-art labeling-free methods in 50 out of 52 cases and selection-labeling-based methods in 96 out of 128 cases.

Much research on automated program debugging often assumes that bug fix location(s) indicate the faults' root causes and that root causes of faults lie within single code elements (statements). It is also often assumed that the number of statements a developer would need to inspect before finding the first faulty statement reflects debugging effort. Although intuitive, these three assumptions are typically used (55% of experiments in surveyed publications make at least one of these three assumptions) without any consideration of their effects on the debugger's effectiveness and potential impact on developers in practice. To deal with this issue, we perform controlled experimentation, split testing in particular, using 352 bugs from 46 open-source C programs, 19 Automated Fault Localization (AFL) techniques (18 statistical debugging formulas and dynamic slicing), two (2) state-of-the-art automated program repair (APR) techniques (GenProg and Angelix) and 76 professional developers. Our results show that these assumptions conceal the difficulty of debugging. They make AFL techniques appear to be (up to 38%) more effective, and make APR tools appear to be (2X) less effective. We also find that most developers (83%) consider these assumptions to be unsuitable for debuggers and, perhaps worse, that they may inhibit development productivity. The majority (66%) of developers prefer debugging diagnoses without these assumptions twice as much as with the assumptions. Our findings motivate the need to assess debuggers conservatively, i.e., without these assumptions.

Assertion inference techniques aim at automatically inferring sets of program assertions that capture the exhibited software behavior, often by generating and filtering assertions through dynamic test executions and mutation testing. Although powerful, such techniques are computationally expensive due to the large number of mutants that require execution. In this study, we introduce the notion of Assertion Inferring Mutants, and demonstrate that these mutants are sufficient for assertion inference and correspond to a small subset (12.95%) of the entire mutant set. Moreover, these mutants are significantly different (71.59%) from Subsuming Mutants that are frequently cited by mutation testing literature. We also show that Assertion Inferring Mutants can be statically approximated via a learning-based method. Given the widespread adoption of encoder-decoder architecture for prediction tasks, we demonstrate that it predicts Assertion Inferring Mutants with 0.79 Precision and 0.49 Recall. Its evaluation on 46 projects showcases that it enables a comparable inference capability (missing only 12.49% assertions) with a complete mutation analysis, while significantly reducing the execution cost (achieving 46.29 times faster inference). Moreover, it enables assertion inference techniques to scale on subjects where complete mutation testing is prohibitively expensive and other mutant selection strategies do not lead to an acceptable assertion inference.

Mutation testing (MT) for deep learning (DL) has gained huge attention in the past few years. However, how MT can really help DL is still unclear. In this paper, we introduce one promising direction for the usage of mutants. Specifically, since mutants can be seen as one kind of ensemble model and ensemble model can be used to boost the adversarial attack, we propose MUTEN, which applies the attack on mutants to improve the success rate of well-known attacks against gradient-masking models. Experimental results on MNIST, SVHN, and CIFAR-10 show that MUTEN can increase the success rate of four attacks by up to 45%. Furthermore, experiments on four defense approaches, bit-depth reduction, JPEG compression, Defensive distillation, and Label smoothing, demonstrate that MUTEN can break the defense models effectively by enhancing the attacks with the success rate of up to 96%.

Mutation testing has been demonstrated to be one of the most powerful fault-revealing tools in the tester's tool kit. Much previous work implicitly assumed it to be sufficient to re-compute mutant suites per release. Sadly, this makes mutation results inconsistent; mutant scores from each release cannot be directly compared, making it harder to measure test improvement. Furthermore, regular code change means that a mutant suite's relevance will naturally degrade over time. We measure this degradation in relevance for 143,500 mutants in 4 non-trivial systems, finding that 52% degrade, on average. We introduce a mutant brittleness measure and use it to audit software systems and their mutation suites. We also demonstrate how consistent-by-construction long-standing mutant suites can be identified with a 10x improvement in mutant relevance over an arbitrary test suite. Our results indicate that the research community should avoid the re-computation of mutant suites and focus, instead, on long-standing mutants, thereby improving the consistency and relevance of mutation testing.

Inspired by the great success of Deep Neural Networks (DNNs) in natural language processing (NLP), DNNs have been increasingly applied in source code analysis and attracted significant attention from the software engineering community. Due to its data-driven nature, a DNN model requires massive and high-quality labeled training data to achieve expert-level performance. Collecting such data is often not hard, but the labeling process is notoriously laborious. The task of DNN-based code analysis even worsens the situation because source code labeling also demands sophisticated expertise. Data augmentation has been a popular approach to supplement training data in domains such as computer vision and NLP. However, existing data augmentation approaches in code analysis adopt simple methods, such as data transformation and adversarial example generation, thus bringing limited performance superiority. In this paper, we propose a data augmentation approach MixCode that aims to effectively supplement valid training data, inspired by the recent advance named Mixup in computer vision. Specifically, we first utilize multiple code refactoring methods to generate transformed code that holds consistent labels with the original data. Then, we adapt the Mixup technique to mix the original code with the transformed code to augment the training data. We evaluate MixCode on two programming languages (Java and Python), two code tasks (problem classification and bug detection), four benchmark datasets (JAVA250, Python800, CodRep1, and Refactory), and seven model architectures (including two pretrained models CodeBERT and GraphCodeBERT). Experimental results demonstrate that MixCode outperforms the baseline data augmentation approach by up to 6.24% in accuracy and 26.06% in robustness.

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%.

Context:When software evolves, opportunities for introducing faults appear. Therefore, it is important to test the evolved program behaviors during each evolution cycle. However, while software evolves, its complexity is also evolving, introducing challenges to the testing process. To deal with this issue, testing techniques should be adapted to target the effect of the program changes instead of the entire program functionality. To this end,commit-aware mutation testing, a powerful testing technique, has been proposed. Unfortunately, commit-aware mutation testing is challenging due to the complex program semantics involved. Hence, it is pertinent to understand the characteristics, predictability, and potential of the technique. Objective:We conduct an exploratory study to investigate the properties ofcommit-relevant mutants, i.e., the test elements of commit-aware mutation testing, by proposing a general definition and an experimental approach to identify them. We thus aim at investigating the prevalence, location, and comparative advantages of commit-aware mutation testing over time (i.e., the program evolution). We also investigate the predictive power of several commit-related features in identifying and selecting commit-relevant mutants to understand the essential properties for its best-effort application case. Method:Our commit-relevant definition relies on the notion of observational slicing, approximated by higher-order mutation. Specifically, our approach utilizes the impact of mutants, effects of one mutant on another in capturing and analyzing the implicit interactions between the changed and unchanged code parts. The study analyses millions of mutants (over 10 million), 288 commits, five (5) different open-source software projects involving over 68,213 CPU days of computation and sets a ground truth where we perform our analysis. Results:Our analysis shows that commit-relevant mutants arelocated mainly outside of program commit change(81%), suggesting a limitation in previous work. We also note that effective selection of commit-relevant mutants has the potential of reducing the number of mutants by up to 93%. In addition, we demonstrate that commit relevant mutation testing is significantly more effective and efficient than state-of-the-art baselines, i.e., random mutant selection and analysis of only mutants within the program change. In our analysis of the predictive power of mutants and commit-related features (e.g., number of mutants within a change, mutant type, and commit size) in predicting commit-relevant mutants, we found that mostproxy features do not reliably predict commit-relevant mutants. Conclusion:This empirical study highlights the properties of commit-relevant mutants and demonstrates the importance of identifying and selecting commit-relevant mutants when testing evolving software systems.

Mutation testing research has indicated that a major part of its application cost is due to the large number of low utility mutants that it introduces. Although previous research has identified this issue, no previous study has proposed any effective solution to the problem. Thus, it remains unclear how to mutate and test a given piece of code in a best effort way, i.e., achieving a good trade-off between invested effort and test effectiveness. To achieve this, we proposeCerebro, a machine learning approach thatstaticallyselects subsuming mutants, i.e., the set of mutants that resides on the top of the subsumption hierarchy, based on the mutants’ surrounding code context. We evaluateCerebrousing 48 and 10 programs written in C and Java, respectively, and demonstrate that it preserves the mutation testing benefits while limiting application cost, i.e., reduces all cost application factors such as equivalent mutants, mutant executions, and the mutants requiring analysis. We demonstrate thatCerebrohas strong inter-project prediction ability, which is significantly higher than two baseline methods, i.e., supervised learning on features proposed by state-of-the-art, and random mutant selection. More importantly, our results show thatCerebro’s selected mutants lead to strong tests that are respectively capable of killing 2 times higher than the number of subsuming mutants killed by the baselines when selecting the same number of mutants. At the same time,Cerebroreduces the cost-related factors, as it selects, on average, 68% fewer equivalent mutants, while requiring 90% fewer test executions than the baselines.

Fault seeding is typically used in empirical studies to evaluate and compare test techniques. Central to these techniques lies the hypothesis that artificially seeded faults involve some form of realistic properties and thus provide realistic experimental results. In an attempt to strengthen realism, a recent line of research uses machine learning techniques, such as deep learning and Natural Language Processing, to seed faults that look like (syntactically) real ones, implying that fault realism is related to syntactic similarity. This raises the question of whether seeding syntactically similar faults indeed results in semantically similar faults and, more generally whether syntactically dissimilar faults are far away (semantically) from the real ones. We answer this question by employing 4 state-of-the-art fault-seeding techniques (PiTest - a popular mutation testing tool, IBIR - a tool with manually crafted fault patterns, DeepMutation - a learning-based fault seeded framework and μBERT - a mutation testing tool based on the pre-trained language model CodeBERT) that operate in a fundamentally different way, and demonstrate that syntactic similarity does not reflect semantic similarity. We also show that 65.11%, 76.44%, 61.39% and 9.76% of the real faults of Defects4J V2 are semantically resembled by PiTest, IBIR, μBERT and DeepMutation faults, respectively.

