TAO_Muxed_TMS Class Reference

#include <Muxed_TMS.h>

Inheritance diagram for TAO_Muxed_TMS:

Inheritance graph
Collaboration diagram for TAO_Muxed_TMS:

Collaboration graph

List of all members.

Public Member Functions

 TAO_Muxed_TMS (TAO_Transport *transport)
virtual ~TAO_Muxed_TMS (void)
virtual CORBA::ULong request_id (void)
virtual int bind_dispatcher (CORBA::ULong request_id, TAO_Reply_Dispatcher *rh)
 Bind the dispatcher with the request id.
virtual int unbind_dispatcher (CORBA::ULong request_id)
virtual int dispatch_reply (TAO_Pluggable_Reply_Params &params)
virtual int reply_timed_out (CORBA::ULong request_id)
virtual bool idle_after_send (void)
virtual bool idle_after_reply (void)
virtual void connection_closed (void)

Private Types

< CORBA::ULong,
*, ACE_Hash< CORBA::ULong >
, ACE_Equal_To< CORBA::ULong >
, ACE_Null_Mutex

Private Member Functions

int clear_cache_i (void)

Private Attributes

 Lock to protect the state of the object.
CORBA::ULong request_id_generator_
TAO_ORB_Core *const orb_core_
 Table of <Request ID, Reply Dispatcher> pairs.

Detailed Description

Using this strategy a single connection can have multiple outstanding requests. @ Can the performance of the demuxer be made more predictable, for example, using the request id as an active demux key?
Check the OMG resolutions about bidirectional connections, it is possible that the request ids can only assume even or odd values.

Member Typedef Documentation

Constructor & Destructor Documentation



TAO_Muxed_TMS::~TAO_Muxed_TMS ( void   )  [virtual]


Member Function Documentation

CORBA::ULong TAO_Muxed_TMS::request_id ( void   )  [virtual]

Generate and return an unique request id for the current invocation.

Implements TAO_Transport_Mux_Strategy.

int TAO_Muxed_TMS::bind_dispatcher ( CORBA::ULong  request_id,
TAO_Reply_Dispatcher rh 
) [virtual]

Bind the dispatcher with the request id.

Implements TAO_Transport_Mux_Strategy.

int TAO_Muxed_TMS::unbind_dispatcher ( CORBA::ULong  request_id  )  [virtual]

Unbind the dispatcher, the client is no longer waiting for the request, for example, because the request timedout. The strategy can (must) cleanup any resources associated with the request. A later reply for that request should be ignored.

Implements TAO_Transport_Mux_Strategy.

int TAO_Muxed_TMS::dispatch_reply ( TAO_Pluggable_Reply_Params params  )  [virtual]

Dispatch the reply for <request_id>, cleanup any resources allocated for that request.

Implements TAO_Transport_Mux_Strategy.

int TAO_Muxed_TMS::reply_timed_out ( CORBA::ULong  request_id  )  [virtual]

Dispatch a reply timeout for request request_id

Implements TAO_Transport_Mux_Strategy.

bool TAO_Muxed_TMS::idle_after_send ( void   )  [virtual]

Request has been just sent, but the reply is not received. Idle the transport now. The return value indicates whether idling was successful or not.

Implements TAO_Transport_Mux_Strategy.

bool TAO_Muxed_TMS::idle_after_reply ( void   )  [virtual]

Request is sent and the reply is received. Idle the transport now. The return value indicates whether idling was successful or not.

Implements TAO_Transport_Mux_Strategy.

void TAO_Muxed_TMS::connection_closed ( void   )  [virtual]

The transport object has closed the connection, inform all Reply dispatchers and waiting strategies.

Implements TAO_Transport_Mux_Strategy.

int TAO_Muxed_TMS::clear_cache_i ( void   )  [private]

Member Data Documentation

Lock to protect the state of the object.

Used to generate a different request_id on each call to request_id().

Keep track of the orb core pointer. We need to this to create the Reply Dispatchers.

Table of <Request ID, Reply Dispatcher> pairs.

The documentation for this class was generated from the following files:

Generated on Thu Nov 27 05:10:20 2008 for TAO by  doxygen 1.5.6