Fri Oct 05 05:58:58 2001  Balachandran  <bala@cs.wustl.edu>

        * TAO version 1.2 released.

Thu Oct  4 18:06:59 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/DynamicAny/DynCommon.cpp (insert_wchar):

          Added a missing CORBA::Any::from_wchar to the actual insertion.
          Thanks to Sunny Leung <sleung@opentext.com> for pointing this out.
          This fix closes [BUGID:1048].

Thu Oct  4 10:11:32 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tao/Transport_Cache_Manager.cpp (purge_entry_i): Added a check
	  for a null in the above method. This incidentally close
	  #1046. Thanks to Donald Acton <acton@opentext.com> for
	  reporting this.

Tue Oct  2 11:26:27 2001  Priyanka Gontla  <pgontla@ece.uci.edu>

	* orbsvcs/examples/Security/Send_File/server_key.pem: 
	* orbsvcs/examples/Security/Send_File/server_cert.pem: 
	* orbsvcs/examples/Security/Send_File/client_key.pem: 
	* orbsvcs/examples/Security/Send_File/client_cert.pem: 
	* orbsvcs/examples/Security/Send_File/cacert.pem:
	  Regenerated as the previous ones have expired.
	
Tue Oct  2 12:13:15 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tao/GIOP_Message_Base.cpp: If the read message is less than the
	  size of the GIOP message, then we make the missing_data as -1 so
	  that the message would be automatically queued. This should fix
	  the Blocking_Sync_None tests hanging once in a while on
	  guajira.cs builds.

Tue Oct  2 11:05:10 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tao/Transport.cpp: Removed a debug statement that made no
	  sense. 

Mon Oct  1 14:24:04 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* docs/Options.html: Changed the default in the
	  -ORBTransportMuxStrategy option.  We have been using Muxed for
	  sometime now and it seems to be going fine. In all probability
	  we would go for x.2 with this option. Thanks to Ossama for
	  poiting this out.

Mon Oct  1 07:49:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tao/Strategies/SHMIOP_Endpoint.i:
	* tao/Strategies/DIOP_Endpoint.i: Applied the same changes from
	  this entry "Sat Sep 29 18:07:55 2001  Ossama Othman
	  <ossama@uci.edu>"  to the above classes.

Mon Oct  1 07:10:49 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tao/IIOP_Connection_Handler.cpp:
	* tao/Strategies/UIOP_Connection_Handler.cpp:
	* tao/Strategies/SHMIOP_Connection_Handler.cpp:
	* tao/Strategies/DIOP_Connection_Handler.cpp: 
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp
	  One of the long standing race conditions that came to the fore
	  recently. The problem was that we were returning a -1 to the
	  reactor if the number of upcalls dropped to 0. This was wrong
	  for many reasons with the latest set of changes for bug 575,
	  which enhanced the concurrency on the server side. Imagine a
	  thread being unblocked by notify () dispatching a
	  request. During the process the thread could have resumed the
	  handler. Another thread in the Reactor could come around and
	  tried closing the connection. Before the thread from the Reactor 
	  finishes with the cleanup job, the thread that was dispatching
	  because of notify () could return dropping the upcall count. It
	  would also return with a -1 to the reactor. At this stage we
	  would have two threads trying to do a handle_close () which
	  could be deleterious. We now return with a 0 to the reactor
	  after we handle_close_i (). We are sure that with handle_close_i
	  () the handlers are closed and returning a 0 would be harmless.


Sun Sep 30 09:26:30 2001  Ossama Othman  <ossama@uci.edu>

	* tests/File_IO/server.cpp (main):

	  Fixed compile-time problems found in VxWorks builds.  VxWorks
	  does not define the POSIX file mode contants.

	  (svc):

	  Print the exception if one is caught and return -1.  An
	  exception should be considered an error in this case.

Sun Sep 30 09:08:59 2001  Ossama Othman  <ossama@uci.edu>

	* orbsvcs/tests/Security/MT_SSLIOP/client.dsp:
	* orbsvcs/tests/Security/MT_SSLIOP/server.dsp:

	  Added missing custom build step for the test IDL file.  Fixes
	  builds on Windows.

Sat Sep 29 18:58:06 2001  Ossama Othman  <ossama@uci.edu>

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.cpp
	  (TAO_SSLIOP_Endpoint):

	  Invalidate the object_addr until the first attempt to use it is
	  made.  Needed for the double checked locking optimization
	  described below.

	  (addr_to_string):

	  Fixed this method so that it returns the correct stringified
	  address.  Previously it was returning the stringified non-SSL
	  INET_Addr instead of the stringified SSL INET_Addr.

Sat Sep 29 18:07:55 2001  Ossama Othman  <ossama@uci.edu>

	* tao/IIOP_Endpoint.i (object_addr):

	  Use the double-checked locking optimization to remove a lock
	  from the critical path.  The lock will only be acquired/released
	  once per initialization of the IIOP_Endpoint instead of
	  once per invocation of this method.

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.i (object_addr):

	  Removed erroneous comment about deadlocking due to acquiring a
	  lock in this method.  The underlying IIOP_Endpoint and the
	  SSLIOP_Endpoint each have their own lock instance.  No deadlock
	  will occur.

	  Acquire a lock before setting the port number in the
	  SSLIOP_Endpoint.  Double checked locking is used to prevent the
	  lock from being acquired during each invocation.  Fixes a race
	  condition.  [Bug 943]

Sat Sep 29 14:06:30 2001  Ossama Othman  <ossama@uci.edu>

	* orbsvcs/tests/Security/MT_SSLIOP/MT_SSLIOP.dsw:
	* orbsvcs/tests/Security/MT_SSLIOP/client.dsp:
	* orbsvcs/tests/Security/MT_SSLIOP/server.dsp:

	  New MSVC++ workspace/project files for this test.

Fri Sep 28 15:40:26 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/ast/ast_expression.cpp:
        * TAO_IDL/be/be_constant.cpp:

          Fixed code generation for wstring constants. Thanks to
          Andreas Huggel <huggel_andreas@isoftel.com> for pointing
          out this problem.

Fri Sep 28 13:52:41 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tao/IIOP_Connection_Handler.cpp (handle_input): Fixed a unused
	  variable warning.

Fri Sep 28 12:33:17 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp:
	* tao/IIOP_Connection_Handler.cpp (handle_close): 
	* tao/Strategies/UIOP_Connection_Handler.cpp:
	* tao/Strategies/DIOP_Connection_Handler.cpp:
	* tao/Strategies/SHMIOP_Connection_Handler.cpp: Added some debug
	  statements.

Fri Sep 28 11:41:50 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tests/File_IO/server.cpp (main): Made the program create a big
	  file for use.  

	* tests/File_IO/big.txt: Removed the file from the repo. Thanks to
	  Ossama for driving some sense in me.

Fri Sep 28 10:17:24 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/Smart_Proxies/SmartProxies_Static.dsp:

          Fixed a setting in the release configuration.

Fri Sep 28 06:17:27 2001  Douglas C. Schmidt  <schmidt@macarena.cs.wustl.edu>

	* ACE-INSTALL.html: Improved the installation instructions for BCB 4/5.
	  Thanks to Chris Kohloff for contributing this and to Jan Kalin
 	  <jan.kalin@zag.si> for motivating this.
	
Fri Sep 28 06:40:37 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp: Fixed a
	  cut and paste error :(.  This should clear of many of the
	  builds. 

Thu Sep 27 22:26:23 2001  Ossama Othman  <ossama@uci.edu>

	* docs/interceptors.html:

	  The Portable Interceptors specification is now part of the CORBA
	  2.5 specification.  Updated accordingly.

Thu Sep 27 17:31:10 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tao/IIOP_Connection_Handler.cpp (handle_close): 
	* tao/IIOP_Connection_Handler.h: We always depended on the
	  handle_close () to cleanup memory, the cache and the
	  transport. The cleanup would not start if the number of upcalls
	  is greater than zero. This could create problems for multi-threaded
	  cases like this one

	  - One of the threads is woken up by the Reactor to handle
            upcalls. 

	  - After the upcall is dispatched another thread could be woken
	    up to close the connection. The second thread would call
	    handle_close () religiously, but cleanup wouldnt have
	    occurred. 
	  
	  We need to start cleanup only after the upcall count reaches
	  0. We now have a method handle_close_i () which does the cleanup
	  activity. This method is called whenever there is a possibility
	  of the upcall count reaching zero. 

	* tao/Strategies/UIOP_Connection_Handler.h:
	* tao/Strategies/UIOP_Connection_Handler.cpp:
	* tao/Strategies/DIOP_Connection_Handler.h:
	* tao/Strategies/DIOP_Connection_Handler.cpp:
	* tao/Strategies/SHMIOP_Connection_Handler.h:
	* tao/Strategies/SHMIOP_Connection_Handler.cpp: 
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.h:
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp:
	  Applied the same fix to the above protocols too.
	
Thu Sep 27 17:12:06 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tests/File_IO/client.cpp: 
	* tests/File_IO/server.cpp: scaled this program down as it was
	  creating problems  with our daily builds on many
	  platforms. Retained the essence of the prgoram though. We now run
	  only 2 threads on the server end 4 threads on the client
	  end. We now request only 128 K of data (ie. twice the normal
	  socket buffer size). Hopefully this will go well on all
	  platforms. 

Thu Sep 27 16:23:34 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/Smart_Proxies/SmartProxies_Static.dsp:

          New project file.

        * TAOACE_Static.dsw:

          Added new project file to workspace. Thanks to Nicolas Vincent
          <Vincent.Nicolas@Radiometer.dk> for sending in the files.

Thu Sep 27 13:11:56 2001  Bala  <bala@cs.wustl.edu>

        * TAO version 1.1.21 released.

Thu Sep 27 02:41:32 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

	* orbsvcs/tests/Simple_Naming/client.cpp: Improved debug
	  statements.  Also replace the %8.8x and ACE_OS::thr_self combo
	  with %t.

Wed Sep 26 15:30:31 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

	* performance-tests/RTCorba/Thread_Pool/client.cpp: Allow the user
	  to set the invocation thread priority at thread creating time or
	  after thread creation.  Also improved parse_args().

Tue Sep 25 23:45:30 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

	* performance-tests/RTCorba/Thread_Pool/client.cpp: 

	  - Include tests/RTCORBA/Linear_Priority/readers.cpp to reuse
	    common file parsing code.  This way we can removed the
	    get_rates() function.

	  - Invocation priorities can either be automatically
	    calculated (as before) or can be explicitly specified
	    through the <invocation_priorities_file> option.

	* performance-tests/RTCorba/Thread_Pool/server.cpp: 

	  - Include tests/RTCORBA/Linear_Priority/readers.cpp to reuse
	    common file parsing code.  Auto lane priorities and bands
	    calculations were moved to readers.cpp.

          - Lane priorities and bands can either be automatically
	    calculated (as before) or can be explicitly specified
	    through the <bands_file> and <lanes_file> options.

	* performance-tests/RTCorba/Thread_Pool/empty_file
	* performance-tests/RTCorba/Thread_Pool/invocation_priorities 
	* performance-tests/RTCorba/Thread_Pool/lanes 
	* performance-tests/RTCorba/Thread_Pool/bands

	  New files.
	
	* tests/RTCORBA/Linear_Priority/client.cpp (main): Renamed
	  Priorities to Short_Array.

	* tests/RTCORBA/Linear_Priority/readers.cpp: Generalized some
	  methods.

Tue Sep 25 17:40:08 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

	* tests/AMI_Buffering/client.cpp (run_buffer_size): 
	* tests/Oneway_Buffering/client.cpp: Changed all the ACE_ERROR
	  statements to ACE_DEBUG statements. Changed the ERROR in the
	  debug messages as DEBUG. We know that this test shows a problem
	  with the ORB/Reactor. The error statements create too much of
	  noise for the builds czars todo any good work. The bug #982 and
	  #1031 that has been captured in the bugzilla talks about the
	  problem.  

Tue Sep 25 11:32:39 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

	* tao/Strategies/advanced_resource.cpp (TAO_Resource_Factory_Changer): 

	  Last night I removed:

	  ACE_STATIC_SVC_REQUIRE(TAO_Advanced_Resource_Factory)
	  
	  but forgot to add the replacement:

	  ACE_Service_Config::process_directive (ace_svc_desc_TAO_Advanced_Resource_Factory);

	  This caused all apps using the Advanced_Resource_Factory to
	  fail.

Tue Sep 25 00:17:57 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

	* auto_run_tests.lst: The following tests were removed from the
	  single threaded run list since they call
	  RT_Protocols_Hooks::get_thread_priority is not supported on
	  single threaded platforms:

	  - tests/RTCORBA/Client_Protocol
	  - tests/RTCORBA/Explicit_Binding
	  - tests/RTCORBA/Private_Connection

Mon Sep 24 19:23:29 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

	* tao/TAO_Internal.cpp (open_services_i):
	* tao/TypeCodeFactory/TypeCodeFactory_Adapter_Impl.cpp (Initializer): 
	* tao/TypeCodeFactory/TypeCodeFactory_Loader.cpp (Initializer): 
	* tao/Strategies/advanced_resource.cpp (TAO_Resource_Factory_Changer): 
	* tao/RTPortableServer/RTPortableServer.cpp (TAO_RTPortableServer_Initializer): 
	* tao/RTCORBA/RTCORBA.cpp (TAO_RTCORBA_Initializer): 
	* tao/PortableServer/PortableServer.cpp (init): 
	* tao/IORTable/IORTable.cpp (init): 
	* tao/IORManipulation/IORManip_Loader.cpp (Initializer): 
	* tao/IFR_Client/IFR_Client_Adapter_Impl.cpp (Initializer): 
	* tao/DynamicInterface/Dynamic_Adapter_Impl.cpp (Initializer): 
	* tao/DynamicAny/DynamicAny.cpp (Initializer): 
	* tao/BiDir_GIOP/BiDirGIOP.cpp (Initializer): 

          - Changed 

            ACE_Service_Config::static_svcs ()->insert (&blah) 

            to

            ACE_Service_Config::process_directive (blah) 

            The reason for this change is that when the ORB is loaded
            into a program as a DLL through the Service Configurator,
            Service_Config::open has already been run.  This prevents
            the static services in TAO to be loaded into the Service
            Repository since Service_Config::open only executes once.
            However, Service_Config::process_directive works
            irrespective of whether Service_Config::open has already
            being run.

	* tao/Strategies/TAO_Strategies_Internal.cpp: 
	* tao/Strategies/TAO_Strategies_Internal.h: 

	  Removed unused files.
	
	* tao/Strategies/Makefile.bor: 
	* tao/Strategies/Makefile: 
	* tao/Strategies/TAO_Strategies.dsp: 
	* tao/Strategies/TAO_Strategies_Static.dsp: 

	  Removed TAO_Strategies_Internal.

	* tao/Strategies/advanced_resource.h:
	  "ACE_STATIC_SVC_REQUIRE(TAO_Advanced_Resource_Factory)" removed
	  as it was not needed.

	* tao/default_resource.cpp (get_parser_names): Removed unnecessary
	  lookup of parsers. 

	* tao/orbconf.h: Removed default defines for
	  TAO_DEFAULT_RESOURCE_FACTORY_ARGS,
	  TAO_ADVANCED_RESOURCE_FACTORY_ARGS,
	  TAO_DEFAULT_CLIENT_STRATEGY_FACTORY_ARGS, and
	  TAO_DEFAULT_SERVER_STRATEGY_FACTORY_ARGS since they are we
	  horribly out of date and the defaults work fine.  User can still
	  define them as needed.

	* tao/TAO_Internal.cpp: 

	  - Since TAO_DEFAULT_RESOURCE_FACTORY_ARGS,
	    TAO_DEFAULT_CLIENT_STRATEGY_FACTORY_ARGS, and
	    TAO_DEFAULT_SERVER_STRATEGY_FACTORY_ARGS are no longer
	    defined by default, the initialization of the
	    corresponding statics in TAO_Internal had to be modified.

          - For consistency, <server_strategy_args_> and
            <client_strategy_args_> got renamed to
            <server_strategy_factory_args_> and
            <client_strategy_factory_args_> respectively.

	* tao/ORB_Core.cpp: Fixed some incorrect comments.

	* tao/MCAST_Parser.h: Removed extra OS.h include.

	* examples/Simple/time-date/Time_Date.cpp: Removed extra
	  TAO_Singleton.h include.

Mon Sep 24 14:12:11 2001  Priyanka Gontla  <pgontla@ece.uci.edu>

	* tao/MCAST_Parser.cpp (assign_to_variables):
	  Modified so that if the nic address isnt specified in the mcast
	  format, the kernel chooses the nic. This is because the nic is
	  platform specific and we cant default it to eth0.

	  Thanks to Ossama for pointing it out.

	* orbsvcs/orbsvcs/IOR_Multicast.cpp (handle_input): Added
	  MCASTServer in the list of multicast services that are valid. This
	  is needed for the test (orbsvcs/tests/IOR_MCAST).

Mon Sep 24 12:22:32 2001  Priyanka Gontla  <pgontla@ece.uci.edu>

	* tao/params.cpp:
	  Made mcast://:::: as the default multicast address. This address
	  will default to
	  mcast://ACE_DEFAULT_MULTICAST_ADDR:
	          TAO_DEFAULT_NAME_SERVER_REQUEST_PORT:eth0:1.

	  With this change, to get the previous multicast behaviour, the user
	  doesnt have to use any extra variables.

	* orbsvcs/tests/Simple_Naming/run_test.pl:
	  Reverted the change to run_test.pl

Mon Sep 24 12:17:28 2001  Ossama Othman  <ossama@uci.edu>

	* tao/PICurrent.cpp (~TAO_PICurrent_Impl, get_slot, set_slot):

	  Use C-style casts to get around Sun CC 4.2 static_cast brain
	  damage.  Sun CC 4.2 could not handle statically casting from a
	  "void *&" to a "CORBA::Any *".  Thanks to Bala for the tip.

	  Note that we could get around this problem by making the slot
	  table contain CORBA::Any pointers instead void pointers, but
	  that would increase footprint.

Mon Sep 24 11:47:34 2001  Ossama Othman  <ossama@uci.edu>

	* tao/ORB.cpp:

	  Fixed problem where ORBInitializer variables were not declared
	  in builds where CORBA messaging was disabled.

	* tao/PICurrent.cpp:

	  Include "tao/debug.h" to pull in the declaration of the
	  TAO_debug_level variable.

	* tao/PICurrent_ORBInitializer.h:

	  Moved `LocalObject.h' include outside of the
	  TAO_HAS_INTERCEPTORS block to pull in TAO-specific configuration
	  headers.  TAO_HAS_INTERCEPTORS was previously not pulled in.

Mon Sep 24 11:03:17 2001  Ossama Othman  <ossama@uci.edu>

	* tao/PICurrent_ORBInitializer.h:
	* tao/PICurrent_ORBInitializer.cpp:

	  Only make code "visible" if the TAO_HAS_INTERCEPTORS
	  preprocessor macro is set to "1".  Fixes build problems in
	  configurations where request interceptors are disabled.

	* tao/PICurrent.h:
	* tao/PICurrent.cpp:

	  "Nuked" trailing whitespace.

	* tao/Makefile:
	* tao/*/Makefile:

	  Updated dependencies.

Mon Sep 24 01:21:16 2001  Ossama Othman  <ossama@uci.edu>

	* tao/PICurrent.h:
	* tao/PICurrent.inl:
	* tao/PICurrent.cpp:

	  Implementation of the "PICurrent" object, i.e. the
	  PortableInterceptor::Current interface.

	* tao/PICurrent_ORBInitializer.h:
	* tao/PICurrent_ORBInitializer.cpp:

	  ORBInitializer that registers the PICurrent object with each
	  ORB.

	* tao/Makefile:
	* tao/Makefile.bor:
	* tao/TAO.dsp:
	* tao/TAO_Static.dsp:

	  Added new PICurrent source files to these Makefiles and project
	  files.

	* tao/ClientInterceptorAdapter.h:
	* tao/PortableServer/ServerInterceptorAdapter.h:

	  Updated list of authors.

	* tao/ClientRequestInfo.cpp (get_slot):
	* tao/ORBInitInfo.cpp (allocated_slot_id):

	  Implemented these methods.

	* tao/ORB.cpp (init_orb_globals):

	  Register the PICurrent ORB initializer.

	* tao/ORB_Core.h:

	  Added TSS portion of the PICurrent object to the ORB Core TSS
	  resources object.

	  Cache the PICurrent object reference for speed/efficiency
	  reasons.

	* tao/ORB_Core.i (pi_current):

	  New methods to set and get the cached thread-scope PICurrent
	  object reference.

	* tao/ORB_Core.cpp (TAO_ORB_Core):

	  Initialize the pointer to the PICurrent object reference to
	  zero.

	* tao/PolicyFactory_Registry.cpp (~TAO_PolicyFactory_Registry):

	  Close the factory map for the sake of completeness.

	* tao/TAO_Server_Request.h (TAO_ServerRequest):

	  The "request scope" PICurrent object implementation for the
	  server side is now held in this class.

	* tao/TAO_Server_Request.i (rs_pi_current):

	  Accessor for the request scope PICurrent object implementation.

	* tao/TAO_Server_Request.cpp (TAO_ServerRequest):

	  Added request scope PICurrent attribute to the base member
	  initializer list.

	* tao/PortableServer/ServerInterceptorAdapter.cpp:

	  Instantiate a "PICurrent" guard before invoking the
	  receive_request_service_contexts() interception point.  Upon
	  leaving the scope, the request PICurrent will be logically
	  copied to the thread scope PICurrent.

	* tao/PortableServer/ServerRequestInfo.h:
	* tao/PortableServer/ServerRequestInfo.inl (server_request):

	  Added accessor to return a reference to the cached
	  TAO_ServerRequest object.

	* tao/PortableServer/ServerRequestInfo.cpp (get_server_policy):

	  Implemented this method.  However, see comments in the
	  implementation for important details.

	  (get_slot, set_slot):

	  Also implemented these methods.  They are used to get and set
	  data in the "request scope" "PICurrent" object.

	* TAO_IDL/be/be_codegen.cpp (start_server_skeletons):

	  Generate code to include the `tao/PICurrent.h' header.

	* TAO_IDL/be/be_visitor_operation/operation_ss.cpp
	  (visit_operation):

	  Generate code that instantiates a "TAO_PICurrent_Guard" before
	  the code that invokes the receive_request() interception point
	  and the upcall.  The thread scope PICurrent must be copied to
	  the request scope PICurrent object when the upcall is done.

	* tests/Portable_Interceptors/PICurrent/ServerORBInitializer.cpp:
	* tests/Portable_Interceptors/PICurrent/ServerORBInitializer.h:
	* tests/Portable_Interceptors/PICurrent/ServerRequestInterceptor.cpp:
	* tests/Portable_Interceptors/PICurrent/ServerRequestInterceptor.h:
	* tests/Portable_Interceptors/PICurrent/client.cpp:
	* tests/Portable_Interceptors/PICurrent/server.cpp:
	* tests/Portable_Interceptors/PICurrent/test.idl:
	* tests/Portable_Interceptors/PICurrent/test_i.cpp:
	* tests/Portable_Interceptors/PICurrent/test_i.h:
	* tests/Portable_Interceptors/PICurrent/Makefile:
	* tests/Portable_Interceptors/PICurrent/Makefile.bor:
	* tests/Portable_Interceptors/PICurrent/client.bor:
	* tests/Portable_Interceptors/PICurrent/server.bor:
	* tests/Portable_Interceptors/PICurrent/PICurrent.dsw:
	* tests/Portable_Interceptors/PICurrent/client.dsp:
	* tests/Portable_Interceptors/PICurrent/server.dsp:

	  Test that verifies PICurrent support is working properly.

	* tests/Portable_Interceptors/Makefile:
	* tests/Portable_Interceptors/Makefile.bor:

	  Added PICurrent test directory to the build list.

	* docs/interceptors.html:

	  Updated PortableInterceptor release notes.

Sun Sep 23 20:34:23 2001  Ossama Othman  <ossama@uci.edu>

	* tao/Typecode.cpp (BadKind, Bounds):

	  Fixed bugs where the RepositoryId for these TypeCode user
	  exceptions contained the incorrect prefix, i.e. "omg.orb"
	  instead of "omg.org."

Sun Sep 23 20:11:45 2001  Christopher Kohlhoff  <chris@kohlhoff.com>

        * tests/Collocation/Collocation.bor:
        * tests/Collocation/Diamond.bor:
          Updated Borland makefiles to find the example's libraries in the
          common bin directory, since all libraries are now built into there.

Sun Sep 23 15:01:08 2001  Ossama Othman  <ossama@uci.edu>

	* tao/Invocation.cpp (location_forward):

	  Use the "CORBA::Object_var::out()" method instead of the
	  "inout()" method to make sure the previous forward object
	  reference is released before assigning another object reference
	  Thanks to Gergely Timar <timarok@elender.hu> for the fix.

Sun Sep 23 14:13:35 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/fe/idl.ll:
        * TAO_IDL/fe/lex.yy.cpp:

          Modified parse_line_and_file() to make it more robust
          when parsing #<line number>. Thanks to Carlos for
          tracking down the solution.

Sat Sep 22 11:47:48 2001  Ossama Othman  <ossama@uci.edu>

	* tao/Profile.cpp (policies):

	  Use the lame "ignore check_for_ace_check" comment hack to
	  silence fuzz.

Fri Sep 21 22:31:33 2001  Ossama Othman  <ossama@uci.edu>

	* tao/default_resource.cpp (get_parser_names):

	  Fixed bug where the "corbaname" parser's name was not properly
	  inserted into the list of names of registered parsers.  Instead
	  the "corbaloc" IOR's parser name was inserted.  "corbaname" IORs
	  should now be supported by default, as was originally designed.

Fri Sep 21 22:23:11 2001  Ossama Othman  <ossama@uci.edu>

	* tao/Profile.cpp (policies):

	  Fixed "unused parameter" warning in the minimum CORBA builds.

Fri Sep 21 21:39:22 2001  Ossama Othman  <ossama@uci.edu>

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Util.h:
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Util.cpp:

	  New SSLIOP utility class containing code common to both the
	  SSLIOP client side and the server side.

	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Acceptor.h:
	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connector.h:
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Acceptor.h:
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.h:

	  Added new attributed that contains the state to be passed to
	  each new TAO_SSLIOP_Connection_Handler.

	  Doxygenated this header.

	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Acceptor.cpp (open_i):
	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connector.cpp (open):
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Acceptor.cpp (ssliop_open_i):
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.cpp (open):

	  Invoke the new TAO_SSLIOP_Util::setup_handler_state() method,
	  and pass the created state to the creation strategy.

	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Transport.h:

	  Cache a reference to the SSLIOP::Current object to avoid any
	  reference resolution overhead in the critical path.

	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Transport.cpp (handle_input_i):

	  Pass cached SSLIOP::Current reference to the
	  TAO_Null_SSL_State_Guard.  Saves an
	  ORB::resolve_initial_references() in th critical path.

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Current.h:

	  Generated and added "TAO_SSLIOP_Current" downcast and reference
	  count related methods.  For example, it is now possible to
	  _narrow() a SSLIOP::Current object reference to
	  TAO_SSLIOP_Current object reference.  This makes it possible to
	  gain access to the methods in TAO_SSLIOP_Current not defined in
	  the SSLIOP::Current IDL without relying on dynamic_cast, which
	  may not work properly when using compilers that do not support
	  RTTI.

	  (orb_id_):

	  No longer any need for this attribute.  (See below.)

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Current.cpp (TAO_SSLIOP_Current):

	  This constructor now accepts a pointer to the ORB Core.  This
	  obviates the need to cache the ORBid.

	  (init):

	  Removed this method.  It is no longer needed since a pointer to
	  the ORB Core is now cached in the TAO_SSLIOP_Current instance at
	  ORB bootstrap time.

	  (setup, teardown):

	  These methods now accept parameters that contain the state that
	  will be used when setting up the TSS portion of the
	  SSLIOP::Current object.  This makes these mades thread-safe and
	  reentrant without introducing any locks.  [Bug 943]

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Current.inl (implementation):

	  Don't call init().  That method has been removed.

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_ORBInitializer.cpp (pre_init):

	  Retrieve a pointer to the ORB Core from the TAO_ORBInitInfo
	  object and pass it to the TAO_SSLIOP_Current constructor when
	  creating the TAO_SSLIOP_Current object.

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.h:
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp
	  (TAO_SSLIOP_Connection_Handler_State):

	  New class that contains state that must be passed to a
	  TAO_SSLIOP_Connection_Handler up on creation.

	  (TAO_SSL_State_Guard):

	  No need to export this class.  It is only used internally by the
	  TAO_SSLIOP library.

	  This class now stores additional state.  Specifically, it
	  contains a pointer to the "previous" SSLIOP::Current
	  implementation that was stored in TSS, the current
	  SSLIOP::Current implementation, and a flag that specifies
	  whether or not the TSS portion of SSLIOP::Current was setup.
	  This state is passed to the connection handler each time a
	  request is handled.  Since this class is instantiated each time
	  a request is handled, the encapsulated state is
	  request-specific, meaning that SSL state setup in TSS is
	  reentrant and thread-safe.  [Bug 943]

	  No longer any need to cache a pointer to the ORB Core.  Adjusted
	  the constructor signature and removed the attribute
	  accordingly.

	  (TAO_SSLIOP_Connection_Handler):

	  This connection handler no longer retains an instance of the
	  TAO_SSLIOP_Current_Impl object.  That object was subject to a
	  race condition.  [Bug 943]

	  (current):

	  New methods to get and set the cached TAO_SSLIOP_Current
	  reference.  There is no longer any need to perform a
	  dynamic_cast on the SSLIOP::Current object reference to gain
	  access to TAO_SSLIOP_Current-specific methods.

	  (setup_ssl_state):

	  It is no longer necessary to retrieve the SSLIOP::Current object
	  reference during the first attempt to set up the TSS portion of
	  the SSLIOP::Current object.  That object reference is now
	  retrieved and cached at ORB bootstrap time.  (See the
	  description of the changes to SSLIOP_ORBInitializer.cpp above
	  for details.)

	  (setup_ssl_state, teardown_ssl_state):

	  These methods now accept parameters that contain the state that
	  will be used when setting up the TSS portion of the
	  SSLIOP::Current object.  This makes these mades thread-safe and
	  reentrant without introducing any locks.  [Bug 943]

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.i:

	  Updated the calls to the setup/teardown_ssl_state() in the
	  TAO_SSL_State_Guard constructor and destructor to match the new
	  method signatures.

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Transport.cpp (handle_input_i):

	  The signature of the TAO_SSL_State_Guard constructor changed.
	  Adjusted the instantiation accordingly.

	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connection_Handler.h:
	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connection_Handler.cpp:

	  Changes analogous to those made in the SSLIOP_Connection_Handler
	  counterparts.  See above for details.

	* orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connection_Handler.inl:

	  Inlined the TAO_Null_SSL_State_Guard constructor and destructor
	  for speed/efficiency reasons.

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Current_Impl.h:

	  Corrected destructor comment.

	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Current_Impl.inl:
	* orbsvcs/orbsvcs/SSLIOP/SSLIOP_Current_Impl.cpp:

	  Inlined the TAO_SSLIOP_Current_Impl constructor for
	  speed/efficiency reasons, i.e. since it is instantiated in the
	  critical path.

	* orbsvcs/orbsvcs/Makefile.SSLIOP (FILES):
	* orbsvcs/orbsvcs/SSLIOP.bor (OBJFILES):

	  Added new SSLIOP_Util sources to the list of files to	compile
	  and link.

Fri Sep 21 13:01:53 2001  Carlos O'Ryan  <coryan@uci.edu>

        * docs/tutorials/Quoter/Event_Service/Makefile:
        * docs/tutorials/Quoter/Naming_Service/Makefile:
        * docs/tutorials/Quoter/RT_Event_Service/Makefile:
          Fixed dependencies on the tutorial.

Fri Sep 21 12:36:36 2001  Ossama Othman  <ossama@uci.edu>

        * tao/Profile.cpp (policies):

	  Added missing CORBA::Environment parameter to the call to
	  CORBA::Policy::policy_type().  Removes a TSS access during each
	  policy insertion into the TAG_POLICIES IOR tagged component.
	  Also fixes a potential problem where an emulated exception would
	  not be caught if thrown during a remote
	  CORBA::Policy::policy_type() invocation.

	* tao/ClientInterceptorAdapter.cpp (receive_exception):
	* tao/PortableServer/ServerInterceptorAdapter.cpp (send_exception):

	  Only re-throw the exception if it hasn't been transformed
	  (e.g. to a LOCATION_FORWARD) by the receive_exception() and
	  send_exception() interception points, respectively.

	* tao/PortableServer/POA.cpp (client_exposed_policies):

	  Improved exception-safety of this method by storing the
	  allocated CORBA::PolicyList in a CORBA::PolicyList_var.  Fixes a
	  memory leak that occurred if an exception was thrown within this
	  method.

	* TAO_IDL/be/be_visitor_operation/remote_proxy_impl_cs.cpp
	  (gen_marshal_and_invoke):
	* TAO_IDL/be/be_visitor_operation/operation_ss.cpp
	  (visit_operation):

	  Do not call the generated instantiation of the RequestInfo
	  object "ri."  Doing so will cause conflicts with IDL that
	  uses an operation parameter of the same name.  The generated
	  code now uses "_tao_ri" to avoid namespace pollution.

	* tests/Portable_Interceptors/Request_Interceptor_Flow/Request_Interceptor_Flow.dsw:
	* tests/Portable_Interceptors/Request_Interceptor_Flow/client.dsp:
	* tests/Portable_Interceptors/Request_Interceptor_Flow/server.dsp:

          Corrected spelling.  "Inteceptor" --> "Interceptor"

Thu Sep 20 15:48:33 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * performance-tests/RTCorba/Thread_Pool/client.cpp: Added ability
          to adjust timeout for max throughput run.