Vulnerability to adversarial attacks is a well-known weakness of Deep Neural networks. While most of the studies focus on single-task neural networks with computer vision datasets, very little research has considered complex multi-task models that are common in real applications. In this paper, we evaluate the design choices that impact the robustness of multi-task deep learning networks. We provide evidence that blindly adding auxiliary tasks, or weighing the tasks provides a false sense of robustness. Thereby, we tone down the claim made by previous research and study the different factors which may affect robustness. In particular, we show that the choice of the task to incorporate in the loss function are important factors that can be leveraged to yield more robust models. We provide the appendix, all our algorithms, models, and open source-code at https://github.com/yamizi/taskaugment

Active learning is an established technique to reduce the labeling cost for building high-quality machine learning models. However, state-of-the-art approaches focus on maximizing the clean performance (e.g. accuracy) but disregarding robustness. In this work, we propose Robust Active Learning, an active learning process that integrates adversarial training, the most established method to produce robust models. First, we conduct an empirical study to evaluate the effectiveness of existing approaches and uncover the characteristics of data. Then, we propose a novel approach, density-based robust sampling with entropy (DRE), to target both clean performance and robustness. Our experiments are conducted on 11 acquisition functions, 4 datasets, 6 DNN architectures, and 15105 trained DNNs.

Graph-based Semi-Supervised Learning (GSSL) is a practical solution to learn from a limited amount of labelled data together with a vast amount of unlabelled data. However, due to their reliance on the known labels to infer the unknown labels, these algorithms are sensitive to data quality. It is therefore essential to study the potential threats related to the labelled data, more specifically, label poisoning. In this paper, we propose a novel data poisoning method which efficiently approximates the result of label inference to identify the inputs which, if poisoned, would produce the highest number of incorrectly inferred labels. We extensively evaluate our approach on three classification problems under 24 different experimental settings each. Compared to the state of the art, our influence-driven attack produces an average increase of error rate 50% higher, while being faster by multiple orders of magnitude. Moreover, our method can inform engineers of inputs that deserve investigation (relabelling them) before training the learning model. We show that relabelling one-third of the poisoned inputs (selected based on their influence) reduces the poisoning effect by 50%.

Machine translation plays an essential role in people's daily international communication. However, machine translation systems are far from perfect. To tackle this problem, researchers have proposed several approaches to testing machine translation. A promising trend among these approaches is to use word replacement, where only one word in the original sentence is replaced with another word to form a sentence pair. However, precise control of the impact of word replacement remains an outstanding issue in these approaches.

Much research on software testing makes an implicit assumption that test failures are deterministic such that they always witness the presence of the same defects. However, this assumption is not always true because some test failures are due to so-called flaky tests, i.e., tests with non-deterministic outcomes. To help testing researchers better investigate flakiness, we introduce a test flakiness assessment and experimentation platform, called FlakiMe. FlakiMe supports the seeding of a (controllable) degree of flakiness into the behaviour of a given test suite. Thereby, FlakiMe equips researchers with ways to investigate the impact of test flakiness on their techniques under laboratory-controlled conditions. To demonstrate the application of FlakiMe, we use it to assess the impact of flakiness on mutation testing and program repair (the PRAPR and ARJA methods). These results indicate that a 10% flakiness is sufficient to affect the mutation score, but the effect size is modest (2% - 5% ), while it reduces the number of patches produced for repair by 20% up to 100% of repair problems; a devastating impact on this application of testing. Our experiments with FlakiMe demonstrate that flakiness affects different testing applications in very different ways, thereby motivating the need for a laboratory-controllable flakiness impact assessment platform and approach such as FlakiMe.

Flaky tests are defined as tests that manifest non-deterministic behaviour by passing and failing intermittently for the same version of the code. These tests cripple continuous integration with false alerts that waste developers’ time and break their trust in regression testing. To mitigate the effects of flakiness, both researchers and industrial experts proposed strategies and tools to detect and isolate flaky tests. However, flaky tests are rarely fixed as developers struggle to localise and understand their causes. Additionally, developers working with large codebases often need to know the sources of non-determinism to preserve code quality, i.e., avoid introducing technical debt linked with non-deterministic behaviour, and to avoid introducing new flaky tests. To aid with these tasks, we propose re-targeting Fault Localisation techniques to the flaky component localisation problem, i.e., pinpointing program classes that cause the non-deterministic behaviour of flaky tests. In particular, we employ Spectrum-Based Fault Localisation (SBFL), a coverage-based fault localisation technique commonly adopted for its simplicity and effectiveness. We also utilise other data sources, such as change history and static code metrics, to further improve the localisation. Our results show that augmenting SBFL with change and code metrics ranks flaky classes in the top-1 and top-5 suggestions, in 26% and 47% of the cases. Overall, we successfully reduced the average number of classes inspected to locate the first flaky class to 19% of the total number of classes covered by flaky tests. Our results also show that localisation methods are effective in major flakiness categories, such as concurrency and asynchronous waits, indicating their general ability to identify flaky components.

Test flakiness forms a major testing concern. Flaky tests manifest non-deterministic outcomes that cripple continu-ous integration and lead developers to investigate false alerts. Industrial reports indicate that on a large scale, the accrual of flaky tests breaks the trust in test suites and entails significant computational cost. To alleviate this, practitioners are constrained to identify flaky tests and investigate their impact. To shed light on such mitigation mechanisms, we interview 14 practitioners with the aim to identify (i) the sources of flakiness within the testing ecosystem, (ii) the impacts of flakiness, (iii) the measures adopted by practitioners when addressing flakiness, and (iv) the automation opportunities for these measures. Our analysis shows that, besides the tests and code, flakiness stems from interactions between the system components, the testing infrastructure, and external factors. We also highlight the impact of flakiness on testing practices and product quality and show that the adoption of guidelines together with a stable infrastructure are key measures in mitigating the problem.

Many software engineering techniques, such as fault localization, operate based on relevance relationships between tests and code. These relationships are often inferred using dynamic test execution information that approximates the link between relevant code units and asserted, by the tests, program behaviour. Unfortunately, in practice, dynamic information is not always available due to the instrumentation overheads or the nature of the production environments. This paper proposes CEMENT, a static technique that automatically infers such test and code relationships given the projects' evolution. The key idea is that developers make relevant changes to tests and code units around the same period, and thus their co-evolution reflects probable links between them. We evaluate CEMENT on 15 open-source projects, showing that it captures relevant links. We further conduct a fault localization case study and demonstrate that CEMENT can perform comparably to a current Information Retrieval-based Fault Localization (IRFL) approach. Additional analysis reveals that CEMENT and the IRFL localize different faults, suggesting complementarity. We leverage this complementarity by combining these two through voting, identifying 82% (32) of the faults located by at least one of them. These results suggest that test and code evolutionary coupling can effectively support test and debugging activities.

peer reviewed

Much of recent software-engineering research has investigated the naturalness of code, the fact that code, in small code snippets, is repetitive and can be predicted using statistical language models like n-gram. Although powerful, training such models on large code corpus can be tedious, time consuming and sensitive to code patterns (and practices) encountered during training. Consequently, these models are often trained on a small corpus and thus only estimate the language naturalness relative to a specific style of programming or type of project. To overcome these issues, we investigate the use of pre-trained generative language models to infer code naturalness. Pre-trained models are often built on big data, are easy to use in an out-of-the-box way and include powerful learning associations mechanisms. Our key idea is to quantify code naturalness through its predictability, by using state-of-the-art generative pre-trained language models. Thus, we suggest to infer naturalness by masking (omitting) code tokens, one at a time, of code-sequences, and checking the models’ ability to predict them. We explore three different predictability metrics; a) measuring the number of exact matches of the predictions, b) computing the embedding similarity between the original and predicted code, i.e., similarity at the vector space, and c) computing the confidence of the model when doing the token completion task regardless of the outcome. We implement this workflow, named CODEBERT-NT, and evaluate its capability to prioritize buggy lines over non-buggy ones when ranking code based on its naturalness. Our results, on 2,510 buggy versions of 40 projects from the SmartShark dataset, show that CODEBERTNT outperforms both, random-uniform and complexity-based ranking techniques, and yields comparable results to the n-gram models.

