@nadako Thanks for chiming in!
mostly LL(1) grammar (there’s one edge case with semicolon before else that requires two token peek)
Interesting. After learning a bit about each of those, I was expecting Haxe to use LR (or its variations) or PEG, which seem to be the most powerful and pragmatic ones and the ones that are generated by most parser generators. I also wonder what is the main advantage of OCaml + sedlex + camlp4 vs the usual lex/yacc route (which AFAIK, generates LALR parsers)? It looks like the grammar for Haxe is then not available in a declarative format (as it would be if a tool like yacc/bison was used)?
Also, from what I’ve read, RD parsers are not very performant, but perhaps the Haxe OCaml implementation is better in this aspect?
Well, I confess that I have never “dumpster dived” into the source code yet, but this web page suggests that it might be good ol’ LEX/YACC.
Looks like it doesn’t use the OCaml impl. of Lex/Yacc, as stated by @nadako.