The annotation parser does not recognize dfdl:lengthUnits. This is problematic for bit fields that are not aligned to bytes.

However, behavior is correct when specifying the length of a string in a wide encoding, e.g., UTF-16. Need to investigate how the processor decides to read length * 2 bytes in that case.

For bit fields that are not aligned to bytes, the processor needs to make some provision for extracting the bits and computing the correct integer value. I'm not sure if anything of the sort is currently implemented.