Studying and exposing software vulnerabilities is important to ensure software security, safety, and reliability. Software engineers often inject vulnerabilities into their programs to test the reliability of their test suites, vulnerability detectors, and security measures. However, state-of-the-art vulnerability injection methods only capture code syntax/patterns, they do not learn the intent of the vulnerability and are limited to the syntax of the original dataset. To address this challenge, we propose the first intent-based vulnerability injection method that learns both the program syntax and vulnerability intent. Our approach applies a combination of NLP methods and semantic-preserving program mutations (at the bytecode level) to inject code vulnerabilities. Given a dataset of known vulnerabilities (containing benign and vulnerable code pairs), our approach proceeds by employing semantic-preserving program mutations to transform the existing dataset to semantically similar code. Then, it learns the intent of the vulnerability via neural machine translation (Seq2Seq) models. The key insight is to employ Seq2Seq to learn the intent (context) of the vulnerable code in a manner that is agnostic of the specific program instance. We evaluate the performance of our approach using 1275 vulnerabilities belonging to five (5) CWEs from the Juliet test suite. We examine the effectiveness of our approach in producing compilable and vulnerable code. Our results show that IntJECT is effective, almost all (99%) of the code produced by our approach is vulnerable and compilable. We also demonstrate that the vulnerable programs generated by IntJECT are semantically similar to the withheld original vulnerable code. Finally, we show that our mutation-based data transformation approach outperforms its alternatives, namely data obfuscation and using the original data.

An established way to improve the transferability of black-box evasion attacks is to craft the adversarial examples on an ensemble-based surrogate to increase diversity. We argue that transferability is fundamentally related to uncertainty. Based on a state-of-the-art Bayesian Deep Learning technique, we propose a new method to efficiently build a surrogate by sampling approximately from the posterior distribution of neural network weights, which represents the belief about the value of each parameter. Our extensive experiments on ImageNet, CIFAR-10 and MNIST show that our approach improves the success rates of four state-of-the-art attacks significantly (up to 83.2 percentage points), in both intra-architecture and inter-architecture transferability. On ImageNet, our approach can reach 94% of success rate while reducing training computations from 11.6 to 2.4 exaflops, compared to an ensemble of independently trained DNNs. Our vanilla surrogate achieves 87.5% of the time higher transferability than three test-time techniques designed for this purpose. Our work demonstrates that the way to train a surrogate has been overlooked, although it is an important element of transfer-based attacks. We are, therefore, the first to review the effectiveness of several training methods in increasing transferability. We provide new directions to better understand the transferability phenomenon and offer a simple but strong baseline for future work.

Abstract Vulnerability prediction refers to the problem of identifying system components that are most likely to be vulnerable. Typically, this problem is tackled by training binary classifiers on historical data. Unfortunately, recent research has shown that such approaches underperform due to the following two reasons: a) the imbalanced nature of the problem, and b) the inherently noisy historical data, i.e., most vulnerabilities are discovered much later than they are introduced. This misleads classifiers as they learn to recognize actual vulnerable components as non-vulnerable. To tackle these issues, we propose TROVON , a technique that learns from known vulnerable components rather than from vulnerable and non-vulnerable components, as typically performed. We perform this by contrasting the known vulnerable, and their respective fixed components. This way, TROVON manages to learn from the things we know, i.e., vulnerabilities, hence reducing the effects of noisy and unbalanced data. We evaluate TROVON by comparing it with existing techniques on three security-critical open source systems, i.e., Linux Kernel, OpenSSL, and Wireshark, with historical vulnerabilities that have been reported in the National Vulnerability Database (NVD). Our evaluation demonstrates that the prediction capability of TROVON significantly outperforms existing vulnerability prediction techniques such as Software Metrics , Imports , Function Calls , Text Mining , Devign , LSTM , and LSTM-RF with an improvement of 40.84% in Matthews Correlation Coefficient (MCC) score under Clean Training Data Settings , and an improvement of 35.52% under Realistic Training Data Settings .

Abstract Applying mutation testing to test subtle program changes, such as program patches or other small-scale code modifications, requires using mutants that capture the delta of the altered behaviours. To address this issue, we introduce the concept of commit-relevant mutants, which are the mutants that interact with the behaviours of the system affected by a particular commit. Therefore, commit-aware mutation testing, is a test assessment metric tailored to a specific commit. By analysing 83 commits from 25 projects involving 2,253,610 mutants in both C and Java, we identify the commit-relevant mutants and explore their relationship with other categories of mutants. Our results show that commit-relevant mutants represent a small subset of all mutants, which differs from the other classes of mutants (subsuming and hard-to-kill), and that the commit-relevant mutation score is weakly correlated with the traditional mutation score (Kendall/Pearson 0.15-0.4). Moreover, commit-aware mutation analysis provides insights about the testing of a commit, which can be more efficient than the classical mutation analysis; in our experiments, by analysing the same number of mutants, commit-aware mutants have better fault-revelation potential (30% higher chances of revealing commit-introducing faults) than traditional mutants. We also illustrate a possible application of commit-aware mutation testing as a metric to evaluate test case prioritisation.

We are pleased to introduce the first part of the special issue on Mutation Testing of the Journal of Software Testing, Verification & Reliability. The issue features three papers selected by a review board of 40 mutation testing experts. Our board was selected from an extensive pool of experts comprising of reviewers from the Mutation workshop series, authors of published manuscripts in established journals related to mutation testing as well as the specific subdomain where mutation was applied. Mutation Testing was introduced in 1970s as a rigorous and automated alternative to fault injection. With over four decades of research and development, it is now accepted as a premier measure of the fault reveling potential of test suites. Mutation Testing is a fault-based testing technique that introduces candidate faults by making syntactic changes. The fault revealing ability of a test suite is then measured as the ratio of mutants it can detect against the total number of detectable mutants. A test suite is mutation testing adequate if the test suite detects all detectable mutants. While mutation testing is usually limited to simple faults, the twin axioms of mutation testing—finite neighborhood and coupling effect—provide confidence that a test suite that is mutation adequate can also reveal complex faults composed of simple faults. While the field of mutation testing has seen active research, numerous important problems remain open. These include (1) how to adapt mutation testing to non-traditional fields such as machine learning, (2) how to apply mutation testing for automatic test generation such as fuzzers, and (3) how to effectively solve the question of equivalent mutants. We thank the editors, authors, and reviewers for their efforts to successfully complete this special issue.

We are pleased to introduce the second part of the special issue on Mutation Testing of the Journal of Software Testing, Verification & Reliability. We thank the editors, authors, and reviewers for their efforts to successfully complete this special issue.

Similar to traditional software that is constantly under evolution, deep neural networks need to evolve upon the rapid growth of test data for continuous enhancement (e.g., adapting to distribution shift in a new environment for deployment). However, it is labor intensive to manually label all of the collected test data. Test selection solves this problem by strategically choosing a small set to label. Via retraining with the selected set, deep neural networks will achieve competitive accuracy. Unfortunately, existing selection metrics involve three main limitations: (1) using different retraining processes, (2) ignoring data distribution shifts, and (3) being insufficiently evaluated. To fill this gap, we first conduct a systemically empirical study to reveal the impact of the retraining process and data distribution on model enhancement. Then based on our findings, we propose DAT, a novel distribution-aware test selection metric. Experimental results reveal that retraining using both the training and selected data outperforms using only the selected data. None of the selection metrics perform the best under various data distributions. By contrast, DAT effectively alleviates the impact of distribution shifts and outperforms the compared metrics by up to five times and 30.09% accuracy improvement for model enhancement on simulated and in-the-wild distribution shift scenarios, respectively.

To effectively test program changes using mutation testing, one needs to use mutants that are relevant to the altered program behaviours. In view of this, we introduce MuDelta, an approach that identifies commit-relevant mutants; mutants that affect and are affected by the changed program behaviours. Our approach uses machine learning applied on a combined scheme of graph and vector-based representations of static code features. Our results, from 50 commits in 21 Coreutils programs, demonstrate a strong prediction ability of our approach; yielding 0.80 (ROC) and 0.50 (PR Curve) AUC values with 0.63 and 0.32 precision and recall values. These predictions are significantly higher than random guesses, 0.20 (PR-Curve) AUC, 0.21 and 0.21 precision and recall, and subsequently lead to strong relevant tests that kill 45%more relevant mutants than randomly sampled mutants (either sampled from those residing on the changed component(s) or from the changed lines). Our results also show that MuDelta selects mutants with 27% higher fault revealing ability in fault introducing commits. Taken together, our results corroborate the conclusion that commit-based mutation testing is suitable and promising for evolving software.

Deep neural networks (DNNs) are increasingly deployed as integral parts of software systems. However, due to the complex interconnections among hidden layers and massive hyperparameters, DNNs must be trained using a large number of labeled inputs, which calls for extensive human effort for collecting and labeling data. Spontaneously, to alleviate this growing demand, multiple state-of-the-art studies have developed different metrics to select a small yet informative dataset for the model training. These research works have demonstrated that DNN models can achieve competitive performance using a carefully selected small set of data. However, the literature lacks proper investigation of the limitations of data selection metrics, which is crucial to apply them in practice. In this paper, we fill this gap and conduct an extensive empirical study to explore the limits of data selection metrics. Our study involves 15 data selection metrics evaluated over 5 datasets (2 image classification tasks and 3 text classification tasks), 10 DNN architectures, and 20 labeling budgets (ratio of training data being labeled). Our findings reveal that, while data selection metrics are usually effective in producing accurate models, they may induce a loss of model robustness (against adversarial examples) and resilience to compression. Overall, we demonstrate the existence of a trade-off between labeling effort and different model qualities. This paves the way for future research in devising data selection metrics considering multiple quality criteria.

