enhancing xtext for general purpose languages

Enhancing Xtext for General Purpose

LanguagesAdolfo Sánchez-Barbudo Herrera

asbh500@york.ac.uk

Outline

★ EngD Project○ Context○ Motivation○ Scope

★ Research○ Problems○ Directions○ Evaluation

★ Current Status

EngD Project Context

New impl.

Eclipse OCL

Eclipse QVTo

2005 2007 2010 2013

Alignment

Extension

Classic

OCL / QVT

3/13

EngD Project Motivation

● New QVTo implementationo Xtext-based parsers and editorso Extending “new” OCL implementation

● Problemo Xtext provides strong support for textual DSLso Eclipse OCL required significant hand written

code● Main goal

o Enhancing Xtext grammars -> Reduce the need of hand written code for QVTo (and other languages)

4/13

EngD Project Scope

● Restrictiono Xtext as a language workbench.

● Other language workbenches not considered

Spoofax

5/13

Research Problem (I)

● Gap between fixed CS and AS

Bridge

6/13

Research Problem (II)

● Name Resolution● o Qualified accesso Nested scopeso Inheritanceo i.e. Name visibility rules

7/13

Research Problem (III)

● Syntax rewriteso Syntax sugar resolution, disambiguation rules, etc.

● Examples (OCL expression):o abc => NameExpCS

abc => VariableExp self.abc => PropertyCallExp

o self.aProp->size() aProp is single-valued ?

● implicit collection conversion => self.aProp.asSet()->size()

8/13

Research Directions (I)

● Higher level of abstraction DSLs o Name resolutiono Syntax rewrites (CS2AS bridge, disambiguation, etc.)

9/13

Research Directions (II)

● Activities dependency analysis

CS 2 AS

NameResolution Disambiguation

10/13

Research Evaluation

● Compare amount of LOC for manual artefactso Measurements of LOC for hand-written java codeo Measurements of LOC for equivalent DSL instances

● Compare execution times of test caseso Measurements based on current Eclipse OCL impl.o Measurements based on future Eclipse OCL impl.o Backed up by: Theoretical algorithm analysis

11/13

Current Status (I)● OCL Documents as intermediate artefacts

12/13

o High expressiveness + Eclipse OCL2Java generatoro OCL descriptions as part of OMG specifications

Current Status (II) 13/13

● DSLs still need to be researched

Questions & Discussion

?e-mail: asbh500@york.ac.uktwitter: @adolfosbh

Discussion - Related Work

● Spoofaxo NaBL (name resolution) & Stratego (syntax rewrites)

● JastAdd (& JastEMF)o Attribute grammars: name resol. & syntax rewrites

● Gra2Molo DSTL to bridge CS and AS

● Xsemanticso A DSL for writing semantic rules for Xtext languages

Discussion - Future Work

● Finish source code generatorso Syntax rewrites and name resolution descriptionso From Complete OCL documents to Java codeo Improve generated source code based on activities

dependency analysis● DSLs

o Design high level of abstraction DSLso Implement generators (to generate Complete OCL)

● Evaluate research project contributions

Discussions - Dep. Analysis 14/14

Acknowledgements

Prof. Richard F. Paige from University of YorkDr. Edward Willink from Willink TransformationsEngineering & Physical Sciences Research Council ES group colleaguesAnonymous reviewers

References[1] OMG. OCL, V2.4. http://www.omg.org/spec/OCL/2.4, January 2013[2] OMG. QVT, V1.2. http://www.omg.org/spec/QVT/1.2, May 2014.[3] M. Eysholdt and H. Behrens. Xtext: Implement your language faster than the quick and dirty way. In OOPSLA, 2010.[4] Edward D. Willink. Re-engineering Eclipse MDT/OCL for Xtext. Electronic Communications of the EASST, 2010.[5] Alfred V. Aho, Monica S. Lam, Ravi Sethi, and Jeffrey D. Ullman. Compilers: principles, techniques, & tools. Pearson Education Inc., 2007.[6] Gabriel Konat, Lennart Kats, Guido Wachsmuth, and Eelco Visser. Declarative name binding and scope rules. In Software Language Engineering, volume 7745. Springer Berlin Heidelberg, 2013.[7] Lennart Kats and Eelco Visser. The spoofax language workbench: rules for declarative specification of languages and IDEs. In ACM Sigplan Notices, volume 45, 2010.[8] Torbjorn Ekman and Gorel Hedin. Modular name analysis for Java using JastAdd. In Generative and � �Transformational Techniques in SE. Springer, 2006.[9] Christoff Burger, Sven Karol, and Christian Wende. Applying attribute grammars for metamodel semantics. In �Proceedings of the International Workshop on Formalization of Modeling Languages, 2010.[10] Javier Canovas Izquierdo and Jesus Garcia Molina. Extracting models from source code in software modernization. Software & Systems Modeling, 2012.