There are two main approaches to partial evaluation [6], a well-known technique for program specialisation. Online partial evaluators basically include an augmented interpreter that tries to evaluate the program constructs as much as possible—using the partially known input data—while still ensuring the termination

PET is a prototype Partial Evaluation-based Test case generation tool for a subset of Java bytecode programs. It performs white-box test generation by means of two consecutive Partial Evaluations (PE). The first PE decompiles the Java bytecode program into an equivalent CLP (Constraint Logic Programming) counterpart. The second PE generates a test-case generator from the CLP program. This generator captures interesting test coverage criteria and it is able to generate further test cases on demand. For the first PE, PET incorporates an existing tool which decompiles bytecode to CLP. The main contribution of this work is the implementation of the second PE and the proof of concept of the approach. This has required the development of a partial evaluator for CLP with appropriate control strategies to ensure the required coverage criteria and to generate test-case generators. PET can be downloaded as free software from its web site, where a repository of examples and a web interface are also provided. Though PET has to be extended to be applicable to larger programs, we argue that it provides some evidence that the approach can be of practical interest.

Decompiling low-level code to a high-level intermediate representation facilitates the development of analyzers, model checkers, etc. which reason about properties of the low-level code (e.g., bytecode,.NET). Interpretive decompilation consists in partially evaluating an interpreter for the low-level language (written in the high-level language) w.r.t. the code to be decompiled. There have been proofs-ofconcept that interpretive decompilation is feasible, but there remain important open issues when it comes to decompile a real language: does the approach scale up? is the quality of decompiled programs comparable to that obtained by ad-hoc decompilers? do decompiled programs preserve the structure of the original programs? This paper addresses these issues by presenting, to the best of our knowledge, the first modular scheme to enable interpretive decompilation of low-level code to a high-level representation, namely, we decompile bytecode into Prolog. We introduce two notions of optimality. The first one requires that each method/block is decompiled just once. The second one requires that each program point is traversed at most once during decompilation. We demonstrate the impact of our modular approach and optimality issues on a series of realistic benchmarks. Decompilation times and decompiled program sizes are linear with the size of the input bytecode program. This demostrates empirically the scalability of modular decompilation of low-level code by partial evaluation. 1

Test data generation (TDG) is the process of automatically generating test-cases for interesting test coverage criteria. The coverage criteria measure how well the program is exercised by a test suite. Examples of coverage criteria are: statement coverage which requires that each line of the code is executed; path coverage which

Testing is a vital part of the software development process. Test Case Generation (TCG) is the process of automatically generating a collection of test-cases which are applied to a system under test. White-box TCG is usually performed by means of symbolic execution, i.e., instead of executing the program on normal values (e.g., numbers), the program is executed on symbolic values representing arbitrary values. When dealing with an objectoriented (OO) imperative language, symbolic execution becomes challenging as, among other things, it must be able to backtrack, complex heap-allocated data structures should be created during the TCG process and features like inheritance, virtual invocations and exceptions have to be taken into account. Due to its inherent symbolic execution mechanism, we pursue in this paper that Constraint Logic Programming (CLP) has a promising application field in TCG. We will support our claim by developing a fully CLP-based framework to TCG of an OO imperative language, and by assessing it on a corresponding implementation on a set of challenging Java programs.

Decompiling low-level code to a high-level intermediate representation facilitates the development of analyzers, model checkers, etc. which reason about properties of the low-level code (e.g., bytecode,.NET). Interpretive decompilation consists in partially evaluating an interpreter for the low-level language (written in the high-level language) w.r.t. the code to be decompiled. There have been proofs-ofconcept that interpretive decompilation is feasible, but there remain important open issues when it comes to decompile a real language: does the approach scale up? is the quality of decompiled programs comparable to that obtained by ad-hoc decompilers? do decompiled programs preserve the structure of the original programs? This paper addresses these issues by presenting, to the best of our knowledge, the first modular scheme to enable interpretive decompilation of low-level code to a high-level representation, namely, we decompile bytecode into Prolog. We introduce two notions of optimality. The first one requires that each method/block is decompiled just once. The second one requires that each program point is traversed at most once during decompilation. We demonstrate the impact of our modular approach and optimality issues on a series of realistic benchmarks. Decompilation times and decompiled program sizes are linear with the size of the input bytecode program. This demostrates empirically the scalability of modular decompilation of low-level code by partial evaluation. 1.

Abstract. Test Data Generation (TDG) aims at automatically obtaining test inputs which can then be used by a software testing tool to validate the functional behaviour of the program. In this paper, we propose resource-aware TDG, whose purpose is to generate test cases (from which the test inputs are obtained) with associated resource consumptions. The framework is parametric w.r.t. the notion of resource (it can measure memory, steps, etc.) and allows using software testing to detect bugs related to non-functional aspects of the program. As a further step, we introduce resource-driven TDG whose purpose is to guide the TDG process by taking resource consumption into account. Interestingly, given a resource policy, TDG is guided to generate test cases that adhere to the policy and avoid the generation of test cases which violate it. 1

Abstract not found

Abstract. A range of methodologies and techniques are available to guide the design and implementation of language extensions and domainspecific languages. A simple yet powerful technique is based on source-tosource transformations interleaved across the compilation passes of a base language. Despite being a successful approach, it has the main drawback that the input source code is lost in the process. When considering the whole workflow of program development (warning and error reporting, debugging, or even program analysis), program translations are no more powerful than a glorified macro language. In this paper, we propose an augmented approach to language extensions for Prolog, where symbolic annotations are included in the target program. These annotations allow selectively reversing the translated code. We illustrate the approach by showing that coupling it with minimal extensions to a generic Prolog debugger allows us to provide users with a familiar, source-level view during the debugging of programs which use a variety of language extensions, such as functional notation, DCGs, or CLP{Q,R}.