We define a practical algorithm for distributed rational tree unification and prove its correctness in both the off-line and on-line cases. We derive the distributed algorithm from a centralized one, showing clearly the trade-offs between local and distributed execution. The algorithm is used to realize logic variables in the Mozart Programming System, which implements the Oz language (see

Publication interne n˚1725 — Juin 2005 — 17 pages Abstract: Functional testing based on a formal specification consists in deriving test cases from a formal model to detect faults within an implementation. In our work, we investigate the use of Constraint Handling Rules (CHRs) to automate functional test cases generation based on a formal model. Our case study is a model of the Java Card Virtual Machine (JCVM) specification written in a subset of the Coq language. In this paper we define an automated translation from this model into CHRs in order to generate test cases for the JCVM. We also propose several test purposes based on rewriting rules coverage and automatic non-conformity detection. The key point of our approach resides in the use of deep guards to model faithfully the semantic of our formal model of the JCVM. Finally, we propose an overall functional test case generation approach based on CHRs that could be applied to other formal models. Key-words: test case generation, functional testing, constraint solving, rewriting rules,

Two variants of an incremental and adaptive entailment algorithm are presented. Both variants decide incrementally whether a fixed conjunction of syntactical equations over the rational trees is entailed by another conjunction in rational solved form which may change dynamically. In the case of entailment, an extension of the rational solved form is computed which is equivalent to the conjunction of both considered conjunctions. Contrary to other incremental entailment algorithms, our approach uses a simplified form of justifications used in truth maintenance systems to support solution adaptation after changes of the entailing rational solved form. In detail, previously made decisions and extensions are adapted after arbitrary additions and deletions of global variable bindings. An application domain of the adaptive entailment algorithm is dynamic constraint processing based on Fruhwirth's Constraint Handling Rules, where head matching and guard entailment depends on the currently v...

Interpretation Frameworks Partner Responsible. INRIA The work in this task was intented to propose an abstract interpretation framework for the family of Concurrent Constraint (cc) languages proposed by Saraswat. Our framework is based on a denotational semantics of concurrent constraint languages, where each agent is seen as a closure operator over the lattice defined by the ACCLAIM Final Report 33 constraint system. In this framework the closure operators are defined directly as functions over stores instead of defining them via the set of their fixpoints. This will make it possible to derive a practical algorithm reusing and only slightly modifying a reexecution-based framework for Prolog. Deliverables due 2Q93. D3.1/1 D3.1/1 A report describing a framework for the abstract interpretation of cc languages. This deliverable will provide the basis for several applications to program analysis and optimisation. (INRIA, public report). Deliverables due 2Q95. None An extra deliverable o...

. Two variants of an incremental and adaptive entailment algorithm are presented. Both variants decide incrementally whether a fixed conjunction of syntactical equations over the rational trees is entailed by another conjunction in rational solved form which may change dynamically. In the case of entailment, an extension of the rational solved form is computed which is equivalent to the conjunction of both considered conjunctions. Contrary to other incremental entailment algorithms, our approach uses a simplified form of justifications used in truth maintenance systems to support solution adaptation after changes of the entailing rational solved form. In detail, previously made decisions and extensions are adapted after arbitrary additions and deletions of global variable bindings. An application domain of the adaptive entailment algorithm is dynamic constraint processing based on Fruhwirth's Constraint Handling Rules, where head matching and guard entailment depends on the currently v...

this paper, I propose a way of reconciling constrained sorts with the finite domain view of disjunction. My approach is both semantically sound and lends itself to an efficient implementation that preserves incrementality.