diff options
author | elliott_c <ocielliottc@users.noreply.github.com> | 2003-01-03 13:51:06 +0000 |
---|---|---|
committer | elliott_c <ocielliottc@users.noreply.github.com> | 2003-01-03 13:51:06 +0000 |
commit | f5922d948fd6b3dc4ef400ec129ad0d666d03c4a (patch) | |
tree | f9576e90f9db172ddff482ea41439b806dc27b89 | |
parent | 9f09a39fb2627c9bcff8b1cfc9caf9052b7c1823 (diff) | |
download | ATCD-f5922d948fd6b3dc4ef400ec129ad0d666d03c4a.tar.gz |
ChangeLogTag: Fri Jan 3 07:48:01 2003 Chad Elliott <elliott_c@ociweb.com>
-rw-r--r-- | TAO/ChangeLog | 793 | ||||
-rw-r--r-- | TAO/tao/DynamicAny/DynamicAnyC.h | 2 | ||||
-rw-r--r-- | TAO/tao/RTCORBA/RTCORBAC.h | 2 |
3 files changed, 402 insertions, 395 deletions
diff --git a/TAO/ChangeLog b/TAO/ChangeLog index 9a7e521ee6d..a54b89399d4 100644 --- a/TAO/ChangeLog +++ b/TAO/ChangeLog @@ -1,3 +1,10 @@ +Fri Jan 3 07:48:01 2003 Chad Elliott <elliott_c@ociweb.com> + + * tao/DynamicAny/DynamicAnyC.h: + * tao/RTCORBA/RTCORBAC.h: + + Changed the name of the AIX template instantiation macro. + Thu Jan 2 17:51:49 2003 Jeff Parsons <j.parsons@vanderbilt.edu> * TAO_IDL/be/be_visitor_valuetype/arglist.cpp: @@ -7,10 +14,10 @@ Thu Jan 2 17:51:49 2003 Jeff Parsons <j.parsons@vanderbilt.edu> * TAO_IDL/be_include/be_codegen.h: * TAO_IDL/be_include/be_visitor_valuetype/valuetype_cs.h: - Moved generation of AMH exception holder raise_* method - body from the stub header file to the stub source file. Also - added explicit include of ace/Auto_Ptr when generating AMH - classes (needed to correct AIX build errors). + Moved generation of AMH exception holder raise_* method + body from the stub header file to the stub source file. Also + added explicit include of ace/Auto_Ptr when generating AMH + classes (needed to correct AIX build errors). Thu Jan 2 13:33:01 2003 Chad Elliott <elliott_c@ociweb.com> @@ -40,261 +47,261 @@ Thu Jan 2 13:33:01 2003 Chad Elliott <elliott_c@ociweb.com> Thu Jan 2 10:33:29 2003 Boris S Kolpackov <boris@isis-server.isis.vanderbilt.edu> - * tao/PortableServer/ImplRepoC.h: - * tao/PortableServer/ImplRepoS.cpp: - * tao/PortableServer/ImplRepoS.i: - * tao/PortableServer/ImplRepoS_T.h: + * tao/PortableServer/ImplRepoC.h: + * tao/PortableServer/ImplRepoS.cpp: + * tao/PortableServer/ImplRepoS.i: + * tao/PortableServer/ImplRepoS_T.h: * tao/PortableServer/ImplRepoC.cpp: - * tao/PortableServer/ImplRepoC.i: - * tao/PortableServer/ImplRepoS.h: - * tao/PortableServer/ImplRepoS_T.cpp: - * tao/PortableServer/ImplRepoS_T.i: - * tao/PortableServer/diffs/ImplRepo.diff: Regenerated - ImplRepo.pidl which apparently fixed bug 1413. ImplRepo.diff - was regenerate as well. + * tao/PortableServer/ImplRepoC.i: + * tao/PortableServer/ImplRepoS.h: + * tao/PortableServer/ImplRepoS_T.cpp: + * tao/PortableServer/ImplRepoS_T.i: + * tao/PortableServer/diffs/ImplRepo.diff: Regenerated + ImplRepo.pidl which apparently fixed bug 1413. ImplRepo.diff + was regenerate as well. Thu Jan 2 00:15:57 2003 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tests/Bug_1020_Regression/ORB_Task.cpp: - * tests/Bug_1020_Regression/client.cpp: - * tests/Bug_1270_Regression/ORB_Task.cpp: - * tests/Bug_1270_Regression/client.cpp: - * tests/Bug_1361_Regression/ORB_Task.cpp: - * tests/Bug_1361_Regression/client.cpp: - * tests/Nested_Upcall_Crash/Client_Peer.cpp: - * tests/Nested_Upcall_Crash/client.cpp: - * tests/Nested_Upcall_Crash/server.cpp: - * tests/Server_Connection_Purging/client.cpp: Added missing - includes to keep the compilation on AIX happy. Missing header - files were tao/Environment.h and ace/Reactor.h. - - * tao/Utils/Servant_Var.inl: - * tao/Utils/Servant_Var.h: #included ace/CORBA_Macros.h and - tao/Environment.h to keep the AIX compilers happy. - - Thanks to Jon Wackley for donating these patches. + * tests/Bug_1020_Regression/ORB_Task.cpp: + * tests/Bug_1020_Regression/client.cpp: + * tests/Bug_1270_Regression/ORB_Task.cpp: + * tests/Bug_1270_Regression/client.cpp: + * tests/Bug_1361_Regression/ORB_Task.cpp: + * tests/Bug_1361_Regression/client.cpp: + * tests/Nested_Upcall_Crash/Client_Peer.cpp: + * tests/Nested_Upcall_Crash/client.cpp: + * tests/Nested_Upcall_Crash/server.cpp: + * tests/Server_Connection_Purging/client.cpp: Added missing + includes to keep the compilation on AIX happy. Missing header + files were tao/Environment.h and ace/Reactor.h. + + * tao/Utils/Servant_Var.inl: + * tao/Utils/Servant_Var.h: #included ace/CORBA_Macros.h and + tao/Environment.h to keep the AIX compilers happy. + + Thanks to Jon Wackley for donating these patches. Tue Dec 31 23:53:17 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * TAO_IDL/include/idl_global.h: Added missing include to get it - compiling on AIX. + * TAO_IDL/include/idl_global.h: Added missing include to get it + compiling on AIX. - * tests/Bug_1269_Regression/ORB_Task.cpp: #included - tao/Environment.h to get it compiling on AIX. + * tests/Bug_1269_Regression/ORB_Task.cpp: #included + tao/Environment.h to get it compiling on AIX. - Thanks to Jon Wackley for the patches. + Thanks to Jon Wackley for the patches. Tue Dec 31 21:03:37 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * performance-tests/Makefile: Added the new directory to the list - of DIRS - * performance-tests/Memory/Makefile: Added a new Makefile to - compile the sub-directories. + * performance-tests/Makefile: Added the new directory to the list + of DIRS + * performance-tests/Memory/Makefile: Added a new Makefile to + compile the sub-directories. Tue Dec 31 20:55:48 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * performance-tests/Memory/Single_Threaded: A simple test for - tracking memory usage. This is going to be compiled only on Linux + * performance-tests/Memory/Single_Threaded: A simple test for + tracking memory usage. This is going to be compiled only on Linux - * performance-tests/Memory/Single_Threaded/Memory_Growth.mpc: - * performance-tests/Memory/Single_Threaded/Makefile: - * performance-tests/Memory/Single_Threaded/Memory_Growth.cpp: - * performance-tests/Memory/Single_Threaded/Memory_Growth.h: - * performance-tests/Memory/Single_Threaded/README: - * performance-tests/Memory/Single_Threaded/Test.idl: - * performance-tests/Memory/Single_Threaded/client.cpp: - * performance-tests/Memory/Single_Threaded/server.cpp: Files for - the new test. + * performance-tests/Memory/Single_Threaded/Memory_Growth.mpc: + * performance-tests/Memory/Single_Threaded/Makefile: + * performance-tests/Memory/Single_Threaded/Memory_Growth.cpp: + * performance-tests/Memory/Single_Threaded/Memory_Growth.h: + * performance-tests/Memory/Single_Threaded/README: + * performance-tests/Memory/Single_Threaded/Test.idl: + * performance-tests/Memory/Single_Threaded/client.cpp: + * performance-tests/Memory/Single_Threaded/server.cpp: Files for + the new test. Tue Dec 31 19:33:09 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tao/Condition.h: - * tao/Condition.cpp: Added missing includes for the AIX compiler. - Thanks to Jon Wackley for providing this patch. - - * tao/Leader_Follower.cpp (wait_for_event): Cached the transport - id before waiting as a follower or drive the event loop as - leader, and use the cached id for all the debug statements. The - transport object could dissappear during the process of waiting - for connection establishment or connection completion and calls - on the transport object could crash. Thanks to Duane Binder for - reporting this problem. + * tao/Condition.h: + * tao/Condition.cpp: Added missing includes for the AIX compiler. + Thanks to Jon Wackley for providing this patch. + + * tao/Leader_Follower.cpp (wait_for_event): Cached the transport + id before waiting as a follower or drive the event loop as + leader, and use the cached id for all the debug statements. The + transport object could dissappear during the process of waiting + for connection establishment or connection completion and calls + on the transport object could crash. Thanks to Duane Binder for + reporting this problem. Tue Dec 31 14:03:53 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * tests/RTCORBA/Profile_And_Endpoint_Selection/lanes_r: - * tests/RTCORBA/Profile_And_Endpoint_Selection/bands_r: - * tests/RTCORBA/Profile_And_Endpoint_Selection/invocation_priorities_r: - * tests/RTCORBA/Profile_And_Endpoint_Selection/run_test.pl: - Changed to run the tests with reduced numbered priority lanes on - Solaris to avoid exhausting the limited SysV semaphores and - causing unnecessary warnings. Thanks to Irfan for the new set - of configuration files. + * tests/RTCORBA/Profile_And_Endpoint_Selection/lanes_r: + * tests/RTCORBA/Profile_And_Endpoint_Selection/bands_r: + * tests/RTCORBA/Profile_And_Endpoint_Selection/invocation_priorities_r: + * tests/RTCORBA/Profile_And_Endpoint_Selection/run_test.pl: + Changed to run the tests with reduced numbered priority lanes on + Solaris to avoid exhausting the limited SysV semaphores and + causing unnecessary warnings. Thanks to Irfan for the new set + of configuration files. Mon Dec 30 12:10:49 2002 Balachandran Natarajan <bala@cs.wustl.edu> - * examples/AMH/Sink_Server/st_server.dsp: - * examples/AMH/Sink_Server/client.dsp: - * examples/AMH/Sink_Server/mt_server.dsp: Changed the include path - information in the release builds. + * examples/AMH/Sink_Server/st_server.dsp: + * examples/AMH/Sink_Server/client.dsp: + * examples/AMH/Sink_Server/mt_server.dsp: Changed the include path + information in the release builds. Mon Dec 30 08:49:31 2002 Jeff Parsons <j.parsons@vanderbilt.edu> * TAO_IDL/util/utl_global.cpp: - Fixed bug in #pragma prefix when the prefix directive is followed by - an #include directive. Thanks to Adee Ran <adeeran@yahoo.com> for - reporting the