Industrial reports indicate that flaky tests are one of the primary concerns of software testing mainly due to the false signals they provide. To deal with this issue, researchers have developed tools and techniques aiming at (automatically) identifying flaky tests with encouraging results. However, to reach industrial adoption and practice, these techniques need to be replicated and evaluated extensively on multiple datasets, occasions and settings. In view of this, we perform a replication study of a recently proposed method that predicts flaky tests based on their vocabulary. We thus replicate the original study on three different dimensions. First, we replicate the approach on the same subjects as in the original study but using a different evaluation methodology, i.e., we adopt a time-sensitive selection of training and test sets to better reflect the envisioned use case. Second, we consolidate the findings of the initial study by building a new dataset of 837 flaky tests from 9 projects in a different programming language, i.e., Python while the original study was in Java, which comforts the generalisability of the results. Third, we propose an extension to the original approach by experimenting with different features extracted from the Code Under Test. We find that a more robust validation consistently decreases performance on the reported results of the original study, but, fortunately, the model remains capable to decently predict flaky tests. We find re-assuring results that the vocabulary-based models can also be used to predict test flakiness in Python. Finally, we find that the information lying in the Code Under Test has a limited impact on the performance of the vocabulary-based models.

Current Java Virtual Machine (JVM) fuzzers aim at generating syntactically valid Java programs, without targeting any particular use of the standard Java library. While effective, such fuzzers fail to discover specific kinds of bugs or vulnerabilities, such as type confusion, that are related to the standard API usage. To deal with this issue, we introduce a mutation-based feedback-guided black-box JVM fuzzer, called Confuzzion. Confuzzion, as the name suggests, targets security-relevant object-oriented flaws with a particular focus on type confusion vulnerabilities. We show that in less than 4 hours, on commodity hardware and without any predefined initialization seed, Confuzzion automatically generates Java programs that reveal JVM vulnerabilities, i.e., the Common Vulnerabilities and Exposures CVE-2017-3272. We also show that state-of-the-art fuzzers or even traditional automatic testing techniques are not capable of detecting such faults, even after 48 hours of execution in the same environment. To the best of our knowledge, Confuzzion is the first fuzzer able to detect JVM type confusion vulnerabilities.

Abstract We propose a new Statistical Model Checking (SMC) method to identify bugs in variability-intensive systems (VIS). The state-space of such systems is exponential in the number of variants, which makes the verification problem harder than for classical systems. To reduce verification time, we propose to combine SMC with featured transition systems (FTS)—a model that represents jointly the state spaces of all variants. Our new methods allow the sampling of executions from one or more (potentially all) variants. We investigate their utility in two complementary use cases. The first case considers the problem of finding all variants that violate a given property expressed in Linear-Time Logic (LTL) within a given simulation budget. To achieve this, we perform random walks in the featured transition system seeking accepting lassos. We show that our method allows us to find bugs much faster (up to 16 times according to our experiments) than exhaustive methods. As any simulation-based approach, however, the risk of Type-1 error exists. We provide a lower bound and an upper bound for the number of simulations to perform to achieve the desired level of confidence. Our empirical study involving 59 properties over three case studies reveals that our method manages to discover all variants violating 41 of the properties. This indicates that SMC can act as a coarse-grained analysis method to quickly identify the set of buggy variants. The second case complements the first one. In case the coarse-grained analysis reveals that no variant can guarantee to satisfy an intended property in all their executions, one should identify the variant that minimizes the probability of violating this property. Thus, we propose a fine-grained SMC method that quickly identifies promising variants and accurately estimates their violation probability. We evaluate different selection strategies and reveal that a genetic algorithm combined with elitist selection yields the best results.

We introduce SEMu , a Dynamic Symbolic Execution technique that generates test inputs capable of killing stubborn mutants (killable mutants that remain undetected after a reasonable amount of testing). SEMu aims at mutant propagation (triggering erroneous states to the program output) by incrementally searching for divergent program behaviors between the original and the mutant versions. We model the mutant killing problem as a symbolic execution search within a specific area in the programs’ symbolic tree. In this framework, the search area is defined and controlled by parameters that allow scalable and cost-effective mutant killing. We integrate SEMu in KLEE and experimented with Coreutils (a benchmark frequently used in symbolic execution studies). Our results show that our modeling plays an important role in mutant killing. Perhaps more importantly, our results also show that, within a two-hour time limit, SEMu kills 37% of the stubborn mutants, where KLEE kills none and where the mutant infection strategy (strategy suggested by previous research) kills 17%.

Testing of deep learning models is challenging due to the excessive number and complexity of the computations involved. As a result, test data selection is performed manually and in an ad hoc way. This raises the question of how we can automatically select candidate data to test deep learning models. Recent research has focused on defining metrics to measure the thoroughness of a test suite and to rely on such metrics to guide the generation of new tests. However, the problem of selecting/prioritising test inputs (e.g., to be labelled manually by humans) remains open. In this article, we perform an in-depth empirical comparison of a set of test selection metrics based on the notion of model uncertainty (model confidence on specific inputs). Intuitively, the more uncertain we are about a candidate sample, the more likely it is that this sample triggers a misclassification. Similarly, we hypothesise that the samples for which we are the most uncertain are the most informative and should be used in priority to improve the model by retraining. We evaluate these metrics on five models and three widely used image classification problems involving real and artificial (adversarial) data produced by five generation algorithms. We show that uncertainty-based metrics have a strong ability to identify misclassified inputs, being three times stronger than surprise adequacy and outperforming coverage-related metrics. We also show that these metrics lead to faster improvement in classification accuracy during retraining: up to two times faster than random selection and other state-of-the-art metrics on all models we considered.

We propose a new Statistical Model Checking (SMC) method to discover bugs in variability-intensive systems (VIS). The state-space of such systems is exponential in the number of variants, which makes the verification problem harder than for classical systems. To reduce verification time, we sample executions from a featured transition system – a model that represents jointly the state spaces of all variants. The combination of this compact representation and the inherent efficiency of SMC allows us to find bugs much faster (up to 16 times according to our experiments) than other methods. As any simulation-based approach, however, the risk of Type-1 error exists. We provide a lower bound and an upper bound for the number of simulations to perform to achieve the desired level of confidence. Our empirical study involving 59 properties over three case studies reveals that our method manages to discover all variants violating 41 of the properties. This indicates that SMC can act as a low-cost-high-reward method for verifying VIS.

This paper presents TransRepair, a fully automatic approach for testing and repairing the consistency of machine translation systems. TransRepair combines mutation with metamorphic testing to detect inconsistency bugs (without access to human oracles). It then adopts probability-reference or cross-reference to post-process the translations, in a grey-box or black-box manner, to repair the inconsistencies. Our evaluation on two state-of-the-art translators, Google Translate and Transformer, indicates that TransRepair has a high precision (99%) on generating input pairs with consistent translations. With these tests, using automatic consistency metrics and manual assessment, we find that Google Translate and Transformer have approximately 36% and 40% inconsistency bugs. Black-box repair fixes 28% and 19% bugs on average for Google Translate and Transformer. Grey-box repair fixes 30% bugs on average for Transformer. Manual inspection indicates that the translations repaired by our approach improve consistency in 87% of cases (degrading it in 2%), and that our repairs have better translation acceptability in 27% of the cases (worse in 8%).

In Continuous Integration, developers want to know how well they have tested their changes. Unfortunately, in these cases, the use of mutation testing is suboptimal since mutants affect the entire set of program behaviours and not the changed ones. Thus, the extent to which mutation testing can be used to test committed changes is questionable. To deal with this issue, we define commit-relevant mutants; a set of mutants that affect the changed program behaviours and represent the commit-relevant test requirements. We identify such mutants in a controlled way, and check their relationship with traditional mutation score (score based on the entire set of mutants or on the mutants located on the commits). We conduct experiments in both C and Java, using 83 commits, 2,253,610 mutants from 25 projects. Our findings reveal that there is a relatively weak correlation (Kendall/Pearson 0.15-0.4) between the sought (commit-relevant) and traditional mutation scores, indicating the need for a commit-aware test assessment metric. Our analysis also shows that traditional mutation is far from the envisioned case as it loses approximately 50%-60% of the commit-relevant mutants when analysing 5-25 mutants. More importantly, our results demonstrate that traditional mutation has approximately 30% lower chances of revealing commit-introducing faults than commit-aware mutation testing.

The rapid spread of the Coronavirus SARS-2 is a major challenge that led almost all governments worldwide to take drastic measures to respond to the tragedy. Chief among those measures is the massive lockdown of entire countries and cities, which beyond its global economic impact has created some deep social and psychological tensions within populations. While the adopted mitigation measures (including the lockdown) have generally proven useful, policymakers are now facing a critical question: how and when to lift the mitigation measures? A carefully-planned exit strategy is indeed necessary to recover from the pandemic without risking a new outbreak. Classically, exit strategies rely on mathematical modeling to predict the effect of public health interventions. Such models are unfortunately known to be sensitive to some key parameters, which are usually set based on rules-of-thumb.

Credit scoring systems are critical FinTech applications that concern the analysis of the creditworthiness of a person or organization. While decisions were previously based on human expertise, they are now increasingly relying on data analysis and machine learning. In this paper, we assess the ability of state-of-the-art adversarial machine learning to craft attacks on a real-world credit scoring system. Interestingly, we find that, while these techniques can generate large numbers of adversarial data, these are practically useless as they all violate domain-specific constraints. In other words, the generated examples are all false positives as they cannot occur in practice. To circumvent this limitation, we propose CoEvA2, a search-based method that generates valid adversarial examples (satisfying the domain constraints). CoEvA2 utilizes multi-objective search in order to simultaneously handle constraints, perform the attack and maximize the overdraft amount requested. We evaluate CoEvA2 on a major bank's real-world system by checking its ability to craft valid attacks. CoEvA2 generates thousands of valid adversarial examples, revealing a high risk for the banking system. Fortunately, by improving the system through adversarial training (based on the produced examples), we increase its robustness and make our attack fail.

