Regular Expression Types for Strings in a Text Processing Language

Abstract

We present λre, a minimal functional calculus with regular expression types for strings, in order to establish a theoretical foundation of using regular expressions as types of strings in text processing languages. The major technical novelties in this paper (with respect to other work such as XDuce) are (1) the use of regular expression effects to statically analyze the shape of the output of an even diverging program and (2) the treatment of as-patterns in non-tail positions. We expect that our approach will be of help for bringing the merits of static typing into scripting languages such as Perl, Python, and Ruby without sacrificing too much of their expressiveness and flexibility.