+ low = (parser->encoding == YAML_UTF16LE_ENCODING ? 0 : 1);
+ high = (parser->encoding == YAML_UTF16LE_ENCODING ? 1 : 0);
+
+ /*
+ * The UTF-16 encoding is not as simple as one might
+ * naively think. Check RFC 2781
+ * (http://www.ietf.org/rfc/rfc2781.txt).
+ *
+ * Normally, two subsequent bytes describe a Unicode
+ * character. However a special technique (called a
+ * surrogate pair) is used for specifying character
+ * values larger than 0xFFFF.
+ *
+ * A surrogate pair consists of two pseudo-characters:
+ * high surrogate area (0xD800-0xDBFF)
+ * low surrogate area (0xDC00-0xDFFF)
+ *
+ * The following formulas are used for decoding
+ * and encoding characters using surrogate pairs:
+ *
+ * U = U' + 0x10000 (0x01 00 00 <= U <= 0x10 FF FF)
+ * U' = yyyyyyyyyyxxxxxxxxxx (0 <= U' <= 0x0F FF FF)
+ * W1 = 110110yyyyyyyyyy
+ * W2 = 110111xxxxxxxxxx
+ *
+ * where U is the character value, W1 is the high surrogate
+ * area, W2 is the low surrogate area.
+ */
+
+ /* Check for incomplete UTF-16 character. */