I am doing a trade-study of ‘parser generators’ and very (very) much like what I see in Boost.Spirit (direct C++ generics!). So, I'm willing to jump thru' a few flaming hoops to give it a go. On my checklist is a grammar (1) validation
and (2) visualization tool.
The nature of the grammar source seems to preclude any general tool
driven by BNF. What are the best practices to fill the gap for (1) and (2)? My
worst-case so far: My grammar will be generally stable in the long-term. I could use the ANTLRWorks IDE against a Qi-compatible-constrained EBNF source, then (carefully) translate to
Qi-compatible source. (I guess I could even build an ANTLR parser to do the
conversion - have to ponder that one a bit.)