Section 8.1 details an extended command to generate a md5 hash of a file.
Is it correct to assume that this is a hash of the file when opened in
binary mode?
As the specification recognizes the need for text and binary transfers, it
may be beneficial to have a version of the md5-hash available that operates
on text files and observes the appropriate convention for end-of-line so
that this check can also be used for text files.
It is possible that the "md5-hash-handle" variant could handle this
distinction, though it needs to be made clear that the hash is generated
based upon the method that file was opened.