Software Vulnerability Analysis Across Programming Language and Program Representation Landscapes: A Survey
Zhuoyun Qian, Fangtian Zhong, Qin Hu, Yili Jiang, Jiaqi Huang, Mengfei Ren, Jiguo Yu · 2025
Read Software Vulnerability Analysis Across Programming Language and Program Representation Landscapes: A Survey by Zhuoyun Qian, Fangtian Zhong, Qin Hu, Yili Jiang, Jiaqi Huang, Mengfei Ren, Jiguo Yu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-20244.txt.
Introduction
Zhuoyun Qian and six co-authors survey software vulnerability analysis in a paper from two thousand twenty-five. Modern systems are written in many languages and often hide flaws that attackers exploit. Many detection tools exist, but no study had compared them across the whole landscape.
The survey spans programming languages, levels of program representation, bug types, and detection techniques. It sets out current practice in finding flaws before release, with each method's strengths and limits. Relationships among language, abstraction, flaw type, and approach are drawn together in one frame. The authors close with open challenges and future directions. Their aim is a map that helps security researchers see where the field stands and where it must go.
Catalog record
| CREATOR | Zhuoyun Qian, Fangtian Zhong, Qin Hu, Yili Jiang, Jiaqi Huang, Mengfei Ren, Jiguo Yu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 111 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-20244
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Software Vulnerability Analysis Across Programming Language and Program Representation Landscapes: A Survey Zhuoyun Qian, Fangtian Zhong, Qin Hu, Yili Jiang, Jiaqi Huang, Mengfei Ren, Jiguo Yu Abstract . Modern software systems are developed in diverse programming languages and often harbor critical vulnerabilities that attackers can exploit to compromise security. These vulnerabilities have been actively targeted in real-world attacks, causing substantial harm to users and cyberinfrastructure. Since many of these flaws originate from the code itself, a variety of techniques have been proposed to detect and mitigate them prior to software deployment. However, a comprehensive comparative study that spans different programming languages, program representations, bug types, and analysis techniques is still lacking. As a result, the relationships among programming languages, abstraction levels, vulnerability types, and detection approaches remain fragmented, and the limitations and research gaps across the landscape are not clearly understood. This article aims to bridge that gap by systematically examining widely used programming languages, levels of program representation, categories of vulnerabilities, and mainstream detection techniques. The survey provides a detailed understanding of current practices in vulnerability discovery, highlighting their strengths, limitations, and distinguishing characteristics. Furthermore, it identifies persistent challenges and outlines promising directions for future research in the field of software security. DOI: XXXXXXX.XXXXXXX Journal: JACM Volume: 37 4 111 8 CCS: Software Security CCS: System security CCS: Vulnerability discovery 1. Introduction 1.1. Background Software vulnerability discovery is the cornerstone of securing modern software systems. It plays a critical role in software security, as vulnerabilities can result in unauthorized access, data breaches, system crashes, privilege escalation, and degraded user experience. By systematically identifying weaknesses in code, configuration, and execution environments, vulnerability discovery enables organizations to assess and mitigate risks before they are exploited by adversaries. In offensive security research, this process reveals exploitable code paths and vulnerable inputs, facilitating the development of targeted mitigations and comprehensive security evaluations. In defensive contexts, it supports proactive protection by uncovering previously unknown vulnerabilities, guiding patch prioritization, and enhancing system resilience against attacks. Vulnerability analysis is typically conducted at three levels: binary , intermediate representation (IR) , and source code , with each offering unique advantages and facing specific challenges. Binary-level analysis operates directly on compiled executables, making it indispensable for assessing closed-source software and deployed applications. Techniques such as binary instrumentation, fuzzing, and taint analysis are commonly used to detect memory corruption, missing-check vulnerabilities, and side-channel leaks. However, this level suffers from limited semantic information, obfuscation challenges, and difficulties in reconstructing accurate control and data flow. IR-level analysis is performed on representations like low-level virtual machine (LLVM) IR, bytecode, assembly, and abstract syntax trees (ASTs), striking a balance between low-level execution details and high-level code semantics. Techniques such as flow-sensitive analysis, data-flow analysis, and symbolic execution are employed to detect concurrency bugs, missing-check vulnerabilities, and memory safety issues. While IR-level analysis benefits from platform independence and better analyzability than binaries, it may