template.py does not understand unicode literals

[Anand Chitipothu]

The following test fails because template.py does not understand unicode literals.

        lambda: t(u'$def with (f)\n$f("x")')(lambda x:x), 'x\n',

And here is the fix.

     def h_literal(self, i):
         item = i[THING]
- if isinstance(item, str) and item[0] in ['"', "'"]:
+ if isinstance(item, (str, unicode)) and item[0] in ['"', "'"]:
             item = item[1:-1]
         elif isinstance(item, (float, int)):
             pass

[Adam Atlas (adam-atlas)]

Actually, instead of `(str, unicode)`, you can just use `basestring`. That's what it's for.

And it might be a few milliseconds faster. :)