Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?
Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur Graph Search.
Type errors reported by compilers can sometimes be cryptic, or difficult to understand. In this paper, we propose a type debugging framework that exposes a high-level representation of the typechecking decision-making process that users normally do not have access to in state-of-the-art compilers. This representation makes it easier for non-experts to analyze complex type errors. Our system is implemented by instrumenting the existing Scala typechecker, but without modifying it. We also provide generic search algorithms that can be used as basic building blocks to build a number of systems, from visual type debugging tools to customized error reporting for normal Scala users as well as to users of domain-specific languages. Using our framework, it is possible to overcome well-known limitations of local type inference by providing precise type error messages to mainline Scala users or Scala DSL users alike. In addition, the framework provides better user feedback for non-trivial type errors from existing Scala libraries and DSLs.
Mathias Josef Payer, Jianhao Xu