Thu Sep 20 13:23:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/BiDirectional_NestedUpcall/client.cpp:
        * tests/BiDirectional_NestedUpcall/server.cpp: Created the servant
          object on the heap instead of on the stack.

        * tests/BiDirectional_NestedUpcall/test_i.h: Inherited the
          implementation from the Refcounted servant base.

          The above changes should prevent the client from crashing if the
          server crashes during the upcall.

Thu Sep 20 13:23:59 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * orbsvcs/tests/IOR_MCast/server.bor: Added missing iortable
          library.

        * orbsvcs/tests/IOR_MCast/MCast_Server_i.cpp (connect_server):
          Fixed missing throw spec.

          Thanks to Johnny Willemsen <jwillemsen@remedy.nl> for the above
          fixes.

Thu Sep 20 12:40:34 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * performance-tests/RTCorba/Thread_Pool/client.cpp:

          - Previously, each thread kept it's own start time.  This
            meant that if a low priority thread did not reach the
            point where it recorded it's start time before it got
            preempted by a thread of higher priority, then it is
            possible that this low priority thread may start after the
            high priority thread is done and hence would not notice
            that it is late at all.  Therefore, the start time is made
            global and is the same for all threads.

          - We now keep track of the invocations that were missed
            because we didn't make the deadline.  This will help note
            any particularities in the test.

          - Some of the threads were being given an unfair advantage
            because a connection had already been setup for them
            before they started pacing.  Therefore, all threads (paced
            and continuous) now do all the accounting/connection setup
            and then synchronize with the main thread.  When all the
            worker threads are ready to go, the main thread signals
            them.  This also gives the main thread a chance to mark
            the start time for the test.

Wed Sep 19 17:43:42 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * tests/RTCORBA/Linear_Priority/server.cpp:
        * tests/RTCORBA/Persistent_IOR/server.cpp:
        * tests/RTCORBA/Policy_Combinations/server.cpp:

          Added missing ACE_UNUSED_ARGs for asserted <result>.

        * tao/TAO_Internal.cpp: Added explicit template instantiations for
          ACE_Dynamic_Service<ACE_Service_Object>.

        * tests/Exposed_Policies/Policy_Tester.cpp (init): Changed
          ACE_CHECK to ACE_CHECK_RETURN.

        * tests/RTCORBA/Collocation/Collocation.cpp
          (create_poa_and_servant_with_tp_policy and
          create_servant_in_root_poa): Fixed misplaced ACE_CHECKs.

Wed Sep 19 14:15:59 2001  Nanbor Wang  <nanbor@cs.wustl.edu>

        * tao/TAO.dsp:
        * tao/TAO_Static.dsp:
        * tao/BiDir_GIOP/TAO_BiDir_GIOP.dsp:
        * tao/DynamicAny/TAO_DynamicAny.dsp:
        * tao/IFR_Client/TAO_IFR_Client.dsp:
        * tao/IFR_Client/TAO_IFR_Client_Static.dsp:
        * tao/IORManipulation/TAO_IORManip.dsp:
        * tao/IORManipulation/TAO_IORManip_Static.dsp:
        * tao/IORTable/TAO_IORTable.dsp:
        * tao/IORTable/TAO_IORTable_Static.dsp:
        * tao/PortableServer/TAO_PortableServer.dsp:
        * tao/PortableServer/TAO_PortableServer_Static.dsp:
        * tao/RTCORBA/TAO_RTCORBA.dsp:
        * tao/RTCORBA/TAO_RTCORBA_Static.dsp:
        * tao/RTPortableServer/TAO_RTPortableServer.dsp:
        * tao/RTPortableServer/TAO_RTPortableServer_Static.dsp:
        * tao/SmartProxies/SmartProxies.dsp:
        * tao/Strategies/TAO_Strategies.dsp:
        * tao/Strategies/TAO_Strategies_Static.dsp:
        * tao/TypeCodeFactory/TypeCodeFactory.dsp: Enabled generation of
          debug info even for release configurations.  Thanks to Jonathan
          Reis <reis@stentor.com> for suggesting this.

Wed Sep 19 12:06:04 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * examples/Simple/time-date/time-date.bor:
        * orbsvcs/tests/Concurrency/CC_client.bor:
        * orbsvcs/tests/Time/client.bor:
        * orbsvcs/tests/ior_corbaname/client.bor:
        * orbsvcs/tests/ior_corbaname/server.bor:
          Corrected the SVCUTILS to SVC_UTILS

        * orbsvcs/tests/IOR_MCast
          Added BCB makefiles

        * orbsvcs/tests/Makefile.bor
          Added IOR_MCast and missing LoadBalancing

Wed Sep 19 02:58:46 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * orbsvcs/orbsvcs/Makefile.av (.obj/Transport.o): Fixed minor
          error in the dependencies (though I am not sure how this error
          got in since the dependencies are auto generated).

Tue Sep 18 17:18:35 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * performance-tests/RTCorba/Thread_Pool/client.cpp: Added priority
          information to max throughput testing.  Also,
          <Continuous_Worker::time_for_test_> has the max time of all the
          continuous workers.

Tue Sep 18 15:55:18 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * tests/RTCORBA/README: Updated with information about new tests.

        * tests/TAO_Tests.dsw: Added Cache_Growth_Test and
          Exposed_Policies.

        * tests/Cache_Growth_Test/Hello.cpp (get_string): transport_cache
          is now in the lane resources.

Tue Sep 18 13:05:09 2001  Carlos O'Ryan  <coryan@uci.edu>

        * */Makefile:
        * */Makefile.*:
          Updated dependencies, some of the paths were absolute leading to
          ambiguities in parallel builds.

Tue Sep 18 12:07:40 2001  Carlos O'Ryan  <coryan@uci.edu>

        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPullSupplier.h:
          Missed one file in the last set of changes.

Tue Sep 18 10:44:58 2001  Carlos O'Ryan  <coryan@uci.edu>

        * orbsvcs/orbsvcs/CosEvent/CEC_ConsumerAdmin.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ConsumerControl.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Default_Factory.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Defaults.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Dispatching.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Dispatching_Task.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_EventChannel.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Factory.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_MT_Dispatching.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPullConsumer.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPullSupplier.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPushConsumer.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPushSupplier.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Pulling_Strategy.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Reactive_ConsumerControl.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Reactive_Pulling_Strategy.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Reactive_SupplierControl.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_SupplierAdmin.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_SupplierControl.h:
          Fixed problems with the @file declarations, the automatically
          generated entries did not correct my mistakes in the old-style
          =FILE comments

Mon Sep 17 19:00:00 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * The following changes are to add support for thread pools and
          lanes to TAO.

        * tao/RTPortableServer/RT_POA.cpp:

          - (thread_pool): Added member and accessor.

          - (valid_priority): Renamed to validate_priority().  Instead
            of trying to match acceptor priority and bands, we check
            the following:

            If this POA is using a thread pool with lanes, make sure
            the priority matches one of the thread lanes.  Note that
            in this case, bands do not matter since matching the lanes
            priority is a stricter condition than meeting the band
            ranges.  In addition, when the POA was created, the bands
            had to match the lanes.

            If we are dealing with a thread pool without lanes, check
            if we have bands.  If we do have bands, make sure that the
            priority is matching one of the bands.

          - (endpoint_count): Counts the potentially relevant
            endpoints for this POA.

          - (parse_rt_policies): Extract the POA's thread pool from
            the policies specified by the user.

          - (key_to_stub_i): Instead of trying to create different
            kinds of acceptor filters, we check the following:

            If this POA is using the default thread pool or a thread
            pool without lanes, create the IOR with the acceptors in
            the thread pool.

            If this POA has the SERVER_DECLARED policy, create the IOR
            with the acceptors in the only thread lane that matches
            the priority of the object.

            If this POA has the CLIENT_PROPAGATED policy, create the
            IOR with the acceptors in the thread lanes that matches
            the bands in this POA.  If there are no bands, all the
            thread lanes are used.

        * tao/RTPortableServer/RT_Policy_Validator.cpp:

          - (validate_impl): Removed call to ORB_Core::open(). Added
            thread-pool validation.

          - (legal_policy_impl): Added <THREADPOOL_POLICY_TYPE> to the
            legal policies.

          - (validate_server_protocol): Previously, if we found one
            protocol that matched, we returned success.  Now we return
            return success only if we match all the protocols
            specified by the user.

          - (validate_lifespan): If a POA is using a RTCORBA thread
            pool, make sure the lifespan policy is not persistent
            since we cannot support it right now.

          - (merge_policies_impl): Merging of policies specified at
            the ORB level was not done before. Check if the user has
            specified the priority model, server protocol, and thread
            pool policies.  If not, check if the policy has been
            specified at the ORB level.  If so, we'll use that policy.

          - (validate_priorities): Revised this function to do the
            following tests:

            Initialize <rt_priority_model> to NOT_SPECIFIED rather
            than CLIENT_PROPAGATED.

            If priority model was not specified, then we better not
            have a thread pool with lanes since this configuration
            does not make sense.

            If priority banded connections are set, make sure that:

             0. A priority model was specified.

             1. There is at least one band.

             2a. low is not < RTCORBA::minPriority
             2b. low <= high
             2c. high is not > RTCORBA::maxPriority

             3. If priority model is SERVER_DECLARED, server_priority
                must match one of the bands.

             4. If this POA has a thread pool with lanes, then for
                each band, there must be at least one thread lane that
                can service it, i.e., whose priority falls into the
                band's range.

            If priority banded connections are not set, and the
            priority model is SERVER_DECLARED, make sure we have at
            least one thread lane that can provide service for the
            specified SERVER_DECLARED priority.

        * tao/RTPortableServer/RT_Acceptor_Filters.h

          - (fill_mprofile): Renamed fill_mprofile() to fill_profile()
            and added a <priority> paramter to this function.

          - (encode_endpoints): Also added a new function
            encode_endpoints() that encodes the endpoints in the
            profiles into the TAO_TAG_ENDPOINTS tag component of
            profiles.

          - (TAO_Priority_Acceptor_Filter; TAO_Bands_Acceptor_Filter;
            validate_acceptor): Removed: POA decides with acceptor
            registries to include in the stub.  Therefore, these
            classes and functions are not longer required.

        * tao/RTPortableServer/RT_Servant_Dispatcher.cpp:

          - (~RT_Priority_Model_Processing): Removed code duplication
            and called post_invoke() directly.

          - (pre_invoke): Changed the upcall thread preprocessing to
            do the following:

            Don't mess with the priority of threads in lanes.

            For the SERVER_DECLARED PriorityModel processing, use the
            request associated with the servant.

            Previously, for the CLIENT_PROPAGATED policy, if the
            server priority was <TAO_INVALID_PRIORITY>, then the
            client propagated priority was ignored.  This was fixed.

            Handle cases where the priority model policy was not
            specified.

            Distinguish between invalid target priorities and where
            the original == target.

        * tao/RTPortableServer/RT_Collocation_Resolver.cpp:
        * tao/RTPortableServer/RT_Collocation_Resolver.h:
        * tao/RTPortableServer/RT_Collocation_Resolver.i:

          This class decides/resolves whether an object is collocated
          with the servant.  The following check are done:

          First we do the good ol' basic check: make sure that the
          servant is in the same ORB that created this object.

          Then we lookup the thread pool used by the POA where the
          servant is located.  At the same time we figure out which
          thread pool the calling thread belongs to by looking at it's
          TSS resources.

          If the POA's thread pool and the calling thread's thread
          pool don't match, then the object is not collocated.

          If they do match and they don't have lanes, we are
          collocated.  Note that this includes the default thread
          pool.

          Then we inspect the priority model policy of the target POA.
          If it is CLIENT_PROPAGATED, then the object is not
          collocated.

          Note that the priority model policy cannot be NOT_SPECIFIED
          because NOT_SPECIFIED is not allowed with thread pool with
          lanes.

          If it is SERVER_DECLARED, then we lookup the servant's
          priority.  If that matches the current thread's priority,
          then the object is not collocated.  Otherwise, it is not
          collocated.

        * tao/RTPortableServer/RT_POA_Initializer.cpp
          (init_rt_default_policies): Removed: it was merging the ORB
          level policies with the default POA policies at initialization
          time.  It correct thing to do would be to merge the ORB level
          policies with the user provided policies when a POA is being
          created.

        * tao/RTPortableServer/RT_Object_Adapter_Factory.cpp:

          - (create): No need to setup the POA Extension Initializer.
            Setup the servant dispatcher and the policy validator on
            the Object Adapter after creating it.

          - (Constructor): Set the name of the collocation resolver to
            be RT_Collocation_Resolver and add it to the service
            configurator.

        * tao/RTPortableServer/RT_POA_Initializer.cpp:

          Renamed TAO_RT_POA_Initializer::init() to
          TAO_RTPortableServer_Initializer::TAO_RTPortableServer_Initializer().

        * tao/RTPortableServer/RT_POA.h:

          Include RTPortableServerC.h instead of RTPortableServer.h.

        * tao/RTPortableServer/Makefile:
        * tao/RTPortableServer/Makefile.bor:
        * tao/RTPortableServer/TAO_RTPortableServer.dsp:
        * tao/RTPortableServer/TAO_RTPortableServer_Static.dsp:

          Removed RT_POA_Initializer.  Added RT_Collocation_Resolver.

        * tao/RTCORBA/Thread_Pool.cpp:

          - (set_tss_resources): Each lane thread on startup sets its
            thread lane pointer in TSS.  This makes it easy to access
            each thread's lane resources and also makes it easy to
            identify which threads belong to thread pools and which
            don't.

          - (validate_and_map_priority): When a thread lane is opened,
            its lane priority is validated and mapped to native
            priority before the threads are spawned.

          - (open): Open the acceptor registry when a thread lane is
            opened.

          - (fini): Finalize resources.

          - (resources): Added TAO_Thread_Lane_Resources accessor.

          - (with_lanes): Since a thread-pool without lanes is
            implemented as a thread-pool with one lane, it is
            difficult to tell them apart. Added a <with_lanes_> flag
            to tell them apart.

          - (constructors): Make sure <allow_request_buffering> and
            <allow_borrowing> are disabled since we do not support
            them.

          - (create_dynamic_threads): In addition to the default flags
            (THR_NEW_LWP and THR_JOINABLE), add in flags to set the
            scope and scheduling policies when creating RT threads.

          - (native_priority): Each lane remembers its native priority
            in addition to its CORBA priority.

          - (destroy_threadpool_i): We now shutdown the reactors in
            the thread pool, wait for the threads in the pool to exit,
            finalize the resources in the pool, and only then delete
            the thread pool.

          - (TAO_RT_New_Leader_Generator::no_leaders_available): This
            method is called by the leader followers class when it is
            out of threads to run the event loop.

            We grab the Thread Pool Manager lock, check if the current
            number of threads in this lane does not exceed the
            <static> + <dynamic> number of threads specified by the
            user.  If it does not exceed <static> + <dynamic>, one
            dynamic thread is created.

        * tao/RTCORBA/RT_Protocols_Hooks.cpp
          (set_default_server_protocol_policy): Don't include all the
          protocols that the ORB knows about in the default server
          protocol policy; only include protocols that the user has
          opened.

        * tao/RTCORBA/RT_Protocols_Hooks.h: Removed
          ACE_STATIC_SVC_REQUIRE directive since it is not needed.

        * tao/RTCORBA/RT_ORB_Loader.cpp:

          - Changed base class from TAO_Object_Loader to
            ACE_Service_Object.

          - (create_object): Removed.

          - (Initializer): Removed.

          - (init): Handle the scheduling policy variable in terms of
            THR_SCHED_* values instead of ACE_SCHED_* values since
            ACE_OS::thr_create expects THR_SCHED_* values.

            However, ACE_Sched_Params::priority_min() and
            ACE_Sched_Params::priority_max() expect ACE_SCHED_*
            values.  Therefore, a conversion from THR_SCHED_* values
            to ACE_SCHED_* values is done in
            TAO_RT_ORBInitializer::pre_init().

            Also, added the ability for the user to specify the
            scheduling scope of the RT threads through the new
            ORBScopePolicy option.  The option supports the PROCESS
            and SYSTEM scopes with PROCESS being the default.

        * tao/RTCORBA/RT_ORB.cpp:

          - (TAO_RT_ORB): Cache the <tp_manager> from the
            TAO_RT_Thread_Lane_Resources_Manager.

          - (pre_init): Register the
            RT_Thread_Lane_Resources_Manager_Factory with the service
            configurator.

         - (TAO_RT_CORBA_Priority_Normalizer): Removed since it was no
           longer needed.

        * tao/RTCORBA/RT_Mutex.cpp (try_lock): Make sure that
          mutex::tryacquire() returning -1 with errno == EBUSY is not
          flagged as an error.

        * tao/RTCORBA/RT_Invocation_Endpoint_Selectors.cpp:

          - Added TAO_RT_Default_Endpoint_Selector.  Pretty much the
            same as TAO_Default_Endpoint_Selector except private
            connections are taken into account.

        * tao/RTCORBA/RT_Endpoint_Selector_Factory.cpp:

          - Replaced TAO_Default_Endpoint_Selector with
            TAO_RT_Default_Endpoint_Selector so that we are sure to
            take private connection into account.

        * tao/RTCORBA/Linear_Priority_Mapping.cpp:

          - Removed special code in the constructors that was added to
            handle incorrect priorities reported by
            ACE_Sched_Params::priority_min() and
            ACE_Sched_Params::priority_max().  This code was not
            needed after Joe's fixes.

          - Changed <policy_> type from int to long.

        * tao/RTCORBA/Direct_Priority_Mapping.cpp: Simplified this
          class such that it is really direct, i.e., corba priority
          maps verbatim to native priority and vice versa.

        * tao/RTCORBA/Continuous_Priority_Mapping.cpp: Maps the first
          n CORBA priorities to the range of native priorities, where
          n is the number of native priorities.

          The lowest native priority is mapped to CORBA priority 0,
          next higher native priority is mapped to CORBA priority 1,
          and so on. Since in all the operating systems where TAO is
          supported the native priority set contains less than 32767
          priorities, part of the CORBA priority range is left unused.
          Consider NT as an example. NT native priorities -15 -2 -1 0
          1 2 15 are mapped to CORBA priorities 0 1 2 3 4 5 6,
          respectively, and the rest of the CORBA priority range is
          not used.

          This class was previously called Direct_Priority_Mapping.

        * tao/RTCORBA/RT_ORBInitializer.cpp (pre_init): Set the
          scheduling policy and scope policy into ORB Parameters.
          Also, converted THR_SCHED_* values into ACE_SCHED_* values.

        * tao/RTCORBA/RTCORBA.cpp:

          - (init): Removed: functionality moved to the constructor.

          - (TAO_RTCORBA_Initializer): Removed the registration of
            TAO_RT_Protocols_Hooks into the service configurator - it
            is already done in TAO_RT_ORBInitializer::pre_init.

        * tao/RTCORBA/RT_Thread_Lane_Resources_Manager.cpp
        * tao/RTCORBA/RT_Thread_Lane_Resources_Manager.h
        * tao/RTCORBA/RT_Thread_Lane_Resources_Manager.i

          RT manager which manages multiple set of resources
          associated with each thread lane and has a default set of
          lane resources.

        * tao/RTCORBA/RT_Current.h:
        * tao/RTCORBA/RT_Mutex.h:
        * tao/RTCORBA/RT_ORB.h:
        * tao/RTCORBA/RT_ORBInitializer.cpp:
        * tao/RTCORBA/RT_Policy_i.h:
        * tao/RTCORBA/Thread_Pool.h
        * tao/RTPortableServer/RTPortableServerC.h
        * tao/RTPortableServer/RT_Acceptor_Filters.h:
        * tao/RTPortableServer/RT_Servant_Dispatcher.h:

          Include RTCORBA.h instead of RTCORBAC.h.

        * tao/RTCORBA/Makefile.bor:
        * tao/RTCORBA/Makefile:
        * tao/RTCORBA/TAO_RTCORBA.dsp:
        * tao/RTCORBA/TAO_RTCORBA_Static.dsp:

          Removed Pool_Per_Endpoint; added
          RT_Thread_Lane_Resources_Manager and
          Continuous_Priority_Mapping.

        * tao/ORB_Core.cpp:

          - (Reactor_Registry): Removed from the ORB Core.  Each
            thread lane how manages its own Leader/Follower object
            which has its own Reactor.

          - (Transport_Cache): Moved from the ORB Core to the thread
            lane.

          - (Acceptor_Registry): Moved from the ORB Core to the thread
            lane.

          - (Thread_Lane_Resources_Manager): The ORB Core now has a
            pointer to a Thread_Lane_Resources_Manager.  It'll either
            be the Default_Thread_Lane_Resources_Manager that manages
            the default thread lane, or it'll be the
            RT_Thread_Lane_Resources_Manager that manages the default
            thread lane and the RT thread lanes.

         - (Thread_Lane_Resources_Manager_Factory):
           Thread_Lane_Resources_Manager are created using this
           factory.  This factory is accessed through the Service
           Configurator.

         - (reactor_): Removed unused member.

         - (inherit_from_parent_thread): Removed dead code.

         - (create_stub): No need to pass ORB_Core::create_stub() a
            pointer to the ORB_Core.

         - (create_stub_object): Simplified this function. MProfile
           creation is going to be done by the POA.

         - (is_collocated): This function now checks with the lane
           resources if the acceptor registry has been created.

         - (leader_follower, lf_strategy): Access the Leader/Followers
           and LF Strategy object from the thread lane resources
           rather than from the Reactor Registry.

         - (open): Removed: thread lanes resourced are opened when
           thread pools are created; thread lanes resourced are opened
           when the Root POA is created.

         - (shutdown): Added missing ACE_CHECKs.

         - (reactor): Removed acceptor based reactor selection.

         - (poa_extension_initializer): Removed since it was no longer
           needed.

         - (TAO_CORBA_Priority_Normalizer): Removed since it was no
           longer needed.

         - (resource_factory_from_service_config_,
            client_factory_from_service_config_,
            server_factory_from_service_config_):

            Removed since it was no longer needed.

         - (resource_factory_, stub_factory_,
            endpoint_selector_factory_, protocols_hooks_,
            client_factory_, server_factory_):

            Simplified the creation of these factories.  Now there is
            no need to include these files:

            Default_Stub_Factory.h
            Default_Endpoint_Selector_Factory.h
            Default_Protocols_Hooks.h

         - (collocation_strategy): When trying to select the
           collocation strategy to use for a method call on an object,
           it is no longer sufficient to check the <is_collocated>
           flag on the object.   The decision is left to the
           collocation resolver, which get interesting in the
           collocated case.

         - (is_collocated): Don't ask only the current thread lane
           resources to see if the mprofile matches the acceptors;
           instead ask all the thread lane resources in the ORB.  The
           functionality is shifted over to the
           Thread_Lane_Resources_Manager which knows how many lanes
           there are in this ORB.

         - (TAO_ORB_Core_TSS_Resources): Removed <transport_cache_>,
           <reactor_registry_>, and <reactor_registry_cookie_t>; added
           <lane_> from/to the per-ORB TSS resources.

        * tao/ORB.cpp (create_stub_object): Removed unused method.  In
          addition, stub creation really belongs to the POA.

        * tao/param.cpp: Added the scheduling policy and scope policy
          for RTCORBA threads created by the ORB.

        * tao/Tagged_Profile.cpp (extract_object_key):
        * tao/IIOP_Transport.cpp (set_bidir_context_info):

          Acceptor registry is now accessed through the lane resources
          rather than through the ORB Core.

        * tao/Transport.cpp:
        * tao/Acceptor_Impl.cpp (make_svc_handler):
        * tao/IIOP_Connector.cpp:
        * tao/IIOP_Connection_Handler.cpp (add_transport_to_cache):
        * tao/Strategies/SHMIOP_Connector.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp (add_transport_to_cache):
        * tao/Strategies/DIOP_Connection_Handler.cpp (add_transport_to_cache):
        * tao/Strategies/UIOP_Connector.cpp:
        * tao/Strategies/UIOP_Connection_Handler.cpp (add_transport_to_cache):
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp (add_transport_to_cache):
        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connector.cpp (add_transport_to_cache):

          Transport cache is now accessed through the lane resources
          rather than through the ORB Core.

        * tao/Pluggable.cpp:
        * tao/IIOP_Acceptor.cpp:
        * tao/Strategies/DIOP_Acceptor.cpp:
        * tao/Strategies/SHMIOP_Acceptor.cpp:
        * tao/Strategies/UIOP_Acceptor.cpp:
        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Acceptor.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Acceptor.cpp:

          - Priority: Since endpoints are no longer associated with
            priorities, the priority is explicitly passed in from the
            thread lane when creating the shared profile.  If a
            "priority" option is specified by the user, it'll be
            flagged as an error.

          - (create_profile): Renamed create_mprofile() to
            create_profile() while adding a <priority> parameter and
            removing the <share_profile> parameter. Creation of a
            shared profile can be done if we have a valid <priority>.
            Otherwise, we create a new profile.

          - (create_new_profile): Renamed create_new_profiles() to
            create_new_profile() while adding a <priority> parameter.

          - (create_shared_profile): Added a <priority> parameter.

          - (open, open_i, open_default): Reactor is specified
            explicitly rather than coming implicitly from the ORB
            Core.

        * tao/Thread_Lane_Resources.cpp
        * tao/Thread_Lane_Resources.h
        * tao/Thread_Lane_Resources.i

          New class representing a thread lane's resources.  The class
          has the following resources:

          - Acceptor Registry
          - Transport Cache
          - Leader/Follower (and hence the Reactor also)

          These resources use to be in the ORB Core.  However, with
          the introduction of thread lanes, these resources were moved
          to this class.  These resources will be managed by two kinds
          of managers: (a) default (non-RT) manager which only has one
          set of lane resources and approximates these resources being
          in the ORB Core.  (b) RT manager which manages multiple set
          of resources associated with each thread lane and has a
          default set of lane resources.

        * tao/Thread_Lane_Resources_Manager.cpp
        * tao/Thread_Lane_Resources_Manager.h
        * tao/Thread_Lane_Resources_Manager.i

          Abstract manager and manager factory.

        * tao/Default_Thread_Lane_Resources_Manager.cpp
        * tao/Default_Thread_Lane_Resources_Manager.h
        * tao/Default_Thread_Lane_Resources_Manager.i

          Default (non-RT) manager which only has one set of lane
          resources and approximates these resources being in the ORB
          Core.

        * tao/Endpoint.h (TAO_Endpoint): Changed default <priority> value
          in constructor to <TAO_INVALID_PRIORITY> instead of -1.

        * tao/Protocols_Hooks.cpp:
        * tao/Default_Protocols_Hooks.cpp:

          - (set_default_server_protocol_policy): Added no-op new
            method.  It'll be useful with RTCORBA.

        * tao/Acceptor_Registry.cpp:

          - (open, open_default, open_i): The reactor is explicitly
            passed to these methods instead of getting it from the ORB
            Core.

          - (open, open_i): The acceptor endpoints used by thread
            lanes are selected randomly from the ones available on the
            local machine.  There is no way for the user to associate
            thread lanes with endpoints because the thread lanes are
            unnamed.

            A new parameter <ignore_address> was added to open() and
            open_i() that ignores the address specified in the
            -ORBEndpoint option.  <ignore_address> is 1 for lane
            acceptors and 0 for default acceptors.

          - (make_mprofile): This function was removed because it was
            doing too much, i.e., calling fill_mprofile() which was
            creating and encoding the mprofile.  Since the POA has to
            now deal with multiple acceptor registries, this
            functionality was moved to the POA.

        * tao/Default_Acceptor_Filter.h:
        * tao/Acceptor_Filter.h:

          - (fill_mprofile): Renamed fill_mprofile() to fill_profile()
            and added a <priority> paramter to this function.

          - (encode_endpoints): Also added a new function
            encode_endpoints() that encodes the endpoints in the
            profiles into the TAO_TAG_ENDPOINTS tag component of
            profiles.

        * tao/Transport_Cache_Manager.cpp (open): Removed.

        * tao/TAO_Internal.cpp (open_services_i): Insert the
          Default_Thread_Lane_Resources_Manager_Factory and the
          Default_Collocation_Resolver into the service configurator.
          Also, cast RT_ORB_Loader to ACE_Service_Object instead of
          TAO_Object_Loader.

        * tao/default_resource.cpp:
        * tao/Resource_Factory.cpp:
        * tao/Strategies/advanced_resource.cpp:

          - (reactor_registry): Removed since it was no longer needed.

          - (ORBReactorRegistry): Report error since this option is no
            longer supported.

        * tao/ORB.cpp (url_ior_string_to_object):
        * tao/CORBALOC_Parser.cpp (make_stub_from_mprofile):
        * tao/Stub.cpp (set_policy_overrides):
        * tao/Object.cpp (operator>>):
        * tao/IORManipulation/IORManipulation.cpp:

          No need to pass ORB_Core::create_stub() a pointer to the
          ORB_Core.

        * tao/Collocation_Resolver.cpp:
        * tao/Collocation_Resolver.h:
        * tao/Collocation_Resolver.i:
        * tao/Default_Collocation_Resolver.cpp:
        * tao/Default_Collocation_Resolver.h:
        * tao/Default_Collocation_Resolver.i:

          This class decides/resolves whether an object is collocated
          with the servant.  The default resolver simply does what was
          done before which is to check the <is_collocated_> flag on
          the object.  The RT resolver is more interesting.

        * tao/Leader_Follower.cpp (elect_new_leader): When we are out
          of leader threads and there are no event loop threads
          waiting and there are no followers available, we call on the
          new leader generator class to create us a new leader thread.

        * tao/Makefile:
        * tao/Makefile.bor:
        * tao/Makefile.am:
        * tao/TAO.dsp:
        * tao/TAO_Static.dsp:

          - Added new files:

            Thread_Lane_Resources
            Thread_Lane_Resources_Manager
            Default_Thread_Lane_Resources_Manager
            Collocation_Resolver
            Default_Collocation_Resolver

          - Removed these files:

            Reactor_Registry
            Single_Reactor
            POA_Extension_Initializer

        * tao/Makefile:
        * tao/Makefile.bor:
        * tao/Makefile.am:

          - Removed the following *S_T.* files.  Since the *S.* are
            not needed, these will also not be needed:

            CONV_FRAMES_T
            GIOPS_T
            IOPS_T
            PollableS_T
            TAOS_T
            TimeBaseS_T

        * tao/TAO.dsp:
        * tao/TAO_Static.dsp:

          - Removed these files:

            TimeBaseS
            CONV_FRAMES
            Connector_Impl
            TAO_Singleton

        * tao/PortableServer/POA.cpp:

          - (create_POA_i): Before a POA is created (including the
            RootPOA), any relevant policies at the ORB level should be
            merged with the policies passed to create_POA() by the
            user.

          - (servant_to_id_i, servant_to_reference): Use the
            <server_priority> from the priority model policy rather
            than <TAO_INVALID_PRIORITY> for servants that are not
            registered with explicit priorities.

          - (create_stub_object): The POA is now responsible for
            selecting which profiles to create the stub with.  This
            change is more relevant for the RT POA.

          - (thread_pool): No-op accessor added.  This change is more
            relevant for the RT POA.

          - (friend): Added TAO_RT_Collocation_Resolver as a friend so
            that it can call lookup_servant_i().  This is safe because
            Servant_Upcall has the lock held.

        * tao/PortableServer/Object_Adapter.cpp:

          - (open):

            Make sure that the default resources are open when the
            RootPOA is created.

            Make sure that the correct default Server Protocol Policy
            is set after we open the default resources.  Previously
            all the protocols supported were included in the default
            Server Protocol Policy.  This should be restricted to only
            the protocols opened by default (such as IIOP) and any
            other asked for by the user (such as SHMIOP).

            Before a POA is created (including the RootPOA), any
            relevant policies at the ORB level should be merged with
            the policies passed to create_POA() by the user.

            No need to access the POA Initializer.
            TAO_RT_Object_Adapter_Factory will set things up for us.

          - (Servant_Upcall::lookup_POA): New method added to simply
            lookup the POA. This method is needed for looking up the
            POA when making collocation decisions.

        * tao/PortableServer/POA_Cached_Policies.cpp
          (TAO_POA_Cached_Policies): Changed the default value of
          <priority_model_> from
          TAO_POA_Cached_Policies::CLIENT_PROPAGATED to
          TAO_POA_Cached_Policies::NOT_SPECIFIED.

        * tao/PortableServer/Default_Policy_Validator.cpp:

          - (merge_policies_impl): No-op method was added.

          - (Default_Policy_Validator): Pass-through-to-base-class
            constructor was added.

        * tao/PortableServer/Default_Acceptor_Filter.cpp:

          - (fill_mprofile): Got renamed to fill_profile().

          - (encode_endpoints): No-op method was added.

        * tao/PortableServer/Policy_Validator.cpp (merge_policies): Merge
          any relevant policies at the ORB level into the current set.

        * tao/PortableServer/POA_Policy_Set.i (policies): Added accessor
          to underlying TAO_Policy_Set implementation.

        * tao/Strategies/Makefile.bor:
        * tao/Strategies/Makefile:
        * tao/Strategies/TAO_Strategies.dsp:
        * tao/Strategies/TAO_Strategies_Static.dsp:

          Removed Reactor_Per_Priority.

        * TAO_IDL/be/be_visitor_interface/strategized_proxy_broker_ss.cpp:
        * tao/PortableServer/ImplRepoS.cpp:
        * tao/PortableServer/MessagingS.cpp:
        * tao/PortableServer/PolicyS.cpp:
        * tao/PortableServer/Strategized_Object_Proxy_Broker.cpp:
        * orbsvcs/IFR_Service/IFR_BaseS.cpp:
        * orbsvcs/IFR_Service/IFR_BasicS.cpp:
        * orbsvcs/IFR_Service/IFR_ComponentsS.cpp:
        * orbsvcs/IFR_Service/IFR_ExtendedS.cpp:
        * tao/Domain/DomainS.cpp:

          Passed environment variable to
          TAO_ORB_Core::collocation_strategy().

