Syntax

The syntax I actually build using iterations. One part at a time I made the generator build its own parser. Formally the syntax is

Exp \rightarrow "prod" Identifier "rarrow" ProdRight Exp Exp \rightarrow "prec" String Num Exp Exp \rightarrow "assoc" String "colon" StringTuple Exp Exp \rightarrow "token" Token Exp Exp \rightarrow "bang" "token" BangToken Exp Exp \rightarrow "group" Group Exp Exp \rightarrow ProdRight \rightarrow String ProdRight ProdRight \rightarrow Identifier ProdRight ProdRight \rightarrow "mid" ProdRight ProdRight \rightarrow Token \rightarrow StringTuple "as" StringTuple Token \rightarrow "f_cap" StringTuple "as" StringTuple BangToken \rightarrow StringTuple Group \rightarrow String "lbrace" StringTuple "rbrace" Identifier \rightarrow "identifier" StringTuple \rightarrow String "comma" StringTuple StringTuple \rightarrow String String \rightarrow "string" Num \rightarrow "num"

The syntax does not contain any need for associativity or precedence. So in order to let the parser generate itself, we insert into it

token -cap "string", "identifiers", "num" as "\g[^\g]*\g", "[A-Z][a-zA-Z']*", "[0-9]*" token "bang" as "!" token "rarrow" as "->" token "lbrace" as "\\{" token "rbrace" as "\\}" token "comma" as "," token "colon" as ":" token "mid" as "\\|" token "f_cap" as "-cap" #token "f_exp" as "-exp" #token "f_uexp" as "-uexp" !token "[\\n\\t\\r ]+", "#[^\\n]*\\n" group "definition" {"prod","prec","assoc","group","bang"} group "parentheses" {"lbrace","rbrace"} group "literal" {"string","num"} group "flag" {"f_cap"} prod Exp -> "prod" Identifier "rarrow" ProdRight Exp prod Exp -> "prec" String Num Exp prod Exp -> "assoc" String "colon" StringTuple Exp prod Exp -> "token" Token Exp prod Exp -> "bang" "token" BangToken Exp prod Exp -> "group" Group Exp prod Exp -> prod ProdRight -> String ProdRight prod ProdRight -> Identifier ProdRight prod ProdRight -> "mid" ProdRight prod ProdRight -> prod Token -> StringTuple "as" StringTuple prod Token -> "f_cap" StringTuple "as" StringTuple prod BangToken -> StringTuple prod Group -> String "lbrace" StringTuple "rbrace" prod Identifier -> "identifier" prod StringTuple -> String "comma" StringTuple prod StringTuple -> String prod String -> "string" prod Num -> "num"

And that is it. The semantics is explained in next chapter.