[h323plus] H.pres status

[Mark Stripes]

Simon,

Thanks for the impressively quick turnaround - and for revealing an
edge of C++ that I hadn't stumbled on before: "not" as a reserved
keyword!

On Thu, Feb 21, 2008 at 12:32 AM, Simon Horne <s.horne at packetizer.com> wrote:
>
>  NOTE: H.460.presence is still very much *in development*. The code has been
>  slightly modified from AVD-3293 after some consultation and the h323plus
>  implementation will be resubmitted at the next ITU meeting.

Caveat noted.

I can confirm that updating to CVS head and adding h460p(res).(h|cxx)
to the linux Makefile results in successful compilation and linking.

I'm a little stuck, though, when it comes to modifying simple_h323
application to play with the concepts as the p[wt]lib stuff is all new
to me and doesn't seem to behave the way I thought it would.  Do you
have available a concrete example of an H.460 extension external to
the core h323plus library being added as a feature by an application
and then being used?

I'm grounding my investigation on Simple 1.20.0 and using the pattern
evident from your previous submission to the opal effort (the only
place that I could find relevant documentation).   In
SimpleH323Process::Initialise, having invoked LoadBaseFeatureSet(), I
then create an instance of H460_FeatureOID3 and attach the
SimpleH323EndPoint instance thus:

BOOL SimpleH323Process::Initialise(PArgList & args)
{
...
  // Load the base featureSet
  LoadBaseFeatureSet();
  // mstr
  PresenceEnable(2, "WELL_KNOWN_STRING");
  H460_FeatureOID3 * presenceFeature = new H460_FeatureOID3;
  presenceFeature->AttachEndPoint(this);
  features.LoadFeatureSet(H460_Feature::FeatureBase);
  features.AddFeature(presenceFeature);
  features.AttachEndPoint(this); // seemingly redundant
...
}

The debug trace shows that an instance is created and that the new
feature is loaded, but when the Available Features are dumped as trace
there is nothing shown and, more importantly, when RAS tries to RRQ my
GK, the FeatureSet created is empty.

  0:00.005       StripesPres            Version 1.20.0 by H323Plus on
Unix Linux (2.6.22-14-generic-i686) at 2008/2/25 15:14:30.355
  0:00.006       StripesPres    PWLib   File handle high water mark
set: 6 Thread unblock pipe
  0:00.006       StripesPres    PWLib   Created thread 0x807b120 H323 Cleaner
  0:00.006       StripesPres    PWLib   Thread high water mark set: 2
  0:00.006       StripesPres    H323    Created endpoint.
  0:00.006       StripesPres    H460    Endpoint Attached
  0:00.006       StripesPres    mstr    PresenceEnable
  0:00.007       StripesPres    OID3    Instance Created
  0:00.007       StripesPres    OID3    Presence Handler created!
  0:00.007       StripesPres    OID3    Attaching Endpoint
  0:00.007       StripesPres    H460    Endpoint Attached
  0:00.007       StripesPres    H460    Loaded OID 1.3.6.1.4.1.17090.0.3
...
  0:00.038       StripesPres    Trans   Making request: registrationRequest
  0:00.038       StripesPres    H460    Create FeatureSet Reg Request PDU
  0:00.038       StripesPres    H460    FeatureSet for Reg Request PDU
{
     replacementFeatureSet = FALSE
   }

It seems that despite adding an instance of my (well, OK, 'your')
H460_FeatureOID3 through H460_FeatureSet::AddFeature, when it comes to
processing H460_FeatureSet::Features, the presence implementation is
no longer there.

Now I'm not going to call for a code review ;-)  but a pointer to an
extension example or some documentation about how extensions can be
built and used would be a wonderful help, particularly as I move
forward and consider h323plus as a component in my application.

Many thanks,

Mark.

-- 
"I don't care to belong to a club that accepts people like me as
members" - Groucho