Abstract Mutant selection refers to the problem of choosing, among a large number of mutants, the (few) ones that should be used by the testers. In view of this, we investigate the problem of selecting the fault revealing mutants, i.e., the mutants that are killable and lead to test cases that uncover unknown program faults. We formulate two variants of this problem: the fault revealing mutant selection and the fault revealing mutant prioritization. We argue and show that these problems can be tackled through a set of ‘static’ program features and propose a machine learning approach, named FaRM, that learns to select and rank killable and fault revealing mutants. Experimental results involving 1,692 real faults show the practical benefits of our approach in both examined problems. Our results show that FaRM achieves a good trade-off between application cost and effectiveness (measured in terms of faults revealed). We also show that FaRM outperforms all the existing mutant selection methods, i.e., the random mutant sampling, the selective mutation and defect prediction (mutating the code areas pointed by defect prediction). In particular, our results show that with respect to mutant selection, our approach reveals 23% to 34% more faults than any of the baseline methods, while, with respect to mutant prioritization, it achieves higher average percentage of revealed faults with a median difference between 4% and 9% (from the random mutant orderings).

Software bugs constitute a frequent and common issue of software development. To deal with this problem, modern software development methodologies introduce dedicated quality assurance procedures. At the same time researchers aim at developing techniques capable of supporting the early discovery and fix of bugs. One important factor that guides such research attempts is the characteristics of software bugs and bug fixes. In this paper, we present an industrial study on the characteristics and differences between pre-release bugs, i.e. bugs detected during software development, and post-release bugs, i.e. bugs that escaped to production. Understanding such differences is of paramount importance as it will improve our understanding on the testing and debugging support that practitioners require from the research community, on the validity of the assumptions of several research techniques, and, most importantly, on the reasons why bugs escape to production. To this end, we analyze 37 industrial projects from BGL BNP Paribas and document the differences between pre-release bugs and post-release bugs. Our findings suggest that post-release bugs are more complex to fix, requiring developers to modify several source code files, written in different programming languages, and configuration files, as well. We also find that approximately 82% of the post-release bugs involve code additions and can be characterized as 'omission' bugs. Finally, we conclude the paper with a discussion on the implications of our study and provide guidance to future research directions.

Many companies rely on software testing to verify that their software products meet their requirements. However, test quality and, in particular, the quality of end-to-end testing is relatively hard to achieve. The problem becomes challenging when software evolves, as end-to-end test suites need to adapt and conform to the evolved software. Unfortunately, end-to-end tests are particularly fragile as any change in the application interface, e.g., application flow, location or name of graphical user interface elements, necessitates a change in the tests. This paper presents an industrial case study on the evolution of Keyword-Driven test suites, also known as Keyword-Driven Testing (KDT). Our aim is to demonstrate the problem of test maintenance, identify the benefits of Keyword-Driven Testing and overall improve the understanding of test code evolution (at the acceptance testing level). This information will support the development of automatic techniques, such as test refactoring and repair, and will motivate future research. To this end, we identify, collect and analyze test code changes across the evolution of industrial KDT test suites for a period of eight months. We show that the problem of test maintenance is largely due to test fragility (most commonly-performed changes are due to locator and synchronization issues) and test clones (over 30% of keywords are duplicated). We also show that the better test design of KDT test suites has the potential for drastically reducing (approximately 70%) the number of test code changes required to support software evolution. To further validate our results, we interview testers from BGL BNP Paribas and report their perceptions on the advantages and challenges of keyword-driven testing.

Machine-learning-based anomaly detection systems can be vulnerable to new kinds of deceptions, known as training attacks, which exploit the live learning mechanism of these systems by progressively injecting small portions of abnormal data. The injected data seamlessly swift the learned states to a point where harmful data can pass unnoticed. We focus on the systematic testing of these attacks in the context of intrusion detection systems (IDS). We propose a search-based approach to test IDS by making training attacks. Going a step further, we also propose searching for countermeasures, learning from the successful attacks and thereby increasing the resilience of the tested IDS. We evaluate our approach on a denial-of-service attack detection scenario and a dataset recording the network traffic of a real-world system. Our experiments show that our search-based attack scheme generates successful attacks bypassing the current state-of-the-art defences. We also show that our approach is capable of generating attack patterns for all configuration states of the studied IDS and that it is capable of providing appropriate countermeasures. By co-evolving our attack and defence mechanisms we succeeded at improving the defence of the IDS under test by making it resilient to 49 out of 50 independently generated attacks.

Programs expecting structured inputs often consist of both a syntactic analysis stage, which parses raw input, and a semantic analysis stage, which conducts checks on the parsed input and executes the core logic of the program. Generator-based testing tools in the lineage of QuickCheck are a promising way to generate random syntactically valid test inputs for these programs. We present Zest, a technique which automatically guides QuickCheck-like random input generators to better explore the semantic analysis stage of test programs. Zest converts random-input generators into deterministic parametric input generators. We present the key insight that mutations in the untyped parameter domain map to structural mutations in the input domain. Zest leverages program feedback in the form of code coverage and input validity to perform feedback-directed parameter search. We evaluate Zest against AFL and QuickCheck on five Java programs: Maven, Ant, BCEL, Closure, and Rhino. Zest covers 1.03x-2.81x as many branches within the benchmarks' semantic analysis stages as baseline techniques. Further, we find 10 new bugs in the semantic analysis stages of these benchmarks. Zest is the most effective technique in finding these bugs reliably and quickly, requiring at most 10 minutes on average to find each bug.

Automation of acceptance test suites becomes necessary in the context of agile software development practices, which require rapid feedback on the quality of code changes. To this end, companies try to automate their acceptance tests as much as possible. Unfortunately, the growth of the automated test suites, by several automation testers, gives rise to potential test smells, i.e., poorly designed test code, being introduced in the test code base, which in turn may increase the cost of maintaining the code and creating new one. In this paper, we investigate this problem in the context of our industrial partner, BGL BNP Paribas, and introduce Ukwikora, an automated tool that statically analyzes acceptance test suites, enabling the continuous inspection of the test code base. Ukwikora targets code written in the Robot Framework syntax, a popular framework for writing Keyword-Driven tests. Ukwikora has been successfully deployed at BGL BNP Paribas, detecting issues otherwise unknown to the automation testers, such as the presence of duplicated test code, dead test code and dependency issues among the tests. The success of our case study reinforces the need for additional research and tooling for acceptance test suites.

Program mutation makes small syntactic alterations to programs' code in order to artificially create faulty programs (mutants). Mutants creation (generation) tools are often characterized by their mutation operators and the way they create and represent the mutants. This paper presents Mart, a mutants generation tool, for LLVM bitcode, that supports the fine-grained definition of mutation operators (as matching rule - replacing pattern pair; uses 816 defined pairs by default) and the restriction of the code parts to mutate. New operators are implemented in Mart by implementing their matching rules and replacing patterns. Mart also implements in-memory Trivial Compiler Equivalence to eliminate equivalent and duplicate mutants during mutants generation. Mart generates mutant code as separated mutant files, meta-mutants file, weak mutation and mutant coverage instrumented files. Mart is publicly available (https://github.com/thierry-tct/mart). Mart has been applied to generate mutants for several research experiments and generated more than 4,000,000 mutants.

Previous work on vulnerability prediction assume that predictive models are trained with respect to perfect labelling information (includes labels from future, as yet undiscovered vulnerabilities). In this paper we present results from a comprehensive empirical study of 1,898 real-world vulnerabilities reported in 74 releases of three security-critical open source systems (Linux Kernel, OpenSSL and Wiresark). Our study investigates the effectiveness of three previously proposed vulnerability prediction approaches, in two settings: with and without the unrealistic labelling assumption. The results reveal that the unrealistic labelling assumption can profoundly mis- lead the scientific conclusions drawn; suggesting highly effective and deployable prediction results vanish when we fully account for realistically available labelling in the experimental methodology. More precisely, MCC mean values of predictive effectiveness drop from 0.77, 0.65 and 0.43 to 0.08, 0.22, 0.10 for Linux Kernel, OpenSSL and Wiresark, respectively. Similar results are also obtained for precision, recall and other assessments of predictive efficacy. The community therefore needs to upgrade experimental and empirical methodology for vulnerability prediction evaluation and development to ensure robust and actionable scientific findings.

Summary In this paper, we propose a new test case prioritization technique that combines both mutation‐based and diversity‐aware approaches. The diversity‐aware mutation‐based technique relies on the notion of mutant distinguishment, which aims to distinguish one mutant's behaviour from another, rather than from the original program. The relative cost and effectiveness of the mutation‐based prioritization techniques (i.e., using both the traditional mutant kill and the proposed mutant distinguishment) are empirically investigated with 352 real faults and 553,477 developer‐written test cases. The empirical evaluation considers both the traditional and the diversity‐aware mutation criteria in various settings: single‐objective greedy, hybrid, and multi‐objective optimization. The results show that there is no single dominant technique across all the studied faults. To this end, the reason why each one of the mutation‐based prioritization criteria performs poorly is discussed, using a graphical model called Mutant Distinguishment Graph that demonstrates the distribution of the fault‐detecting test cases with respect to mutant kills and distinguishment. © 2018 John Wiley & Sons, Ltd.