lack complete context in the source code. Source-code level analysis focuses on human-readable source code, allowing early detection of logic errors, insecure coding patterns, and API misuse. Common techniques include static analysis and machine learning-based vulnerability detection. This level provides rich syntactic information, enabling precise reasoning about code behavior. However, it may be limited in capturing vulnerabilities that manifest only during execution or after compiler transformations. Furthermore, the programming language itself significantly influences the nature of vulnerabilities and the detection strategies employed. For example, C/C++ programs are particularly prone to memory safety issues, performance bottlenecks, concurrency bugs, and logic flaws. In contrast, JavaScript frequently exhibits logic bugs, missing-check vulnerabilities, and concurrency-related issues due to its dynamic and event-driven nature. As such, a comprehensive understanding of software vulnerabilities requires cross-language, cross-representation, and multi-bug-type perspectives. 1.2. Classification by Bugs Figure 1. Classification of Bugs The classification of software bugs, as illustrated in Fig. 1 , spans seven major categories: performance bugs, memory safety bugs, concurrency bugs, missing-check bugs, logic bugs, resource management bugs, and side channel vulnerabilities, each representing distinct causes of system instability or inefficiency. Performance bugs occur when software exhibits slow execution or excessive resource use without functional errors, often due to synchronization issues, configuration problems, or inefficient algorithms, frequently observed in C/C++, Java, and Android. Memory safety bugs arise from improper memory operations like use-after-free, buffer overflows, and invalid access, which may lead to crashes or data corruption, especially in C/C++ and Rust. Concurrency bugs result from incorrect handling of multiple threads or processes, leading to race conditions, atomicity violations, or inconsistent states—common across C/C++, Java, Android, and JavaScript. Missing-check bugs stem from the failure to validate inputs, permissions, or security conditions, enabling vulnerabilities such as access control violations and component hijacking across C/C++, Java, Android, PHP, and JavaScript. Logic bugs are caused by flawed control or computation logic, including API misuse, semantic errors, and improper comparisons, impacting correctness in languages like C/C++, Java, JavaScript, Python, and PHP. Resource management bugs involve the incorrect handling or accounting of system resources (e.g., memory), leading to leaks or data loss in Android, C/C++, and Python. Side-channel vulnerabilities are security flaws where attackers infer sensitive information through indirect behaviors such as timing, cache access, or traffic patterns, with examples like SSL/TLS leaks in C/C++, resource usage leaks in Java, and network traffic leaks in JavaScript. 1.3. Related Work Prior surveys have explored specific aspects of software vulnerability detection, often focusing on narrow domains or limited bug types. For instance, ( Johari and Sharma, 2012 ) evaluates static and dynamic analysis techniques for detecting SQL injection (SQLi) and cross-site scripting (XSS) vulnerabilities, while ( Ullah et al., 2020 ) integrates rule-based filtering with machine learning to detect attacks in SS7 telecommunication networks. Fuzzing has been extensively studied in ( Zhao et al., 2024 ) , which categorizes modern fuzzing strategies, including machine learning-assisted fuzzing, constraint-solving integration, and directed fuzzing for improving code coverage and vulnerability discovery efficiency. ( Ji et al., 2018 ) provides an overview of automated techniques for vulnerability detection, exploitation, and patching, emphasizing the growing role of machine learning in software security automation. Similarly, ( Zuo and Rhee, 2024 ) investigates the mining of patch commits to identify silent vulnerabilities in open-source software, leveraging NLP and graph-based machine learning approaches. Machine learning has indeed become a transformative force in vulnerability detection. ( Zhu et al., 2023 ; Jia et al., 2021b ) surveys deep learning-based techniques, analyzing models such as CNNs, RNNs, and Transformers for software vulnerability detection. ( Kaur et al., 2023 ) focuses specifically on LSTM-based models for XSS detection. In contrast to these focused studies without generalizability and reusability of findings, our survey provides a broad, technique-driven overview of software vulnerabilities across multiple programming languages, program representations, and bug types. This holistic perspective fills an important gap in the current literature by integrating language diversity, representation levels, and vulnerability taxonomies into a single cohesive framework. The differences between existing surveys and ours are presented in Table 1 . Table 1. Comparison of Surveys Paper Multi-Programming Languages Multi-Program Representations Multi-Bug Types Multi-Analysis Methods ( Johari and Sharma, 2012 ) ✓ \times \times ✓ ( Ullah et al., 2020 ) \times \times \times \times ( Zhao et al., 2024 ) \times \times ✓ \times ( Ji et al., 2018