The algorithm used is Generalized LL parsing (http://dotat.at/tmp/gll.pdf http://dotat.at/tmp/gll.pdf), which promises linear-time parsing of LL grammars and cu

The algorithm used is Generalized LL parsing (http://dotat.at/tmp/gll.pdf http://dotat.at/tmp/gll.pdf), which promises linear-time parsing of LL grammars and cubic-time parsing of arbitrary grammars. For some reason this isn't mentioned until the very end.