PyUnicode_Substring() now accepts end bigger than string length

Fix also a bug: call PyUnicode_READY() before reading string length.
1 file changed