grammar-protos.p6

# In a grammar, how do I reuse a proto'ed definition? Here for
# example, I want <ld> to use some <elem>:

#use Grammar::Debugger;

grammar Foo {

    proto token elem {*}
    token elem:sym<let> { <[a..z]> }
    token elem:sym<dig> { <[0..9]> }

    token anys { <elem>+ }

        # A <let> followed by a <dig>. None of these work. Some fail
        # to match, some fail to compile.
   # token ld {         <let>         <dig> }
   # token ld {      sym<let>      sym<dig> }
   # token ld { elem:sym<let> elem:sym<dig> }
   # token ld {     elem<let>     elem<dig> }

    token TOP { <anys> | \( <ld> \) }
};

    # Matches <TOP> with <anys>.
#say Foo.parse('bar42');

    # Can't get this to match <TOP> with parenthesized <ld>.
say Foo.parse('(k9)');

=finish
- --------------------------------------------------------------------

Two solutions were suggested by ShimmerFairy#raku@freenode.net.irc

1. Factor out the names of the required tokens:

    token let { <[a..z]> }
    token dig { <[0..9]> }
    token elem:sym<let> { <let> }
    token elem:sym<dig> { <dig> }
    token ld { <let> <dig> }

2. Define <ld> like this:

    token ld { <elem:sym<let>> <elem:sym<dig>> }

Did you try escaping the left paren? Inside double or single quotes?

Tom, it turns out the parentheses were not a problem. I actually needed to fix my token definitions, and ShimmerFairy pointed out a couple of ways to do it, shown in the current revised version of the gist.

Hm, great. The sym thing is something I've avoided as hard to understand, but I'm at a point in my own project I need to buckle down and study it more closely. @ShimmerFairy has a lot of Raku experience and it would be nice if she could add more to our docs about the solution and more.