Infoscience

Report

Traversal Query Language For Scala.Meta

With the rise of metaprogramming in Scala, manipulating ASTs has become a daily job. Yet the standard API provides only low-level mechanisms to transform or to collect information on those data structures. Moreover, those mechanisms often force the programmer to manipulate state in order to retrieve information on these ASTs. In this report, we try to solve those problems by introducing TQL, a high-level combinator Scala library to transform and query data structures in a purely functional way. Parser combinators allow to combine several small parsers to build a bigger one in an expressive way. In this report, we argue that we can apply the same concept to data structure manipulation and construct complicated traversers on top of smaller ones. Yet combinators may feel unnatural or too complicated for certain usage. We therefore built a library on top of TQL to manipulate data structures as a collection. We then put TQL in practice to scala.meta ASTs, and describe the challenges we face when traversing a real-word data structure, especially performance-wise.

Related material