        * orbsvcs/ImplRepo_Service/ImplRepo_i.cpp:

          - make_mprofile() was replaced by endpoint_count() and
            fill_profile().

          - Access the acceptor registry through the lane resources
            rather than through the ORB Core.

          - Include "tao/Thread_Lane_Resources.h".

        * tests/RTCORBA/Policy_Combinations: New test added that
          combines and tests several RT policies in different ways.

          - The IDL interface has two methods: method() and
            prioritized_method().

          - The client sets its thread to the default priority, calls
            method() which returns a priority.  If the priority
            returned is different from <TAO_INVALID_PRIORITY>, the
            client sets its thread to the priority returned by the
            server and calls prioritized_method()

          - The servant returns the <client_priority_> member in
            method().  In prioritized_method(), it makes sure that the
            thread running the upcall is at priority
            <server_priority_>.  These two members allow several
            different policies to be tested with the same code.

          - In addition to testing servants in the RootPOA and in a
            simple child POA, the following policy combinations are
            tested in this example:
            ________________________________________________________________________________

            Thread-pool    BANDS   PRIORITY MODEL    Client Priority    Server Priority
            ________________________________________________________________________________

            Default Pool     NO   CLIENT_PROPAGATED        3                  3
            Default Pool     NO   SERVER_DECLARED          3                  5
            Without Lanes    NO   CLIENT_PROPAGATED        1                  1
            Without Lanes    NO   SERVER_DECLARED          1                  5
            With Lanes       NO   CLIENT_PROPAGATED        2                  2
            With Lanes       NO   SERVER_DECLARED          3                  5
            Default Pool    YES   CLIENT_PROPAGATED        3                  3
            Default Pool    YES   SERVER_DECLARED          1                  5
            Without Lanes   YES   CLIENT_PROPAGATED        1                  1
            Without Lanes   YES   SERVER_DECLARED          3                  5
            With Lanes      YES   CLIENT_PROPAGATED        3                  2
            With Lanes      YES   SERVER_DECLARED          1                  5

            where the priorities are relative to the minimum priority
            of the thread scheduling policy used for the test.

          - This test uses the -ORBPriorityMapping continuous option.

        * tests/RTCORBA/Linear_Priority: This is a test for the Linear
          Priority mapping in TAO.  In addition, this test combines
          and tests several RT policies in different ways.  This test
          can also be used for testing the different scheduling
          policies (e.g., FIFO, RR, OTHER) by using the ORBSchedPolicy
          and ORBScopePolicy options in svc.conf file.

          - The server can be setup to use bands or no bands and
            thread lanes or no lanes.

          - The client can be setup to use bands or no bands and
            different invocation priorities.  It creates a thread for
            each invocation priority and issues multiple requests to
            the server from each thread.

          - The following policy combinations are tested in this
            example:
            __________________________________________________

            Server-side   Thread   Client-side  Multi-priority
               Bands       Lanes      Bands        Clients
            __________________________________________________

                NO          NO          NO           NO
                YES         NO          NO           NO
                NO          YES         NO           NO
                YES         YES         NO           NO
                NO          NO          YES          NO
                NO          YES         YES          NO
                NO          NO          NO           YES
                YES         NO          NO           YES
                YES         YES         NO           YES
                NO          NO          YES          YES
                NO          YES         YES          YES

        * tests/RTCORBA/Thread_Pool:

          - server.cpp: Associated the thread pools with POAs.
            Produced three servants, one that uses the default thread
            pool, the second uses an RT thread pool without lanes, and
            the third that uses an RT thread pool with lanes. Also
            added the use of CLIENT_PROPAGATED priority model.

          - client.cpp: Changed the default ior file from <ior> to
            <ior_1>.

          - run_test.pl: Spawn multiple clients for each servant.  The
            servants with multiple threads in their thread-pool
            respond faster than the servants with a single thread
            thread-pool.

          - test_i.cpp (method): Added debugging output to see which
            lane/pool/thread is being used to run the upcall.

          - README: Updated to reflect changes in the test.

        * tests/RTCORBA/Destroy_Thread_Pool: New test added that tests
          the creation and destruction of thread pools (with and
          without lanes).

        * tests/RTCORBA/Persistent_IOR: New test added that:

          - For POAs using the default thread pool, the user can
            continue to use -ORBEndpoint option.

          - Makes sure that if the user tries to combine thread pools
            and persistence, an InvalidPolicy exception is thrown.

          - Makes sure that for POAs using thread pools, the user can
            specify the protocol selection, protocol version, and
            other protocol specific options (basically everything
            other than the endpoint address) through the -ORBEndpoint
            option.

        * tests/RTCORBA/Collocation: This is a new test for
            collocation when using thread pools and lanes.  There are
            multiple servants in this test:

            (a) One servant is in the Root POA which is service by the default
                thread pool.

            (b) Another servant is in a POA which is serviced by a
                thread pool without lanes.  Since a thread pool
                without lanes is really a thread pool with one lane,
                all requests will be handled by this one lane.
                Therefore, setting the priority model policy was not
                necessary for this POA.

            (c) There are two servants is in a POA which is serviced
                by a thread pool with lanes, with the
                CLIENT_PROPAGATED policy.  Depending on the priority
                of the caller, either one of the lanes processes the
                request.

            (d) There are two more servants is in a POA which is
                serviced by a thread pool with lanes, with the
                SERVER_DECLARED policy.  One servant is registered to
                match one lane and the other servant matches the other
                lane.

             All six servants are put into a global array.  The main
             thread starts off the testing by calling start() on each
             servant in the array.  The implementation of start()
             calls method() on each servant in the array.  Since the
             upcalls will be handled by the lane corresponding to each
             servant, all six lanes ends up calling into one another.

             The main thread then changes its priority and calls
             start() on each servant in the array.  This makes sure to
             exercise the second lane in the thread pool with lanes
             and the CLIENT_PROPAGATED priority.

             The test also converts the objects into strings and back
             to make sure that collocation works as expected for
             imported objects.

             This test has enough asserts to make sure that the
             correct thread is making the upcall.

        * tests/RTCORBA/Banded_Connections:

          - server.cpp:

            - We cache the RT_Current in the constructor instead of
              having to look it up on every call.

            - Simplified some of the options and some parts of the code.

            - The POA priority is now deduced from the lane
              priorities.  Lane priorities are not deduced from the
              bands.

            - The old Reactor-per-Priority scheme was replaced by RT
              Thread-Pools.

            - Removed dependency on advance resources in the
              strategies library.

            - Added check to make sure multiple priorities are
              supported.

          - client.cpp:

            - Simplified some of the options and some parts of the
              code.

            - Client thread priorities are now deduced from the bands.

            - Added check to make sure multiple priorities are
              supported.

          - test.idl (Test): test_method() now passes an additional
            boolean parameter <client_propagated> to indicate what
            kind of invocation it thinks it is making.  This is a
            useful sanity check.

          - server.conf: The service configurator directives in this
            file were no longer needed; therefore this file was
            removed.

          - run_test.pl: Removed specification of endpoint priorities
            since they are no longer needed or supported.  Also
            simplified some of the other options used in this test.

          - client.dsp:
          - server.dsp:

            Removed the static build configurations.

          - README: Updated to reflect changes in the test.

        * tests/RTCORBA/Server_Declared:

          - server.cpp:

            - The old Reactor-per-Priority scheme was replaced by a
              simple ORB::run().

            - Exception test for an invalid servant priority was
              removed since there is already a test that does this.

            - Added check to make sure multiple priorities are
              supported.

            - Removed dependency on advance resources in the
              strategies library.

          - client.cpp: Added check to make sure multiple priorities
            are supported.

          - run_test.pl: Removed specification of endpoint priorities
            since they are no longer needed or supported.  Also
            simplified some of the other options used in this test.

          - server.conf: The service configurator directives in this
            file were no longer needed; therefore this file was
            removed.

          - README: Updated to reflect changes in the test.

        * tests/RTCORBA/MT_Client_Protocol_Priority

          - server.cpp:

            - The old Reactor-per-Priority scheme was replaced by RT
              Thread-Pools.

            - Added check to make sure multiple priorities are
              supported.

          - client.cpp:

            - Added check to make sure multiple priorities are
              supported.

            - Removed dependency on advance resources in the
              strategies library.

          - run_test.pl: Removed specification of endpoint priorities
            since they are no longer needed or supported.

          - server.conf: Removed some of the service configurator
            directives in this file that were no longer needed.

          - README: Updated to reflect changes in the test.

        * tests/RTCORBA/check_supported_priorities.cpp: Common check
          used by several RT tests to make sure multiple priorities
          are supported.

        * tests/RTCORBA/Linear_Priority/client.cpp:
        * tests/RTCORBA/Linear_Priority/server.cpp:

          No need to use ACE_DEFAULT_THREAD_PRIORITY. 0 is a valid
          CORBA priority for any range when using Linear mapping.

        * tests/RTCORBA/Destroy_Thread_Pool/Destroy_Thread_Pool.cpp:
        * tests/RTCORBA/Persistent_IOR/server.cpp:
        * tests/RTCORBA/Policy_Combinations/client.cpp:
        * tests/RTCORBA/Policy_Combinations/server.cpp:
        * tests/RTCORBA/Thread_Pool/server.cpp:

          Removed dependency on ACE_DEFAULT_THREAD_PRIORITY by getting
          the main thread's priority from the RT Current.

        * tests/RTCORBA/Banded_Connections/bands.nt:
        * tests/RTCORBA/MT_Client_Protocol_Priority/run_test.pl:
        * tests/RTCORBA/Server_Declared/run_test.pl:

          Reworked priority values for NT after the direct mapping
          became simple.  The only valid CORBA priorities with direct
          mapping on NT are: 0, 1, 2, and 15.  For bands, only 0, 1,
          and 2 are counted because of the emptiness between 2 and 15.

        * tests/RTCORBA/RTCORBA_tests.dsw:
        * tests/RTCORBA/Makefile:
        * tests/RTCORBA/Makefile.bor:

          Added new tests ORB_init, Policy_Combinations,
          Destroy_Thread_Pool, Persistent_IOR, Collocation, and
          Linear_Priority.

        * tests/RTCORBA/Banded_Connections/Makefile:
        * tests/RTCORBA/Banded_Connections/client.bor:
        * tests/RTCORBA/Banded_Connections/client.dsp:
        * tests/RTCORBA/Banded_Connections/server.bor:
        * tests/RTCORBA/Banded_Connections/server.dsp:
        * tests/RTCORBA/Client_Propagated/Makefile:
        * tests/RTCORBA/Client_Propagated/client.bor:
        * tests/RTCORBA/Client_Propagated/client.dsp:
        * tests/RTCORBA/Client_Propagated/server.bor:
        * tests/RTCORBA/Client_Propagated/server.dsp:
        * tests/RTCORBA/Linear_Priority/client.dsp:
        * tests/RTCORBA/Linear_Priority/server.dsp:
        * tests/RTCORBA/Policy_Combinations/client.dsp:
        * tests/RTCORBA/Policy_Combinations/server.dsp:
        * tests/RTCORBA/Private_Connection/client.dsp:
        * tests/RTCORBA/Private_Connection/server.dsp:
        * tests/RTCORBA/Client_Propagated/Makefile:
        * tests/RTCORBA/Client_Propagated/client.bor:
        * tests/RTCORBA/Client_Propagated/client.dsp:
        * tests/RTCORBA/Client_Propagated/server.bor:
        * tests/RTCORBA/Client_Propagated/server.dsp:

          Removed Strategies library since it is not needed.

        * tests/RTCORBA/Client_Protocol/client.bor:
        * tests/RTCORBA/Client_Protocol/client.dsp:
        * tests/RTCORBA/Explicit_Binding/client.bor:
        * tests/RTCORBA/Explicit_Binding/client.dsp:
        * tests/RTCORBA/MT_Client_Protocol_Priority/client.bor:
        * tests/RTCORBA/MT_Client_Protocol_Priority/client.dsp:
        * tests/RTCORBA/Private_Connection/client.bor:
        * tests/RTCORBA/RTMutex/server.dsp:
        * tests/RTCORBA/Server_Protocol/client.bor:
        * tests/RTCORBA/Server_Protocol/client.dsp:

          Removed PortableServer and RTPortableServer libraries since
          they are not needed.

        * tests/TAO_Tests.dsw:
        * tests/BiDirectional/test_i.cpp:
        * tests/BiDirectional_NestedUpcall/test_i.cpp:
        * tests/Collocation/Coll_Test_Stubs.bor:
        * tests/Collocation/Coll_Test_Stubs.dsp:
        * tests/Collocation/Coll_Tester.cpp:
        * tests/Collocation/Coll_Tester.h:
        * tests/Collocation/Collocation.bor:
        * tests/Collocation/Collocation.cpp:
        * tests/Collocation/Collocation.dsp:
        * tests/Collocation/Collocation.dsw:
        * tests/Collocation/Collocation_Test_Stubs.bor:
        * tests/Collocation/Collocation_Test_Stubs.dsp:
        * tests/Collocation/Collocation_Tester.cpp:
        * tests/Collocation/Collocation_Tester.h:
        * tests/Collocation/Diamond.bor:
        * tests/Collocation/Diamond.dsp:
        * tests/Collocation/Makefile.bor:
        * tests/Collocation/Makefile.test:
        * tests/Collocation/README:
        * tests/Connection_Purging/client.bor:
        * tests/Connection_Purging/client.cpp:
        * tests/Connection_Purging/client.dsp:
        * tests/Connection_Purging/server.bor:
        * tests/Connection_Purging/server.cpp:
        * tests/Connection_Purging/server.dsp:
        * tests/DLL_ORB/Test_Client_Module.dsp:
        * tests/DLL_ORB/Test_Server_Module.dsp:
        * tests/DLL_ORB/client.dsp:
        * tests/DLL_ORB/server.dsp:
        * tests/Exposed_Policies/Makefile:
        * tests/Exposed_Policies/Policy_Tester.cpp:
        * tests/Exposed_Policies/Policy_Verifier.cpp:
        * tests/Exposed_Policies/run_test.pl:
        * tests/Exposed_Policies/server.conf:
        * tests/Exposed_Policies/svc.conf:
        * tests/Faults/middle.dsp:
        * tests/Multiple/Client.dsp:
        * tests/Multiple/Server.dsp:
        * tests/Oneways_Invoking_Twoways/Oneways_Invoking_Twoways.dsw:

          Made a bunch of small changes to make code and makefiles
          consistent.

        * docs/rtcorba/features.html:
        * docs/rtcorba/status.html:
        * docs/rtcorba/issues.html:

          Updated RTCORBA documentation to reflect new thread-pool
          support.  Also added new issues that need to be addressed.

        * docs/Options.html: Added documentation for new
          -ORBScopePolicy option and updated documentation for
          -ORBPriorityMapping option.

Mon Sep 17 13:18:35 2001  Carlos O'Ryan  <coryan@uci.edu>

        * orbsvcs/orbsvcs/CosEventComm.idl:
        * orbsvcs/orbsvcs/CosEventChannelAdmin.idl:
        * orbsvcs/orbsvcs/CosEvent/CEC_ConsumerAdmin.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ConsumerControl.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Default_Factory.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Defaults.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Dispatching.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Dispatching_Task.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_EventChannel.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Event_Loader.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Factory.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_MT_Dispatching.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPullConsumer.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPullSupplier.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPushConsumer.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_ProxyPushSupplier.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Pulling_Strategy.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Reactive_ConsumerControl.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Reactive_Pulling_Strategy.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_Reactive_SupplierControl.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_SupplierAdmin.h:
        * orbsvcs/orbsvcs/CosEvent/CEC_SupplierControl.h:
          Use Doxygen to document the COS Event Service implementation and
          its IDL interfaces.

        * orbsvcs/orbsvcs/RtecBase.idl:
        * orbsvcs/orbsvcs/RtecDefaultEventData.idl:
        * orbsvcs/orbsvcs/RtecEventChannelAdmin.idl:
        * orbsvcs/orbsvcs/RtecEventComm.idl:
        * orbsvcs/orbsvcs/RtecUDPAdmin.idl:
          Use Doxygen to document TAO's Real-time Event Service interface.

        * orbsvcs/orbsvcs/CosNotification.idl:
        * orbsvcs/orbsvcs/CosNotifyChannelAdmin.idl:
        * orbsvcs/orbsvcs/CosNotifyComm.idl:
        * orbsvcs/orbsvcs/CosNotifyFilter.idl:
          Use Doxygen to document the OMG Notification Service IDL files.

Sun Sep 16 11:22:12 2001  Balachandran  <bala@cs.wustl.edu>

        * TAO version 1.1.20 released.

Fri Sep 14 11:35:55 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp (consolidate_message): Used the size () of the
          message block instead of the length (). Thanks to Mahesh
          Varadarajan <mahesh.varadarajan@divatv.com> for
          reporting this.

Fri Sep 14 07:08:41 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Makefile.bor: Added Cache_Growth_Test to this
          Makefile. Thanks to Johnny Willemsen for pointing this out.

Thu Sep 13 13:46:54 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * docs/Options.html: Changed -ORBCacheMax in the documentation to
          be -ORBConnectionCacheMax. Thanks to Jean-christophe Dubois
          <jcd@one.com>  for reporting this.

Thu Sep 13 13:18:48 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Cache_Growth_Test: A new test for testing whether the
          Cache size grows when old clients start disconnecting  and new
          clients start connecting.

        * tests/Cache_Growth_Test/run_test.pl:
        * tests/Cache_Growth_Test/Makefile:
        * tests/Cache_Growth_Test/Makefile.bor:
        * tests/Cache_Growth_Test/server.dsp:
        * tests/Cache_Growth_Test/Cache_Grow.dsw:
        * tests/Cache_Growth_Test/server.cpp:
        * tests/Cache_Growth_Test/client.cpp:
        * tests/Cache_Growth_Test/server.bor:
        * tests/Cache_Growth_Test/client.bor:
        * tests/Cache_Growth_Test/README:
        * tests/Cache_Growth_Test/Hello.h:
        * tests/Cache_Growth_Test/Hello.cpp: Files for the test.

        * tests/Makefile: Added the new test.

Thu Sep 13 12:46:45 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.h:
        * tao/Transport.cpp: Changed the name of the method mark_invalid
          () to purge_entry (). The mark_invalid () does not remove the
          handler from cache. Just leaving the cache entry after the
          connection is closed leads to increase in run time memory
          requirements. See bug #1025 for details. Thanks to
          Jean-Christophe Dubois <jcd@one.com> and Kevin Cline
          <krc@one.com> for reporting this. Thanks to Chris Cleeland and
          Chad Elliott for reviewing the fix.

        * tao/IIOP_Connection_Handler.cpp:
        * tao/Strategies/DIOP_Connection_Handler.cpp:
        * tao/Strategies/UIOP_Connection_Handler.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp (handle_close):

          Purge the entry while  closing down the connection instead of
          instead of marking it invalid.

Thu Sep 13 11:33:35 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/BiDirectional/test_i.cpp (call_client):
        * tests/BiDirectional_NestedUpcall/test_i.cpp: Instead of just
          using ACE_ASSERT, added an extra debug statement with reason why
          the program is going to abort.

Thu Sep 13 11:05:13 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp:
        * tao/IIOP_Connection_Handler.cpp (process_listen_point_list):
          When creating an IIOP Endpoint, passed in the value of the
          ORBDottedDecimal address value from the ORB_Core. This is a real
          stupid mistake :(. Added a debug statement. Thanks to Werner
          Buchert <w.buchert@medat.de> for reporting this problem.

Wed Sep 12 17:05:00 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/DynamicAny/DynAny_i.cpp: Fixed a compile problem with g++.

Wed Sep 12 16:29:49 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Leader_Follower.cpp (wait_for_event): If the follower times
          out, we set the state in the LF_Event to indicate an error.

        * tao/LF_Event.h:
        * tao/LF_Event.cpp:
        * tao/LF_Event.inl: Added a new private method that allows setting
          the state without holding the lock. Moreover, before we signal
          the follower in state_changed () we check whether the LF_Event
          has reached a final state or not. Further, we also check whether
          we have a valid follower_ pointer before we signal the
          follower. This fix should fix the problems MT_Timeout tests. The
          problem was that the leader thread was processing the followers
          reply when the follower timedout.

Wed Sep 12 12:33:15 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/DynamicAny/DynAny_i.cpp:

          Added 'const' to the char* passed to the extraction
          operator in the change below.

Wed Sep 12 10:13:31 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/DynamicAny/DynAny_i.cpp:

          Strange bug that seems to happen only on Win32, where when
          a DynAny containing a string is copied from another, its type
          code member gets the bound copied correctly, but its member
          any's type code does not. It crops up only in the equal()
          method where the strings are extracted from the member anys
          and compared. The CORBA::Any::to_string struct is left out
          of the extraction from the lhs any. There is no danger of
          type mismatch, since the type codes (which do retain bound
          information) have already been compared for equivalence.

Tue Sep 11 18:07:46 2001  Yamuna Krishnamurthy  <yamuna@cs.wustl.edu>

        * orbsvcs/orbsvcs/AV/Fill_ACE_QoS.h:
        * orbsvcs/orbsvcs/AV/Fill_ACE_QoS.cpp:

          Fixed compile errors with RAPI enabled.

Tue Sep 11 15:46:44 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/File_IO/client.cpp:
        * tests/File_IO/server.cpp: Reduced the number of threads a bit
          for Sun CC  alone. Bug 957 is in action again.

Mon Sep 10 20:00:25 2001  Priyanka Gontla  <pgontla@ece.uci.edu>

        * tao/MCAST_Parser.cpp:
          Fixed the errors on Win.

Mon Sep 10 13:34:43 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/ORB_Core.cpp (output_cdr_dblock_allocator): Added a comment
          in the code.

Mon Sep 10 12:23:30 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/DynamicAny/DynCommon.cpp (insert_string, get_string):

          Fixed a problem in these methods with bounded strings.
          Thanks to Russell Mora <rd.mora@econz.co.nz> for
          reporting this bug.

        * tests/DynAny_Test/da_tests.idl:
        * tests/DynAny_Test/test_dynsequence.cpp:

          Added a bounded string example to the DynSequence test.

Mon Sep 10 12:17:12 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/Makefile:
          Moved the libraries around to get proper link line. It was
          busted. Further added a realclean so that we remove strange
          generated files out of the directory. This shoudl fix one of the
          compile errors in LYNX_PPC builds.

Mon Sep 10 07:41:05 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Xt_Stopwatch/Makefile: Updated dependencies.

Sun Sep  9 11:09:04 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Makefile: Updated dependencies.

Fri Sep  7 17:16:50 2001  Anand Krishnan  <anandk@cs.wustl.edu>

        * orbsvcs/tests/Security/MT_SSLIOP/run_test.pl:
          Actually committing the file.

Fri Sep  7 12:56:45 2001  Craig Rodrigues  <crodrigu@bbn.com>

       * orbsvcs/AV/Protocol_Factory.cpp: Add default constructor.
         Thanks to Joe Loyall <jloyall@bbn.com> for finding this.

Thu Sep  6 20:07:25 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Makefile.bor: Added MCAST_Parser to the Makefile.

Thu Sep  6 12:56:55 2001  Priyanka Gontla  <gontla_p@ociweb.com>

        * orbsvcs/tests/Simple_Naming/run_test.pl (client):
          Modified the perl script to adhere to the new way of providing
          the reference to a multicasted service.

Thu Sep  6 07:17:13 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/MCAST_Parser.cpp: Fixed compile errors in Minimum builds.

Wed Sep  5 22:28:41 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/MCAST_Parser.cpp: Fixed a compile error with g++.
        * tao/ORB.cpp: Fixed a warning with g++.

Wed Sep  5 19:51:36 2001  Priyanka Gontla  <gontla_p@ociweb.com>

        * tao/TAO_Static.dsp:
        * tao/TAO.dsp :
          Added the new MCAST_Parser files.

Wed Sep  5 17:26:12 2001 Yamuna Krishnamurthy  <yamuna@cs.wustl.edu>

        * orbsvcs/orbsvcs/AV/AVStreams_i.cpp:

          Made changes to TAO_StreamEndPoint::change_qos to allow passing
          a null AVStreams::QoS to the TAO_AV_Flow_Handler::change_qos.

        * orbsvcs/orbsvcs/AV/AVStreams_i.i:

          Added a missing parameter to the ACE_DEBUG statement to print
          the flow name.

        * orbsvcs/orbsvcs/AV/QoS_UDP.cpp:

          Facilitated the passing of null AVStreams::QoS to the negotiator.

Wed Sep  5 17:26:12 2001  Anand Krishnan  <anandk@cs.wustl.edu>

        * orbsvcs/tests/Security/MT_SSLIOP/run_test.pl:
        * orbsvcs/tests/Security/MT_SSLIOP/server.cpp:
        * orbsvcs/tests/Security/MT_SSLIOP/Server_Worker.h:
        * orbsvcs/tests/Security/MT_SSLIOP/Server_Worker.cpp:
        * orbsvcs/tests/Security/MT_SSLIOP/client.cpp:
        * orbsvcs/tests/Security/MT_SSLIOP/Client_Worker.cpp:
        * orbsvcs/tests/Security/MT_SSLIOP/Client_Worker.h:
        * orbsvcs/tests/Security/MT_SSLIOP/Makefile:
        * orbsvcs/tests/Security/MT_SSLIOP/test_i.cpp:
        * orbsvcs/tests/Security/MT_SSLIOP/test_i.h:
        * orbsvcs/tests/Security/MT_SSLIOP/test_i.i:
        * orbsvcs/tests/Security/MT_SSLIOP/client.conf:
        * orbsvcs/tests/Security/MT_SSLIOP/server.conf:
        * orbsvcs/tests/Security/MT_SSLIOP/selfsigncert.pem:
        * orbsvcs/tests/Security/MT_SSLIOP/pvtkey.pem:
        * orbsvcs/tests/Security/MT_SSLIOP/test.idl: A test for a simple
          multi-threaded SSLIOP test. This test will not be included in
          the daily builds for the upcoming beta, but will be there for
          the next beta.

Wed Sep  5 12:10:40 2001  Priyanka Gontla  <gontla_p@ociweb.com>

        This set of changes are for bug 977.

        * tao/MCAST_Parser.i:
        * tao/MCAST_Parser.h:
        * tao/MCAST_Parser.cpp:

                The Parser for the new IP multicast format. The multicast
        format is  mcast://mcast_address:mcast_port:nic_address:ttl.
        All the multicast requests are now dealt via this parser. The default
        multicast address is 224.9.9.2. The default multicast port is
        10013 ( the same port that we used for NameService .. no big
        reason .. just a simple choice), default nic is eth0 and default
        TTL value is 1.

        * tao/TAO_Internal.cpp:
          Add MCAST protocol to the list of services that have to be
          initiated.

        * tao/ORB.h :
        * tao/ORB.cpp :
          Moved the multicast_to_service and multicast_query methods to
          MCAST_Parser. Modified ::resolve_service accordingly.

        * tao/ORB_Core.cpp :
          Check for mcast: format when ORBDefaultInitRef option is used
          and set the object delimiter to '/' if it is mcast protocol.

        * tao/default_resource.cpp:
          Modify the total no. of parsers to check for: from 4 to 5.
          and dynamically load the MCAST_Parser too.

        * tao/Makefile:
          Added MCAST_Parser

        * orbsvcs/tests/IOR_MCast/README :
        * orbsvcs/tests/IOR_MCast/Makefile :
        * orbsvcs/tests/IOR_MCast/MCast.idl :
        * orbsvcs/tests/IOR_MCast/MCast_Server_i.h :
        * orbsvcs/tests/IOR_MCast/MCast_Server_i.cpp :
        * orbsvcs/tests/IOR_MCast/client.cpp :
        * orbsvcs/tests/IOR_MCast/ior_mcast_client_i.h :
        * orbsvcs/tests/IOR_MCast/ior_mcast_client_i.cpp :
        * orbsvcs/tests/IOR_MCast/server.cpp :
        * orbsvcs/tests/IOR_MCast/server_i.h :
        * orbsvcs/tests/IOR_MCast/server_i.cpp :

          Simple test to test the new MCAST_Parser.

Wed Sep  5 12:35:33 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Endpoint.h: Added a lock that is used to lock when a thread
          does remote object address lookups. Theoretically speaking we
          shouldnt be putting the lock in the TAO_Endpoint class. But as
          the lock is required for most of the protocols supported in TAO
          we have pushed it here. We also believe that other protocols may
          need a lock during lookups.

        * tao/IIOP_Endpoint.i: Hold a lock in object_addr (). There was a
          subtle race condition in that method. In object_addr (void) the
          code first checks for object_addr_.get_type() != AF_INET and, if
          so, it calls ACE_INET_Addr::set(). ACE_INET_Addr::set() sets
          type field to AF_INET, zeros the inet_addr structure and then
          does a hostname lookup.

          If two threads enter the object_addr() method the first may end
          up blocked for a while in the hostname lookup.  The second
          thread will see that AF_INET is set and return an (zero'ed)
          ACE_INET_Addr structure.

          The race can happen from TAO_IIOP_Connector::connect() when
          multiple threads attempt to talk to the same object.  You will
          sometimes see transient exceptions as one of the threads tries
          to connect to the (invalid) inet_addr. This could potentially
          fix #189, but it hasnt been tested out properly. Thanks to
          <pphillip@opentext.com> for nailining the probelm and suggesting
          this fix. This should fix #1017.

        * tao/Strategies/DIOP_Endpoint.i:
        * tao/Strategies/SHMIOP_Endpoint.i: Applied the same fix.

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.i: Just added a comment
          telling that we shouldnt be holding the lock as the lock is being
          held in the IIOP class.

        * tao/IIOP_Transport.cpp:
        * tao/Strategies/DIOP_Transport.cpp:
        * tao/Strategies/SHMIOP_Transport.cpp:
        * tao/Strategies/UIOP_Transport.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Transport.cpp: Do not print out
          error messages if the recv () returns from a timeout. This extra
          print statements in a thread-per-connection case, when a thread
          timedout, was causing more confusion than it tried addressing.

Wed  Sep  5 7:20:17 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/tests/AVStreams/Multicast/ftp.dsp:
        * orbsvcs/tests/AVStreams/Multicast/server.dsp: Added the
          path to teh strategies library. I am not sure when I will get
          the  Win32 stuff right every time:(

Tue Sep  4 20:11:17 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/tests/AVStreams/Multicast/ftp.dsp:
        * orbsvcs/tests/AVStreams/Multicast/server.dsp: Added the
          strategies library  to the set of libraries that needs to be
          linked.

Tue Sep  4 15:11:17 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Makefile:
          Generated dependencies again. Also fixed a cut and paste error
          in IDL names. This created problems during 'make
          realclean'. This should fix compile errors in NoInline builds.

Tue Sep  4 11:19:44 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/File_IO/client.cpp: Changed the 'oflag' arguments for the
          file so that it works in our daily builds. We only pass the
          O_RDONLY value to the remote call. This should fix the exception
          problems seen in our daily builds.

Tue Sep  4 11:06:55 2001  Chad Elliott <elliott_c@ociweb.com>

        * orbsvcs/orbsvcs/CosEvent/CEC_ConsumerAdmin.cpp:
        * orbsvcs/orbsvcs/CosEvent/CEC_Default_Factory.cpp:
        * orbsvcs/orbsvcs/CosEvent/CEC_SupplierAdmin.cpp:
        * orbsvcs/orbsvcs/Event/EC_ConsumerAdmin.cpp:
        * orbsvcs/orbsvcs/Event/EC_SupplierAdmin.cpp:

          Added includes for ESF templates.  Fix for GHS on VxWorks.
          Thanks, Priyanka <pgontla@ece.uci.edu>, for figuring this one
          out.

Tue Sep  4 09:19:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/Makefile.av: Updated dependencies again to fix a
          compilation error.

Tue Sep  4 07:53:23 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * */Makefile: Updated dependencies. We wantde this done before
          1.2. I dont think we should be having anymore big changes to
          regenerate this again.

Tue Sep  4 07:22:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.inl:
          Changed the typename declaration to class. The usage of typename
          chokes SunCC42.

Mon Sep  3 18:52:03 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/tests/AVStreams/Multicast/Makefile:
        * orbsvcs/tests/AVStreams/Multicast/ftp.bor:
        * orbsvcs/tests/AVStreams/Multicast/server.bor:
        * orbsvcs/tests/AVStreams/Multicast/server.h:
        * orbsvcs/tests/AVStreams/Multicast/server.cpp:
        * orbsvcs/tests/AVStreams/Multicast/ftp.h:
        * orbsvcs/tests/AVStreams/Multicast/ftp.cpp:
          Use service configurator to select reactor type.  Fix some typos.

        * orbsvcs/tests/AVStreams/Multicast/components_svc.conf: (Added)

Fri Aug 31 19:28:26 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Faults/server.cpp:
        * tests/Faults/test_i.h: The servant implementation inherits from
          the PortableServer::RefCountServantBase. Changes were also done
          in the server.cpp file to create the servant on the heap. Thanks
          to Chris Kohlhoff for pointing this out.

Fri Aug 31 18:58:27 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/File_IO/client.cpp: Fixed a warning.

Fri Aug 31 18:53:28 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/IIOP_Connection_Handler.cpp:
        * tao/Strategies/UIOP_Connection_Handler.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp: Reverted the
          change "Tue Aug 28 16:56:07 2001  Balachandran Natarajan
          <bala@cs.wustl.edu>". Put the sockets back in blocking mode for
          a RW connection handler. We had severe fall in performance. This
          should get that back. We need to come up with a better way to
          handle the problem for which the fix was put in.

