Abstract
Pattern completeness is the property that the left-hand sides of a functional
program or term rewrite system cover all cases w.r.t. pattern matching. We
verify a recent (abstract) decision procedure for pattern completeness that
covers the general case, i.e., in particular without the usual restriction of
left-linearity. In two refinement steps, we further develop an executable
version of that abstract algorithm. On our example suite, this verified
implementation is faster than other implementations that are based on
alternative (unverified) approaches, including the complement algorithm, tree
automata encodings, and even the pattern completeness check of the GHC Haskell
compiler.
License
Topics
Related publications
- Thiemann, R., & Yamada, A. (2024). A Verified Algorithm for Deciding Pattern Completeness. Schloss Dagstuhl – Leibniz-Zentrum für Informatik. https://doi.org/10.4230/LIPICS.FSCD.2024.27