Automatic evaluation of top-down predictive parsing

Abstract

We develop efficient methods to check whether two given Context-Free Grammars (CFGs) are transformed into parsers that recognize the same language and construct the same Abstract Syntax Trees (ASTs) for each input. In this setting, we consider a model of top-down predictive parser generator with directives for AST construction that is a simplified variant of PCCTS/ANTLR3. As an application, we implement an evaluator for an online judge with educational purposes in the context of a Compilers course.