Fri Aug 31 16:32:14 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.h:
          Changed the typename declaration to class. The usage of typename
          chokes SunCC42.

Fri Aug 31 08:31:09 2001  Chad Elliott  <elliott_c@ociweb.com>

        * tests/TestUtils/TestCombinedThreads.cpp:

          Fix for the VxWorks based thread join.  Return 0 even if the
          thread id could not be verified.

Fri Aug 31 07:52:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * orbsvcs/tests/Concurrency/*.bor:
        * orbsvcs/tests/LoadBalancing/*.bor:
        * orbsvcs/tests/Makefile.bor:
          Added BCB makefiles for these tests

Thu Aug 30 15:25:13 2001  Ossama Othman  <ossama@uci.edu>

        * tests/DLL_ORB/client.cpp (main):
        * tests/DLL_ORB/server.cpp (main):

          Reverted the change that pre-loaded the default resource factory
          before the ORB.  Chris Kohlhoff's slick fix for the Server
          Configurator obviates the need to this sort of pre-loading.

Thu Aug 30 07:59:40 2001  Carlos O'Ryan  <coryan@uci.edu>

        * TAOACE.dsw:
          Remove bogus dependency for the RT Event Service

Wed Aug 29 14:35:31 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.h:

          "Doxygen-ated" this header.

Wed Aug 29 14:46:25 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Makefile:
        * orbsvcs/performance-tests/EC_Federated_Latency/Makefile: Will
          not build with old g++ compilers.

Wed Aug 29 14:27:27 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/Throughput/svc.conf: Used blocked flushing
          strategy. Using a RW connection handler with a LF flushing
          strategy can be really really tricky. It may not be that tricky,
          it wouldnt work :-). This should fix the failure problem
          observed in Solaris builds.

Wed Aug 29 13:13:23 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.inl:
          Fixed warnings in TRU 64 builds.

Wed Aug 29 06:21:25 2001  Douglas C. Schmidt  <schmidt@macarena.cs.wustl.edu>

        * tao/ORB.h (CORBA_ORB): Improved the documentation for
          ORB::destroy().  Thanks to Wilfried Reinoehl
          <Wilfried.Reinoehl@WiredMinds.de> for reporting this.

Tue Aug 28 16:56:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/IIOP_Connection_Handler.cpp:
        * tao/Strategies/UIOP_Connection_Handler.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp: Put the sockets in
          non-blocking mode even for wait on read strategy. Need to figure
          out how much of performance impact does this have. If it has a
          penalty then we can revert this change. This solves one problem
          that we could have. If someone uses RW connection handler
          through a svc conf file, we could inadvertently turn the server
          sockets into blocking mode which could create problems.

Tue Aug 28 11:31:58 2001  Ossama Othman  <ossama@uci.edu>

        * tests/DLL_ORB/client.cpp (main):
        * tests/DLL_ORB/server.cpp (main):

          Load the Resource_Factory before the test module.  This forces
          the Resource_Factory to exist longer than the dynamically loaded
          ORB, thus allowing the ORB to make invocations on it during
          destruction.  Fixes a segmentation fault in both the client and
          server binaries.

Tue Aug 28 12:28:58 2001  Chad Elliott  <elliott_c@ociweb.com>

        * tests/TestUtils/TestCombinedThreads.cpp:

          Reformat the code to follow the style guidelines.

Tue Aug 28 10:40:06 2001  Chad Elliott  <elliott_c@ociweb.com>

        * tests/TestUtils:
        * tests/TestUtils/Makefile:
        * tests/TestUtils/TestCombinedThreads.h:
        * tests/TestUtils/TestCombinedThreads.cpp:

          Added the TestUtils VxWorks testing library.  This is used to
          create "one-button" TAO tests with VxWorks.

        * tests/Makefile:

          Added TestUtils to the Makefile.  This is compiled only for
          VxWorks.

Tue Aug 28 08:22:07 2001  Chad Elliott  <elliott_c@ociweb.com>

        * tests/AMI_Buffering/run_test.pl:
        * tests/Big_Oneways/run_test.pl:
        * tests/Big_Request_Muxing/run_test.pl:
        * tests/Big_Twoways/run_test.pl:
        * tests/Blocking_Sync_None/run_test.pl:
        * tests/Client_Leaks/run_test.pl:
        * tests/Crash_On_Write/run_test.pl:
        * tests/Crashed_Callback/run_test.pl:
        * tests/Exposed_Policies/run_test.pl:
        * tests/Hello/run_test.pl:
        * tests/MT_Server/run_test.pl:
        * tests/MT_Timeout/run_test.pl:
        * tests/Muxing/run_test.pl:
        * tests/Oneway_Buffering/run_buffer_size.pl:
        * tests/Oneway_Buffering/run_message_count.pl:
        * tests/Oneway_Buffering/run_test.pl:
        * tests/Oneway_Buffering/run_timeout.pl:
        * tests/Param_Test/run_test.pl:
        * tests/RTCORBA/Server_Declared/run_test.pl:
        * tests/RTCORBA/Server_Protocol/run_test.pl:
        * tests/RTCORBA/Thread_Pool/run_test.pl:
        * tests/Reliable_Oneways/run_sync_none.pl:
        * tests/Reliable_Oneways/run_sync_with_server.pl:
        * tests/Reliable_Oneways/run_sync_with_target.pl:
        * tests/Reliable_Oneways/run_sync_with_transport.pl:
        * tests/Reliable_Oneways/run_test.pl:
        * tests/Server_Leaks/run_test.pl:
        * tests/Single_Read/run_test.pl:
        * tests/Timed_Buffered_Oneways/run_test.pl:

          Very minor bug fixes and consistency modifications.

Mon Aug 27 07:36:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * tests/Makefile.bor: added new File_IO test
        * examples/simple/time-date/time-date.bor: changed targetdir
          to ./ to that svc.conf can find the dll
        * orbsvcs/tests/ior_corba_name/*.bor: Added BCB makefiles
        * orbsvcs/tests/Time/*.bor: Added BCB makefiles
        * orbsvcs/tests/Makefile.bor: added ior_corbaname and time

Sun Aug 26 08:20:30 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.inl:
          Removed the default value of the argument. This should fix
          warnings in GCC 3.0 builds.

Sat Aug 25 23:38:26 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/server.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/client.cpp:
        * orbsvcs/performance-tests/EC_Latency/server.cpp:
        * orbsvcs/performance-tests/EC_Latency/client.cpp: #included
          "tao/Messaging.h". Fixes build problems in no interceptors
          build.

Sat Aug 25 09:49:06 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/File_IO/server.dsp:
        * tests/File_IO/client.dsp: Fixed the Win32 builds.

Sat Aug 25 09:04:06 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Makefile:
        * orbsvcs/performance-tests/EC_Latency/Makefile: Will not build
          if corba_messaging == 0.

Fri Aug 24 23:16:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/File_IO/File.idl:
        * tests/File_IO/File_i.cpp:
        * tests/File_IO/File_i.h:
        * tests/File_IO/client.cpp
        * tests/File_IO/run_test.pl:
        * tests/File_IO/server.cpp: Fixed Fuzz errors.

Fri Aug 24 19:20:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

       * orbsvcs/orbsvcs/AV/AVStreams_i.cpp:
         Remove duplicate template instantiations.

Fri Aug 24 16:15:36 2001  Carlos O'Ryan  <coryan@uci.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Peer.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/Coordinator.cpp:
          More fixes for platforms without native C++ exceptions.

Fri Aug 24 14:48:28 2001  Carlos O'Ryan  <coryan@uci.edu>

        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.cpp:
          Fixed the code to deal with platforms without C++ native
          exceptions.

Fri Aug 24 14:35:59 2001  Carlos O'Ryan  <coryan@uci.edu>

        * orbsvcs/orbsvcs/Event/EC_Gateway.cpp:
          Comment out debug messages
          Don't publish special event types for the gateway, they are
          not used for filtering supplier data anyway.

        * orbsvcs/orbsvcs/Event/EC_ObserverStrategy.cpp:
          Use a temporary reference to make the code easier to read.

        * orbsvcs/Makefile:
        * orbsvcs/performance-tests/Makefile:
          Add new directory for the orbsvcs performance tests.

        * orbsvcs/performance-tests/EC_Latency/Makefile:
        * orbsvcs/performance-tests/EC_Latency/README:
        * orbsvcs/performance-tests/EC_Latency/run_test.pl:
        * orbsvcs/performance-tests/EC_Latency/Client_Task.h:
        * orbsvcs/performance-tests/EC_Latency/Client_Task.cpp:
        * orbsvcs/performance-tests/EC_Latency/Consumer.h:
        * orbsvcs/performance-tests/EC_Latency/Consumer.cpp:
        * orbsvcs/performance-tests/EC_Latency/Supplier.h:
        * orbsvcs/performance-tests/EC_Latency/Supplier.cpp:
        * orbsvcs/performance-tests/EC_Latency/client.cpp:
        * orbsvcs/performance-tests/EC_Latency/server.cpp:
          A latency test for the real-time event channel that can be
          automatically ran in the nightly regression testsuite.

        * orbsvcs/performance-tests/EC_Federated_Latency/README:
        * orbsvcs/performance-tests/EC_Federated_Latency/Makefile:
        * orbsvcs/performance-tests/EC_Federated_Latency/run_test.pl:
        * orbsvcs/performance-tests/EC_Federated_Latency/Control.idl:
        * orbsvcs/performance-tests/EC_Federated_Latency/ECFL_Configuration.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Consumer.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Consumer.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/Coordinator.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Coordinator.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/Loopback.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Loopback.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/Loopback_Consumer.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Loopback_Consumer.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/Loopback_Supplier.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Loopback_Supplier.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/Peer.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Peer.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.inl:
        * orbsvcs/performance-tests/EC_Federated_Latency/Servant_var.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/Supplier.h:
        * orbsvcs/performance-tests/EC_Federated_Latency/Supplier.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/client.conf:
        * orbsvcs/performance-tests/EC_Federated_Latency/client.cpp:
        * orbsvcs/performance-tests/EC_Federated_Latency/server.cpp:
          Finally!  A test to measure latency over a federated event
          channel, initial numbers look good.

Fri Aug 24 13:36:19 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/File_IO/run_test.pl:
        * tests/File_IO/File_i.cpp: Removed some vague debug statements.
        * tests/File_IO/client.cpp: Reduced teh number of threads to 10,
          instead of 20.

Fri Aug 24 12:44:45 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/File_IO/: New test to our test suite. This test actually
          brought the problem mentioned in bug #1002 to the surface. The
          problem was burried in other tests and this test actually
          could reproduce the bug every time consistently.

        * tests/File_IO/File_IO.dsw:
        * tests/File_IO/File_i.h:
        * tests/File_IO/server.bor:
        * tests/File_IO/Makefile:
        * tests/File_IO/Makefile.bor:
        * tests/File_IO/client.dsp:
        * tests/File_IO/server.cpp:
        * tests/File_IO/File.idl:
        * tests/File_IO/client.bor:
        * tests/File_IO/File_i.cpp:
        * tests/File_IO/README:
        * tests/File_IO/client.cpp:
        * tests/File_IO/run_test.pl:
        * tests/File_IO/big.txt:
        * tests/File_IO/server.dsp: Files for the new test. Thanks to
          Jonathan Reis <reis@stentor.com> for providing this test.

Fri Aug 24 08:24:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * examples/simple/time-date/time-date.bor:
          Added target, so that time-date dll name is appended with
          decorator string

Thu Aug 23 09:49:54 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/include/utl_err.h:
        * TAO_IDL/util/utl_err.cpp:
        * TAO_IDL/util/utl_identifier.cpp:

          Added new error message saying that an idenifier has
          more than one leading underscore, and a check for
          that condition in the UTL_Identifier contstructor.
          Thanks to Daniel Manfis <dany_100@yahoo.com> for
          reporting this bug.

Thu Aug 23 09:15:24 2001  Paul Calabrese <calabrese_p@ociweb.com>

        * tests/Strategies/run_test.pl:

          Use the PerlACE::LocalFile() function for the svc.conf
          files.

Thu Aug 23 01:54:42 2001  Christopher Kohlhoff  <chris@kohlhoff.com>

        * tests/DLL_ORB/Test_Client_Module.bor:
        * tests/DLL_ORB/Test_Server_Module.bor:
        * tests/DLL_ORB/client.bor:
        * tests/DLL_ORB/server.bor:
          Fixed Borland makefiles to build the DLL_ORB test correctly.

Wed Aug 22 17:20:06 2001  Carlos O'Ryan  <coryan@uci.edu>

        * orbsvcs/orbsvcs/Event/EC_Default_Factory.h:
        * orbsvcs/orbsvcs/Event/EC_Default_Factory.cpp:
          The factory now owns the memory for its orbid_ field, some
          applications pass temporary data in the argc/argv arguments.
          Thanks to J. Russell Noseworthy <rnosewor@objectsciences.com>
          for pointing this out.

Wed Aug 22 14:12:33 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_operation/interceptors_cs.cpp:

          Made the generation of the interceptors arguments()
          method contingent on Any support. If the Any operators
          are not generated by the IDL compiler, CORBA::NO_IMPLEMENT
          is generated for the body of this method. Thanks to
          Gergely Timar <timarok@elender.hu> for reporting this
          bug.

Wed Aug 22 10:35:27 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAOACE_Static.dsw:

          Added dependencies to projects so the entire workspace
          can be batch built. Thanks to Frank Wolf <fwolf@gum.de>
          for pointing out some of these.

Wed Aug 22 08:16:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/orbsvcs/AV/AVStreams_i.h:
        * orbsvcs/orbsvcs/AV/AVStreams_i.i:
        * orbsvcs/orbsvcs/AV/AVStreams_i.cpp:
          Remove TAO_String_Hash_Key class, replace all uses with
          ACE_CString.  Only reason for keeping the TAO_String_Hash_Key
          class was to have a string class with a hash function so it
          could be put in an ACE_Hash_Map_Manager container.  ACE_CString
          already has this.

        * orbsvcs/orbsvcs/AV/Fill_ACE_QoS.h:
          Change include.

Tue Aug 21 22:58:30 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Sequence_T.cpp: Fixed a place where nil () was not made as
          tao_nil (). This should fix many of our build problems.

Tue Aug 21 13:55:31 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_interface.cpp:

          Added prefix "tao_" to the recently added static
          methods in the generated interface _var class, to
          avoid possible name clashes.

        * TAO_IDL/be/be_visitor_interface/interface_ch.cpp:
        * TAO_IDL/be/be_visitor_interface/interface_cs.cpp:

         Cosmetic changes to generated code and comments.

        * tao/CurrentC.cpp:
        * tao/CurrentC.h:
        * tao/CurrentC.i:
        * tao/DomainC.cpp:
        * tao/DomainC.h:
        * tao/DomainC.i:
        * tao/IOPC.cpp:
        * tao/IOPC.h:
        * tao/IOPC.i:
        * tao/MessagingC.cpp:
        * tao/MessagingC.h:
        * tao/Object.h:
        * tao/Object.i:
        * tao/PolicyC.cpp:
        * tao/PolicyC.h:
        * tao/PollableC.cpp:
        * tao/PollableC.h:
        * tao/PortableInterceptorC.cpp:
        * tao/PortableInterceptorC.h:
        * tao/Sequence_T.cpp:
        * tao/TAOC.cpp:
        * tao/TAOC.h:
        * tao/BiDir_GIOP/BiDirPolicyC.cpp:
        * tao/BiDir_GIOP/BiDirPolicyC.h:
        * tao/DynamicAny/DynamicAnyC.cpp:
        * tao/DynamicAny/DynamicAnyC.h:
        * tao/IFR_Client/IFR_BaseC.cpp:
        * tao/IFR_Client/IFR_BaseC.h:
        * tao/IFR_Client/IFR_BasicC.cpp:
        * tao/IFR_Client/IFR_BasicC.h:
        * tao/IFR_Client/IFR_ComponentsC.cpp:
        * tao/IFR_Client/IFR_ComponentsC.h:
        * tao/IFR_Client/IFR_ExtendedC.cpp:
        * tao/IFR_Client/IFR_ExtendedC.h:
        * tao/IORManipulation/IORC.cpp:
        * tao/IORManipulation/IORC.h:
        * tao/IORTable/IORTableC.cpp:
        * tao/IORTable/IORTableC.h:
        * tao/PortableServer/ImplRepoC.cpp:
        * tao/PortableServer/ImplRepoC.h:
        * tao/PortableServer/PortableServerC.cpp:
        * tao/PortableServer/PortableServerC.h:
        * tao/RTCORBA/RTCORBAC.cpp:
        * tao/RTCORBA/RTCORBAC.h:
        * tao/RTPortableServer/RTPortableServerC.cpp:
        * tao/RTPortableServer/RTPortableServerC.h:
        * tao/TypeCodeFactory/TypeCodeFactoryC.cpp:
        * tao/TypeCodeFactory/TypeCodeFactoryC.h:
        * tao/diffs/Current.pidl.diff:
        * tao/diffs/CurrentC.cpp.diff:
        * tao/diffs/CurrentC.h.diff:
        * tao/diffs/CurrentC.i.diff:
        * tao/diffs/DomainC.h.diff:

          Updated these files to reflect the changes made above to
          the IDL compiler, in addition to the recent change of
          adding a static int member to interface classes and
          using it's address in _unchecked_narrow() and
          _tao_Queryinterface, instead of the address of the
          _narrow() method.

Tue Aug 21 09:41:17 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/IORManipulation/TAO_IORManip.dsp:
        * tao/TypeCodeFactory/TypeCodeFactory_Static.dsp:

          Changed the names of the libraries in some configurations
          to be more consistent. Thanks to Frank Wolf <WolfF@Becker.de>
          for pointing these out.

Mon Aug 20 11:27:37 2001  Paul Calabrese <calabrese_p@ociweb.com>

        * tests/Strategies/Strategies.dsw:
        * tests/Strategies/client.dsp:
        * tests/Strategies/server.dsp:
        * tests/Strategies/run_test.pl:

          Added MSVC project files and a run_test.pl for this test.

        * tests/Strategies/advanced_svc.conf:

          Modified this file for use on Windows platforms (commented out
          the UIOP protocol factory).

Mon Aug 20 10:27:56 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/TAO.dsw:

          Added dependencies to several projects. Thanks to
          Frank Wolf <fwolf@gum.de> for pointing out this lack
          in Domain.dsp and Domain_Static.dsp.

Mon Aug 20 09:39:06 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * orbsvcs/IFR_Service/tmplinst.cpp:

          Removed explicit instantiation of some template classes
          that were already present in the TAO_PortableServer
          library, to which the IFR_Service executable is linked.
          These redundancies were caught by the new RedHat static
          build.

Mon Aug 20 08:37:00 2001  Chad Elliott  <elliott_c@ociweb.com>

        * orbsvcs/orbsvcs/Trader/constraint.l:
        * orbsvcs/orbsvcs/Trader/Constraint_l.cpp:

          Explicitly allow a leading sign on integers as specified by
          section B.2.5 of the OMG Trading Service specification.

Sun Aug 19 13:41:07 2001  Ossama Othman  <ossama@uci.edu>

        * tests/DLL_ORB/Test_i.cpp (shutdown):

          Do not wait for completion when shutting down the ORB since
          ORB::shutdown() is being called in the context of an upcall.
          Thanks to Craig Rodrigues <crodrigu@bbn.com> for finding the
          problem and providing the fix.

Sat Aug 18 08:59:32 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/ORBInitInfo.cpp: Fixed a compiler error in the static
          builds.

Sat Aug 18 07:09:09 2001  Douglas C. Schmidt  <schmidt@ace.cs.wustl.edu>

        * Removed the "announcement" file since it was hopelessly
          out of date!  Thanks to Kitty for noticing this.

Fri Aug 17 19:26:00 2001  Craig Rodrigues <crodrigu@bbn.com>

       * orbsvcs/orbsvcs/Property/CosPropertyService_i.h

         Add default values for CORBA::Environment variables.

Fri Aug 17 13:01:36 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.h (TAO_SSLIOP_Endpoint):

          Corrected class documentation.

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.cpp
          (TAO_SSLIOP_Endpoint):

          Initialize the SSL port to zero instead of the IANA assigned
          port of 684 if one wasn't provided.  This is necessary to allow
          INS over plain IIOP to function properly.  Thanks to Dmitry
          Goldshtain <dmitry.goldshtain@iname.com> for providing the fix.

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.cpp (connect):

          Verify that the SSL port is non-zero before attempting to
          establish trust in the target.  Otherwise, throw an
          exception.  It is not possible to establish trust in the target
          without first retrieving the certificate from the target during
          SSLIOP session establishment.

          If the SSL port is zero and protection is desired, then do not
          attempt to make a secure invocation since it is not possible to
          make one.  Instead throw an exception.

          In both of the above cases, the SSL port will be zero if no
          SSLIOP tagged component was available in the IOR.

Fri Aug 17 14:57:18 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/fe/fe_init.cpp:

          Removed code that had been commented out.

Fri Aug 17 14:14:21 2001  Paul Calabrese <calabrese_p@ociweb.com>

        * docs/Options.html:

          Clean up the Advanced_Resource_Factory section.

Fri Aug 17 08:18:40 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/RTPortableServer/TAO_RTPortableServer.dsp:
        * tao/DynamicInterface/TAO_DynamicInterface.dsp:

          Fixed setting in the MFC versions of these projects.
          Thanks to Frank Wolf <fwolf@gum.de> for pointing
          these out.

Fri Aug 17 10:59:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * tao/tests/Makefile.bor
        * tao/tests/DLL_Orb/*.bor
          Added BCB makefiles for this test

Thu Aug 16 10:40:56 2001  Ossama Othman  <ossama@uci.edu>

        * tao/ORBInitInfo.h (TAO_ORBInitInfo):

          The CodecFactory reference is now a "_var" since the
          CodecFactory is now reference counted.

        * tao/ORBInitInfo.cpp (register_initial_reference,
          resolve_initial_references):

          Removed the restriction that prevented
          register_initial_reference() and resolve_initial_references()
          from being called in ORBInitializer::post_init() and
          ORBInitializer::pre_init(), respectively.  The Portable
          Interceptors FTF relaxed those restrictions.

          (codec_factory):

          Since the CodecFactory is now reference counted, create a new
          CodecFactory on the heap and cache it.  Fixes a seg fault
          reported by Greg J. Hall <Greg.Hall@Australia.Boeing.com>.

Thu Aug 16 11:38:02 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/TypeCodeFactory/TypeCodeFactory.dsp:

          Fixed typo in MFC release configuration.

        * tao/TypeCodeFactory/TypeCodeFactory_Static.dsp:
        * tao/TAO.dsw:

          Removed unnecessary "_Static" substring from static
          release and debug library names. Also fixed path for
          TypeCodeFactory_Static.dsp in the workspace. Thanks to
          Frank Wolf <fwolf@gum.de> for pointing these out.

Thu Aug 16 10:04:33 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_interface.cpp:

          Removed an unecessary "::" added to code generation
          in

          Tue Aug 14 10:56:01 2001  Jeff Parsons <parsons@cs.wustl.edu>

          caught by Sun Forte, which has a problem with "(::".

Wed Aug 15 11:37:24 2001  Ossama Othman  <ossama@uci.edu>

        * tao/GIOP_Message_Lite.cpp (process_request):

          Removed statement that initialized a reference to a
          CORBA::Environment from TAO_default_environment().
          TAO_default_environment() incurs a TSS access.  This removes a
          TSS access from the critical path when using TAO's GIOPlite
          implementation.

          Verify that forward object reference was successfully marshaled
          into the CDR stream before continuing.

        * tao/GIOP_Message_Base.cpp (process_request):

          Verify that forward object reference was successfully marshaled
          into the CDR stream before continuing.

        * tao/Object.cpp (operator>>):

          Fixed improper emulated exception code.  It unnecessarily
          introduced a TSS access by using TAO_default_environment().

Wed Aug 15 14:26:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/Single_Threaded_Latency/svc.conf: Add the
          exclusive Transport_Mux_Strategy to this test.

Wed Aug 15 12:11:04 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tests/IDL_Test/constants.idl:

          Dropped the last digit from the mantissa of a constant
          double declaration equal to the IEEE max value for a
          double. This max value has 16 decimal places, while the
          generated output on many platforms has only 15, rounded
          up. The problem arises with gcc 3.0, which outputs a
          warning, since the rounded value is greater than the
          allowed maximum. Thanks to Ossama Othman <ossama@uci.edu>
          for pointing out the problem.

Wed Aug 15 12:06:20 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_State.cpp:
        * tao/GIOP_Message_State.h: Cosmetic changes to the debug
          statements and comments.

Wed Aug 15 11:35:18 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp: Added a debugging statement that is going to
          create noise if TAO receives a ERROR message on a connection.

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_State.cpp: Added debugging statments to inform
          the user, that TAO got a GIOP_ERROR_MESSAGE on a
          connection. Looks like some ORB's send ERROR messages if the
          size of the data exceeded a limit. In such cases it is better
          that the user knows why the test fails or hangs instead of
          digging around the ORB code.

Wed Aug 15 11:25:15 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/DynamicAny/DynCommon.cpp (insert_reference):

          Added a check for a nil object reference before attempting
          to access its repo id and calling _is_a(). Thanks to
          Philippe Merle <Philippe.Merle@lifl.fr> for pointing out
          this bug.

Tue Aug 14 15:16:32 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/SmartProxies/SmartProxies.dsp (CFG): Fixed a problem with
          MFC configuration.

Tue Aug 14 14:50:32 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/SmartProxies/SmartProxies.dsp (RSC): Added MFC
          configurations. Thanks to Garry Shammis <Garry.Shamis@kbcfp.com>
          for giving us the patch.

        * tao/DynamicInterface/DII_Reply_Dispatcher.cpp (dispatch_reply):
        * tao/Synch_Reply_Dispatcher.cpp:
        * tao/Asynch_Reply_Dispatcher.cpp: Made some cosmetic changes.

Tue Aug 14 12:29:08 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Transport.cpp (recv_i):

          Use the timeout value when attempting to read from the SSL
          connection.  Fixes some thread-per-connection problems.  Thanks
          to Bala for pointing out the problem.

Tue Aug 14 14:18:36 2001  Paul Calabrese <calabrese_p@ociweb.com>

        * tao/orbconf.h:

          Add the TAO_ADVANCED_RESOURCE_FACTORY_ARGS flag.  This
          fixes a problem with the TAO_Strategies library when
          TAO_PLATFORM_SVC_CONF_FILE_NOTSUP is defined.

        * docs/configurations.html:

          Change Resource_Factory to Advanced_Resource_Factory in a
          few places.  Clean up some of the text.  Add the
          TAO_ADVANCED_RESOURCE_FACTORY_ARGS flag.

Tue Aug 14 12:10:09 2001  Carlos O'Ryan  <coryan@uci.edu>

        * docs/Options.html:
          Fixed typo.

Tue Aug 14 10:56:01 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_interface.cpp:
        * TAO_IDL/be/be_visitor_operation/operation_ss.cpp:
        * TAO_IDL/be/be_visitor_operation/rettype.cpp:
        * TAO_IDL/be/be_visitor_operation/rettype_vardecl_ss.cpp:

          Added global qualifier '::' to the generation of the
          full names for an operation return type in the stub header
          file, return value allocation in the skeleton body, and
          ancestors in _tao_QueryInterface(). This addition helps
          some compilers when dealing with nested classes. Thanks
          to Gary Duzan <gduzan@bbn.com> for reporting the bug.

        * tests/IDL_Test/nested_scope.idl:

          Added Gary Duzan's example to the test file.

Mon Aug 13 20:09:24 2001  Balachandran  <bala@cs.wustl.edu>

        * TAO version 1.1.19 released.

Mon Aug 13 12:05:17 2001  Chad Elliott  <elliott_c@ociweb.com>

        * orbsvcs/orbsvcs/Notify/Notify_ProxyConsumer_T.cpp:

          Explicitly reference the event_manager_ with this-> to avoid a
          warning on HP-UX with aCC.

Sun Aug 12 17:55:57 2001  Ossama Othman  <ossama@uci.edu>

        * tao/CodecFactory.h (TAO_CodecFactory):
        * tao/Messaging_PolicyFactory.h (TAO_Messaging_PolicyFactory):

          Instances of these objects are now reference counted to make
          sure they exist long enough for the ORB to release them.  This
          fixes some dynamic unloading problems.

        * tao/CodecFactory_ORBInitializer.h:
        * tao/Messaging_ORBInitializer.h:

          Maintain a reference (i.e., a "_var") to the CodecFactory and
          Messaging_PolicyFactory, respectively.  They will be destroyed
          when all outstanding references are released.

        * tao/CodecFactory_ORBInitializer.cpp (pre_init):
        * tao/Messaging_ORBInitializer.cpp (register_policy_factories):

          Initialize a CodecFactory and Messaging_PolicyFactory,
          respectively, on the heap.  These instances are now reference
          counted.

        * tao/default_resource.h (TAO_Default_Resource_Factory):

          No longer any need to inherit from ACE_Cleanup.  The Service
          Configurator will once again manage the default resource
          factory's memory.

        * tao/default_resource.cpp:

          Removed code that was commented out during the previous changes
          to this file.  That code will not be used.

        * tests/DLL_ORB/Test_Client_Module.cpp:

          Removed "tabs" from this source file.

        * tests/DLL_ORB/Test_Server_Module.cpp (init):

          No longer any need to use the "-ORBSkipServiceConfigOpen" option
          (at least for this single-threaded test).

        * tests/DLL_ORB/client.cpp (main):

          Removed debugging ORB debugging options.  They shouldn't be in
          "released" version of this test.

Sun Aug 12 08:46:01 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Native_Exceptions/test_i.cpp: Fixed warning with g++.

Sun Aug 12 00:29:27 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Native_Exceptions/test_i.cpp: Fixed warning with g++.

Sat Aug 11 17:02:44 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_State.cpp:
        * tao/GIOP_Message_State.inl:
        * tao/Incoming_Message_Queue.inl:
        * tao/Transport.cpp:  A chance for us to test the fragmentation
          support on the server side. Fixed quite a few bugs with our
          fragmentation support. The tests were conducted against ORBIX
          2K. The bugs fixed were mostly the way we handle the fragments.

Sat Aug 11 11:34:47 2001  Carlos O'Ryan  <coryan@uci.edu>

        * orbsvcs/orbsvcs/Event_Utilities.cpp (insert):
          Fixed event initialization, one of the insert() methods was
          setting an event before growing the publications list.
          Thanks to Jody Hagins <jody@atdesk.com> for pointing this out.

Sat Aug 11 08:31:37 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/IDL_Test/idl_test.dsp: Fixed the relase builds.

Sat Aug 11 08:24:37 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/Notify_ProxyConsumer_T.cpp:
        * orbsvcs/orbsvcs/Notify_ProxySupplier_T.cpp: Fixed warnings with
          HPUX's aCC.

Sat Aug 11 08:14:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Native_Exceptions/test_i.cpp: Fixed warning with KCC.

Fri Aug 10 10:59:42 2001  Chad Elliott  <elliott_c@ociweb.com>

        * tests/Native_Exceptions/test_i.cpp:
        * tests/RTCORBA/Thread_Pool/server.cpp:

          Changes to avoid warnings on HP-UX with aCC.

Fri Aug 10 08:07:57 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Oneways_Invoking_Twoways/Receiver_i.cpp: Added a debugging
          statement.

Thu Aug  9 16:15:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/RTCORBA/ORB_init/ORB_init.cpp: Fixed a warning in KCC
          builds on Linux.

Thu Aug  9 11:10:59 2001  Ossama Othman  <ossama@uci.edu>

        * tao/default_resource.cpp:

          Temporarily reverted previous change.  The default resource
          factory is still destroyed before the ORB core is destroyed.
          Another solution is needed.

Thu Aug  9 11:46:27 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/CDR.h: Fixed the ambigous constructor problem. The InputCDR
          class had two constructors which were  similar if the default
          values for the arguments were exercised. Thanks to Torbjorn
          Backstrom <torbjorn.k.backstrom@volvo.com>  for reporting this.

Wed Aug  8 22:53:37 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/DLL_ORB.cpp: Fixed a unused variable warning.

Wed Aug  8 16:15:59 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tests/AMI_Buffering/client.cpp:
        * tests/Oneway_Buffering/client.cpp:
          The liveness test was blowing the stack on the server.  On each
          request the server calls the 'admin' interface to report the
          number of bytes received.  While waiting for a response on that
          call the server receives another call (the client is pushing
          oneways or AMI calls like crazy).
          This results in an segfault when the stack is overflowed, now we
          sync the server every so many calls.

        * tests/AMI_Buffering/svc.conf:
          Removed, the default configuration works now.

        * tests/Big_Oneways/Session.cpp:
          Made the test more silent.

Wed Aug 08 16:14:01 2001  Ossama Othman  <ossama@uci.edu>

        * tao/default_resource.h (TAO_Default_Resource_Factory):

          The default resource factory now also inherits from ACE_Cleanup
          so that it may be registered for destruction with the
          TAO_Singleton_Manager.  This generally only necessary for the
          default resource factory since it may be created during ORB
          initialization, and it must exist long enough for the ORB
          Core to make some calls on it during ORB Core finalization.
          Resource factories that are statically or dynamically loaded
          (i.e. registered with the Service Repository) before the ORB is
          initialized need not inherit from ACE_Cleanup, nor is it
          necessary for them to register with the TAO_Singleton_Manager.

        * tao/default_resource.cpp (_make_TAO_Default_Resource_Factory):

          A custom factory definition is used instead of the commonly used
          one created by the ACE_FACTORY_DEFINE macro.  This is necessary
          to get around dynamic unloading issues.  The default resource
          factory must exist long enough for the ORB to make the a
          resource factory reclaim the reactor, for example.  The easiest
          (and probably best) way to do that is register the
          TAO_Default_Resource_Factory for destruction with the
          TAO_Singleton_Manager.  This fixes a segmentation fault (memory
          access violation) that occurred when an ORB that was initialized
          within a dynamically loaded object (e.g. within an
          ACE_Service_Object instance) was finalized/destroyed.

