Enabledness-based program abstractions for behavior validation

Abstract

Code artifacts that have nontrivial requirements with respect to the ordering in which their methods or procedures ought to be called are common and appear, for instance, in the form of API implementations and objects. This work addresses the problem of validating if API implementations provide their intended behavior when descriptions of this behavior are informal, partial, or nonexistent. The proposed approach addresses this problem by generating abstract behavior models which resemble typestates. These models are statically computed and encode all admissible sequences of method calls. The level of abstraction at which such models are constructed has shown to be useful for validating code artifacts and identifying findings which led to the discovery of bugs, adjustment of the requirements expected by the engineer to the requirements implicit in the code, and the improvement of available documentation.Fil: de Caso, Guido. Universidad de Buenos Aires; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Braberman, Victor Adrian. Universidad de Buenos Aires; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Garbervetsky, Diego David. Universidad de Buenos Aires; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Uchitel, Sebastian. Universidad de Buenos Aires; Argentina. Imperial College London; Reino Unido. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentin

Automatic mining of specifications from invocation traces and method invariants

“…Static [17,42] and hybrid approaches [15,46] provide two alternatives to specification mining based strictly on invocation traces. Shoham et al [42] infer, from client-side code, FSM models that over-approximate the actual invocation sequences.…”

“…By contrast, our algorithms work on traces generated from multiple parts of the code and, potentially, from multiple applications. De Caso et al [17] statically analyze C programs for invariants and use Contractor to create models that allow more behavior than the ground truth.…”

The chemical approach to typestate-oriented programming

“…This feature is crucial for the effectiveness of the approach given the use of structural subtyping and the richness of types. We are comforted by the fact that inference of object protocols has been investigated in a number of works (a detailed survey is given in [13]), some of which use specification languages inspired to regular expressions [25] as we do.…”

“…1 The code defining the Channel class has a straightforward correspondence with its formal counterpart. The class consists of a set of event declarations specifying the messages that can be targeted to instances of the class (lines 3-10), the reaction rules specifying the behavior of instances of the class (lines [11][12][13][14][15][16][17][18][19][20], as well as the initial state of each instance (lines [21][22]. The main program creates a channel that is shared by two asynchronous processes that exchange integers and strings in full-duplex (lines [37][38][39].…”

“…The messages representing the state of the channel (lines 3-6) are private to enforce encapsulation of the state. The public interface (lines 7-10) is represented by synchronous events, which means that invoking a Channel's public operation (lines 27-28 and 31-32) suspends the calling thread until the matching reaction has fired and a result has been returned (lines 13,15,17,19). In the formal model, all message sends are asynchronous and sequentiality is encoded with explicit continuations (see the reference c in Listing 5).…”