Oct. 5, 2009
2:07 a.m.
2009/10/4 INADA Naoki <songofacandy@gmail.com>:
-1 That requires keeping formatting information around in every string instance. Adding new "format_string" class avoids it. unicode("foo") <=> u"foo" format_string("foo") <=> f"foo"
This way's pros: * Many libraries can use one transition way. * Transition stage syncs to Python version. "library A uses {} and library B uses %" problem not happen in transition. * We have experience same way on unicode.
And all are outweighted by the introduction of a new string class. -- Regards, Benjamin