          (ACE_STATIC_SVC_DEFINE):

          Do not set the ACE_Service_Type::DELETE_OBJ flag within the
          TAO_Default_Resource_Factory-specific declaration/expansion of
          this macro.  The "payload" object, i.e. the default resource
          factory, is now managed by the TAO_Singleton manager instead of
          the Service Repository.

        * tao/DLL_ORB.h (TAO_DLL_ORB):

          Updated class documentation to mention the fact that this class
          is deprecated, and why it should not be used.

        * tao/DLL_ORB.cpp (init):

          Added debugging message that mentions that the TAO_DLL_ORB class
          is deprecated.

        * tao/TAO_Internal.cpp (open_i):
        * tao/TAO_Singleton_Manager.h (TAO_Singleton_Manager):

          Updated and corrected class documentation.

        * tests/DLL_ORB/Test_Client_Module.cpp (fini):

          Explicitly release the reference to the target object.  This is
          a bit of a hack.  The ORB Core's lifetime is tied to the
          lifetime of an object reference.  We need to wipe out all object
          references before we call fini() on the TAO_Singleton_Manager.
          This is a tough problem to correct.  Fortunately, this is not a
          problem when a resource factory is loaded into the Service
          Repository before the ORB is initialized, i.e. when the default
          resource factory is not used.

Wed Aug  8 15:03:58 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/DIOP_Transport.cpp (send_request): There was no
          reason to idle the transport after a send. In other protocols,
          where the handles are cached the call idle_after_send () does
          certain things based on the muxed strategy. But for DIOP, we
          dont cache the handles and the call started doing something
          really bad like closing the handles. We now just return a 0
          after sending the request. This should fix the problem in the
          full builds with DIOP.

Wed Aug  8 12:16:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/Thread_Per_Connection_Latency/svc.conf:
        * performance-tests/Thread_Pool_Latency/svc.conf: Made the client
          to use Wait_On_Read_Write as we see a big performance drop if we
          use the Wait_On_LF. To make the test work without any problems,
          we now use the exclusive transport mux strategy. Hopefully this
          will get our performance back to normal.

        * tao/Muxed_TMS.cpp: Fixed a subtle problem in connection_closed
          (). While iterating through the reply_disptacher table to close
          the  reply dispatchers, a wrong comparision was made in the for
          loop. This would not allow any of the reply dispatchers to be
          closed and in turn prevent the transport cleaning up the
          resources. This lead to the client hanging on select () even
          after the connection is closed by the server. I am surprised
          that this has not been found out so far. God save people who
          have been using the Muxed strategy :-).

          BTW, the above fix should fix problems seen in the Faults and
          Crashed_Callback tests in the daily builds.

Tue Aug 07 21:54:31 2001  Ossama Othman  <ossama@uci.edu>

        * tests/IORManipulation/IORTest.cpp (main):

          Updated URL IORs used in test to be of the standard "corbaloc"
          form.  Fixes a CORBA::INV_OBJREF exception.

Tue Aug  7 23:29:34 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/Thread_Per_Connection_Latency/svc.conf:  Did
          the same change as the last checkin, ie. removed the RW
          connection handler. But beware, we *could* have a performance
          drop in our stats page. If wehave performance drop, then we
          change the Transport_Mux_Strategy for the following test. Even
          if we do change the Mux_Strategy and the connection handler to
          RW, it seems that this would be in odd with what we preached a
          few days back at the TAO workshop.

Tue Aug  7 23:02:09 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/Thread_Pool_Latency/svc.conf: Looks like our
          past seems to haunt us. Once upon a time this svc.conf file was
          used in Latency measurements. When the tests were changed to a
          new directory the svc.conf file was copied. The svc.conf file
          had a RW connection handler. That was enough to cause the test
          to hang. Further removed the -ORBReactor type as it was trying
          to load a TP reactor. This should fix the problem in the daily
          builds with this test.

Tue Aug  7 14:37:10 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp (handle_input_i): Added the transport
          information  to a debugging output.

        * tao/orbconf.h: Fixed a simple typo.

Mon Aug  6 09:49:39 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/RTPortableServer/TAO_RTPortableServer.dsp: Fixed a build
          problem in Release builds. Should have nailed the problem by
          now.

Sun Aug  5 16:58:39 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/DLL_ORB/Makefile: Thanks to Ossama for helping to fix the
          the warnings in the daily builds.

Sun Aug  5 16:00:28 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Oneway_Buffering/Test.idl: Changed the oneway in the
          Test.idl to a twoway call.

Sun Aug 05 15:12:31 2001  Balachandran Natarajan <bala@cs.wustl.edu>

        * tao/RTPortableServer/TAO_RTPortableServer.dsp: Fixed a build
          problem in Release builds.

Sun Aug 05 15:03:31 2001  Balachandran Natarajan <bala@cs.wustl.edu>

        * tao/TAO_Static.dsp: Added the new flushing strategies to the
          static builds.

Fri Aug 04 3:33:31 2001  Yamuna Krishnamurthy  <yamuna@cs.wustl.edu>

        * TAO/orbsvcs/orbsvcs/AV/QoS_UDP.h:
        * TAO/orbsvcs/orbsvcs/AV/QoS_UDP.cpp:
        * TAO/orbsvcs/orbsvcs/AV/QoS_UDP.i:

          Added a helper class to create qos sessions, activate qos
          handlers and set qos on a session. Fixed a bug in translation
          from ACE_Flow_Spec to AVStreams::streamQoS. Fixed addressing
          problems.

Fri Aug 03 20:50:42 2001  Ossama Othman  <ossama@uci.edu>

        * tests/README:

          Added summary about the new DLL_ORB test.

Fri Aug 03 20:33:51 2001  Ossama Othman  <ossama@uci.edu>

        * tests/DLL_ORB/Makefile (MAKEFLAGS):

          Force non-parallel build of this test since the IDL file rules
          in both `Makefile.Test_Client_Module' and
          `Makefile.Test_Server_Module' cause corruption of the generated
          stubs and skeletons when these Makefiles are run in parallel.
          Fixes build problems in some of our parallel build enabled
          nightly builds.

Fri Aug 03 10:16:37 2001  Carlos O'Ryan  <coryan@uci.edu>

        * docs/Options.html:
          Add documentation for the new Leader/Followers flushing
          strategy.

Fri Aug  3 10:37:06 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_operation/operation_cs.cpp (visit_operation):

          Added ACE_UNUSED_ARG generation to satisfy strict compilers,
          for the recently modified cases where the operation returns
          a long long or a long double.

Fri Aug  3 07:59:24 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/DLL_ORB/Test_Client_Module.cpp: Changed the order of
          #includes.

Fri Aug  3 07:50:49 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/TAO_Singleton_Manager.h: Changed the #ifndef to
          TAO_SINGLETON_MANAGER_H instead of TAO_OBJECT_MANAGER_H.

Fri Aug  3 07:30:02 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/DLL_ORB/Test_Client_Module.cpp:
        * tests/DLL_ORB/Test_Server_Module.cpp: Added missing
          #includes. This should take care of the problem with g++.

Thu Aug 02 20:39:19 2001  Ossama Othman  <ossama@uci.edu>

        * tao/Stub.h (TAO_Exception_Data):

          The "id" member should be "const char *" not "char *".  Fixes a
          warning exhibited by some strict compilers (e.g. g++ with
          "-pedantic" flag enabled).  Conversion from a string constant to
          a "char *" is deprecated.

Thu Aug  2 17:21:27 2001  Ossama Othman  <ossama@uci.edu>

        * tests/Makefile:

          Added DLL_ORB test directory to the list of directories to
          build.

Thu Aug 02 17:09:53 2001  Ossama Othman  <ossama@uci.edu>

        * tests/DLL_ORB/*:

          New test that dynamically loads a shared object that initializes
          an ORB upon initialization of that shared object, and destroys
          that ORB upon finalization of the shared object.  [Bug 832]

Thu Aug  2 11:44:51 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Leader_Follower.cpp: Added a #include. This should take care
          of the No AMI builds.

Thu Aug  2 10:17:12 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_operation/operation_cs.cpp (visit_operation):

          Recent changes to this method made it necessary now to unalias
          the return type before using it.

Thu Aug  2 09:09:26 2001  Douglas C. Schmidt  <schmidt@tango.doc.wustl.edu>

        * tests/Smart_Proxies/On_Demand/client.cpp (parse_args): Fixed a
          but where there was a missing ":" in the get_opts string.
          Thanks to Goran Lowkrantz <goran.lowkrantz@ismobile.com> for
          reporting this.  Fixes bugid 995.

Wed Aug  1 18:17:38 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tao/PortableServer/ImplRepoC.h:
        * tao/PortableServer/ImplRepoC.i:
        * tao/PortableServer/ImplRepoC.cpp:
        * tao/PortableServer/ImplRepoS.h:
        * tao/PortableServer/ImplRepoS.i:
        * tao/PortableServer/ImplRepoS.cpp:
        * tao/PortableServer/ImplRepoS_T.h:
        * tao/PortableServer/ImplRepoS_T.i:
        * tao/PortableServer/ImplRepoS_T.cpp:
        * tao/PortableServer/diffs/ImplRepo.diff:
          After Jeff's changes it was necessary to regenerate the ImplRepo
          pre-compiled IDL.

        * tao/diffs/ImplRepoC.cpp.diff:
        * tao/diffs/ImplRepoC.h.diff:
        * tao/diffs/ImplRepoC.i.diff:
        * tao/diffs/ImplRepoS.cpp.diff:
        * tao/diffs/ImplRepoS.h.diff:
        * tao/diffs/ImplRepoS.i.diff:
        * tao/diffs/ImplRepoS_T.cpp.diff:
        * tao/diffs/ImplRepoS_T.h.diff:
        * tao/diffs/ImplRepoS_T.i.diff:
          Move diffs to $TAO_ROOT/tao/PortableServer/ImplRepo.diff

Wed Aug  1 16:05:36 2001  Carlos O'Ryan  <coryan@uci.edu>

        * Merged changes from the fix_886 branch

        * tao/Transport.cpp:
          There was a subtle race condition in the handle_output() method:
          the state of the queue was checked in drain_queue(), while
          holding the mutex, if the queue was empty the decision was made
          to call cancel_output().
          However, that was performed *outside* the context of the mutex,
          so another thread could attempt to send data, queue it,
          schedule_output() only to find out that is was cancelled right
          after it did...
          Before the patches below it was not possible to move the
          cancel_output() to the context of the mutex, the ORB would
          deadlock for other reasons. I took us (Bala and myself) a couple
          of days to track this one down, obviously I still don't know how
          to write MT-safe code :-)

        Tue Jul 31 12:55:07 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tao/Follower.h:
        * tao/Follower.inl:
        * tao/Follower.cpp:
        * tao/Follower_Auto_Ptr.h:
        * tao/Follower_Auto_Ptr.inl:
        * tao/Follower_Auto_Ptr.cpp:
        * tao/LF_Follower.h:
        * tao/LF_Follower.inl:
        * tao/LF_Follower.cpp:
        * tao/LF_Follower_Auto_Ptr.h:
        * tao/LF_Follower_Auto_Ptr.inl:
        * tao/LF_Follower_Auto_Ptr.cpp:
        * tao/Makefile:
        * tao/Makefile.am:
        * tao/Makefile.bor:
        * tao/TAO.dsp:
        * tao/TAO_Static.dsp:
          Renamed TAO_Follower to TAO_LF_Follower and
          TAO_Follower_Auto_Ptr to TAO_LF_Follower_Auto_Ptr, I think this
          is more consistent with the other files in the project.

        * tao/LF_Follower_Auto_Adder.h:
        * tao/LF_Follower_Auto_Adder.inl:
        * tao/LF_Follower_Auto_Adder.cpp:
          Remove unused code from the Auto_Adder files.

        * tao/LF_Event.h:
        * tao/LF_Event.inl:
        * tao/LF_Event.cpp:
        * tao/LF_Event_Binder.h:
        * tao/LF_Event_Binder.inl:
        * tao/Leader_Follower.h:
        * tao/Leader_Follower.i:
        * tao/Leader_Follower.cpp:
          Use the new class names and #includes.

        Mon Jul 30 14:41:43 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tao/LF_Follower_Auto_Adder.h:
        * tao/LF_Follower_Auto_Adder.inl:
        * tao/LF_Follower_Auto_Adder.cpp:
          Automatically manipulate the L/F follower set: its constructor
          inserts a follower into the set and the destructor removes it.

        * tao/Leader_Follower.cpp:
          Use the new LF_Follower_Auto_Adder class.

        * tao/Makefile:
        * tao/Makefile.am:
        * tao/Makefile.bor:
        * tao/TAO.dsp:
        * tao/TAO_Static.dsp:
          Add the new files to Makefiles and projects.

        Sun Jul 29 11:58:51 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tao/Leader_Follower.cpp:
          If an error is detected while waiting as a follower the loop
          should return -1.

        * tao/Transport.cpp:
          Merged in some bug fixes from the main trunk.

        Sun Jul 29 09:00:57 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tao/LF_Event.cpp:
          Fixed the error_detected() function

        Fri Jul 27 17:34:40 2001  Carlos O'Ryan  <coryan@cs.wustl.edu>

        * tao/Leader_Follower.h:
        * tao/default_resource.cpp:
          Fixed warnings and compilation errors for gcc-2.7.2

        Fri Jul 27 10:59:50 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tao/LF_Event.h:
        * tao/LF_Event.inl:
          Add new method to unbind a LF_Event and its Follower:
          Reply_Dispatchers can be used multiple times to wait for several
          replys, mostly when a LOCATION_FORWARD message is received.

        * tao/LF_Event.cpp:
          Modify the state machine: the state can go back to ACTIVE from
          CONNECTION_CLOSED or SUCCESSFUL.  This represents the location
          forward scenario described above.

        * tao/LF_Event_Binder.h:
        * tao/LF_Event_Binder.inl:
        * tao/LF_Event_Binder.cpp:
          Helper class to automate the bind/unbind calls to a LF_Event.

        * tao/Leader_Follower.cpp:
          Use LF_Event_Binder to handle the bind/unbind calls into the
          LF_Event.

        * tao/Makefile:
        * tao/Makefile.bor:
        * tao/TAO.dsp:
        * tao/TAO_Static.dsp:
          Add new files to the projects and Makefiles.

        * tao/Invocation.cpp:
        * tao/Wait_On_Read.cpp:
        * tao/Wait_On_Reactor.cpp:
        * tao/Synch_Reply_Dispatcher.h:
        * tao/Synch_Reply_Dispatcher.cpp:
          Remove the reply_received() flag from Synch_Reply_Dispatcher,
          the LF_Event state is enough to know what happens.

        Thu Jul 26 18:00:12 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tao/Strategies/advanced_resource.cpp:
          Fixed typo in last commit.

        Thu Jul 26 16:50:46 2001  Carlos O'Ryan  <coryan@uci.edu>

        * Part of the fixes for
          http://ace.cs.wustl.edu/bugzilla/show_bug.cgi?id=886

          the changes also close the following bug:

          http://ace.cs.wustl.edu/bugzilla/show_bug.cgi?id=296

        * tao/Leader_Follower_Flushing_Strategy.h:
        * tao/Leader_Follower_Flushing_Strategy.cpp:
          New flushing strategy that participates in the Leader/Followers
          protocol.
          To support this several changes to the Leader/Followers
          implementation were required.  The most important involved using
          some abstract representation for the events that the
          Leader/Followers wait for, in the old days there were only reply
          events, so there was no need to abstract anything, but now the
          Leader/Followers set can wait for both 'message flushed' events,
          as well as 'reply received'.
          With this explicit representation for events at hand it was
          easier to encapsulate the Leader/Followers wait loop in
          TAO_Leader_Follower class, instead of hidden in
          Wait_On_Leader_Follower.
          To match the events that L/F waits for and the threads waiting
          for them we addd a class that represents a Follower thread.
          These TAO_Follower objects had to implement an intrusive list
          for fast addition into the follower set, once that intrusive
          list was implemented adding a free list was trivial, and thus we
          could solve bug 296 easily too.

        * tao/Asynch_Queued_Message.cpp:
        * tao/Synch_Queued_Message.cpp:
          Use the TAO_LF_Event methods to signal any waiters when the
          state changes.

        * tao/Follower.h:
        * tao/Follower.inl:
        * tao/Follower.cpp:
          This class represents a thread playing the Follower role.  It
          contains the condition variable used by the thread.
          The class provides the necessary hooks to implement an intrusive
          linked list.

        * tao/Invocation.cpp:
          The waiting strategy wants the complete Synch_Reply_Dispatcher,
          not just the reply_received flag.

        * tao/LF_Event.h:
        * tao/LF_Event.inl:
        * tao/LF_Event.cpp:
          New class to represent events that the Leader/Followers loop
          waits for.  Used as a base class for both TAO_Queued_Message and
          for TAO_Synch_Reply.

        * tao/LF_Event_Loop_Thread_Helper.h:
        * tao/LF_Event_Loop_Thread_Helper.inl:
        * tao/LF_Event_Loop_Thread_Helper.cpp:
          Move helper class to its own file, no sense in exposing it to
          everybody through the Leader_Follower.h file.

        * tao/Leader_Follower.h:
        * tao/Leader_Follower.i:
        * tao/Leader_Follower.cpp:
          Add free list for TAO_Follower, as well as allocation and
          deallocation methods.
          Move Leader/Followers main loop to this class.
          Move LF_Strategy and friends to their own files.

        * tao/ORB_Core.h:
        * tao/ORB_Core.i:
        * tao/ORB_Core.cpp:
          Removed the TSS Leader/Followers condition variable, the
          Leader/Followers free list implements the same optimization with
          less problems (i.e. without bug 296).

        * tao/Queued_Message.h:
        * tao/Queued_Message.cpp:
        * tao/Synch_Reply_Dispatcher.h:
        * tao/Synch_Reply_Dispatcher.cpp:
          This class derives from TAO_LF_Event now.  Any state or methods
          required to detect timeouts, closed connections or transmition
          errors are in the base class.

        * tao/Reply_Dispatcher.h:
        * tao/Asynch_Reply_Dispatcher.h:
        * tao/Asynch_Reply_Dispatcher.cpp:
          Remove the dispatcher_bound() calls, they are no longer required
          to match follower threads and their reply dispatchers, this is
          now done in the TAO_LF_Event::bind() method, called from
          TAO_Leader_Follower::wait_for_event()

        * tao/Transport.h:
        * tao/Transport.cpp:

        * tao/Transport_Mux_Strategy.h:
        * tao/Transport_Mux_Strategy.cpp:
        * tao/Muxed_TMS.cpp:
        * tao/Exclusive_TMS.cpp:
          Since there is no need to call dispatcher_bound() anymore the
          bind_dispatcher() methods were simplified.

        * tao/Wait_On_Leader_Follower.h:
        * tao/Wait_On_Leader_Follower.cpp:
        * tao/Wait_On_Reactor.h:
        * tao/Wait_On_Reactor.cpp:
        * tao/Wait_On_Read.h:
        * tao/Wait_On_Read.cpp:
        * tao/Wait_Strategy.h:
        * tao/Wait_Strategy.cpp:
          Use a TAO_Synch_Reply_Dispatcher to wait for a reply.  The hack
          using a reply_received flag + a cond.var. was too ugly, plus it
          was tightly coupling the Leader/Followers loop to the reply
          dispatching logic.

        * tao/default_resource.h:
        * tao/default_resource.cpp:
          Made Leader_Follower_Flushing_Strategy the default.

        * tao/orbconf.h:
        * tao/default_client.cpp:
          Made Muxed_TMS the default

        * tao/LF_Strategy.h:
        * tao/LF_Strategy.inl:
        * tao/LF_Strategy.cpp:
        * tao/LF_Strategy_Complete.h:
        * tao/LF_Strategy_Complete.inl:
        * tao/LF_Strategy_Complete.cpp:
          Move the LF_Strategy classes to their own files, no sense in
          exposing them to everybody through the Leader_Follower.h file.

        * tao/Follower_Auto_Ptr.h:
        * tao/Follower_Auto_Ptr.inl:
        * tao/Follower_Auto_Ptr.cpp:
          Helper class to automatically allocate and deallocate
          TAO_Follower objects from the Leader/Followers set.

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_Lite.cpp:
        * tao/Reactor_Registry.cpp:
          Must #include the "LF_Strategy.h" file explicitly.

        * tao/TAO.dsp:
        * tao/TAO_Static.dsp:
        * tao/Makefile:
        * tao/Makefile.bor:
        * tao/Strategies/TAO_Strategies.dsp:
        * tao/Strategies/TAO_Strategies_Static.dsp:
        * tao/Strategies/Makefile:
        * tao/Strategies/Makefile.bor:
          Add new files to the projects and Makefile


        * tao/Strategies/advanced_resource.cpp:
        * tao/Strategies/LF_Strategy_Null.h:
        * tao/Strategies/LF_Strategy_Null.inl:
        * tao/Strategies/LF_Strategy_Null.cpp:
          Move the Null Leader/Follower Strategy to the TAO_Strategies
          library, it was in TAO, but was only used here.

        * tao/RTPortableServer/TAO_RTPortableServer.dsp:
          Fixed missing libraries in link line.

        * tao/TAO.dsw:
          Add missing dependencies for RTPortableServer and RTCORBA


Wed Aug  1 13:15:10 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_operation/ami_exception_holder_operation_cs.cpp:
        * TAO_IDL/be/be_visitor_operation/exceptlist_cs.cpp:
        * TAO_IDL/be/be_visitor_operation/interceptors_cs.cpp:
        * TAO_IDL/be/be_visitor_operation/interceptors_exceptlist.cpp:
        * TAO_IDL/be/be_visitor_operation/interceptors_ss.cpp:
        * tao/Stub.h:
        * tao/Invocation.cpp:

          Modified the struct TAO_Exception_Data to contain the repository id
          instead of the type code, since the type code was used only to
          furnish the repo id string. The twoway invoke() method and generated
          code have been modified accordingly. Also, the interceptor method
          exceptions() now creates a simple array of type codes instead of
          an array of TAO_Exception_Data structs. If type codes are suppressed
          in generated code, the interceptor methods result() (which returns an
          Any) and exception() (from which the spec requires a list of type
          codes as the return value) will throw CORBA::NO_IMPLEMENT(). If only
          Anys are suppressed in generated code, only result() will throw the
          exception. Thanks to Russell Mora <rd.mora@econz.co.nz> for reporting
          the clash of type code suppression and exceptions, and to Ossama Othman
          <othman@cs.wustl.edu> and Carlos O'Ryan <coryan@ece.uci.edu> for
          their suggestions.

Wed Aug  1 11:32:26 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/UIOP_Connection_Handler.cpp:Fixed a compile error
          with g++.

Wed Aug  1 09:36:37 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_operation/operation_cs.cpp:

          Fixed an ACE_CHECK_RETURN in generated operation body where
          the return type is CORBA::LongLong or CORBA::LongDouble.

Wed Aug 1 8:38:06 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/orbconf.h: Removed the #define TAO_DOESNT_YADA_YADA as there
          is no use for it.

        * tao/Connection_Handler.h: Removed the fetch_handle () as there
          is no use for it.

        * tao/IIOP_Connection_Handler.h:
        * tao/IIOP_Connection_Handler.cpp:
        * tao/Strategies/DIOP_Connection_Handler.cpp:
        * tao/Strategies/DIOP_Connection_Handler.h:
        * tao/Strategies/SHMIOP_Connection_Handler.h:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp:
        * tao/Strategies/UIOP_Connection_Handler.h:
        * tao/Strategies/UIOP_Connection_Handler.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp: Removed
          the implementation of fetch_handle (). Thanks to Carlos, the
          foot-print police!!, for motivating this change.

Wed Aug 1 8:02:06 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/IIOP_Connection_Handler.h:
        * tao/IIOP_Connection_Handler.cpp: Removed the flag <resume_flag_>
          that was  introduced so that the input and output datapath would
          behave differently. The output data path would allow the reactor
          to resume the handler whereas the input data path would resume
          the handle. This was creating more confusion than
          necessary. Hence zapped the the flag and made the input and
          output data path consistent. Thanks to Carlos for motivating
          this stuff.

        * tao/Strategies/DIOP_Connection_Handler.cpp:
        * tao/Strategies/DIOP_Connection_Handler.h:
        * tao/Strategies/SHMIOP_Connection_Handler.h:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp:
        * tao/Strategies/UIOP_Connection_Handler.h:
        * tao/Strategies/UIOP_Connection_Handler.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp: Applied the
          same changes to the above mentioned protocols.

Tue Jul 31 12:53:06 2001  Carlos O'Ryan  <coryan@uci.edu>

        * tao/Transport.cpp (register_handler):
          Fixed race condition, the register_handler_i() method may use
          the connection_handler_, but we do not check if it is nil before
          calling.

Tue Jul 31 13:40:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Oneways_Invoking_Twoways/client.dsp: Fixed a compile error
          in release builds.

Tue Jul 31 10:39:18 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tests/IDL_Test/gperf.idl:

          New file in IDL_Test containing examples sent in by
          Karl Proese <karl.proese@mchp.siemens.de> and Vsevolod Novikov
          <novikov@df.nnov.rfnet.ru> that uncovered bugs in gperf.

        * tests/IDL_Test/idl_test.dsp:
        * tests/IDL_Test/Makefile:
        * tests/IDL_Test/Makefile.bor:

          Updated project and makefiles to include the new generated files.

Tue Jul 31 08:25:46 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/fe/idl.ll:
        * TAO_IDL/fe/idl.yy:
        * TAO_IDL/fe/lex.yy.cpp:
        * TAO_IDL/fe/lex.yy.cpp.diff:
        * TAO_IDL/fe/y.tab.cpp:
        * TAO_IDL/fe/y.tab.cpp.diff:
        * TAO_IDL/fe/y.tab.h:

          Removed OBV-related token IDL_INIT and replaced it with
          IDL_FACTORY, which is returned when the string 'factory' is seen
          in an IDL file. This string signals the declaration of an value
          type's initializing member function. Also changed the production
          rule 'init_decl' to be closer to the correct grammar for this
          type of function, although this feature is still completely
          unimplemented. And finally, modified two production rules to
          eliminate long-standing shift/reduce error messages when
          generating y.tab.* from idl.yy. Thanks to Russ Noseworthy
          <j.russell.noseworthy@objectsciences.com> for reporting that an
          'init' identifier string in an IDL file caused an error when
          compiled with the -Gv option (value types) enabled.

Tue Jul 31 06:58:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        This checkin is for fixing a race condition while trying to
        manipulate the number of upcalls. This was not a problem before
        575 fix, as the manipulation was done when there was an implicit
        synchronisation in the TP Reactor. As the implicit synchronisation
        has been broken, we had a race condition. The surpsising element
        was the fact that it took sometime to figure out this race
        condition. We have now added a lock that will be held by the
        thread  before the variable is manipulated.

        * tao/Connection_Handler.cpp:
        * tao/Connection_Handler.h:
        * tao/Connection_Handler.inl: Added a lock to the class. Also
          added three methods, incr_pending_upcalls (),
          decr_pending_upcalls () and pending_upcalls (). The first two
          does the manipulation of the pending_upcalls_ variable after
          holding the lock.

        * tao/IIOP_Connection_Handler.cpp:
        * tao/IIOP_Connection_Handler.h: Removed he peding_upcalls_
          variable and called the incr_pending_upcalls () and
          decr_pending_upcalls () to achieve what needs to be done.

        * tao/Strategies/DIOP_Connection_Handler.cpp
        * tao/Strategies/DIOP_Connection_Handler.h
        * tao/Strategies/SHMIOP_Connection_Handler.cpp
        * tao/Strategies/SHMIOP_Connection_Handler.h
        * tao/Strategies/UIOP_Connection_Handler.cpp
        * tao/Strategies/UIOP_Connection_Handler.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp:
          Replicated the changes from IIOP to the above protocols.

Sun Jul 29 19:31:34 2001  Douglas C. Schmidt  <schmidt@tango.doc.wustl.edu>

        * tao/RTCORBA/RT_Mutex.cpp (try_lock): Since we fixed the timed
          ACE_OS::mutex_lock() to set errno to ETIME we can cleanup the
          code here to remove the checks for errno == EBUSY and errno ==
          ETIMEDOUT.

Sun Jul 29 20:00:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * tao/PortableServer/ORB_Manager.h:

          Improved comments and converted to doxygen format.

Sun Jul 29 10:15:13 2001  Ossama Othman  <ossama@uci.edu>

        * tao/TAO_Internal.cpp (open_services_i):

          Reverted my change that prevented the default resource factory
          from being inserted into the Service Repository.  The resource
          factory must be inserted into the Service Repository so that
          Service Configurator directives such as 'static Resource_Factory
          "-ORBResources global"' actually work properly.  An alternative
          solution for the dynamic loading problem related the default
          resource factory is necessary.

Sun Jul 29 00:59:28 2001  Ossama Othman  <ossama@uci.edu>

        * docs/Options.html:

          Updated documentation for the "-ORBSkipServiceConfigOpen" ORB
          option.  It is no longer necessary the Service Configurator is
          now reentrant and thread-safe.  This option is deprecated, and
          will be removed in releases of TAO in the near future.

Sun Jul 29 00:22:30 2001  Ossama Othman  <ossama@uci.edu>

        * examples/Simple/time-date/svc.conf:

          The support for nested Service Configurator directive processing
          that was just added to ACE made it possible to greatly simply
          this `svc.conf' file.  It is no longer necessary to pre-load
          services that are by default loaded by the ORB.  It is also no
          longer necessary to use the "-ORBSkipServiceConfigOpen" ORB
          option since the Service Configurator is now
          reentrant/thread-safe.

          Instead of the nine Service Configurator directives that were
          listed in this file, only three of them are now necessary.

Sat Jul 28 23:29:21 2001  Ossama Othman  <ossama@uci.edu>

        * tao/TAO_Internal.cpp (open_services_i):

          We cannot insert the default resource factory into the Service
          Repository before the ORB is created since it will be finalized
          before the ORB is finalized.  The ORB requires that a resource
          factory exist in order to reclaim the reactor during
          finalization.

          This was only a problem when the ORB was dynamically
          loaded/unloaded.

Wed Jul 25 23:48:58 2001  Douglas C. Schmidt  <schmidt@ace.cs.wustl.edu>

        * tests/Blocking_Sync_None/client.cpp (main): Changed an
          error messages so that the "Right Thing[TM]" will happen when
          this test fails, i.e., the build system will automagically
          detect it.  Thanks to Johnny Willemsen for reportng this.

Fri Jul 27 21:48:12 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/Cubit/TAO/IDL_Cubit/run_test.pl: Removed all
          the references to the GIOP lite protocol from this script.

        * performance-tests/Cubit/TAO/IDL_Cubit/run_test_lite.pl: We dont
          seem to be having any sort of tests that are run for GIOP
          Lite. Now we will run this in our nightly builds. This test uses
          IIOP_Lite & UIOP_Lite to run the IDL_Cubit test. Having a
          protocol that we are not sure whether it works may be a bad
          idea.

Fri Jul 27 21:38:58 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp: Added a TAO_debug_level guard around
          the places where dump_msg () is called. Though dump_msg ()
          prints out information only if the TAO_debug_level is set, this
          extra guard will prevent us from calling this method when we are
          trying to do performance measurements.

        * tao/GIOP_Message_Lite.cpp: For some reason this class never had
          a hexdump. Added the hexdump in dump_msg ().

Fri Jul 27 16:33:24 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/DIOP_Connection_Handler.cpp:
        * tao/Strategies/UIOP_Connection_Handler.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp: Enabled GIOP lite
          flag to pass through to their transport object.

Fri Jul 27 16:32:46 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/IIOP_Connection_Handler.cpp: Enabled GIOP lite flag to pass
          through to their transport object.

        * tao/GIOP_Message_Lite.cpp: Transfered the reply that has been
          received to the another CDR which is used to dispatch the
          reply. This was actually fixed in my branch. Looks like this got
          missed when the branch was merged to the main trunk.  Added some
          cosmetic fixes to the debugging output. Thanks to Paul Calabrese
          for alerting me about this miss.

Fri Jul 27 14:43:46 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/BiDirectional_NestedUpcall/svc.conf: Removed from the
          repository.  It is no longer needed as this test will work with
          a TP Reactor (the default one).

Fri Jul 27 12:58:51 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp: Removed a debug statement.

Fri Jul 27 12:25:49 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp: Reverted the change made in this Wed Jul 25
          22:10:21 2001  Balachandran Natarajan  <bala@cs.wustl.edu>. We
          need to resume the handle as soon the reply is ready for
          dispatching.  The situation where it can create problems can be
          easily seen in $TAO_ROOT/tests/LongUpcalls/run_ami_test.pl.

          Further, the fix that was done earlier was to resume the handle
          after dispatching the reply. It was thought that it would fix a
          race condition. I had a long discussion with Irfan yesterday,
          and looks like the possibility of a race condition is not
          there at all.

Fri Jul 27 11:09:25 2001  Chad Elliott  <elliott_c@ociweb.com>

        * tests/RTCORBA/Banded_Connections/bands.hpux:

          Added to fix the runtime problem on HP-UX.

        * tests/RTCORBA/Banded_Connections/run_test.pl:
        * tests/RTCORBA/Client_Propagated/client.cpp:
        * tests/RTCORBA/Client_Propagated/server.cpp:
        * tests/RTCORBA/MT_Client_Protocol_Priority/run_test.pl:
        * tests/RTCORBA/Server_Declared/run_test.pl:

          Corrections to fix some of the RTCORBA runtime test problems for
          HP-UX.  These tests still have some problems (especially with
          SHMIOP).

