Full contract verification for ATL using symbolic execution

Abstract

The Atlas Transformation Language (ATL) is currently one of the most-used model transformation languages and has become a de-facto standard in modeldriven engineering for implementing model transformations. At the same time, it is understood by the community that enhancing methods for exhaustively verifying such transformations allows for a more widespread adoption of model-driven engineering in industry. A variety of proposals for the verification of ATL transformations have arisen in the past few years. However, the majority of these techniques are either based on nonexhaustive testing or on proof methods that require human assistance and/or are not complete.In this paper, we describe our method for statically verifying the declarative subset of ATL model transformations. This verification is performed by translating the transformation (including features like filters, OCL expressions, and lazy rules) into our model transformation language DSLTrans. As we handle only the declarative portion of ATL, and DSLTrans is Turingincomplete, this reduction in expressivity allows us to use a symbolic-execution approach to generate representations of all possible input models to the transform-

Fixing Multiple Type Errors in Model Transformations with Alternative Oracles to Test Cases.

“…Their approach uses matching functions that automatically create the alignment between specifications and implementations. Oakes et al (Oakes et al 2018) presented one method to fully verify pre-/post-condition contracts on declarative portion of ATL model transformations. Their approach transforms the declarative portion of ATL transformations into DSLTrans and uses a symbolic-execution to produce a set of path conditions, which represent all possible executions to the transformation.…”

Spectrum-Based Fault Localization in Model Transformations

“…We have slightly tweaked some transformations to increase their variability. For instance, in the BibTeX2DocBook, we have integrated the helpers within the rules, since the same transformation with the same behavior can be written with and without helpers [80], or in the CPL2SPL we have included some rules to transform features that were not included in the original transformation. All transformations are available on our website [100] and briefly described in the following:…”

“…Some works propose debugging model transformations by bringing them to a different domain such as Maude [102], DSLTrans [80], or Colored Petri Nets [110], where some specific analysis can be applied. The problem with these approaches is that the user needs to be familiar with such domains; besides, their performance and scalability can be worse than that of the original model transformation [102].…”

“…Among them, The ATLas transformation language (ATL) [61,84] has come to prominence in the MDE community both in the academic and the industrial arenas, so the testing of ATL transformations is of prime importance. This success is due to ATL's flexibility, support of the main metamodeling standards, usability that relies on strong tool integration within the Eclipse world, and a supportive development community [80]. To implement our approach and achieve automation, we have built a prototype for debugging ATL model transformations.…”

Towards Interactive, Test-driven Development of Model Transformations.

“…The fact that model transformation languages are higher-level that general purpose language could, in principle, provide opportunities for supporting verification methods which are more difficult to provide in general-purpose language languages. For instance, several transformation-specific analysis have been proposed for different languages (Cuadrado et al 2018a) (Cheng & Tisi 2017) (Oakes et al 2018), mechanisms for generating test cases automatically (González & Cabot 2012) (Guerra & Soeken 2015) (Gogolla & Vallecillo 2011), visualizations (Guana & Stroulia 2014), etc. An scenario which has been less explored until now is to enhance the development environment with features to help the transformation developer construct meaningful transformation rules interactively by means of (semi-)automatically generated test cases.…”