Enhancement / Question: How 4-byte utf-16 characters are handled in VersionInfo

[maxzhenzhera]

Given

1. Strings in VersionInfo have utf-16-le encoding
2. To parse a string in VersionInfo get_string_u_at_rva used
   

   pefile/pefile.py

   Lines 6476 to 6517 in 4b3b1e2

   	def get_string_u_at_rva(self, rva, max_length=2**16, encoding=None):
   	"""Get an Unicode string located at the given address."""
   	if max_length == 0:
   	return b""
   	# If the RVA is invalid let the exception reach the callers. All
   	# call-sites of get_string_u_at_rva() will handle it.
   	data = self.get_data(rva, 2)
   	# max_length is the maximum count of 16bit characters needs to be
   	# doubled to get size in bytes
   	max_length <<= 1
   	requested = min(max_length, 256)
   	data = self.get_data(rva, requested)
   	# try to find null-termination
   	null_index = -1
   	while True:
   	null_index = data.find(b"\x00\x00", null_index + 1)
   	if null_index == -1:
   	data_length = len(data)
   	if data_length < requested or data_length == max_length:
   	null_index = len(data) >> 1
   	break
   	# Request remaining part of data limited by max_length
   	data += self.get_data(rva + data_length, max_length - data_length)
   	null_index = requested - 1
   	requested = max_length
   	elif null_index % 2 == 0:
   	null_index >>= 1
   	break
   	# convert selected part of the string to unicode
   	uchrs = struct.unpack("<{:d}H".format(null_index), data[: null_index * 2])
   	s = "".join(map(chr, uchrs))
   	if encoding:
   	return s.encode(encoding, "backslashreplace_")
   	return s.encode("utf-8", "backslashreplace_")

3. In that part where "decoding" goes we can see the handling of 2-byte chunks
   

   pefile/pefile.py

   Lines 6510 to 6512 in 4b3b1e2

   	# convert selected part of the string to unicode
   	uchrs = struct.unpack("<{:d}H".format(null_index), data[: null_index * 2])
   	s = "".join(map(chr, uchrs))

Problem

Therefore, if the VersionInfo string contains a 4-byte utf-16 character - it will not be treated properly.
It will result in 2 different forcefully casted Unicode characters.

Question

Am I wrong or do not know something?
Or it should be fixed in pefile?

I understand that frequency meeting characters taking 4-byte size might not be big.
But at the end of the day, it is not handled.