Bytecode Manipulation Research Articles

Design and Implementation of a Java Bytecode Manipulation Library for Clojure

Design and Implementation of a Java Bytecode Manipulation Library for Clojure

Introduction to dynamic program analysis with DiSL

Dynamic program analysis (DPA) tools assist in many software engineering and development tasks, such as profiling, program comprehension, and performance model construction and calibration. On the Java platform, many DPA tools are implemented either using aspect-oriented programming (AOP), or rely on bytecode instrumentation to modify the base program code. The pointcut/advice model found in AOP enables rapid tool development, but does not allow expressing certain instrumentations due to limitations of mainstream AOP languages—developers thus use bytecode manipulation to gain more expressiveness and performance. However, while the existing bytecode manipulation libraries handle some low-level details, they still make tool development tedious and error-prone. Targeting this issue, we provide the first complete presentation of DiSL, an open-source instrumentation framework that reconciles the conciseness of the AOP pointcut/advice model and the expressiveness and performance achievable with bytecode manipulation libraries. Specifically, we extend our previous work to provide an overview of the DiSL architecture, advanced features, and the programming model. We also include case studies illustrating successful deployment of DiSL-based DPA tools.

JRebel Tool Demo

JRebel started as an academical project that became a successful commercial product used by thousands of developers worldwide. It extends the Java Virtual Machine with a mechanism that allows seamless class reloading. It uses bytecode manipulation extensively, both for the just-in-time class translator and numerous integrations with the Java SE and EE APIs. In this live demo we will show how it can be used in real-life projects to cut development time by 8 to 18 per cent.

User-changeable visibility

A trait is a unit of behaviour that can be composed with other traits and used by classes. Traits offer an alternative to multiple inheritance. Conflict resolution of traits, while flexible, does not completely handle accidental method name conflicts: if a trait with method m is composed with another trait defining a different method m then resolving the conflict may prove delicate or infeasible in cases where both versions of m are still needed. In this paper we present freezeable traits , which provide an expressive composition mechanism to support unanticipated method composition conflicts. Our solution introduces private trait methods and lets the class composer change method visibility at composition time (from public to private and vice versa). Moreover two class composers may use different composition policies for the same trait, something which is not possible in mainstream languages. This approach respects the two main design principles of traits: the class composer is empowered and traits can be flattened away. We present an implementation of freezable traits in Smalltalk. As a side-effect of this implementation we introduced private (early-bound and invisible) methods to Smalltalk by distinguishing object-sends from self-sends. Our implementation uses compile-time bytecode manipulation and, as such, introduces no run-time penalties.

Injecting roles in Java agents through runtime bytecode manipulation

Agents are problem-solving entities that can be used to develop complex and distributed systems because they are autonomous, mobile, reactive, social, and proactive. Today's trends in agent technology include the development of applications as multi-agent systems, where several agents interact within the same application. In these systems, the interactions among agents must be carefully considered. Roles constitute a powerful paradigm for modeling interactions, allowing algorithmic issues and interaction-dependent issues to be handled independently. In this paper, we present the RoleX interaction infrastructure, which enables JavaTM agents to dynamically assume and use roles at runtime. Our approach is based on using bytecode manipulation to add (or remove) Java members from agents, changing their capabilities. We detail the main component of RoleX, the Role Loader, which performs the bytecode manipulation that allows agents to dynamically assume and release roles.

Visualizing Program Analysis with the Soot-Eclipse Plugin

Our integration of the Soot bytecode manipulation framework into the Eclipse IDE forms a powerful tool for graphically visualizing both the progress and output of program analyses. We demonstrate several examples of the visualizations that we have developed, and explain how they are useful for both compiler research and teaching.

Exploiting runtime bytecode manipulation to add roles to Java agents

Thanks to their sociality, agents can interact with other agents in a cooperative or competitive way. Such interactions must be carefully taken into consideration in the development of agent-based applications. A good paradigm for modeling such interactions is the one based on the concept of roles, which is fully exploited in the BRAIN framework. The use of roles achieves several advantages, from separation of concerns between the algorithmic issues and the interaction ones, to the reuse of solutions and experiences in different applications. In this paper, we propose an interaction infrastructure for enabling Java agents to dynamically assume roles at runtime and then to use them. Our approach is based on the capability of modifying the bytecode of Java agents at runtime in order to add the members of role classes. An application example and a comparison with other approaches show the effectiveness of our approach.