Inappropriate method used to read string-type field values

[chaoslawful]

com.g414.haildb.TupleStorage.loadString used getString() to convert JNA pointer to Java string like this:

        return HailDB.ib_col_get_value(tupl.tupl, index).getString(0);

The method getString() assumed that every string buffer read from InnoDB is null-ended, which holds for string values written by com.g414.haildb.TupleStorage.storeString but not for real InnoDB databases. So when I use the package to parse a InnoDB backup data file, there are always unexpected trailing characters beyond the real data.

I wonder if TupleStorage.loadString could use getByteArray() instead of getString() to read exactly the same size of string data as field length specified, like the following code snippets doing:

        try {
            String res = new String(HailDB.ib_col_get_value(tupl.tupl, index).getByteArray(0, len), "utf-8");
            return res;
        } catch (UnsupportedEncodingException e) {
            return "";
        }

Of course TupleStorage.storeString also need to be changed to stop writing a extra null if this modification is made. By doing this, this project could be used to correctly parse offline MySQL InnoDB backups instead of just as a private storage engine.