I have now written a new parser. It does not yet generate any AST but implements all the productions of the PIM3/4 grammar.
The grammar is embedded in the source code and I would appreciate if others could go over it to verify that the code matches the grammar. The idea is that two sets of eyes can see more than one set of eyes. Any oversight at this stage will have a knock on effect later on, so it is important to apply scrutiny.
The code isn't yet on the official repo, I intend to move it there when I have a working executable that can be used for syntax checking.
For now it is in my bitbucket repo and the parser is at:
Note that the skip-set references in the match-token and match-set calls are mostly tentative at this stage, that is, right now, they are not very fine grained but most will skip to the next production in the event of an error (SKIP_TO_FOLLOW_OF...). The idea is to refine them when we have a syntax checker executable to experiment with the error-recovery setup.
regards
benjamin