Fri Jul 27 10:50:42 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tests/Smart_Proxies/Benchmark/Makefile:
        * tests/Smart_Proxies/Policy/Makefile:

          Replaced VLDLIBS with TAO_SRVR_LIBS for the server
          target in each Makefile, in order to pull in
          TAO_PortableServer. Thanks to Ekkehard Hoffmann
          <ehoffman@fzi.de> for pointing out the link errors.

Thu Jul 26 22:41:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

       * orbsvcs/orbsvcs/AV/AVStreams_i.h:

         Make TAO_FlowConnection and TAO_MMDevice virtually
         inherit TAO_PropertySet.  Thanks to Rob Ruff <rruff@scires.com>
         for pointing this out.

Thu Jul 26 21:48:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/tests/AVStreams/mpeg/*: REMOVED

          Example uses archaic version of AVStreams and does not
          work well.

        * orbsvcs/tests/AVStreams/mpeg/README.uav: (added)

Thu Jul 26 17:12:00 2001  Ossama Othman  <ossama@uci.edu>

        * tao/default_resource.cpp (get_parser_names):

          Fixed problem where the FILE and DLL parser Service Objects were
          not inserted into the parser registry.  This problem surfaced
          when attempting to dynamically load the ORB, and manifested
          itself as a CORBA::INV_OBJREF exception when attempting to
          destringify an IOR such as "file://foo.ior".  [Bug 744]

Thu Jul 26 09:44:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/orbsvcs/AV/QoS_UDP.cpp:

          Hide more debugging messages behind: if( TAO_debug_level > 0)

Thu Jul 26 07:37:29 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/RTCORBA/Server_Protocol/server.cpp:
        * tests/RTCORBA/Client_Propagated/server.cpp:
        * tests/RTCORBA/Thread_Pool/server.cpp: Fixed warnings in g++.

Wed Jul 25 23:37:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/tests/AVStreams/Full_Profile/Makefile:
        * orbsvcs/tests/AVStreams/Latency/Makefile:
        * orbsvcs/tests/AVStreams/Modify_QoS/Makefile:
        * orbsvcs/tests/AVStreams/Multicast/Makefile:
        * orbsvcs/tests/AVStreams/Multicast_Full_Profile/Makefile:
        * orbsvcs/tests/AVStreams/Pluggable/Makefile:
        * orbsvcs/tests/AVStreams/Simple_Three_Stage/Makefile:
        * orbsvcs/tests/AVStreams/Simple_Two_Stage/Makefile:
        * orbsvcs/orbsvcs/Makefile.av:

          Correct link flags so that ACE_QoS is linked in when
          rapi=1 is specified in platform_macros.GNU.

Wed Jul 25 22:45:10 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/SHMIOP_Transport.cpp:
        * tao/Strategies/DIOP_Transport.cpp: Fixed a compile error that
          came up from my previous change.

Wed Jul 25 22:39:32 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/RTCORBA/Server_Protocol/server.cpp:
        * tests/RTCORBA/Client_Propagated/server.cpp:
        * tests/RTCORBA/Thread_Pool/server.cpp: Added checks & debugging
          statments for a null RootPOA. thanks to Johnny Willemsen for
          pointing this out. Did some minor cosmetic fixes.

Wed Jul 25 22:10:21 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp:
        * tao/Transport.h: Fixed a subtle problem that seems to have lead
          to the  Muxing tests failing randomly. The problem is something
          like this
          - multiple client threads can try to share a connection
          - because of the above, more than one message are sent on the
            same connection
          - if the server is multi-threaded, the messages can be processed
            concurrently
          - there may be a possibility of more than two replies coming on
            the same connection.
          - one of the client threads can pick up both the replies
          - one of the replies would be queued up and the first one can be
            its own
          - after queueing up the second it would wake up another thread
          - if the woken up thread does not own the reply, it could just
            take the reply and try to transfer ownership to the right
            thread.
          - before the second thread transfers the reply, teh second
            thread would have resumed the handler and because of which one
            of the threads would have gone into the reactor from the LF.
          - at exactly the same instant the seccond thread will have
            difficulty in waking up the thread on select () is it is the
            owner.
          Fixed this problem by not resuming the handle till we dispatch
          the reply. We dont buy anything by resuming the handle before
          dispatching the reply because, the dispatching will not be
          unbounded. The forces that apply to the server thread, which
          resumes the handle before making an upcall does not apply to the
          client threads that reads and processes replies. This fix should
          ideally fix the Muxing test failure on different paltforms. If
          it doesnt, it will atleast prevent the race condition outlined
          above :-)

Wed Jul 25 20:33:21 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * examples/Simple/time-date/Makefile.bor:
        * examples/Simple/time-date/server.bor:
        * examples/Simple/time-date/time_date.bor: Fixed Borland builds
          for this example. This commit is for Johnny Willemsen who is
          away from his work.

Wed Jul 25 12:50:00 2001  Michael Kircher <Michael.Kircher@mchp.siemens.de>

        * tao/Strategies/DIOP_Factory.cpp:

          Changed the return value of requires_explicit_endpoint () to 0
          and documented that this return code is not reflecting that
          the endpoints are not cleaned-up but that we disable it by default
          because DIOP is only suitable for certain use cases, e.g. it only
          supports one-ways.

Wed Jul 25 08:41:39 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_interface/interface_cs.cpp:

          Fixed formatting in generation of _unchecked_narrow().

Tue Jul 25 01:00:00 2001  Craig Rodrigues <crodrigu@bbn.com>

        * orbsvcs/tests/AVStreams/Multicast/run_test.pl:
        * orbsvcs/tests/AVVStreams/Asynch_Three_Stage/run_test.pl:
        * orbsvcs/tests/AVStreams/Asynch_Three_Stage/input:

         Increase the times that the perl scripts expecting the
         the CORBA processes to run for.  Decrease the size
         of the Asynch_Three_Stage input file by 4000 lines.

Tue Jul 24 18:16:13 2001  Krishnakumar B  <kitty@cs.wustl.edu>

        * tests/Exposed_Policies/Policy_Verifier.cpp (init):
        * tests/Exposed_Policies/RT_Properties.cpp:

          Changed handling of the default argument values. Added options
          "-BaseObjectIOR" and "-OverriddenIOR" to enable waiting on the IOR
          file instead of sleeping in the run_test.pl

        * tests/Exposed_Policies/run_test.pl:

          Fixed priorities to handle Tru64. Added the new options
          mentioned above to specify the IOR.

        * tests/Exposed_Policies/POA.cfg:
        * tests/Exposed_Policies/Object.cfg:

          Since the IOR files are passed from the command-line, remove
          them from here.

        * tests/Exposed_Policies/POA.cfg.tru64: *
          tests/Exposed_Policies/Object.cfg.tru64:

          Added new files with proper priority values for Tru64.

          OCI folks, you want to change similarly for HP-UX.


Tue Jul 24 11:12:25 2001  Ossama Othman  <ossama@uci.edu>

        * tao/IIOP_Acceptor.cpp (create_shared_profile,
          create_new_profile):
        * tao/Strategies/DIOP_Acceptor.cpp
          (create_shared_profile, create_new_profile):
        * tao/Strategies/SHMIOP_Acceptor.cpp (create_profile):
        * tao/Strategies/UIOP_Acceptor.cpp (create_profile):

          Do not add any tagged components to the profile if an IIOP 1.0
          endpoint/profile is being created.  Tagged components were
          introduced in IIOP 1.1.  The same convention is adopted for the
          other pluggable protocols distributed with TAO (UIOP, SHMIOP,
          and DIOP).  SSLIOP already does this check since it requires
          tagged components in order to convey security association
          information to the client.  These changes address
          interoperability issues.

Tue Jul 24 12:33:17 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Big_Reply/Client_Task.cpp (validate_connection): Some
          cosmetic changes.

Tue Jul 24 08:38:14 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * orbsvcs/examples/Security/SecurityLevel1/SLevel1_Test.idl:

          Removed 'void' parameter from two operations. Thanks to
          Ugendreshwar Kudupudi <ugenderk@rediffmail.com>
          for reporting the bug.

Mon Jul 23 22:31:18 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Big_Reply/Client_Task.cpp: Added some debufg
          statments. Further fixed a small goof up. A method to
          validate connections was added but never called :(. Fixed that
          by calling validate_connection () before an actual call to the
          server object.

Mon Jul 23 22:24:05 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * examples/Simple/time-date/svc.conf: Removed the entry for
          TAO_RT_Protocol_Hooks. Looks like they have taken some other
          form after the RTCORBA subsetting effort. The test were failing
          in the daily builds but never came up on the scoreboard.

Mon Jul 23 21:40:38 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Lite.cpp: Fixed a warning in Win32 builds.

Mon Jul 23 1:44:23 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/TAO_Static.dsp: Added GIOP_Lite files to the builds.

Mon Jul 23 11:44:30 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        Support for GIOP Lite and GIOP Fragments are being added in this
        checkin. Merged from the branch giop_lite_fragment.

        * tao/Makefile:
        * tao/Makefile.bor: Added GIOP Lite  files to the list of files.

Sat Jul 14 16:42:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/DIOP_Transport.cpp:
        * tao/Strategies/SHMIOP_Transport.cpp:
        * tao/Strategies/UIOP_Transport.cpp: Added support for GIOP lite.

Fri Jul 13 16:54:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Pluggable_Messaging.h:
        * tao/GIOP_Message_Base.h:
        * tao/GIOP_Message_Base.cpp:
        * tao/Transport.cpp:
        * tao/Transport.h:
        * tao/Transport.inl: Added support for GIOP fragments. The GIOP
          fragmentation supportseems to have some copying and allocation
          overhead. Need to look into this when it gets
          important. Further, the fragmentation support hasnt been
          tested at all as we have no way of testing it.

        * tao/IIOP_Transport.cpp: Enabled GIOP lite support
        * tao/TAO.dsp: Added GIOP lite files back to the builds.

        * tao/Connection_Handler.h:
        * tao/orbconf.h: Moved some of the #defines from
          Connection_Handler.h to orbconf.h.

Mon Jul  9 10:23:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Generator_Parser_12.cpp: The long talked about
          alignment for the LocateReply messages have been removed. We
          dont align the messages on an 8 byte boundary as described by
          the CORBA 2.4 spec. This is one of the urgent resolutions
          adopted by the OMG.

Mon Jul  9 09:40:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Lite.h:
        * tao/GIOP_Message_Lite.cpp: Added support for GIOP lite. The
          implementation now  is similar to the regular GIOP.

        * tao/GIOP_Message_Base.cpp: Used the payload_size () in
          GIOP_Message_State  for claculating the payload. We were using
          the message_size () and then subtracting the length of the GIOP
          header in consolidate_node (). This looked cumbersome.



Mon Jul 23 07:46:30 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/orbconf.h: Enabled DIOP for minimum CORBA.

Sun Jul 22 21:53:23 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * $ACE_ROOT/auto_run_tests.lst: Disabled Exposed_Policies test for
          single threaded configurations and Linux.  This test requires
          multiple threads and support for multiple native priorities.

        * tests/Exposed_Policies/Policy_Tester.cpp (Policy_Tester): Fixed
          several things:

          - No need for the Policy_Tester::~Policy_Tester to call
          shutdown().  This is already done by the servant.

          - No need for TRY/CATCH blocks in each function.  This
          unnecessarily stops any exceptions from propagating to higher
          layers.

          - Fixed return values and added error checking.

          - No need to call destroy() on the POAs. ORB::shutdown is
          enough.

          - Removed unnecessary default values from the function
          prototypes.

        * tests/Exposed_Policies/run_test.pl:
        * tests/Exposed_Policies/POA.cfg:
        * tests/Exposed_Policies/Object.cfg:

          Updated config files so that the priorities would be ok for
          Win32.

        * tests/Exposed_Policies/server.cpp: Improved error checking.

        * tests/Exposed_Policies/server.conf: Removed Thread Pool Reactor
          directive from the service config file.

Sun Jul 22 16:44:16 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Big_Reply/Client_Task.cpp:
        * tests/Big_Reply/Client_Task.h: Fixed a compile error in no
          exception builds.

Sun Jul 22 13:36:21 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/orbsvcs/AV/RTP.cpp:
          If a TAO_AV_frame_info is not specified in the
          TAO_AV_RTP_Object::sendframe() method, insert a
          timestamp into the RTP header based on the current system
          time.  The timestamp will be in milliseconds.
          Thanks to Rob Ruff <rruff@scires.com> for pointing this out.

Sun Jul 20 12:30:00 2001  Michael Kircher <Michael.Kircher@mchp.siemens.de>

        * tao/Asynch_Invocation.cpp:
        * tao/Asynch_Invocation.i:

          Changed the code which sets up a reply dispatcher for
          AMI calls to only set up one, if a non-nil reply handler
          got registerd. This change is related to my change on
          Fri Jul 20 08:10:00 2001. It assumes that the ORB
          will drop replies to which no reply dispatcher is registered.

Sun Jul 22 09:43:09 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Big_Reply/Test.idl: Added a no-op ping () operation which
          is used to setup a connection properly.

        * tests/Big_Reply/Client_Task.cpp:
        * tests/Big_Reply/Client_Task.h: Called the ping () method before
          getting the replies.

        * tests/Big_Reply/Big_Reply_i.h:
        * tests/Big_Reply/Big_Reply_i.cpp: Implementation for the ping ()
          method.

Sun Jul 22 09:27:30 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Oneway_Buffering/Test.idl: Changed the operation
          request_received () in Oneway_Buffering_Admin interface to a
          oneway. The reason for doing this has been well documented in
          bug #982.

Fri Jul 20 23:58:12 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp: Fixed a subtle bug. The problem was when
          reading  a big message. We read part of the message into a stack
          allocated buffer. We then grow the buffer to read the rest of
          the message. This was working fine. But in the read after
          grwoing the  buffer, if we get an error we were just returning
          to the reactor. This was right. But we have to queue up the
          message before returning to the reactor incase we get a
          EWOULDBLOCK.

Fri Jul 20 12:58:12 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/IIOP_Connection_Handler.h:
        * tao/IIOP_Connection_Handler.cpp: Removed the resumption of the
          handler in the  handle_output () call. Looks like we dont win
          anything by resuming the handler at this level. We need to get
          back to this after 1.2 if it is required.

        * tao/Connection_Handler.h: Removed some of the #defines
        * tao/orbconf.h: Moved the #defines here from
          Connection_handler.h.

        * tao/Strategies/DIOP_Connection_Handler.h:
        * tao/Strategies/DIOP_Connection_Handler.cpp:
        * tao/Strategies/UIOP_Connection_Handler.h:
        * tao/Strategies/UIOP_Connection_Handler.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.h:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp: Replicated
          IIOP changes to these protocols.


Fri Jul 20 09:25:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * tests/ior_corbaloc/run_test.pl:
        * tests/InterOp-Naming/run_test.pl:

          Do not include ACEutils Perl module in scripts which use
          PerlACE::Run_Test module.  ACEutils prevents
          PerlACE::Run_Test from seeing the -ExeSubDir argument.
          This causes test failures on certain platforms.
          Thanks to Darrell Brunsch for explaining this and pointing
          this out.

Tue Jul 20 08:27:31 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/tests/AVStreams/Component_Switching/receiver.dsp: Fixed
          a typo in the  library names.

Fri Jul 20 08:10:00 2001  Michael Kircher <Michael.Kircher@mchp.siemens.de>

        * tao/ORB_Core.cpp:
        * tao/Leader_Follower.i:
        * tao/Strategies/DIOP_Acceptor.cpp:
        * tao/Strategies/DIOP_Connection_Handler.cpp:

          Removed old, meanwhile unrelevant comments from me.

        * tao/Asynch_Reply_Dispatcher.cpp:

          Added a check for a nil reply handler. In the case of a nil reply
          handler no response is dispatched. Thanks to Andreas Geisler
          <Andreas.Geisler@erl9.siemens.de> for pointing this out.

        * tests/AMI/simple_client.cpp:

          Added a test case for the above change.

        * tests/AMI/client.cpp:

          Did a cosmetic change.

Thu Jul 19 18:48:09 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * tao/ORB.h (run/work_pending/perform_work): Updated documentation
          wrt timeouts.  Thanks to Jean-Christophe Dubois <jcd@one.com>
          for suggesting this.

Thu Jul 19 10:37:55 2001  Paul Calabrese <calabrese_p@ociweb.com>

  * docs/Options.html:

          Document the differences between the default and
          advanced resource factories.

Tue Jul 19 09:30:31 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp (make_queued_data): Created a data block for
          the size that is required instead of the size of the incoming
          data block. This should fix the lingering problems with
          Single_Read tests in our daily builds.

        * tao/GIOP_Message_Base.cpp: Added some comments.

Tue Jul 19 09:34:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/tests/AVStreams/Component_Switching/sender.bor:
        * orbsvcs/tests/AVStreams/Component_Switching/receiver.bor:
        * orbsvcs/tests/AVStreams/Component_Switching/distributer.bor:
          Fixed the Borland makefiles files to link in the strategies library.

Tue Jul 19 07:56:31 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/tests/AVStreams/Component_Switching/sender.dsp:
        * orbsvcs/tests/AVStreams/Component_Switching/receiver.dsp:
        * orbsvcs/tests/AVStreams/Component_Switching/distributer.dsp:
          Fixed the dsp files to link in the strategies library.

Tue Jul 19 12:49:31 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Sequence.i (replace):
        * tao/Sequence.cpp: Set the read and write pointers properly for
          copied message blocks at the positions.

Tue Jul 19 12:27:31 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp (make_queued_data): Fixed a typo and added a
          new line for a DEBUG statement.

Tue Jul 18 11:25:31 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.h:
        * tao/Transport.cpp: Refactored some code in to a new method
          called  get_queued_data ().

Wed Jul 18 23:58:00 2001  Craig Rodrigues <crodrigu@bbn.com>

        * orbsvcs/tests/AVStreams/Component_Switching/sender.cpp:
        * orbsvcs/tests/AVStreams/Component_Switching/receiver.cpp:
        * orbsvcs/tests/AVStreams/Component_Switching/distributer.cpp:

          Add #include "tao/Strategies/advanced_resource.h"
          so that the ace_static_svc_TAO_Advanced_Resource_Factory
          symbol is defined in each binary, and the TAO_Strategies
          library is actually linked in during static builds.
          Fixes failures of this test with static builds.

        * orbsvcs/tests/AVStreams/Component_Switching/components_svc.conf:
          Fix syntax so Service Configurator file can be used
          for static and dynamic builds.

Wed Jul 18 19:22:48 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * tao/Leader_Follower.i (set/reset_client_leader_thread): Fixed a
          subtle bug in the Leader Follower code: <client_leader_thread_>
          was marked 1 in set_client_leader_thread() and 0 in
          reset_client_leader_thread().  This was a problem when the
          client thread was made leader thread multiple times in nested
          upcall situations.  Changing the code to increment
          <client_leader_thread_> in set_client_leader_thread() and
          decrement in reset_client_leader_thread() gave us an accurate
          count of how many times the client was made the leader. Thanks
          to Christian von Mueffling <cvm@aiss.de> and Alex Lehner
          <lehner@aiss.de> for reporting this.

        * tao/Asynch_Reply_Dispatcher.cpp: Removed unnecessary include
          file "Leader_Followers.h".

        * tests/NestedUpcall/Simple/run_test.pl: Changed MT configuration
          to use parameters with which the test was previously failing.

Wed Jul 18 17:44:56 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/driver/drv_preproc.cpp (DRV_pre_proc):
        * orbsvcs/IFR_Service/drv_preproc_ifr.cpp (DRV_pre_proc):

          Added one more flag set in the ACE_Log_Msg instance to
          ensure that the output of the preprocessor dump (-E command
          line option) goes to stdout.

Tue Jul 18 09:11:13 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp: Made sure that the timeout values are passed
          to  the recv  () calls in the handle_input_i (). The
          thread-per-connection uses the timeout value. Thanks to Mike
          Pyle <mike.pyle@burning-glass.com> for influencing this change.

Wed Jul 18 08:55:29 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * orbsvcs/IFR_Service/drv_preproc_ifr.cpp (DRV_pre_proc):

          Dump of preprocessed IDL file (from -E command line option)
          now goes to stdout instead of stderr. Similar to change in
          TAO IDL compiler in yesterday's entry.

Tue Jul 18 07:58:30 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp (consolidate_message): Fixed a subtle that
          hardly showed up.  The problem stemmed from the fact we were
          trying to queue up messages if we had a partial read. The
          Message blocks were duplicated but not the underlying data
          blocks. Fixed them by doing a proper copying onto the message
          queue.

        * tao/Sequence.i:
        * tao/Sequence.cpp: Used the self_flags () instead of flags () to
          check on the origin of the data block.

Tue Jul 17 21:13:30 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * tests/RTCORBA/Server_Protocol/run_test.pl: Updated and fixed
          run_test.pl and add new file server_uiop.conf.

Tue Jul 17 16:24:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/tests/AVStreams/Component_Switching/run_test.pl:
        * orbsvcs/tests/AVStreams/Component_Switching/svc.conf: (removed)
        * orbsvcs/tests/AVStreams/Component_Switching/components_svc.conf: (added)
          Renamed svc.conf to components_svc.conf to fix problems
          in statically compiled builds.

Tue Jul 17 14:09:16 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Single_Read/client.cpp:
        * tests/Single_Read/test_i.cpp: Added some debug statements. This
          test seems to be failing only in one of the Full builds. Not
          sure what could be the problem. The test runs to perfection but
          the srever doesn't shutdown. I havent been able to reproduce the
          error. These debug statements should give me sufficient clue to
          what is happening.

Tue Jul 17 13:48:00 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/UIOP_Transport.cpp: Fixed a warning.

Tue Jul 17 13:44:49 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/POA/Demux/Makefile: Updated dependencies. This
          should fix the compile error with FORTE.

Tue Jul 17 11:54:07 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/driver/drv_preproc.cpp:

          Modified preprocessor dump to go to stdout instead of stderr. This
          feature was requested by Alex Hornby <alex@anvil.co.uk>.

Mon Jul 16 22:24:17 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * examples/PluggableUDP/tests/SimplePerformance/run_test.pl:
        * examples/PluggableUDP/tests/Basic/run_test.pl: Fixed the run
          test scripts to make sure that a DIOP endpoint is specified when
          the server is invoked. This would fix the errors seen in our
          builds.

        * examples/PluggableUDP/tests/Basicsvc.conf: Just removed the file
          from the repository. Looks like it was not needed and confused
          the client too much leading to SEGV's.

Mon Jul 16 22:03:13 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/IIOP_Connection_Handler.cpp:
        * tao/Strategies/DIOP_Connection_Handler.cpp:
        * tao/Strategies/UIOP_Connection_Handler.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp: Fixed,
          what looked like a subtle race condition. If a thread that calls
          handle_input () gets a -1 on read, we return -1 to the
          reactor. This makes the reactor close the handler and remove
          that from its internal map. If we resume the handler for such
          cases, before returning from handle_input (), looked like one
          more thread was woken up by the reactor to read from the
          handle. When this occurs on another thread, the first thread
          went about doing its task of closing the handle and removing the
          handler from its internal map. Bad things started happening.

Mon Jul 16 17:23:57 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/IORManipulation/TAO_IORManip.dsp:

          Further fixes to the MFC versions of this project.

Mon Jul 16 14:25:23 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_State.cpp (read_ulong): Looks like some of the
          compilers  have problems when they try to dereference a pointer
          got out of reinterpret_cast'ing another pointer.  The problem
          showed up on solaris builds. The fix that has been applied is to
          make a local copy on the stack of the data that is needed before
          calling a reinterpret_cast on it.

Mon Jul 16 11:44:30 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/Big_Reply/server.cpp: The reply sent was too huge. The bug
          #957 started showing up. Reduced the size of the reply to
          approximately 4 MB.

Mon Jul 16 11:36:00 2001  Frank Hunleth  <fhunleth@cs.wustl.edu>

        * tao/Strategies/DIOP_Factory.cpp (requires_explicit_endpoint):

          Make DIOP require an explicit endpoint to prevent a DIOP
          endpoint from being automatically created when the Strategies
          library is linked.

Mon Jul 16 11:20:40 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/DyanmicInterface/TAO_DynamicInterface.dsp:
        * tao/IORTable/TAO_IORTable.dsp:
        * tao/IORManipulation/TAO_IORManip.dsp:
        * orbsvcs/orbsvcs/LoadBalancing.dsp:
        * orbsvcs/orbsvcs/CosNotification.dsp:

          Fixed MFC settings for these projects. Thanks to
          truename <shi_bing@netease.com> for reporting the
          problems.

Mon Jul 16 9:17:43 2001  Christopher Kohlhoff <chris@kohlhoff.com>

        * performance-tests/Throughput/client.bor:
        * performance-tests/Throughput/server.bor:
          Corrected definitions of CFLAGS and LIBFILES to fix Borland
          compile error.

Mon Jul 16 07:49:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/Cubit/TAO/IDL_Cubit/collocation_test.dsp:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/Server.dsp:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/Client.dsp:
          These dsp files needed the strategies library in their release
          configuration.  Added them to fix Win32 builds. I thought I had
          made the changes during my last checkin.  Apparently I had not :(.

Mon Jul 16 07:34:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * docs/tutorials/Quoter/Event_Service/Makefile: Updated
          dependencies.

Sun Jul 15 9:59:24 2001  Balachandran Natarajan <bala@cs.wustl.edu>

        * tests/Param_Test/anyop.dsp (InputPath): Set the dependecy for
          the generated code to the IDL compiler. Missed out when
          generated this file.

Sun Jul 15 9:33:34 2001  Balachandran Natarajan <bala@cs.wustl.edu>

        * tao/Sequence.cpp:
        * tao/Sequence.i: A fix at the places where
          ACE_Message_Block::duplicate () is called.  The duplicate ()
          call just allocated memory for a new Message block and then
          duplicated the exisiting data block -- which just increments the
          reference count of the data block. This could be bad for cases
          where the incoming message is on a data block on stack. By
          incrementing the reference count we get nothing for such
          cases. The fix that has been put in does the following

          - checks whether the data block is on stack, if so does a deep
            copy before calling duplicate on the message block.
          - if the data block is already on the heap just calls duplicate
            () on the message block.

        * orbsvcs/orbsvcs/Event/EC_Gateway_UDP.cpp: Added a comment where
          duplicate () is used. But the code that uses duplicate () has
          been commented out .

Sat Jul 14 20:18:36 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * docs/tutorials/Quoter/Event_Service/Makefile: Updated
          dependencies.

Sat Jul 14 18:59:48 2001  Balachandran Natarajan <bala@cs.wustl.edu>

        * performance-tests/Cubit/TAO/IDL_Cubit/collocation_test.dsp:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/Server.dsp:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/Client.dsp:
          These dsp files needed the strategies library. Added them to fix
          Win32 builds.

Sat Jul 14 17:44:28 2001  Balachandran Natarajan <bala@cs.wustl.edu>

        * tests/Param_Test/anyop.dsp (RSC): Looks like this has been
          broken for  a long time now. Fixed the dsp file so that we dont
          any errors during loading.

Sat Jul 14 00:33:28 2001  Paul Calabrese <calabrese_p@ociweb.com>

        * examples/PluggableUDP/tests/SimplePerformance/svc.conf:

          Switch this configuration file to use the
          Advanced_Resource_Factory.  This way the options
          actually do something.

Fri Jul 13 16:07:04 2001  Douglas C. Schmidt  <schmidt@tango.doc.wustl.edu>

        * orbsvcs/ImplRepo_Service/ImplRepo_i.cpp: Make sure
          that this->ior_multicast is non-NULL in ~ImplRepo_i() before
          using it to lookup in the reactor.  Thanks to Victor Chernenko
          <v_chernenko@hotmail.com> for reporting this.

Thu Jul 13 14:30:16 2001  Paul Calabrese <calabrese_p@ociweb.com>

        * examples/AMI/FL_Callback/AMI_Peer.dsp:
        * examples/AMI/FL_Callback/AMI_Progress.dsp:
        * orbsvcs/tests/AVStreams/Latency/control.dsp:
        * orbsvcs/tests/AVStreams/Latency/ping.dsp:
        * orbsvcs/tests/AVStreams/Latency/pong.dsp:
        * performance-tests/Callback/client.dsp:
        * performance-tests/Callback/server.dsp:
        * performance-tests/Cubit/TAO/MT_Cubit/client.dsp:
        * performance-tests/Cubit/TAO/MT_Cubit/server.dsp:
        * tests/Collocation/Collocation.dsp:
        * tests/Exposed_Policies/Client.dsp:
        * tests/Exposed_Policies/Server.dsp:

          Add in missing path for TAO_Strategies library.

Fri Jul 13 10:48:03 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_union/union_cs.cpp:
        * TAO_IDL/be/be_visitor_union_branch/pulbic_assign_cs.cpp:

          Fixed code generation for copy constructor and assignment
          operator for object reference union members. These operations
          were failing if the rhs union was uninitialized. Thanks to
          Matt Cheers <matt.cheers@boeing.com> for sending in the example
          IDL file and application code.

Thu Jul 12 22:22:31 2001  Ossama Othman  <ossama@uci.edu>

        * tao/Connector_Registry.cpp (create_profile):

          Print the hexadecimal value of the unknown profile tag in the
          debugging output rather than the decimal value.  The former is
          generally more useful.

Thu Jul 12 21:36:53 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/Makefile (MKLIST):

          The AV Service needs the CosNaming and CosProperty libraries.
          Add their Makefiles to the build list if they aren't already
          there.

Thu Jul 12 20:40:51 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.cpp (connect):

          Corrected placement of an ACE_CHECK_RETURN statement.

Thu Jul 12 20:29:42 2001  Ossama Othman  <ossama@uci.edu>

        * tao/ORB_Core.i (resolve_rt_orb, resolve_rt_current):

          Fixed broken emulated exception code.  ACE_CHECK_RETURN should
          have been used instead of ACE_TRY_FLAG.

Thu Jul 12 20:21:36 2001  Ossama Othman  <ossama@uci.edu>

        * docs/tutorials/Quoter/Makefile (DIRS):
        * docs/tutorials/Quoter/Simple/Makefile (DIRS):

          Do not build certain directories if AMI is disabled or if
          Minimum CORBA is enabled.  Some tutorials require AMI, or
          features that are not available in the minimum CORBA
          configuration.

Thu Jul 12 19:45:06 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/Naming_Service/NT_Naming_Server.cpp (ConsoleHandler):

          Fixed unused argument warning.

Thu Jul 12 17:30:42 2001  Frank Hunleth  <fhunleth@cs.wustl.edu>

        * tao/orbconf.h:
        * tao/ORB.cpp:
        * tao/ORB_Core.cpp:
        * tao/ORB_Core.h:
        * tao/ORB_Core.i:

          Removed the RTORB and RTCurrent from the hardcoded list of
          initial references.  Now, both get added dynamically when the
          RTCORBA library is loaded using the PortableInterceptors.

          This should fix a segfault that was reported by Max
          Voronoy <M.Voronoy@telesenskscl.com.ua> and investigated by
          Ossama.

        * tao/RTCORBA/RTCORBA.cpp:
        * tao/RTCORBA/RT_ORB.cpp:
        * tao/RTCORBA/RT_ORB.h:
        * tao/RTCORBA/RT_ORBInitializer.cpp:
        * tao/RTCORBA/RT_ORBInitializer.h:
        * tao/RTCORBA/RT_ORB_Loader.cpp:
        * tao/RTCORBA/RT_ORB_Loader.h:
        * tao/RTCORBA/Thread_Pool.cpp:
        * tao/RTCORBA/Thread_Pool.h:

          Code changes necessary to support creating the RTORB and
          RTCurrent in the pre_init method rather than on demand.
          This is necessary to add both objects to the initial
          references list using the PortableInterceptors.

        * tests/RTCORBA/Makefile:
        * tests/RTCORBA/Makefile.bor:
        * tests/RTCORBA/README:
        * tests/RTCORBA/ORB_init/Makefile:
        * tests/RTCORBA/ORB_init/Makefile.bor:
        * tests/RTCORBA/ORB_init/ORB_init.cpp:
        * tests/RTCORBA/ORB_init/ORB_init.dsp:
        * tests/RTCORBA/ORB_init/README:
        * tests/RTCORBA/ORB_init/run_test.pl:

          Added unit test to check for correct processing when
          instantiating multiple RT enabled ORBs.  Looking at some
          of the RT code indicated that there might be a problem,
          and there actually was.  Thanks to Irfan for writing
          the test.

        * tao/TAO_Internal.cpp:

          Added check to automatically initialize RTCORBA if it
          has been linked in.  This should fix a common source
          of RTCORBA errors.  Thanks to Irfan for the suggestion.

        * tests/RTCORBA/Private_Connection/svc.conf:
        * tests/RTCORBA/Server_Declared/server.conf:
        * tests/RTCORBA/Server_Declared/svc.conf:
        * tests/RTCORBA/Server_Protocol/server_iiop.conf:
        * tests/RTCORBA/Server_Protocol/server_reverse.conf:
        * tests/RTCORBA/Server_Protocol/server_reverse_nt.conf:
        * tests/RTCORBA/Server_Protocol/server_shmiop.conf:
        * tests/RTCORBA/Server_Protocol/svc.conf:
        * tests/RTCORBA/Thread_Pool/svc.conf:
        * tests/Exposed_Policies/server.conf:
        * tests/Exposed_Policies/svc.conf:
        * tests/RTCORBA/Banded_Connections/server.conf:
        * tests/RTCORBA/Banded_Connections/svc.conf:
        * tests/RTCORBA/Client_Propagated/svc.conf:
        * tests/RTCORBA/Client_Protocol/svc.conf:
        * tests/RTCORBA/Explicit_Binding/svc.conf:
        * tests/RTCORBA/MT_Client_Protocol_Priority/server.conf:
        * tests/RTCORBA/MT_Client_Protocol_Priority/svc.conf:

          Updated configuration files to remove RT_ORB initialization
          to test out the above change.


