split with no sep specified will not return an empty string, whereas split with sep specified will. The attached patch updates the docstring to reflect this for str.split and unicode.split.
I don't think this really addresses the issue properly... The original docstring read: +Note that not specifying sep (or using None) will cause\n\ +empty strings to be removed from the output. thus ' '.strip()\n\ +returns []. while ' '.strip(' ') returns ['', '']. (Obviously strip is wrong here, it should be split) The committed docstring reads: +If sep is not specified or is None, any whitespace string is a separator and leading and trailing whitespace is stripped before splitting. But the point is not just that leading and trailing whitespace is stripped; ''.strip() returns [] while ''.strip(' ') returns ['']. This is probably the most unexpected corner case; it seems to be better to say that empty strings are removed from the output.
