[krbdev.mit.edu #6086] dump wrong-endian krb5 gssapi mech support

[Ken Raeburn via RT]

The krb5 mechanism code in our gssapi library has support for an acceptor to interact with an 
initiator that uses big-endian encoding for some integer values, contrary to the spec (RFC 
1964).  It looks like this is for backwards compatibility with versions of the mechanism 
implemented by OpenVision prior to the 1.0 beta 4 release.  (From the logs in subversion, the 
update to support both byte orders was checked in before beta4.)

I think it's time to delete that code.  Getting rid of the conditionals will clean up the 
accept_sec_context, wrap and unwrap code.