Thu Jul 12 15:10:52 2001  Ossama Othman  <ossama@uci.edu>

        * TAO_IDL/driver/drv_preproc.cpp (DRV_pre_proc):

          Open the temporary file with the O_EXCL flag to close a symbolic
          link attack vulnerability.

Thu Jul 12 13:31:25 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/Security/EstablishTrustPolicy.h:

          Corrected constructor signature to match the implementation.

        * orbsvcs/orbsvcs/Security/Security_PolicyFactory.cpp
          (create_policy):

          Support creation of the SecurityLevel2::EstablishTrustPolicy.

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.cpp (connect):

          Refactored IIOP-specific and SSLIOP-specific code into new
          iiop_connect() and ssliop_connect() methods, respectively.

          (ssliop_connect):

          Added support for the SecurityLevel2::EstablishTrustPolicy.

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.i (object_addr):

          Cache the SSLIOP-specific ACE_INET_Addr in the
          TAO_SSLIOP_Endpoint rather than initializing one each time an
          invocation is made.  This should improve SSLIOP performance.

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.cpp:

          Removed left over Security::Detect{Replay,Misordering} security
          association bits from the default security association value.
          They were actually unused, but removed them anwyay since SSLIOP
          doesn't support either.  No visible run-time change will occur.

        * orbsvcs/orbsvcs/Makefile.Security (FILES):
        * orbsvcs/orbsvcs/Security.bor (OBJFILES):
        * orbsvcs/orbsvcs/Security.dsp:

          Added EstablishTrustPolicy files to the list of sources.

Thu Jul 12 15:10:46 2001  Paul Calabrese <calabrese_p@ociweb.com>

        * examples/PluggableUDP/tests/Basic/svc.conf:
        * examples/PluggableUDP/tests/Performance/svc.conf:
        * tests/RTCORBA/Server_Protocol/server_iiop.conf:
        * tests/RTCORBA/Server_Protocol/server_reverse.conf:
        * tests/RTCORBA/Server_Protocol/server_reverse_nt.conf:
        * tests/RTCORBA/Server_Protocol/server_shmiop.conf:

          Switch these configuration files to use the
          Advanced_Resource_Factory.  This way their options
          actually do something.

Thu Jul 12 13:07:20 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tests/IDL_Test/interface.idl:
        * tests/IDL_Test/reopened_modules.idl:

          Moved some example code having nested modules from
          interface.idl to reopened_modules.idl. This move hides
          the generated code from both configurations of VxWorks,
          which can't compile C++ code generated from nested IDL
          modules.

Thu Jul 12 07:11:39 2001  Ossama Othman  <ossama@uci.edu>

        * tao/Strategies/advanced_resource.cpp (init):

          Fixed broken code that initialized a stack allocated string
          array with a non-const size.  Use a CORBA::StringSeq instead.

Wed Jul 11 18:00:13 2001  Ossama Othman  <ossama@uci.edu>

        * tao/Invocation_Endpoint_Selectors.cpp (select_endpoint):

          Applied patch from Wayne Erchak <werchak@stentor.com> that fixes
          problem where the endpoint list in a given profile was not
          iterated through during connection failures.  "Fail-over"
          semantics for a non-RTCORBA configured ORB once again work.
          [Bug 927]

        * orbsvcs/orbsvcs/Security/EstablishTrustPolicy.h:
        * orbsvcs/orbsvcs/Security/QOPPolicy.h:

          Added Doxygen comments for the TAO_QOPPolicy and
          TAO_EstablishTrustPolicy classes.

Wed Jul 11 16:03:49 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/Security/EstablishTrustPolicy.h:
        * orbsvcs/orbsvcs/Security/EstablishTrustPolicy.cpp:

          Implementation of the SecurityLevel2::EstablishTrustPolicy.  It
          makes it possible to control whether or not client/target
          authentication is performed.

Wed Jul 11 16:03:58 2001  Paul Calabrese <calabrese_p@ociweb.com>

        * tao/Object_Loader.h:
        * tao/Resource_Factory.cpp:
        * tao/Resource_Factory.h:
        * tao/Server_Strategy_Factory.cpp:
        * tao/Server_Strategy_Factory.h:
        * tao/default_client.cpp:
        * tao/default_client.h:
        * tao/default_resource.cpp:
        * tao/default_resource.h:
        * tao/default_server.cpp:
        * tao/default_server.h:
        * tao/Strategies/advanced_resource.cpp:
        * tao/Strategies/advanced_resource.h:

          Improved processing of service configurator options.  Many
          failures were not being reported to the user.  These changes
          cause many failure to print warning messages but should not
          affect execution of the code.  The following situations now
          result in warning messages:
            - Passing unknown -ORB* options to a factory
              (This includes passing advanced resource factory
              options to the default resource factory)
            - Passing an unknown value to a defined option
            - Passing options to the default resource factory
              (Resource_Factory) when the advanced resource factory
              (Advanced_Resource_Factory) is being used

          I also removed several deprecated options: -ORBEventLoopLock,
          -ORBDemuxStrategy, and -ORBConnectorLock.

        * examples/AMI/FL_Callback/AMI_Peer.dsp:
        * examples/AMI/FL_Callback/AMI_Progress.dsp:
        * examples/AMI/FL_Callback/Makefile:
        * examples/AMI/FL_Callback/peer.conf:
        * examples/AMI/FL_Callback/peer.cpp:
        * examples/AMI/FL_Callback/svc.conf:
        * orbsvcs/tests/AVStreams/Latency/Makefile:
        * orbsvcs/tests/AVStreams/Latency/control.cpp:
        * orbsvcs/tests/AVStreams/Latency/control.dsp:
        * orbsvcs/tests/AVStreams/Latency/ping.cpp:
        * orbsvcs/tests/AVStreams/Latency/ping.dsp:
        * orbsvcs/tests/AVStreams/Latency/pong.cpp:
        * orbsvcs/tests/AVStreams/Latency/pong.dsp:
        * orbsvcs/tests/AVStreams/Latency/svc.conf:
        * orbsvcs/tests/Event/Performance/latency.conf:
        * orbsvcs/tests/Notify/performance-tests/RedGreen/svc.conf:
        * performance-tests/Callback/Makefile:
        * performance-tests/Callback/client.bor:
        * performance-tests/Callback/client.cpp:
        * performance-tests/Callback/client.dsp:
        * performance-tests/Callback/server.bor:
        * performance-tests/Callback/server.cpp:
        * performance-tests/Callback/server.dsp:
        * performance-tests/Callback/svc.conf:
        * performance-tests/Cubit/TAO/IDL_Cubit/collocation_test.cpp:
        * performance-tests/Cubit/TAO/IDL_Cubit/iiop_lite.conf:
        * performance-tests/Cubit/TAO/IDL_Cubit/run_test.pl:
        * performance-tests/Cubit/TAO/IDL_Cubit/svc.mt_server.conf:
        * performance-tests/Cubit/TAO/IDL_Cubit/svc.st_client.conf:
        * performance-tests/Cubit/TAO/IDL_Cubit/svc.st_server.conf:
        * performance-tests/Cubit/TAO/IDL_Cubit/svc.zero_lock.conf:
        * performance-tests/Cubit/TAO/IDL_Cubit/uiop_lite.conf:
        * performance-tests/Cubit/TAO/IDL_Cubit/collocation/svc.conf:
        * performance-tests/Cubit/TAO/MT_Cubit/Makefile:
        * performance-tests/Cubit/TAO/MT_Cubit/client.bor:
        * performance-tests/Cubit/TAO/MT_Cubit/client.cpp:
        * performance-tests/Cubit/TAO/MT_Cubit/client.dsp:
        * performance-tests/Cubit/TAO/MT_Cubit/server.cpp:
        * performance-tests/Cubit/TAO/MT_Cubit/server.dsp:
        * performance-tests/Cubit/TAO/MT_Cubit/svc.conf:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/client.conf:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/server.conf:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/client.cpp:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/server.cpp:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/Makefile:
        * performance-tests/RTCorba/Multiple_Endpoints/Single_Endpoint/client.conf:
        * performance-tests/Throughput/Makefile:
        * performance-tests/Throughput/client.bor:
        * performance-tests/Throughput/client.cpp:
        * performance-tests/Throughput/server.bor:
        * performance-tests/Throughput/server.cpp:
        * performance-tests/Throughput/svc.conf:
        * tests/AMI_Timeouts/svc.conf:
        * tests/Collocation/Collocation.bor:
        * tests/Collocation/Collocation.cpp:
        * tests/Collocation/Collocation.dsp:
        * tests/Collocation/Makefile.test:
        * tests/Collocation/svc.conf:
        * tests/Exposed_Policies/Client.dsp:
        * tests/Exposed_Policies/Makefile:
        * tests/Exposed_Policies/Server.dsp:
        * tests/Exposed_Policies/client.bor:
        * tests/Exposed_Policies/client.cpp:
        * tests/Exposed_Policies/server.bor:
        * tests/Exposed_Policies/server.conf:
        * tests/Exposed_Policies/server.cpp:
        * tests/Exposed_Policies/svc.conf:
        * tests/FL_Cube/Makefile:
        * tests/FL_Cube/client.bor:
        * tests/FL_Cube/client.cpp:
        * tests/FL_Cube/server.bor:
        * tests/FL_Cube/server.cpp:
        * tests/FL_Cube/svc.conf:
        * tests/LongUpcalls/svc.conf:
        * tests/MT_Server/server.conf:
        * tests/Strategies/Makefile:
        * tests/Strategies/README:
        * tests/Strategies/client.bor:
        * tests/Strategies/server.bor:
        * tests/Strategies/svc.conf

          These are changes to TAO tests and examples with
          broken service config files.  The above changes
          caused these files to generate warnings.  The vast
          majority of the problems were the passing of advanced
          resource factory options to the default resource
          factory.  Most were changed to simply use the advanced
          resource factory.  Some were modified to use
          the defaults of the default resource factory.

Wed Jul 11 14:41:20 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/ast/ast_decl.cpp (compute_prefix):

          Fixed function to work if the #pragma prefix string looks like
          '#pragma prefix      "foo.bar"'. The preprocessor takes care
          of the whitespace, if any, between '#pragma' and 'prefix'.

Wed Jul 11 14:38:00 2001  Craig Rodrigues <crodrigu@bbn.com>

       * tao/Strategies/DIOP_Connection_Handler.cpp:
         Only display debugging message if TAO_debug_level is set > 5.

Wed Jul 11 13:49:00 2001  Craig Rodrigues <crodrigu@bbn.com>

        * utils/catior/catior.cpp:
          Add support for printing DIOP (GIOP over UDP) profiles.

Wed Jul 11 08:39:14 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/RTCORBA/RT_Protocols_Hooks.cpp (get_thread_CORBA_priority):

          Replace ACE_CHECK with ACE_CHECK_RETURN (-1) - function must
          return an int.

Wed Jul 11 09:25:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * orbsvcs/tests/Makefile.bor:
          Add Simple_Naming test

Wed Jul 11 02:19:59 2001  Irfan Pyarali  <irfan@cs.wustl.edu>

        * tao/RTPortableServer/RT_Servant_Dispatcher.cpp: Remember the
          native priority of the invoking thread.  Once the invocation
          completes, restore the thread to the original native thread
          priority.

          Previously, when the priority was restored, it was done it terms
          of CORBA priority.  The problem with this approach was that in
          the conversion from native to CORBA and then back to native, we
          can lose precision and hence the native priority of the native
          thread may not be the same as its original native priority.
          Therefore, remembering the original native priority will
          alleviate this problem.

        * tao/Protocols_Hooks:
        * tao/Default_Protocols_Hooks:
        * tao/RT_Protocols_Hooks.cpp:

          Expanded the priority interfaces on the Protocols_Hooks such
          that the user can obtain both the native and CORBA priorities.
          Also, the user can set both.

        * tao/RTCORBA/RT_Invocation_Endpoint_Selectors.cpp (select_endpoint):
        * tao/RTCORBA/RT_Current.cpp (the_priority):
        * tao/Strategies/Reactor_Per_Priority.cpp (reactor):

          Changed to use new priority interfaces.

Tue Jul 10 22:37:48 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/Security/QOPPolicy.cpp (copy):

          Perform a deep copy, not a shallow copy.  The copy is supposed
          to be independent of the original.

Tue Jul 10 15:59:31 2001  Ossama Othman  <ossama@uci.edu>

        * tao/PortableInterceptorC.h:

          Include "PolicyC.h" to pull in some policy related exception
          definitions.  This fixed a problem with some minimum CORBA
          builds with native exception support enabled.  Thanks to
          Sangeetha Ramadurai <Sangeetha.Ramadurai@geind.ge.com> for
          reporting the problem and providing a fix.

Tue Jul 10 15:56:13 2001  Krishnakumar B  <kitty@cs.wustl.edu>

        * tao/GIOP_Message_Lite.h:
        * tao/GIOP_Message_Lite.cpp:
        * tao/Pluggable_Messaging.h:
        * tao/Pluggable_Messaging.cpp:

          Fixed warning about virtual function override. Was caught by the
          Tru64 compiler.

Tue Jul 10 13:09:18 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_valuetype/valuetype.cpp:

          Corrected a generation of skel_export_macro to
          stub_export_macro. Thanks to Alexander Rieger
          <Alexander.Rieger@inka.de> for tracking this down.

Tue Jul 10 11:51:05 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_interface/interface_ch.cpp:
        * TAO_IDL/be/be_visitor_interface/interface_sh.cpp:

          Changed code so that the inheritance list of a class
          declaration consists of fully scoped names. The stub
          generation was using ACE_NESTED_CLASS, and the skeleton
          generation was using a computed relative name. Both
          generated uncompilable code in some cases. Thanks to
          Richard L. Johnson <rich@huey.jpl.nasa.gov> for reporting
          this bug and for sending in the example IDL file.

        * tests/IDL_Test/interface.idl:

          Added example IDL to this file in the IDL test suite.

Tue Jul 10 05:52:09 2001  Douglas C. Schmidt  <schmidt@ace.cs.wustl.edu>

        * tao/Profile.h,
        * tao/Object.h: Fixed typos in comments.  Thanks to Johnny
          Willemsen for reporting this.

        * tao/Connector_Registry.h (TAO_Connector_Registry): Fixed a typo
          where svc.comf should have been svc.conf.  Thanks to Johnny
          Willemsen for reporting this.

        * tao/ORB.cpp: Added an addition expression to the #ifdef for the
          using std::set_unexpected declaration.  Thanks to Scott Plant
          <splant@softhome.net> for reporting this.

Tue Jul 10 00:45:00 2001  Ossama Othman  <ossama@uci.edu>

        * tao/RTCORBA/RT_ORB.cpp (TAO_RT_CORBA_Priority_Normalizer):

          Fixed busted code that did not check if the resolved
          PriorityMappingManager object reference was nil.  Since this
          check was missing, a seg fault would occur when attempting to
          invoke a method on that mapping manager.  Throw a
          CORBA::INTERNAL() exception if the reference is nil.  This at
          leasts lets us identify a problem without seg faulting.

          Fixed broken code that did not use emulated exceptions properly.
          The ACE_TRY/CATCH block was missing.

Tue Jul 10 09:13:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * tao/orbsvcs/orbsvcs/security.bor

          Added the missing files Security_PolicyFactory and QOPPolicy

Mon Jul 09 23:43:29 2001  Ossama Othman  <ossama@uci.edu>

        * tao/params.i (default_init_ref):

          Use CORBA::string_dup() to return a duplicate string instead of
          relying on ACE_CString::rep().  This allows us to portably take
          advantage of CORBA::String_var in TAO_ORB_Core::resolve_rir().

        * tao/ORB_Core.cpp (list_initial_references):

          Corrected long standing bug where the stringified object
          reference was placed in the ObjectIdList instead of the ObjectId
          that represents it.

          Include the ObjectIds stored in the underlying table for
          the ORB::register_initial_reference() mechanism in the returned
          ObjectIdList.  This was a bug.

          (TAO_ORB_Core, fini):

          The rt_priority_mapping_manager_ cached object reference member
          was unused.  Removed its initialization and finalization code.

          (resolve_rir):

          Improved exception-safety of this method by using a
          CORBA::String_var instead of relying solely on delete().

        * tao/ORB_Core.h (rt_priority_mapping_manager_):

          Removed this unused attribute.

        * tao/Object_Ref_Table.h (begin, end):

          Made iterator accessors public so that the ORB_Core can use
          them.

        * tao/Object_Ref_Table.cpp:

          Removed three unnecessary Hash Map related template
          instantiations.

          (current_size):

          Return the current size of the underlying table.

        * tests/InterOp-Naming/README:

          Corrected format for corbaloc IORs.

Mon Jul  9 17:39:21 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * orbsvcs/tests/Security/Secure_Invocation/client.cpp:

          Added .in() to a CORBA::ORB_var passed to a function.

Mon Jul  9 14:56:26 2001  Douglas C. Schmidt  <schmidt@tango.doc.wustl.edu>

        * docs/components.html (bgcolor): Fixed a typo.  Thanks to
          John Ashmun for reporting this.

Mon Jul 09 10:56:30 2001  Ossama Othman  <ossama@uci.edu>

        * tao/DynamicInterface/Server_Request.h:
        * tao/DynamicInterface/Server_Request.inl (_tao_server_request):

          Added this accessor that returns a reference to the underlying
          TAO_ServerRequest object.  Thanks to James Megquier
          <jmegq@bbn.com> and Mouna Seri <seri@crhc.uiuc.edu> for
          suggesting addition of this accessor.

Mon Jul  9 11:16:36 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/IFR_Client/TAO_IFR_Client.dsp:
        * tao/TypeCodeFactory/TypeCodeFactory.dsp:

          Added MFC debug and release versions to these projects.
          Thanks to Francois Bernier <fbernier@gel.ulaval.ca> for
          sending them in.

Mon Jul  9 10:13:54 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_interface.cpp:
        * TAO_IDL/be_visitor_interface/interface_ch.cpp:
        * TAO_IDL/be_visitor_interface/interface_cs.cpp:
        * tao/Object.cpp:
        * tao/Object.h:
        * tao/Object.i:

          Added static int _tao_class_id to CORBA::Object and
          to all classes generated from IDL interfaces. The
          address of this variable is used in all
          _unchecked_narrow() and _tao_QueryInterface() methods.
          Formerly, the address of the _narrow() method was used,
          but on BCB, this address is not usable across DLLs.
          Thanks to Christopher Kohlhoff <chris@kohlhoff.com> for
          suggesting the fix.

Mon Jul  9 15:27:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * tao/orbsvcs/tests/AVStreams/Makefile.bor
        * tao/orbsvcs/tests/AVStreams/Multicast_Full_Profile/*.bor
        * tao/orbsvcs/tests/AVStreams/Full_Profile/*.bor

          Added BCB makefiles for the mentioned AVStreams test

Mon Jul 09 04:11:45 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/orbsvcs/Security/QOPPolicy.h:

          Added missing "ace/post.h" include.

Mon Jul 09 02:43:38 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/tests/Security/Secure_Invocation/Makefile:
        * orbsvcs/tests/Security/Secure_Invocation/client.cpp:
        * orbsvcs/tests/Security/Secure_Invocation/client.dsp:
        * orbsvcs/tests/Security/Secure_Invocation/run_test.pl:

          Updated this test to take advantage of the newly added
          Security::QOPPolicy support.  The test is more self-contained
          now, i.e. it no longer depends so much on the "run_test.pl" Perl
          script to test all features.

Mon Jul 09 02:21:32 2001  Ossama Othman  <ossama@uci.edu>

        * docs/releasenotes/index.html:

          Updated Security Service release notes.

Mon Jul 09 01:51:27 2001  Ossama Othman  <ossama@uci.edu>

        * tao/Invocation.h:
        * tao/Invocation.i:

          Added accessors to the transport and max_wait_time members.

        * tao/Invocation.cpp (perform_call):

          Updated invocation of the connect() method to match the
          signature change described below.

        * tao/Connector_Registry.cpp:
        * tao/Connector_Registry.h:
        * tao/IIOP_Connector.cpp:
        * tao/IIOP_Connector.h:
        * tao/Pluggable.h:
        * tao/Strategies/DIOP_Connector.cpp:
        * tao/Strategies/DIOP_Connector.h:
        * tao/Strategies/SHMIOP_Connector.cpp:
        * tao/Strategies/SHMIOP_Connector.h:
        * tao/Strategies/UIOP_Connector.cpp:
        * tao/Strategies/UIOP_Connector.h:
        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connector.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.h:

          Changed signature of connect() method to accept a pointer to a
          TAO_GIOP_Invocation object.  Some pluggable protocols
          (e.g. SSLIOP) may need access to some of the information in that
          object.

        * orbsvcs/orbsvcs/Security.idl:

          Corrected spelling of Security::SecFeaturePolicy.  "Feature" not
          "Features".

        * orbsvcs/orbsvcs/Security/QOPPolicy.cpp:
        * orbsvcs/orbsvcs/Security/QOPPolicy.h:

          Implementation of the Security::QOPPolicy policy.  This policy
          is used to set the desired invocation Quality-of-Protection
          (QoP).  It can be created using ORB::create_policy(), and used
          in conjunction with the standard policy manipulation CORBA
          features (e.g. PolicyManager, PolicyCurrent), meaning that this
          policy can be set on a per-ORB, per-thread or per-object basis.

          This policy makes it possible to, for example, make both secure
          and insecure invocations within the same client process.

        * orbsvcs/orbsvcs/Security/Security_PolicyFactory.cpp:
        * orbsvcs/orbsvcs/Security/Security_PolicyFactory.h:

          Security policy factory implementation that is registered with
          the ORB's policy factory registry.

        * orbsvcs/orbsvcs/Security/Security_ORBInitializer.cpp (post_init,
          register_policy_factories):

          Register the supported security policy factories with the ORB.

        * orbsvcs/orbsvcs/Security/Security_ORBInitializer.h:

          Added a shared security policy factory member to this
          ORBInitializer.  The factory is reentrant so there is no need
          create one for each ORB.  Just share one between all ORBs.

        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connector.cpp (connect):

          Moved code that obtained the TAO_IIOP_Endpoint from the
          TAO_SSLIOP_Endpoint from this method to the corresponding one in
          TAO_SSLIOP_Connector.  This cleans up the code a bit.  It
          shouldn't differ all that much from the
          IIOP_Connector::connect() implementation.

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connector.cpp (connect):

          Now that both secure and insecure invocations are supported
          within the same client process, an IIOP-only transport
          descriptor must be used instead of the one passed to this method
          since the latter is used for SSLIOP connections.  Doing so
          prevents an IIOP-only cached transport from being associated
          with an SSLIOP connection.

          This fixes a problem that was revealed when support for secure
          and insecure invocations within the same client process was
          added (i.e. via the Security::QOPPolicy support).  An insecure
          cached transport was matched against an SSLIOP connection.

        * tao/IIOP_Endpoint.i:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Endpoint.cpp:

          Cosmetic updates.

        * tao/Exception.cpp:

          Corrected "unknown description" message.

        * tao/Strategies/SHMIOP_Endpoint.h:
        * tao/Strategies/UIOP_Endpoint.h:

          Doxygen-ated these headers.

        * orbsvcs/orbsvcs/Makefile.Security:
        * orbsvcs/orbsvcs/Security.dsp:
        * orbsvcs/orbsvcs/Security/Security.bor:

          Added new QOPPolicy and Security_PolicyFactory filenames to
          these Makefiles and project files.

Mon Jul  9 08:41:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * tao/orbsvcs/tests/AVStreams/Makefile.bor
        * tao/orbsvcs/tests/AVStreams/Pluggable/*.bor
        * tao/orbsvcs/tests/AVStreams/Multicast/*.bor

          Added BCB makefiles for the mentioned AVStreams test

Sat Jul  7 19:44:55 2001  Krishnakumar B  <kitty@cs.wustl.edu>

        * tao/Incoming_Message_Queue.inl (get_queued_data):

          Moved the definition to the beginning to fix warning. Was caught
          by Tru64 cxx compiler.

Fri Jul  6 22:04:51 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/LongUpcalls/run_ami_test.pl: Made the perl script
          runnable.  It was giving weird errors in the builds.

Fri Jul 06 21:53:26 2001  Christopher Kohlhoff  <chris@kohlhoff.com>

        * tao/Makefile.bor:
        * tao/Strategies/Makefile.bor:
          Removed GIOP Lite support and added new source files.

        * orbsvcs/orbsvcs/AV/AV_Core.h:
          Fixed borland #pragmas to correctly reset previous options.

Fri Jul  6 16:58:52 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/Notify_Service/Notify_Service.cpp (shutdown):

          Applied the same fix from Jody's patch described below to this
          method.

Fri Jul  6 15:59:17 2001  Ossama Othman  <ossama@uci.edu>

        * orbsvcs/Notify_Service/Notify_Service.h (naming_):
        * orbsvcs/Notify_Service/Notify_Service.cpp (init,
          resolve_naming_service):

          Integrated patch from Jody Hagins <jody@atdesk.com>.  Take
          advantage of the methods provided by the
          CosNaming::NamingContextExt interface so that the Notification
          Service does the right thing with an option such as
          "-Factory Foo.Bar/My_Event_Channel".

Fri Jul  6 18:02:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/TAO_Static.dsp: Removed support for GIOP Lite for the time
          being. Just want to make sure that the interface and other stuff
          stabilises before we can squeeze this in.

Fri Jul  6 17:55:25 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_argument/pre_invoke_cs.cpp (void_return_type):

          Change a call to base_node_type() on the operation's return type
          node to node_type(), which is the same but does no unaliasing.
          We are checking for a return type of void, which can't be aliased
          anyway. The return from base_node_type() checks for equality to
          the enum value NT_pre_defined, which now returns true for some
          aliased return types, and that causes problems later in the
          function. Thanks to Walter Wehrli <walter.wehrli@ubs.com> for
          sending in the example IDL file that uncovered this bug.

Fri Jul  6 17:13:20 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/ORB_Core.cpp:

          Changed the string passed to
          ACE_Dynamic_Service<TAO_Object_Loader>::instance from
          "TypeCodeFactory" to "TypeCodeFactor_Loader" to match the string
          in the ACE_STATIC_SVC_DEFINE macro in
          TypeCodeFactory_Loader.cpp. Thanks to Francois Bernier
          <fbernier@gel.ulaval.ca> for reporting the bug.

Fri Jul  6 17:10:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp:
        * tao/Transport.cpp: Added support for LocateRequest & LocateReply
          that. It should have been added before the code from my branch
          came to the main trunk, but somehow got missed.

Fri Jul  6 16:01:45 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp (process_queue_head):
        * tao/GIOP_Message_State.cpp: Fixed a problem that has long been
          there. We have been unlucky that we did not get to this before.

Fri Jul 06 14:58:09 2001  Nanbor Wang  <nanbor@cs.wustl.edu>

        To build "CosEvent - Win32 MFC Release":

        * TAOACE.dsw: Make RTEvent a dependent of CosEvent.

        * orbsvcs/orbsvcs/RTEvent.dsp: Set up IDL custom build rules for
          MFC configs.

        * orbsvcs/orbsvcs/CosEvent.dsp: Changed the name for release
          version libarary to XXXmfc.lib.

          Thanks to Kristopher Johnson <kristopher.johnson@transcore.com>
          for providing the fix.  [Bug 898]

Fri Jul  6 13:33:35 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/Makefile.BE:
        * TAO_IDL/Makefile.FE:

          Added lines to force static lib compiliation for mingw. Thanks
          to Cristian Ferretti <cristian_ferretti@yahoo.com> for the
          patches.

Fri Jul  6 13:23:46 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Synch_Reply_Dispatcher.cpp:
        * tao/Asynch_Reply_Dispatcher.cpp: Removed the initialization of a
          buffer  from the constructor. Looks like I used to do it in some
          life of mine and I dont which one it was. :( SunCC 4.2 doesnt
          like that.

        * tao/Resume_Handle.cpp: Added a #include of the reactor to fix
          compile errors in Sun CC4.2.

Fri Jul  6 13:19:25 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_sequence/gen_bounded_obj_sequence_ch.cpp:
        * TAO_IDL/be/be_visitor_sequence/gen_bounded_sequence_ch.cpp:
        * TAO_IDL/be/be_visitor_sequence/gen_bounded_str_sequence_ch.cpp:
        * TAO_IDL/be/be_visitor_sequence/gen_bounded_wstr_sequence_ch.cpp:
          Added generation of 'TAO_EXPORT_MACRO' to class declaration.
          Thanks to Craig Rodrigues and Christopher Kohlhoff for helping to
          track this stuff down.

         * TAO_IDL/be/be_visitor_sequence/gen_unbounded_sequence_ch.cpp:
          Changed generation of 'TAO_EXPORT_NESTED_MACRO' to 'TAO_EXPORT_MACRO'.

Fri Jul  6 12:38:06 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp: Fixed a bug with thread-per-connection
          model. We dont need to send notify () to the reactor as there is
          no reactor.

Fri Jul  6 11:27:00 2001  Craig Rodrigues  <crodrigu@bbn.com>

        * tao/IFR_Client/IFR_BaseC.h:
        * tao/DynamicAny/DynamicAnyC.h:
        * tao/IORManipulation/IORC.h:
          Export more classes to appease Borland C++.

Fri Jul  6 08:27:06 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Incoming_Message_Queue.h:
        * tao/Resume_Handle.h:
        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Transport.h: Fixed fuzz errors.

Fri Jul  6 08:14:48 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/Cubit/TAO/MT_Cubit/Makefile:
        * performance-tests/Cubit/TAO/IDL_Cubit/Makefile: Generated
          dependencies again.

Fri Jul  6 07:43:15 2001  Chad Elliott  <elliott_c@ociweb.com>

        * orbsvcs/IFR_Service/Makefile:

          Disable building this for Visual Age C++ due to the fact that the
          using keyword is not properly supported.

Fri Jul  6 07:38:55 2001  Chad Elliott  <elliott_c@ociweb.com>

        * performance-tests/Cubit/TAO/MT_Cubit/Task_Client.h:

          Fix warning when using gcc on Tru64.

Fri Jul  6 07:38:44 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/SSLIOP.bor:
        * orbsvcs/orbsvcs/SSLIOP.dsp: Added the new file
          IIOP_SSL_Transport.cpp  to the above files.

Fri Jul  6 07:12:44 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/Makefile.av: Generated dependencies again. Looks
          like the last dependency generation did not do a good job. This
          should fix all the compile errors seen in the builds.

Fri Jul  6 06:56:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Makefile.bor:
        * tao/Strategies/Makefile.bor: Fixed the makefiles. Thanks to
          Johnny Willemsen for alerting me on this.

Thu Jul  6 01:06:55 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/DIOP_Connection_Handler.h: Fixed a link error
          on Win32 builds.

Fri Jul  6 00:39:38 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/*/Makefile: Updated dependencies.

Fri Jul  6 00:38:05 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * examples/*/Makefile: Updated dependencies.

Fri Jul  6 00:34:13 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * performance-tests/*/Makefile: Updated dependencies.

Fri Jul  6 00:27:54 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/*/Makefile: Updated dependencies.

Thu Jul  6 00:16:55 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/TAO_Static.dsp:
        * tao/Strategies/TAO_Strategies.dsp:
        * tao/Strategies/TAO_Strategies_Static.dsp: Added the new files
          and removed old ones.

Thu Jul  5 23:57:55 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Makefile:
        * tao/*/Makefile: Updated dependencies.

Thu Jul  5 23:49:34 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/Makefile.SSLIOP: Added IIOP_SSL_Transport to it.

Thu Jul  5 23:44:16 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/GIOP_Message_NonReactive_Base.h:
        * tao/Strategies/GIOP_Message_NonReactive_Base.cpp:
        * tao/Strategies/GIOP_Message_NonReactive_Handler.h:
        * tao/Strategies/GIOP_Message_NonReactive_Handler.cpp:
        * tao/Strategies/GIOP_Message_NonReactive_Handler.inl:
        * tao/GIOP_Message_Reactive_Handler.h:
        * tao/GIOP_Message_Reactive_Handler.cpp:
        * tao/GIOP_Message_Reactive_Handler.inl: Removed them from the
          main trunk. They are no longer needed.

Thu Jul  5 23:42:17 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_State.i: Removed from the main trunk. Replaced
          that with GIOP_Message_State.inl.

Thu Jul  5 23:30:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

          The long awaited fix for bug 575 is finally in!! This has been
          merged  from bug_575_stage_2. The ChangeLog entries start here,

Thu Jul  5 23:00:43 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.h
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Transport.cpp
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Transport.h: Fixed some more
          compilation errors in g++.

Thu Jul  5 22:35:07 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Transport.h:
        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Transport.cpp: Fixed compile
          errors in g++. The new IIOP_SSL_Transport classes now does the
          functionality of the IIOP_SSL_Connection_Handler classes. The
          connection handler classes exist now only to create the
          transport.

Thu Jul  5 21:44:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Transport.h:
        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Transport.cpp: New files for the
          SSLIOP.

        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connection_Handler.h:
        * orbsvcs/orbsvcs/SSLIOP/IIOP_SSL_Connection_Handler.cpp: Moved
          the handle_input () from the connection handler to transport.

        * tao/Strategies/DIOP_*:  Fixed it for the new setup. We now make
          a buffer of dgram size and use that to read messages.

Thu Jul  5 14:44:59 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        Merged with the main trunk and moved it to a new branch by name
        bug_575_stage_2.

Wed Jul  4 18:53:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Transport.cpp (recv_i): Added the
          error checking.

        * tao/Pluggable_Messaging.h: Made reset () as a pure virtual
          function.

        * tao/Strategies/UIOP_Transport.cpp: Removed the close_connection
          () call on TMS.

        * tao/LIST_OF_TODO: Updated..

