add ebnf representation

[alexpantyukhin]

Related to issue: #1

[alecthomas]

This is awesome, thanks!

[alecthomas]

Hmmm, actually this doesn't seem to be working correctly.

The following struct:

type grammar struct {
	ID   string   `@Ident { @"." @Ident }`
	Next *grammar `[ @@ ]`
}

Results in this EBNF:

ID := token("Ident") .
Next := Ident { "." Ident } [ participle.grammar ] .

I'm going to revert this. I'd be happy to take another look if you can fix the bugs and add some tests.

[alexpantyukhin]

Yes, it seems problem there. I will fix it.
I guess that creating intermediate structure between node -> string in ebnf generation process would be useful for testing.

[alexpantyukhin]

@alecthomas .
The problem is that ebnf is basing on the node structure. If run

    fmt.Print(parser.String())

for the grammar which you pasted we wil get the following result:

strct(type=main.grammar, expr=(@(field=ID, node=token("Ident")) { (@(field=ID, node=".") @(field=ID, node=token("Ident"))) } [@(field=Next, node=<>)]))
.