This change looks fine, but wants_regex seems like a reoccurring issue in the parser, perhaps there's a better way to do it? If not, just go ahead and merge this.

Maybe someone can take a look at how is this done in Ruby's parser.

I just read parse.y in ruby/ruby. Ruby's parser has internal state and updates this on parsing, and here is to decide / is division operator or regular expression start. It looks like Crystal, however Ruby's state has more semantic name like EXPR_BEG or EXPR_END. I'm not sure which is better.