Here's a little sample from my Racket-based parser for parsing tokenized Python (uses Matt Might's derivative parser):
; <program> ::= (program <stmt>*) (file_input (@--> (seq! `(rep (or NEWLINE stmt)) ENDMARKER) (λ (seq) `(program ,@seq)))) ; <funcdef> ::= (def (<NAME> <NAME>*) <suite>) (funcdef ($--> (seq! "def" `NAME `parameters ":" `suite) `(def (,($ 0) ,@($ 1)) ,($ 2)))) ; ... more stuff ; <simple_stmt> ::= <small_stmt> | (begin <small_stmt>+) (simple_stmt (>--> (seq! `(seq* small_stmt (rep (seq! ";" `small_stmt))) (opt ";") NEWLINE) [`((,small_stmt ())) small_stmt] [`((,small_stmt ,rep)) `(begin ,small_stmt ,@rep)])) ; ... and a ton more cryptic cases
Because you can never have enough randomly* placed symbols sprinkled throughout your code...
*and by randomly, I mean very carefully
A whole file of this is actually kinda mesmerizing when you look back on what you've written.