NFD » ndn-cxx

@Vince says he cannot start working on this Task anytime soon.

I'll start implementation after #1918 is closed AND the design below is reviewed.

My plan is to pull NFD FaceUri type into ndn-cxx, and extend it with the following:

partial class FaceUri
{
  /** \return true if FaceUri is in canonical form, false otherwise
   *  \note Only the following scheme names are supported: udp, udp4, udp6, tcp, tcp4, tcp6, fd, unix, ether, dev. false will be returned for all other scheme names.
   */
  bool
  isCanonical() const;

  /** \brief asynchronously convert this FaceUri to the canonical form
   *  \note See isCanonical for supported scheme names. onFailure will be invoked for unsupported scheme names.
   */
  void
  startCanonize(function<FaceUri> onSuccess, function<std::string> onFailure, boost::asio::io_service& io, time::nanoseconds timeout);
};

Change-Id: I63a459c0620f9c229e690df05de1333a8dea2bdd imports nfd::ethernet and FaceUri into ndn-cxx.

ether scheme is enabled universally regardless of whether libpcap is enabled, because a library compiled without libpcap shall still be interoperable with a forwarder with libpcap.

unix scheme is enabled universally, but the constructor taking Asio UNIX socket is enabled only if Asio has this feature. Unfortunately this requires #include <boost/asio.hpp> in FaceUri header.

I'll continue with this Task after this Change is approved.

I decide to implement canonization for udp,tcp,ether only, because other schemes can't be used in faces/create command and are rarely used in face query operation.