GRK deployment scripts for VS Cobol II

• liberal.fst The grammar is liberalised to include obsolete constructs or more or less undocumented constructs. This script is accumulated on the basis of approved Cobol code (normally VS Cobol II code).

• seq.fst The grammar is completed and refactored to correctly describe the syntax of possibly nested statement sequences. We did not include this phase in the recovery / re-engineering part because the many details are probably not suitable for a language reference.

• copy.fst The grammar is further extended to include typical occurrences of COPY statements as context-free syntax. In this manner, Cobol source can be parsed even if the COPY statements were not expanded.

• reduce.fst The grammar is reduced, say various nonterminals are eliminated together with, as for all other grammar transformations, comments on the justification of these eliminations.

• permute.fst Simplify grammar to enable a very simple approach for removing permutation phrases. This simple approach parses x||y as (x|y)*. In order to make sense for certain kinds of parsers, we better make sure that x and y do not generate empty string.

• dcg.fst The grammar is transformed to prepare it for its use as a DCG to be executed in Prolog. Most notably, we establish cuts to limit backtracking so to make the generated validating prototype parser more efficient.

• flex.fst The grammar is transformed to prepare it for its use as with a flex-based scanner as done for the BTYACC-based parser. This basically necessitates making the grammar a bit more liberal, e.g., level numbers and numeric literals are not separated.