Wed Jul  4 18:45:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_State.inl (reset):
        * tao/GIOP_Message_State.h:
        * tao/GIOP_Message_State.cpp:
        * tao/GIOP_Message_Base.h:
        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_Base.inl:
        * tao/Incoming_Message_Queue.cpp:
        * tao/Pluggable_Messaging.h:
        * tao/Pluggable_Messaging.cpp: Added documentation.

        * tao/Transport.h:
        * tao/Transport.cpp: Fixed a minor bug. In consolidate_message ()
          we were not checking the queue before processing the message on
          hand. Now we check the queue and process the head of the queue
          if it is not empty. In the process we add the message on hand in
          the queue.

Wed Jul  4 16:21:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/IIOP_Transport.cpp: Removed the connection_closed () from
          the recv_i () call.

        * tao/Transport.h:
        * tao/Transport.cpp: Added a number of comments and cleaned up the
          code for readability.

        * tao/LIST_OF_TODO: Updated the list.

Wed Jul  4 09:20:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.h (TAO_Transport): Added some documentation for
          the incoming_data_path.

        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Transport.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Transport.h:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.cpp:
        * orbsvcs/orbsvcs/SSLIOP/SSLIOP_Connection_Handler.h:  Fixed the
          SSLIOP for the changes. Havent compiled this yet.

Tue Jul  3 17:09:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/SHMIOP_Transport.cpp: Fixed a bug in
          consolidate_message (). The IDL_Cubit tests work fine with the
          SHMIOP & UIOP transports.

Tue Jul  3 16:29:06 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/SHMIOP_Connection_Handler.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.h:
        * tao/Strategies/SHMIOP_Transport.cpp: Fixed warnings and errors
          in g++.

        * tao/Strategies/UIOP_Transport.cpp:
        * tao/Strategies/UIOP_Transport.h:
        * tao/Strategies/UIOP_Connection_Handler.h:
        * tao/Strategies/UIOP_Connection_Handler.cpp: Fixed warnungs and
          errors in g++.

        * tao/IIOP_Connection_Handler.cpp: Used fetch_handle () to get the
          handle instead of the handle passed as an argument of the
          handle_input () call.

        * tao/Connection_Handler.h: Added a new #define and changed
          TAO_CONNECTION_HANDLER_BUF_SIZE as
          TAO_CONNECTION_HANDLER_STACK_BUF_SIZE.


Tue Jul 03 15:45:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Strategies/SHMIOP_Transport.h:
        * tao/Strategies/SHMIOP_Transport.cpp:
        * tao/Strategies/SHMIOP_Connection_Handler.h:
        * tao/Strategies/SHMIOP_Connection_Handler.cpp: Fixed the protocol
          to work  with the latest changes.

        * tao/Strategies/TAO_Strategies.dsp: Removed the
          GIOP_NonReactive_* files.

        * tao/Resume_Handle.cpp: Fixed a bug in resume_handle (). If we
          have a null ORB_Core we just dont resume the handle.

Mon Jul 02 23:16:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/IIOP_Transport.cpp:
        * tao/IIOP_Transport.h: Removed some functions which would no
          longer be used.

        * tao/Transport.h: Added some documentation for the incoming data
          path.

        * tao/Strategies/UIOP_Transport.h:
        * tao/Strategies/UIOP_Transport.cpp:
        * tao/Strategies/UIOP_Connection_Handler.cpp: Brought them up in
          sync with IIOP.

Mon Jul 02 20:28:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        This checkin fixes many problems with multiple calls being read in
        a single read.

        * tao/GIOP_Message_Base.cpp: Changed the way we calculate the
          remaining length that needs to be copied if the last read had
          left with a very small piece of message ie. less than 12 bytes
          in the queue.

        * tao/GIOP_Message_State.cpp: Fixed a check condition before we go
          ahead to parse the header.

        * tao/Incoming_Message_Queue.h:
        * tao/Incoming_Message_Queue.cpp:
        * tao/Incoming_Message_Queue.inl: Added a method set_flags
          (). Further the condition for resuming the handler has been
          relaxed. Clients of this class can decide not to resume handlers
          at all. This comes in handy at times. Also added a operator=
          method.

        * tao/Transport.cpp:
        * tao/Transport.h: One of the bigger problems have been
          solved. The way we now multiple oneways is like this

          - The leader thread uses the transport and reads more messages
          - It splits up the messages in to pieces and queues them up.
          - It just takes one message and processes (the head of the
            queue). The changes above makes sure that only one of the
            message is processed.
          - Before processing if it finds one more message it sends a
            notify () to the reactor. An important point is that it does
            this without resuming the handler.
          - The notify call processes one more message. Before processing
            if it finds one more complete message it just sends another
            notify to the reactor.
          - The thread that reads the last complete message from the queue
            resumes the handler before processing the message.

          By the above process we dont starve any thread and at the
          same time ensures concurrency within the ORB.

        * tao/LIST_OF_TODO: Updated..

Sun Jul 01 18:35:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Asynch_Reply_Dispatcher.h:
        * tao/Asynch_Reply_Dispatcher.cpp:
        * tao/DynamicInterface/DII_Reply_Dispatcher.h:
        * tao/DynamicInterface/DII_Reply_Dispatcher.cpp:  Applied the same
          optimization that was done in the last checkin. The last checkin
          works fine. It is no more a suspect.

Sun Jul 01 18:05:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp: Fixed small problems in getting
          message types.

        * tao/Synch_Reply_Dispatcher.h:
        * tao/Synch_Reply_Dispatcher.cpp:
        * tao/Pluggable_Messaging_Utils.h:
        * tao/Pluggable_Messaging_Utils.cpp: Suspect optimisations added
          to create datablocks on stack.

        * tao/LIST_OF_TODO: Updated list..

Sun Jul 01 09:05:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport_Mux_Strategy.h:
        * tao/Muxed_TMS.h:
        * tao/Muxed_TMS.cpp:
        * tao/Exclusive_TMS.cpp:
        * tao/Exclusive_TMS.h: Removed commented out code. They have been
          commented out for sometime that we dont need them anymore.

        * tao/Incoming_Message_Queue.cpp:
        * tao/Incoming_Message_Queue.inl:
        * tao/Incoming_Message_Queue.h: Added lots of comments. Changed
          the name of the method copy_message () to copy_tail () and
          missing_data () to missing_data_tail (). Added a new static
          method TAO_Queued_Data::release () that releases a node.

        * tao/Transport.cpp: Accomodated the changes to the
          Incoming_Message_Queue to the Transport class.

        * tao/LIST_OF_TODO: Updated ..

Sat Jun 30 13:00:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_Base.h: Fixed a subtle bug while allocating
          memory for the datablock in the queue. Memory was allocated only
          for the exact number of bytes without thinking about the
          alignment. Increased the number of bytes allocated.

          Created the outgoing CDR streams on the stack with memory drawn
          from the TSS. We cannot share the buffers like the way it has
          been done so far.

        * tao/IIOP_Transport.cpp: Increased the debug_level for a debug
          statement.

        * This checkin fixes the AMI problems that have been seen so far.

        * tao/LIST_OF_TODO: Updated the list.

Fri Jun 29 18:30:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp: Fixed a bug in calculating the number
          of bytes that  needs to be copied into queue.

        * tao/Incoming_Message_Queue.cpp:
        * tao/Incoming_Message_Queue.h: Changed the signature of copy
          (). It returns the number of bytes copied instead of a void.

        * tao/Transport.cpp (parse_consolidate_messages):
        * tao/Transport.h: Added a new method parse_consolidate_messages
          () to the class. It does some common functaionalities like
          parsing the messages & consolidating the messages.

Thu Jun 28 18:30:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Resume_Handle.cpp: Added a check for resumable_handlers
          before actually resuming the handle.

Thu Jun 28 18:25:03 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp (process_queue_head):
        * tao/Transport.h: Added a new method that would take the message
          from the head of the incoming queue and send it for processing.

        * tao/GIOP_Message_Base.cpp: Minor formatting.

Thu Jun 28 17:16:50 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_Base.h:
        * tao/Pluggable_Messaging.h: Removed all the arguments that had an
          TAO_ORB_Core as one of the arguments. GIOP classes hold a copy
          of the pointer to the ORB_Core and so it is no
          necessary. Removed the following methods

          - is_message_complete ()
          - message_type ()

          The message_type () method is now local to the GIOP classes and
          it returns the Pluggable Message type from the
          GIOP_Message_State. This method is now used to fill the node
          with the right information type about the message.

        * tao/Transport.cpp: Used the information from the node of the
          queue to determine the type of message before processing it.

Thu Jun 28 15:12:32 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_Base.i: Moved the get_queued_data () from the
          .i  file to the .cpp file.

        * tao/Incoming_Message_Queue.h: #include'd a file.
        * tao/Makefile: Added new files.
        * tao/Resume_Handle.inl: Added the ACE_INLINE macros to the
          methods defined in this file.

Thu Jun 28 09:30:43 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp:
        * tao/GIOP_Message_Base.h:
        * tao/GIOP_Message_Base.i:
        * tao/GIOP_Message_Base.cpp: Made lots of changes to deal with the
          following

          - to parse & queue up messages if we have read multiple messages.
          - to consolidate messages if we had read half of a message
          - to process a consolidated message.

         We now use the node of the Incoming Message Queue to share data
         between the Transport layer and the GIOP layer.

        * tao/Incoming_Message_Queue.h:
        * tao/Incoming_Message_Queue.cpp:
        * tao/Incoming_Message_Queue.inl: Made the TAO_Queued_Data as a
          seperate class. This helped using the node of the queue to be
          shared between the GIOP layer. Removed many of the useless
          methods like wr_ptr (), is_message_complete (), add_message ()
          etc. We have the following methods that are more meaningful

          - is_tail_complete ()
          - is_message_complete ()
          - dequeue_head ()
          - dequeue_tail ()
          - enqueue_tail ()

         Added a static method get_queued_data () to create a node in the
         Message Queue.

         Added the protocol version information to the Queued Data.  We
         also need the message type and that has also been added.

        * tao/Resume_Handle.h:
        * tao/Resume_Handle.cpp:
        * tao/Resume_Handle.inl: This is a utility class that is used to
          resume handlers. This works more or less similar to our
          ACE_GUARD macros, but uses a flag to keep track whether the
          handle has been resumed.

        * tao/IIOP_Connection_Handler.cpp: Installed the Resume_Handle in
          handle_input () methods.

        * tao/Pluggable_Messaging.h: Removed the method  byte_order () and
          added the methods consolidate_node (), get_message_data () and
          extract_next_message ().

        * tao/Wait_On_Read.cpp:
        * tao/Connection_Handler.cpp: Changes to keep in sync with the
          changes to the signature of handle_input_i () in Transport
          class.

        * tao/LIST_OF_TODO: Updated the list of TODO's.

Mon Jun 25 19:21:43 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_Base.h:
        * tao/GIOP_Message_Base.i:
        * tao/GIOP_Message_State.inl:
        * tao/IIOP_Transport.cpp:
        * tao/Incoming_Message_Queue.cpp:
        * tao/Incoming_Message_Queue.h:
        * tao/Incoming_Message_Queue.inl:
        * tao/Pluggable_Messaging.h:
        * tao/Transport.cpp:
        * tao/Transport.h:  An inconsistent checkin only to transfer the
          files to the box at home.

Mon Jun 25 12:10:15 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Incoming_Message_Queue.{h,cpp,inl}:
        * tao/LIST_OF_TODO:
        * tao/GIOP_Message_State.inl: Added these to my branch.

Mon Jun 25 07:54:31 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/* : Merged the files from the main branch.

Mon Jun 25 07:45:38 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * Created this file.

Fri Jun 22 17:00:38 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/LIST_OF_TODO: Updated the list of TODO's.

Mon Jun 18 13:31:38 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Asynch_Reply_Dispatcher.cpp (dispatch_reply):
        * tao/Synch_Reply_Dispatcher.cpp:
        * tao/GIOP_Message_Base.cpp:
        * tao/Transport.cpp:
        * tao/DynamicInterface/DII_Reply_Dispatcher.cpp: Fixed warnings in
          g++ builds.

Sat Jun 17 17:46:23 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        This set of changes comes with complete revamping of the previous
        design. The flaws with the previous design were as follows
        (1) We were unnecessarily penalising large data blocks. We were
            trying to read a particular size of data till the data was
            completely removed from the socket. This was totally
            ridiculous because we were doing more reads than required.

        (2) The message block that was constructed on the stack with a
            buffer from stack never did what we wanted. It was allocating
            a data block on the heap and was thus spoiling whatever
            optimization that we had tried putting in.

        (3) The incoming message Queue is now managed by the TAO_Transport
            object instead of the GIOP classes.

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_Base.h: Removed the references to the incoming
          message queue. Implemention for two methods missing_data () and
          byte_order (). Added an extra argument to the methods
          process_request_message () and process_reply_message (). Used
          the incoming message block to create a input CDR with the
          DONT_DELETE flag so that the data block is not deleted after
          request processing.

        * tao/GIOP_Message_State.cpp: Removed the inclusion of
          Transport.h.

        * tao/Incoming_Message_Queue.h:
        * tao/Incoming_Message_Queue.cpp:
        * tao/Incoming_Message_Queue.inl: Added an argument to the
          add_message (). Further the implementation of add_message () has
          changed a bit. It now adds only a new message to the queue. It
          doesn't modify a half filled queue. The TAO_Transport object
          does that job. So declared the TAO_Transport as the friend class
          of the Incoming_Message_Queue.

          Changed the name of the methods complete_message () as
          is_complete_message (). Removed the methods current_message ()&
          current_byte_order ().

          Added a new method copy_message () which copies messages into
          the half empty nodes.

          Added a method wr_ptr () to access the write pointer of the tail
          node that has halfempty message.

        * tao/Transport.cpp:
        * tao/Transport.h: The Incoming Message Queue is now managed by
          this class. Added the following methods
          - missing_data ()
          - parse_incoming_messages ()
          - check_message_integrity ()
          - consolidate_message ()
          - conslodate_message_queue ()

        * tao/Pluggable_Messaging.h: Added two new virtual functions
          missing_data () and byte_order ().

        * tao/ORB_Core.h:
        * tao/ORB_Core.cpp:
        * tao/ORB_Core.i: Added an accessor for the locking_strategy used
          for the CDR blocks.

        * tao/Synch_Reply_Dispatcher.cpp:
        * tao/Asynch_Reply_Dispatcher.cpp:
        * tao/DynamicInterface/DII_Reply_Dispatcher.cpp: Changed the
          exchange_data_block () to clone_from () which is a new method in
          ACE_InputCDR.

        * tao/LIST_OF_TODO: Updated the list

Sat Jun 16 15:49:23 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Any.cpp:
        * tao/Asynch_Reply_Dispatcher.cpp:
        * tao/CDR.cpp:
        * tao/CDR.h:
        * tao/CDR.i:
        * tao/GIOP_Message_Lite.cpp:
        * tao/IIOP_Profile.cpp:
        * tao/Invocation.cpp:
        * tao/ORB.cpp:
        * tao/Pluggable_Messaging_Utils.cpp:
        * tao/Synch_Reply_Dispatcher.cpp:
        * tao/TAO_Server_Request.cpp:
        * tao/DynamicInterface/DII_Reply_Dispatcher.cpp: Integrated some
          of the changes from the main trunk in this branch.

Wed Jun 13 17:55:24 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Incoming_Message_Queue.cpp:
        * tao/Incoming_Message_Queue.inl: Fixed a bad way to use the tail
          of the circular linked list. This is now replaced by the size of
          the linked list. This fixes quite a few errors.

        * tao/Transport.cpp: On read () if we get errno == EWOULDBLOCK we
          were closing the connection prematurely. This is not right. We
          should only return a 0 to the reactor, so that it can call us
          back when there is data in the socket.

Wed Jun 13 10:45:24 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_Base.cpp (process_reply_message): Fixed the
          reply parsing and generation of the CDR stream that is passed on
          to the higher layers of the ORB.

Wed Jun 13 07:25:24 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Makefile: Added new files.
        * tao/Exclusive_TMS.cpp: Removed some commented code.

Tue Jun  12 18:42:55 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_State.h:
        * tao/GIOP_Message_State.inl: The message_size () now returns the
          payload size + the GIOP header length. Added a new method
          payload_size () that returns the payload size alone.

        * tao/GIOP_Message_Base.cpp: Fixed a warning with g++ builds.

        * tao/Connection_Handler.h:
        * tao/Connection_Handler.cpp: The svc_i () method now calls
          handle_input_i () on the transport instead of the method in the
          same class. The handle_input_i () in the TAO_Connection_Handler
          class has been removed as it is no longer used.

        * tao/IIOP_Connection_Handler.h:
        * tao/IIOP_Connection_Handler.cpp: Removed the implementation of
          handle_input_i ().

        * tao/Incoming_Message_Queue.cpp: When trying to make a new data
          block we dont add the size of GIOP header. The message_size ()
          now returns with that value.

        * tao/Incoming_Message_Queue.inl: Fixed a link error in g++
          builds.

Tue Jun  12 17:42:55 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        First set of checkins for big two ways.

        * tao/Connection_Handler.h: Added #define for the default
          buffer size . Not sure yet whether this is the right place for
          it.

        * tao/Incoming_Message_Queue.h:
        * tao/Incoming_Message_Queue.cpp:
        * tao/Incoming_Message_Queue.inl: Queue up the incoming
          messages. We form a circular linked list of messages that are
          bigger than a particular buffer size. During processing we take
          messages of the queue to pass it onto the higher layers of the
          ORB.

        * tao/GIOP_Message_Base.cpp:
        * tao/GIOP_Message_Base.h: Added support for big two way
          requests. If the message has only been partially read we add the
          message in the queue and then go for the next read. During
          processing we check if the queue has messages before processing
          the message on hand.

        * tao/GIOP_Message_State.h:
        * tao/GIOP_Message_State.cpp:
        * tao/GIOP_Message_State.inl: Added accessor methods.

        * tao/Transport.cpp: Added support for bigger two ways. If we
          receive a two way bigger than a particular size we read and
          queue the message and do a further read to retrieve the rest of
          the message.

        * tao/TAO.dsp: Added new files.
        * tao/default_resource.cpp: For single threaded configuration we
          have been creating a locked data block and that seems to defy
          logic. We now should create a lock free datablock for single
          threaded configurations.

Sat Jun  2 12:02:55 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/* : Merged with the main trunk.

Fri Jun  1 17:22:29 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/GIOP_Message_State.cpp (parse_message_header_i):
        * tao/Connection_Handler.cpp:
        * tao/GIOP_Message_Base.cpp:
        * tao/Makefile:
        * tao/Transport.cpp: Fixed warnings and errors in Linux g++.

Thu Jun 01 13:39:02 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.cpp:
        * tao/Transport.h:  Added the following methods -- handle_input_i
          (), process_parsed_message (). The handle_input_i () creates a
          buffer on the stack. It uses the buffer to read the
          messages. Any errors in read () or dealt here directly. The read
          message is then parsed and processed by the GIOP classes. Moved
          most of the code for processing here as they seem to be common
          between different transport protocols. Once we have received a
          full message, we just resume the handler and go ahead with
          processing the message.

        * tao/IIOP_Connection_Handler.cpp (handle_input ()): Calls
          handle_input_i () on the transport.

        * tao/Connection_Handler.h:
        * tao/Connection_Handler.cpp:  The svc_i () calls the
          handle_input_i () on the transport instead of the same call on
          the connection handler. This way we should be able to share the
          same code among different protocol objects.

        * tao/GIOP_Message_State.h:
        * tao/GIOP_Message_State.cpp: Much of the message parsing and
          state information is stored here.

        * tao/GIOP_Message_Base.h:
        * tao/GIOP_Message_Base.cpp:
        * tao/Pluggable_Messaging.h: made some changes to suit the
          above. But more changes could come in, to suit different
          protocols. We may want to change the interfaces in
          Pluggable_Messaging to give a much cleaner interface to the
          world.

        * tao/Wait_On_Read.cpp: Used the handle_input_i () on the
          transport for processing the incoming data.

        * tao/TAO.dsp:
        * tao/PortableServer/TAO_PortableServer.dsp: Added new files and
          removed files from GIOP_Lite in the first round.

        * tao/GIOP_Message_Reactive_Handler.cpp:
        * tao/GIOP_Message_Reactive_Handler.h: Looks like these files
          would be removed during the final merge.

        The above checkins works for simple two way calls. The
        BiDirectional_NestedUpcall test works with the TP_Reactor. This
        works on Win32.


Thu May 24 12:19:02 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/Transport.h: Added some design forces for the input data
          path.

        End of ChangeLog entries from branch bug_575_stage_2.

Thu Jul  5 22:04:24 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tests/LongUpcalls/run_ami_test.pl: Adding a new perl script to
          this test. This is supposed to fail in the daily builds as this
          needs bug fix 575.

Thu Jul  5 21:16:00 2001  Craig Rodrigues <crodrigu@bbn.com>

         * orbsvcs/orbsvcs/AV/AV_Core.h

           Added some Borland compiler pragma options to prevent
           singleton problems when AV_Core is in a DLL.

Thu Jul  5 20:41:00 2001  Craig Rodrigues <crodrigu@bbn.com>

         * tao/CONV_FRAMEC.h
         * tao/DynamicC.h
         * tao/PolicyC.h
            Added export macros for some Unbounded Sequence types.

         * TAO_IDL/be/be_visitor_sequence/gen_unbounded_obj_sequence_ch.cpp
            Add export macro for unbounded object sequence types

Thu Jul  5 09:22:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * orbsvcs/tests/AVStreams/Makefile.bor
          Added  Asynch_Three_Stage
        * orbsvcs/tests/AVStreams/Asynch_Three_Stage/*.bor
          Added BCB makefiles

Thu Jul  5 04:53:00 2001  Craig Rodrigues <crodrigu@bbn.com>

        * orbsvcs/tests/AVStreams/Component_Switching/Makefile
          Add missing link flag -lTAO_Strategies, needed
          for Advanced_Resource_Factory in svc.conf.

        * orbsvcs/tests/AVStreams/Component_Switching/receiver.cpp
          Fix typo

        * orbsvcs/tests/AVStreams/Component_Switching/run_test.pl
          Add -ORBSkipServiceConfigOpen to startup of Naming_Service,
          it did not like the svc.conf file in this directory.

Wed Jul  4 09:53:01 2001  Frank Hunleth  <fhunleth@cs.wustl.edu>

        * tao/Strategies/TAO_Strategies_Static.dsp:

          Updated static build with DIOP files.

Wed Jul  4 09:52:20 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/DynamicAny/DynCommon.cpp (insert_reference):

          Added a call to _is_a() as a final check to see if there
          is a type mismatch between the dynamic any and the
          object reference argument. Thanks to Jonathan Biggar
          <jon@floorboard.com> for clarifying the spec and to
          Philippe Merle <Philippe.Merle@lifl.fr> for suggesting
          optimizations.

Wed Jul  4 09:10:24 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_visitor_interface/tie_sh.cpp:
        * TAO_IDL/be/be_visitor_root/root_sth.cpp:

          Removed ACE_HAS_USING_KEYWORD guard from each TIE class
          declaration, and added it to the global module reopening
          where all TIE class declarations will occur. This
          includes the TAO_NAMESPACE macro inside the guard, which
          is what we want if the platform does not support
          namespaces.

Wed Jul  4 14:05:12 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * orbsvcs/tests/Makefile.bor
          Updated to include AVStreams

        * orbsvcs/tests/AVStreams/Component_Switching/*.bor
        * orbsvcs/tests/AVStreams/Simple_Three_Stage/*.bor
        * orbsvcs/tests/AVStreams/Simgle_Two_Stage/*.bor
          Updated these BCB makefiles so that these tests are now
          build without errors

Tue Jul  3 20:09:58 2001  Frank Hunleth  <fhunleth@cs.wustl.edu>

        * tao/Strategies/TAO_Strategies.dsp:

          Added DIOP source files to project.

Tue Jul  3 20:09:58 2001  Frank Hunleth  <fhunleth@cs.wustl.edu>

        * examples/PluggableUDP/tests/Basic/Makefile:
        * examples/PluggableUDP/tests/Performance/Makefile:
        * examples/PluggableUDP/tests/SimplePerformance/Makefile:
        * tao/Strategies/Makefile:

          Updated dependencies.

Tue Jul  3 18:25:41 2001  Frank Hunleth  <fhunleth@cs.wustl.edu>

        * examples/PluggableUDP/DIOP/*:

          Moved DIOP source files from PluggableUDP to Strategies.

        * examples/PluggableUDP/README:
        * examples/PluggableUDP/DIOP/README:

          Updated README files to note file location change.

        * examples/PluggableUDP/Makefile:
        * examples/PluggableUDP/Makefile.bor:

          Removed DIOP directory from build.

        * examples/PluggableUDP/tests/Basic/Makefile:
        * examples/PluggableUDP/tests/Basic/client.bor:
        * examples/PluggableUDP/tests/Basic/client.cpp:
        * examples/PluggableUDP/tests/Basic/client.dsp:
        * examples/PluggableUDP/tests/Basic/server.bor:
        * examples/PluggableUDP/tests/Basic/server.cpp:
        * examples/PluggableUDP/tests/Basic/server.dsp:
        * examples/PluggableUDP/tests/Basic/svc.conf:
        * examples/PluggableUDP/tests/Performance/Makefile:
        * examples/PluggableUDP/tests/Performance/client.bor:
        * examples/PluggableUDP/tests/Performance/client.cpp:
        * examples/PluggableUDP/tests/Performance/client.dsp:
        * examples/PluggableUDP/tests/Performance/server.bor:
        * examples/PluggableUDP/tests/Performance/server.cpp:
        * examples/PluggableUDP/tests/Performance/server.dsp:
        * examples/PluggableUDP/tests/Performance/svc.conf:
        * examples/PluggableUDP/tests/SimplePerformance/Makefile:
        * examples/PluggableUDP/tests/SimplePerformance/client.bor:
        * examples/PluggableUDP/tests/SimplePerformance/client.cpp:
        * examples/PluggableUDP/tests/SimplePerformance/client.dsp:
        * examples/PluggableUDP/tests/SimplePerformance/server.bor:
        * examples/PluggableUDP/tests/SimplePerformance/server.cpp:
        * examples/PluggableUDP/tests/SimplePerformance/server.dsp:
        * examples/PluggableUDP/tests/SimplePerformance/svc.conf:

          Updated source files and Makefiles to look for the DIOP
          code in the Strategies library rather than in the DIOP library.

        * tao/orbconf.h:

          Added #define to control whether DIOP is compiled in or not.

        * tao/Strategies/DIOP_Acceptor.cpp:
        * tao/Strategies/DIOP_Acceptor.h:
        * tao/Strategies/DIOP_Acceptor.i:
        * tao/Strategies/DIOP_Connection_Handler.cpp:
        * tao/Strategies/DIOP_Connection_Handler.h:
        * tao/Strategies/DIOP_Connection_Handler.i:
        * tao/Strategies/DIOP_Connector.cpp:
        * tao/Strategies/DIOP_Connector.h:
        * tao/Strategies/DIOP_Endpoint.cpp:
        * tao/Strategies/DIOP_Endpoint.h:
        * tao/Strategies/DIOP_Endpoint.i:
        * tao/Strategies/DIOP_Factory.cpp:
        * tao/Strategies/DIOP_Factory.h:
        * tao/Strategies/DIOP_Profile.cpp:
        * tao/Strategies/DIOP_Profile.h:
        * tao/Strategies/DIOP_Profile.i:
        * tao/Strategies/DIOP_Transport.cpp:
        * tao/Strategies/DIOP_Transport.h:
        * tao/Strategies/DIOP_Transport.i:

          Moved DIOP files to Strategies.  Updated linking specifiers and
          include files to reflect change.

        * tao/Strategies/Makefile:
        * tao/Strategies/Makefile.bor:

          Added DIOP files.

        * tao/Strategies/TAO_Strategies_Internal.cpp:
        * tao/Strategies/advanced_resource.cpp:

          Added service configurator hooks for DIOP.

Tue Jul  3 14:53:00 2001  Craig Rodrigues <crodrigu@bbn.com>

        * orbsvcs/orbsvcs/AV/AV_Core.h
        * orbsvcs/orbsvcs/AV/RTCP.h
        * orbsvcs/orbsvcs/AV/RTP.h
        * orbsvcs/orbsvcs/AV/Transport.h
        * orbsvcs/orbsvcs/AV/UDP.h

          Fix some includes, add TAO_AV_Export in more class
          declarations.

Tue Jul  3 11:27:03 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/be/be_codegen.cpp:
        * TAO_IDL/be/be_visitor_interface/tie_sh.cpp:
        * TAO_IDL/be/be_visitor_interface/tie_si.cpp:

          Removed the #if defined (ACE_HAS_USING_KEYWORD)
          guards from the top and bottom of the *S_T.{h,i}
          files and added their generation on a TIE class by
          TIE class basis, conditionally, if the original
          interface is defined inside a module. TIE class
          code will compile for interfaces declared at global
          scope, even on platforms that do not support
          namespaces. This enhancement was requested by
          Marco Kranawetter <Marco.Kranawetter@icn.siemens.de>.

Tue Jul  3 09:55:22 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tao/DynamicAny/DynUnion_i.cpp:

          Added missing .in() to a CORBA::Any_var.

        * tao/DynamicAny/DynCommon.cpp:

          Removed unnecessary break statement.

Tue Jul  3 08:45:13 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tests/Smart_Proxies/Benchmark/client.dsp:
        * tests/Smart_Proxies/Benchmark/server.dsp:

          Fixed IDL compiler settings in the release version.

Mon Jul  2 08:59:16 2001  Johnny Willemsen <jwillemsen@remedy.nl>

        * orbsvcs/tests/AVStreams/Makefile.bor
        Added BCB makefile

Mon Jul  2 23:13:00 2001  Craig Rodrigues <crodrigu@bbn.com>

        * orbsvcs/orbsvcs/AV/AVStreams_i.h
        * orbsvcs/orbsvcs/AV/Protocol_Factory.h
        * orbsvcs/orbsvcs/AV/Transport.h
        * orbsvcs/orbsvcs/AV/UDP.h

        Added TAO_AV_Export to more classes to fix Borland DLL builds.

        * orbsvcs/tests/AVStreams/Simple_Two_Stage/receiver.bor
        * orbsvcs/tests/AVStreams/Simple_Two_Stage/sender.bor

        Added more linker flags to fix Borland DLL builds.

Mon Jul  2 19:32:09 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tests/Queued_Message_Test/Queued_Message_Test.cpp:

          Changed use of ACE_OS::rand_r(seed) to ACE_OS::srand(seed)
          and ACE_OS::rand(). ACE_OS::rand_r(seed) is not supported
          on Win32 platforms, and was causing the test to hang.

Mon Jul  2 17:06:22 2001  Balachandran Natarajan  <bala@cs.wustl.edu>

        * tao/ORB_Core.cpp: Initialization of the transport_cache ()
          should preceed the preconnects (). Thanks to Irfan for pointing
          this out.

Mon Jul  2 15:53:29 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * orbsvcs/tests/InterfaceRepo/IFR_Test/Admin_Client.cpp:
        * tao/Typecode.cpp:
        * tao/append.cpp:
        * tao/skip.cpp:
        * tao/DynamicAny/DynUnion_i.cpp:

          Modified CORBA::TypeCode::member_label() to return a duplicate
          Any pointer (to be comsumed by the caller), making it consistent
          with other ORB functions that return pseudo-object types. Also
          modified all uses of member_label() in TAO to prevent memory
          leaks. Thanks to Philippe Merle <Philippe.Merle@lifl.fr> for
          pointing out this inconsistency, and to Jonanthan Biggar
          <jon@floorboard.com> for clarifying the spec with regard to
          memory management of pseudo-objects.

Mon Jul  2 14:29:49 2001  Frank Hunleth  <fhunleth@cs.wustl.edu>

        * performance-tests/RTCorba/Multiple_Endpoints/Single_Endpoint/Makefile:
        * performance-tests/RTCorba/Multiple_Endpoints/Orb_Per_Priority/Makefile:

          Removed duplicate ACE and TAO library includes.  Should fix KCC
          linker warnings.

Mon Jul 2  14:50:00 2001  Craig Rodrigues <crodrigu@bbn.com>

        * orbsvcs/AV/AVStreams_i.cpp

          Only display certain debug messages when higher ORB debug level
          is specified.

Mon Jul  2 13:37:45 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * TAO_IDL/fe/idl.ll:
        * TAO_IDL/fe/lex.yy.cpp:

          Added '\x' to the list of escaped characters in the
          regular expression for strings in the TAO IDL lexer,
          and regenerated the C++ file. Legal IDL strings such
          as "\xA" were producing syntax errors. Thanks to
          Jules Colding <dsl11814@vip.cybercity.dk> for reporting
          this bug.

Mon Jul  2 13:18:14 2001  Douglas C. Schmidt  <schmidt@tango.doc.wustl.edu>

        * Moved all the ChangeLog-* files into the new ChangeLogs
          directory.  Thanks to Holger P . Krekel <krekel@merlinux.de> for
          this suggestion.

Mon Jul  2 10:57:26 2001  Jeff Parsons <parsons@cs.wustl.edu>

        * tests/Smart_Proxies/Benchmark/Benchmark.dsw:
        * tests/Smart_Proxies/Benchmark/client.dsp:
        * tests/Smart_Proxies/Benchmark/server.dsp:

          New files to build the smart proxies benchmark test on
          Win32 platforms.

        * tests/Smart_Proxies/Benchmark/client.cpp:
        * tests/Smart_Proxies/Benchmark/Smart_Proxy_Impl.cpp:
        * tests/Smart_Proxies/Benchmark/Smart_Proxy_Impl.h:

          Removed unnecessary TAO_HAS_SMART_PROXIES macro from
          these files.

        * docs/Smart_Proxies.html:

          Updated documentation.