Testing is crucial to ensure the quality of software systems-but testing is an expensive process, so test managers try to minimise the set of tests to run to save computing resources and speed up the testing process and analysis. One problem is that there are different perspectives on what is a good test and it is usually not possible to compare these dimensions. This is a perfect example of a multi-objective optimisation problem, which is hard-especially given the scale of the search space here. In this paper, we propose a novel hybrid algorithm to address this problem. Our method is composed of three steps: a greedy algorithm to find quickly some good solutions, a genetic algorithm to increase the search space covered and a local search algorithm to refine the solutions. We demonstrate through a large scale empirical evaluation that our method is more reliable (better whatever the time budget) and more robust (better whatever the number of dimensions considered)-in the scenario with 4 objectives and a default execution time, we are 178% better in hypervolume on average than the state-of-the-art algorithms.

Background: Code is repetitive and predictable in a way that is similar to the natural language. This means that code is "natural" and this "naturalness" can be captured by natural language modelling techniques. Such models promise to capture the program semantics and identify source code parts that `smell', i.e., they are strange, badly written and are generally error-prone (likely to be defective). Aims: We investigate the use of natural language modelling techniques in mutation testing (a testing technique that uses artificial faults). We thus, seek to identify how well artificial faults simulate real ones and ultimately understand how natural the artificial faults can be. Our intuition is that natural mutants, i.e., mutants that are predictable (follow the implicit coding norms of developers), are semantically useful and generally valuable (to testers). We also expect that mutants located on unnatural code locations (which are generally linked with error-proneness) to be of higher value than those located on natural code locations. Method: Based on this idea, we propose mutant selection strategies that rank mutants according to a) their naturalness (naturalness of the mutated code), b) the naturalness of their locations (naturalness of the original program statements) and c) their impact on the naturalness of the code that they apply to (naturalness differences between original and mutated statements). We empirically evaluate these issues on a benchmark set of 5 open-source projects, involving more than 100k mutants and 230 real faults. Based on the fault set we estimate the utility (i.e. capability to reveal faults) of mutants selected on the basis of their naturalness, and compare it against the utility of randomly selected mutants. Results: Our analysis shows that there is no link between naturalness and the fault revelation utility of mutants. We also demonstrate that the naturalness-based mutant selection performs similar (slightly worse) to the random mutant selection. Conclusions: Our findings are negative but we consider them interesting as they confute a strong intuition, i.e., fault revelation is independent of the mutants' naturalness.

Testing is the primary approach for detecting software defects. A major challenge faced by testers lies in crafting efficient test suites, able to detect a maximum number of bugs with manageable effort. To do so, they rely on coverage criteria, which define some precise test objectives to be covered. However, many common criteria specify a significant number of objectives that occur to be infeasible or redundant in practice, like covering dead code or semantically equal mutants. Such objectives are well-known to be harmful to the design of test suites, impacting both the efficiency and precision of the tester's effort. This work introduces a sound and scalable technique to prune out a significant part of the infeasible and redundant objectives produced by a panel of white-box criteria. In a nutshell, we reduce this task to proving the validity of logical assertions in the code under test. The technique is implemented in a tool that relies on weakest-precondition calculus and SMT solving for proving the assertions. The tool is built on top of the Frama-C verification platform, which we carefully tune for our specific scalability needs. The experiments reveal that the pruning capabilities of the tool can reduce the number of targeted test objectives in a program by up to 27% and scale to real programs of 200K lines, making it possible to automate a painstaking part of their current testing process.

Empirical validation of software testing studies is increasingly relying on mutants. This practice is motivated by the strong correlation between mutant scores and real fault detection that is reported in the literature. In contrast, our study shows that correlations are the results of the confounding effects of the test suite size. In particular, we investigate the relation between two independent variables, mutation score and test suite size, with one dependent variable the detection of (real) faults. We use two data sets, CoreBench and Defects4J, with large C and Java programs and real faults and provide evidence that all correlations between mutation scores and real fault detection are weak when controlling for test suite size. We also find that both independent variables significantly influence the dependent one, with significantly better fits, but overall with relative low prediction power. By measuring the fault detection capability of the top ranked, according to mutation score, test suites (opposed to randomly selected test suites of the same size), we find that achieving higher mutation scores improves significantly the fault detection. Taken together, our data suggest that mutants provide good guidance for improving the fault detection of test suites, but their correlation with fault detection are weak.

Recent research shows that language models, such as n-gram models, are useful at a wide variety of software engineering tasks, e.g., code completion, bug identification, code summarisation, etc. However, such models require the appropriate set of numerous parameters. Moreover, the different ways one can read code essentially yield different models (based on the different sequences of tokens). In this paper, we focus on n-gram models and evaluate how the use of tokenizers, smoothing, unknown threshold and n values impact the predicting ability of these models. Thus, we compare the use of multiple tokenizers and sets of different parameters (smoothing, unknown threshold and n values) with the aim of identifying the most appropriate combinations. Our results show that the Modified Kneser-Ney smoothing technique performs best, while n values are depended on the choice of the tokenizer, with values 4 or 5 offering a good trade-off between entropy and computation time. Interestingly, we find that tokenizers treating the code as simple text are the most robust ones. Finally, we demonstrate that the differences between the tokenizers are of practical importance and have the potential of changing the conclusions of a given experiment.

Natural language processing techniques, in particular n-gram models, have been applied successfully to facilitate a number of software engineering tasks. However, in our related ICSME '18 paper, we have shown that the conclusions of a study can drastically change with respect to how the code is tokenized and how the used n-gram model is parameterized. These choices are thus of utmost importance, and one must carefully make them. To show this and allow the community to benefit from our work, we have developed TUNA (TUning Naturalness-based Analysis), a Java software artifact to perform naturalness-based analyses of source code. To the best of our knowledge, TUNA is the first open-source, end-to-end toolchain to carry out source code analyses based on naturalness.

peer reviewed

Mutation testing realises the idea of fault-based testing, i.e., using artificial defects to guide the testing process. It is used to evaluate the adequacy of test suites and to guide test case generation. It is a potentially powerful form of testing, but it is well-known that its effectiveness is inhibited by the presence of equivalent mutants. We recently studied Trivial Compiler Equivalence (TCE) as a simple, fast and readily applicable technique for identifying equivalent mutants for C programs. In the present work, we augment our findings with further results for the Java programming language. TCE can remove a large portion of all mutants because they are determined to be either equivalent or duplicates of other mutants. In particular, TCE equivalent mutants account for 7.4 and 5.7 percent of all C and Java mutants, while duplicated mutants account for a further 21 percent of all C mutants and 5.4 percent Java mutants, on average. With respect to a benchmark ground truth suite (of known equivalent mutants), approximately 30 percent (for C) and 54 percent (for Java) are TCE equivalent. It is unsurprising that results differ between languages, since mutation characteristics are language-dependent. In the case of Java, our new results suggest that TCE may be particularly effective, finding almost half of all equivalent mutants.

Many studies suggest using coverage concepts, such as branch coverage, as the starting point of testing, while others as the most prominent test quality indicator. Yet the relationship between coverage and fault-revelation remains unknown, yielding uncertainty and controversy. Most previous studies rely on the Clean Program Assumption, that a test suite will obtain similar coverage for both faulty and fixed ('clean') program versions. This assumption may appear intuitive, especially for bugs that denote small semantic deviations. However, we present evidence that the Clean Program Assumption does not always hold, thereby raising a critical threat to the validity of previous results. We then conducted a study using a robust experimental methodology that avoids this threat to validity, from which our primary finding is that strong mutation testing has the highest fault revelation of four widely-used criteria. Our findings also revealed that fault revelation starts to increase significantly only once relatively high levels of coverage are attained.

Mutation analysis is a popular test assessment method. It relies on the mutation score, which indicates how many mutants are revealed by a test suite. Yet, there are mutants whose behaviour is equivalent to the original system, wasting analysis resources and preventing the satisfaction of the full (100%) mutation score. For finite behavioural models, the Equivalent Mutant Problem (EMP) can be addressed through language equivalence of non-deterministic finite automata, which is a well-studied, yet computationally expensive, problem in automata theory. In this paper, we report on our preliminary assessment of a state-of-the-art exact language equivalence tool to handle the EMP against 3 models of size up to 15,000 states on 1170 mutants. We introduce random and mutation-biased simulation heuristics as baselines for comparison. Results show that the exact approach is often more than ten times faster in the weak mutation scenario. For strong mutation, our biased simulations are faster for models larger than 300 states. They can be up to 1,000 times faster while limiting the error of misclassifying non-equivalent mutants as equivalent to 10% on average. We therefore conclude that the approaches can be combined for improved efficiency.

Mutation testing is extensively used in software testing studies. However, popular mutation testing tools use a restrictive set of mutants which does not conform to the community standards and mutation testing literature. This can be problematic since the effectiveness of mutation strongly depends on the used mutants. To investigate this issue we form an extended set of mutants and implement it on a popular mutation testing tool named PIT. We then show that in real-world projects the original mutants of PIT are easier to kill and lead to tests that score statistically lower than those of the extended set of mutants for a range of 35% to 70% of the studied classes. These results raise serious concerns regarding the validity of mutation-based experiments that use PIT. To further show the strengths of the extended mutants we also performed an analysis using a benchmark with mutation-adequate test cases and identified equivalent mutants. Our results confirmed that the extended mutants are more effective than a) the original version of PIT and b) two other popular mutation testing tools (major and muJava). In particular, our results demonstrate that the extended mutants are more effective by 23%, 12% and 7% than the mutants of the original PIT, major and muJava. They also show that the extended mutants are at least as strong as the mutants of all the other three tools together. To support future research, we make the new version of PIT, which is equipped with the extended mutants, publicly available.

