I think we've seen that working string support requires a substantial rewrite for most languages (perhaps ironically, Java is the one that has managed to avoid it). Ruby 1.9 was pretty big; Python 3 was more so.
Perl 6 was the Perl equivalent. Unfortunately we know how that turned out.
Getting Unicode right is more than "working string support". It requires programmer knowledge upgrades. No amount of boilerplate you put in your programs--or in the language itself--will fix a programmer's lack of knowledge.