Vulnerabilities are one of the main concerns faced by practitioners when working with security critical applications. Unfortunately, developers and security teams, even experienced ones, fail to identify many of them with severe consequences. Vulnerabilities are hard to discover since they appear in various forms, caused by many different issues and their identification requires an attacker's mindset. In this paper, we aim at increasing the understanding of vulnerabilities by investigating their characteristics on two major open-source software systems, i.e., the Linux kernel and OpenSSL. In particular, we seek to analyse and build a profile for vulnerable code, which can ultimately help researchers in building automated approaches like vulnerability prediction models. Thus, we examine the location, criticality and category of vulnerable code along with its relation with software metrics. To do so, we collect more than 2,200 vulnerable files accounting for 863 vulnerabilities and compute more than 35 software metrics. Our results indicate that while 9 Common Weakness Enumeration (CWE) types of vulnerabilities are prevalent, only 3 of them are critical in OpenSSL and 2 of them in the Linux kernel. They also indicate that different types of vulnerabilities have different characteristics, i.e., metric profiles, and that vulnerabilities of the same type have different profiles in the two projects we examined. We also found that the file structure of the projects can provide useful information related to the vulnerabilities. Overall, our results demonstrate the need for making project specific approaches that focus on specific types of vulnerabilities.

Model-based mutation analysis is a powerful but expensive testing technique. We tackle its high computation cost by proposing an optimization technique that drastically speeds up the mutant execution process. Central to this approach is the Featured Mutant Model, a modelling framework for mutation analysis inspired by the software product line paradigm. It uses behavioural variability models, viz., Featured Transition Systems, which enable the optimized generation, configuration and execution of mutants. We provide results, based on models with thousands of transitions, suggesting that our technique is fast and scalable. We found that it outperforms previous approaches by several orders of magnitude and that it makes higher-order mutation practically applicable.

Although white-box regression test prioritization has been well-studied, the more recently introduced black-box prioritization approaches have neither been compared against each other nor against more well-established white-box techniques. We present a comprehensive experimental comparison of several test prioritization techniques, including well-established white-box strategies and more recently introduced black-box approaches. We found that Combinatorial Interaction Testing and diversity-based techniques (Input Model Diversity and Input Test Set Diameter) perform best among the black-box approaches. Perhaps surprisingly, we found little difference between black-box and white-box performance (at most 4% fault detection rate difference). We also found the overlap between black- and white-box faults to be high: the first 10% of the prioritized test suites already agree on at least 60% of the faults found. These are positive findings for practicing regression testers who may not have source code available, thereby making white-box techniques inapplicable. We also found evidence that both black-box and white-box prioritization remain robust over multiple system releases.

Mutation testing introduces artificial defects to measure the adequacy of testing. In case candidate tests can distinguish the behaviour of mutants from that of the original program, they are considered of good quality -- otherwise developers need to design new tests. While, this method has been shown to be effective, industry-scale code challenges its applicability due to the sheer number of mutants and test executions it requires. In this paper we present PIT, a practical mutation testing tool for Java, applicable on real-world codebases. PIT is fast since it operates on bytecode and optimises mutant executions. It is also robust and well integrated with development tools, as it can be invoked through a command line interface, Ant or Maven. PIT is also open source and hence, publicly available at \url{http://pitest.org/}

Much research on software testing and test techniques relies on experimental studies based on mutation testing. In this paper we reveal that such studies are vulnerable to a potential threat to validity, leading to possible Type I errors; incorrectly rejecting the Null Hypothesis. Our findings indicate that Type I errors occur, for arbitrary experiments that fail to take countermeasures, approximately 62% of the time. Clearly, a Type I error would potentially compromise any scientific conclusion. We show that the problem derives from such studies’ combined use of both subsuming and subsumed mutants. We collected articles published in the last two years at three leading software engineering conferences. Of those that use mutation-based test assessment, we found that 68% are vulnerable to this threat to validity.

In widely used mobile operating systems a single vulnerability can threaten the security and privacy of billions of users. Therefore, identifying vulnerabilities and fortifying software systems requires constant attention and effort. However, this is costly and it is almost impossible to analyse an entire code base. Thus, it is necessary to prioritize efforts towards the most likely vulnerable areas. A first step in identifying these areas is to profile vulnerabilities based on previously reported ones. To investigate this, we performed a manual analysis of Android vulnerabilities, as reported in the National Vulnerability Database for the period 2008 to 2014. In our analysis, we identified a comprehensive list of issues leading to Android vulnerabilities. We also point out characteristics of the locations where vulnerabilities reside, the complexity of these locations and the complexity to fix the vulnerabilities. To enable future research, we make available all of our data.

To assist the vulnerability identification process, researchers proposed prediction models that highlight (for inspection) the most likely to be vulnerable parts of a system. In this paper we aim at making a reliable replication and comparison of the main vulnerability prediction models. Thus, we seek for determining their effectiveness, i.e., their ability to distinguish between vulnerable and non-vulnerable components, in the context of the Linux Kernel, under different scenarios. To achieve the above-mentioned aims, we mined vulnerabilities reported in the National Vulnerability Database and created a large dataset with all vulnerable components of Linux from 2005 to 2016. Based on this, we then built and evaluated the prediction models. We observe that an approach based on the header files included and on function calls performs best when aiming at future vulnerabilities, while text mining is the best technique when aiming at random instances. We also found that models based on code metrics perform poorly. We show that in the context of the Linux kernel, vulnerability prediction models can be superior to random selection and relatively precise. Thus, we conclude that practitioners have a valuable tool for prioritizing their security inspection efforts.

Mutation testing is considered as one of the most powerful testing methods. It operates by asking testers to design tests that reveal a set of mutants, which are purpose-made injected defects. Evidently, the strength of the method strongly depends on the used mutants. However, this dependence raises concerns regarding the mutation testing practice that is implemented by existing tools. Thus, it is probable that implementation inadequacies can lead to incompetent results. In this paper, we cross-evaluate three popular mutation testing tools for Java, namely MUJAVA, MAJOR and PIT, with respect to their effectiveness. We perform an empirical study of 3,324 manually analysed mutants from real-world projects and we find that there are large differences between the tools' effectiveness, ranging from 76% to 88%, with MUJAVA achieving the best results. We also demonstrate that no tool is able to subsume the others and provide practical recommendations on how to strengthen each one of the studied tools. Finally, our analysis shows that 11%, 12% and 7% of the mutants generated by MUJAVA, MAJOR and PIT are equivalent, respectively.

Software Product Line (SPL) feature selection involves the optimization of multiple objectives in a large and highly constrained search space. We introduce SATIBEA, that augments multi-objective search-based optimization with constraint solving to address this problem, evaluating it on five large real-world SPLs, ranging from 1,244 to 6,888 features with respect to three different solution quality indicators and two diversity metrics. The results indicate that SATIBEA statistically significantly outperforms the current state-of-the-art (p

Identifying equivalent mutants remains the largest impediment to the widespread uptake of mutation testing. Despite being researched for more than three decades, the problem remains. We propose Trivial Compiler Equivalence (TCE) a technique that exploits the use of readily available compiler technology to address this long-standing challenge. TCE is directly applicable to real-world programs and can imbue existing tools with the ability to detect equivalent mutants and a special form of useless mutants called duplicated mutants. We present a thorough empirical study using 6 large open source programs, several orders of magnitude larger than those used in previous work, and 18 benchmark programs with hand-analysis equivalent mutants. Our results reveal that, on large real-world programs, TCE can discard more than 7% and 21% of all the mutants as being equivalent and duplicated mutants respectively. A human- based equivalence verification reveals that TCE has the ability to detect approximately 30% of all the existing equivalent mutants.

In software testing, coverage criteria specify the requirements to be covered by the test cases. However, in practice such criteria are limited due to the well-known infeasibility problem, which concerns elements/requirements that cannot be covered by any test case. To deal with this issue we revisit and improve state-of-the-art static analysis techniques, such as Value Analysis and Weakest Precondition calculus. We propose a lightweight greybox scheme for combining these two techniques in a complementary way. In particular we focus on detecting infeasible test requirements in an automatic and sound way for condition coverage, multiple condition coverage and weak mutation testing criteria. Experimental results show that our method is capable of detecting almost all the infeasible test requirements, 95% on average, in a reasonable amount of time, i.e., less than 40 seconds, making it practical for unit testing.

Summary The equivalent mutant problem is a major hindrance to mutation testing. Being undecidable in general, it is only susceptible to partial solutions. In this paper, mutant classification is utilised for isolating likely to be first‐order equivalent mutants. A new classification technique,Isolating Equivalent Mutants (I‐EQM), is introduced and empirically investigated. The proposed approach employs a dynamic execution scheme that integrates theimpacton the program execution of first‐order mutants with the impact on the output of second‐order mutants. An experimental study, conducted using two independently created sets of manually classified mutants selected from real‐world programs revalidates previously published results and provides evidence for the effectiveness of the proposed technique. Overall, the study shows that I‐EQM substantially improves previous methods by retrieving a considerably higher number of killable mutants, thus, amplifying the quality of the testing process. Copyright © 2014 John Wiley & Sons, Ltd.

Summary Fault localization methods seek to identify faulty program statements based on the information provided by the failing and passing test executions. Spectrum‐based methods are among the most popular ones and assist programmers by assigning suspiciousness values on program statements according to their probability of being faulty. This paper proposes Metallaxis, a fault localization approach based on mutation analysis. The innovative part of Metallaxis is that it uses mutants and links them with the faulty program places. Thus, mutants that are killed mostly by failing tests provide a good indication about the location of a fault. Experimentation using Metallaxis suggests that it is significantly more effective than statement‐based approaches. This is true even in the case where mutation cost‐reduction techniques, such as mutant sampling, are facilitated. Additionally, results from a controlled experiment show that the use of mutation as a testing technique provides benefits to the fault localization process. Therefore, fault localization is significantly improved by using mutation‐based tests instead of block‐based or branch‐based test suites. Finally, evidence in support of the methods’ scalability is also given. Copyright © 2013 John Wiley & Sons, Ltd.

Modern systems tend to be highly configurable. Testing such systems requires selecting test cases from a large input space. Thus, there is a need to systematically sample program inputs in order to reduce the testing effort. In such cases, testing the interactions between program parameters has been identified as an effective way to deal with this problem. In these lines, Combinatorial Interaction Testing (CIT) models the program input interactions and uses this model to select test cases. Going a step further, we apply mutation analysis on the CIT input model to select program test cases. Mutation operates by injecting defects to the program input model and measures the number of defects found by the selected test cases. Experiments performed on four real programs show that measuring the number of model-based defects gives a stronger correlation to code-level faults than measuring the number of the exercised interactions. Therefore, the proposed mutation analysis approach forms a valid and more effective alternative to CIT.

Mutation testing is an effective technique for either improving or generating fault-finding test suites. It creates defective or incorrect program artifacts of the program under test and evaluates the ability of test suites to reveal them. Despite being effective, mutation is costly since it requires assessing the test cases with a large number of defective artifacts. Even worse, some of these artifacts are behaviourally ``equivalent'' to the original one and hence, they unnecessarily increase the testing effort. We adopt a variability perspective on mutation analysis. We model a defective artifact as a transition system with a specific feature selected and consider it as a member of a mutant family. The mutant family is encoded as a Featured Transition System, a compact formalism initially dedicated to model-checking of software product lines. We show how to evaluate a test suite against the set of all candidate defects by using mutant families. We can evaluate all the considered defects at the same time and isolate some equivalent mutants. We can also assist the test generation process and efficiently consider higher-order mutants.

Large Software Product Lines (SPLs) are common in industry, thus introducing the need of practical solutions to test them. To this end, t-wise can help to drastically reduce the number of product configurations to test. Current t-wise approaches for SPLs are restricted to small values of t. In addition, these techniques fail at providing means to finely control the configuration process. In view of this, means for automatically generating and prioritizing product configurations for large SPLs are required. This paper proposes (a) a search-based approach capable of generating product configurations for large SPLs, forming a scalable and flexible alternative to current techniques and (b) prioritization algorithms for any set of product configurations. Both these techniques employ a similarity heuristic. The ability of the proposed techniques is assessed in an empirical study through a comparison with state of the art tools. The comparison focuses on both the product configuration generation and the prioritization aspects. The results demonstrate that existing t-wise tools and prioritization techniques fail to handle large SPLs. On the contrary, the proposed techniques are both effective and scalable. Additionally, the experiments show that the similarity heuristic can be used as a viable alternative to t-wise.

Mass customization of software products requires their efficient tailoring performed through combination of features. Such features and the constraints linking them can be represented by Feature Models (FMs), allowing formal analysis, derivation of specific variants and interactive configuration. Since they are seldom present in existing systems, techniques to re-engineer FMs have been proposed. There are nevertheless error-prone and require human intervention. This paper introduces an automated search-based process to test and fix FMs so that they adequately represent actual products. Preliminary evaluation on the Linux kernel FM exhibit erroneous FM constraints and significant reduction of the inconsistencies.

Fault diagnosis is the process of analyzing programs with the aim of identifying the code fragments that are faulty. It has been identified as one of the most expensive and time consuming tasks of software development. Even worst, this activity is usually accomplished based on manual analysis. To this end, automatic or semi-automatic fault diagnosis approaches are useful in assisting software developers. Hence, they can play an essential role in decreasing the overall development cost. This paper presents Proteum/FL, a mutation analysis tool for diagnosing previously detected faults. Given an ANSI-C program and a set of test cases, Proteum/FL returns a list of program statements ranked according to their likelihood of being faulty. The tool differs from the rest of the mutation analysis and fault diagnosis tools by employing mutation analysis as a means of diagnosing program faults. It therefore demonstrates the effective use of mutation in supporting both testing and debugging activities.

Software Products Lines (SPLs) are families of products sharing common assets representing code or functionalities of a software product. These assets are represented as features, usually organized into Feature Models (FMs) from which the user can configure software products. Generally, few features are sufficient to allow configuring millions of software products. As a result, selecting the products matching given testing objectives is a difficult problem.

In this paper, a technique named I-EQM, able to dynamically isolate first order equivalent mutants, is proposed. I-EQM works by employing a novel dynamic execution scheme that integrates both first and second order mutation. The proposed approach combines the "impact" on the program execution of the first order mutants with the "impact" on the output of second order ones, to isolate likely to be first order equivalent mutants. Experimental results on a benchmark set of manually classified mutants, selected from real word programs, reveals that I-EQM achieves to classify equivalent mutants with a 71% and 82% classification precision and recall respectively. These results improve the previously proposed approaches by selecting (retrieving) a considerably higher number of killable mutants with only a limited loss on the classification precision.

Many fault localization techniques operate by crosscutting coverage information of passed and failed test executions. Generally, their accuracy depends on the utilized coverage elements and on the selected test cases. This paper proposes a novel fault localization method using mutation and investigates its accuracy when using classical test selection criteria such as mutation, branch and block. A controlled experiment shows that (1) the mutation based approach is quite effective at identifying "unknown" faulty program statements. Additionally, the experimental results reveal (2) that the mutation-based test suites are significantly more effective at supporting fault localization than block or branch-based test suites. Further, (3) evidence in support of facilitating mutation alternatives, such as mutant sampling, in order to diminish mutation overheads is also given.

Mutation testing while being a successful fault revealing technique for unit testing, it is a rather expensive one for practical use. To bridge these two aspects there is a need to establish approximation techniques able to reduce its expenses while maintaining its effectiveness. In this paper several second order mutation testing strategies are introduced, assessed and compared along with weak mutation against strong. The experimental results suggest that they both constitute viable alternatives for mutation as they establish considerable effort reductions without greatly affecting the test effectiveness. The experimental assessment of weak mutation suggests that it reduces significantly the number of the produced equivalent mutants on the one hand and that the test criterion it provides is not as weak as is thought to be on the other. Finally, an approximation of the number of first order mutants needed to be killed in order to also kill the original mutant set is presented. The findings indicate that only a small portion of a set of mutants needs to be targeted in order to be killed while the rest can be killed collaterally.

Software testing forms a substantial activity of the software development cycle. Although important, it lacks from being automated mainly because of the various undecidable problems that it encounters. To this extend efficient heuristics have been proposed in order to bypass this problem. One such approach, called symbolic execution, is usually used for automating the test data generation activity. In this paper, an automated symbolic execution tool is proposed. The tool employs an efficient path heuristic, integrated with random testing for producing test cases. The tool handles the path explosion and constraint solving problems efficiently. This is achieved by targeting on specific likely to be feasible paths and by using a linear programming approach for the determination of their feasibility. Preliminary results are very encouraging as they show that a high coverage can be achieved within a limited amount of time-effort.

The automatic test case generation is the principal issue of the software testing activity. Dynamic symbolic execution appears to be a promising approach to this matter as it has been shown to be quite powerful in producing the sought tests. Despite its power, it has only been effectively applied to the entry level criteria of the structural criteria hierarchy such as branch testing. In this paper an extension of this technique is proposed in order to effectively generate test data based on mutation testing. The proposed approach conjoins program transformation and dynamic symbolic execution techniques in order to successfully automate the test generation process. The propositions made in this paper have been incorporated into an automated framework for producing mutation based test cases. Its evaluation on a set of benchmark programs suggests that it is able to produce tests capable of killing most of the non equivalent introduced mutants. The same study also provides some evidence that by employing efficient heuristics it can be possible to perform mutation with reasonable resources.

Mutation testing has been identified as one of the most effective techniques, in detecting faults. However, because of the large number of test elements that it introduces, it is regarded as rather expensive for practical use. Therefore, there is a need for testing strategies that will alleviate this drawback by selecting effective test data that will make the technique more practical. Such a strategy based on path selection is reported in this paper. A significant influence on the efficiency associated with path selection strategies is the number of test paths that must be generated in order to achieve a specified level of coverage, and it is determined by the number of paths that are found to be feasible. Specifically, a path selection strategy is proposed that aims at reducing the effects of infeasible paths and conversely developing effective and efficient mutation based tests. The results obtained from applying the method to a set of program units are reported and analysed presenting the flexibility, feasibility and practicality of the proposed approach.