bug. + Fixed bug in #pragma prefix when the prefix directive is followed by + an #include directive. Thanks to Adee Ran <adeeran@yahoo.com> for + reporting the bug. Mon Dec 30 06:12:53 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * examples/AMH/Sink_Server/st_server.dsp (USERDEP__TEST_): - * examples/AMH/Sink_Server/client.dsp: - * examples/AMH/Sink_Server/mt_server.dsp: Changed the path to the - tao_idl in the external_dependencies section. This should fix - the warnings in the daily builds. + * examples/AMH/Sink_Server/st_server.dsp (USERDEP__TEST_): + * examples/AMH/Sink_Server/client.dsp: + * examples/AMH/Sink_Server/mt_server.dsp: Changed the path to the + tao_idl in the external_dependencies section. This should fix + the warnings in the daily builds. Sun Dec 29 11:29:20 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * examples/AMH/Sink_Server/AMH_Sink_Server.dsw: - * examples/AMH/Sink_Server/client.dsp: - * examples/AMH/Sink_Server/st_server.dsp: - * examples/AMH/Sink_Server/mt_server.dsp: - * tests/AMH_Exceptions/AMH_Exceptions.dsw: - * tests/AMH_Exceptions/server.dsp: - * tests/AMH_Exceptions/client.dsp: Added the above files. No need - to explain why they need to be added ;) + * examples/AMH/Sink_Server/AMH_Sink_Server.dsw: + * examples/AMH/Sink_Server/client.dsp: + * examples/AMH/Sink_Server/st_server.dsp: + * examples/AMH/Sink_Server/mt_server.dsp: + * tests/AMH_Exceptions/AMH_Exceptions.dsw: + * tests/AMH_Exceptions/server.dsp: + * tests/AMH_Exceptions/client.dsp: Added the above files. No need + to explain why they need to be added ;) Sun Dec 29 08:44:06 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * TAO_IDL/ast/ast_array.cpp: - * TAO_IDL/ast/ast_sequence.cpp: - * TAO_IDL/be/be_array.cpp: - * TAO_IDL/be/be_decl.cpp: - * TAO_IDL/be/be_sequence.cpp: - * TAO_IDL/be/be_union_branch.cpp: - * TAO_IDL/be/be_valuetype.cpp: - * TAO_IDL/be/be_visitor_amh_pre_proc.cpp: - * TAO_IDL/be/be_visitor_ami_pre_proc.cpp: - * TAO_IDL/be/be_visitor_argument.cpp: - * TAO_IDL/be/be_visitor_array.cpp: - * TAO_IDL/be/be_visitor_attribute.cpp: - * TAO_IDL/be/be_visitor_component.cpp: - * TAO_IDL/be/be_visitor_decl.cpp: - * TAO_IDL/be/be_visitor_exception.cpp: - * TAO_IDL/be/be_visitor_field.cpp: - * TAO_IDL/be/be_visitor_interface_fwd.cpp: - * TAO_IDL/be/be_visitor_operation.cpp: - * TAO_IDL/be/be_visitor_scope.cpp: - * TAO_IDL/be/be_visitor_sequence.cpp: - * TAO_IDL/be/be_visitor_structure.cpp: - * TAO_IDL/be/be_visitor_typedef.cpp: - * TAO_IDL/be/be_visitor_union.cpp: - * TAO_IDL/be/be_visitor_union_branch.cpp: - * TAO_IDL/be/be_visitor_valuetype_fwd.cpp: #include'd - ace/Log_Msg.h to all the above files. This helps compilation to - go smooth on AIX. - - * TAO_IDL/include/ast_module.h: #include'd ace/Unbounded_Set.h. - - * tao/Condition.h: #include'd tao/orbconf.h and forward declared + * TAO_IDL/ast/ast_array.cpp: + * TAO_IDL/ast/ast_sequence.cpp: + * TAO_IDL/be/be_array.cpp: + * TAO_IDL/be/be_decl.cpp: + * TAO_IDL/be/be_sequence.cpp: + * TAO_IDL/be/be_union_branch.cpp: + * TAO_IDL/be/be_valuetype.cpp: + * TAO_IDL/be/be_visitor_amh_pre_proc.cpp: + * TAO_IDL/be/be_visitor_ami_pre_proc.cpp: + * TAO_IDL/be/be_visitor_argument.cpp: + * TAO_IDL/be/be_visitor_array.cpp: + * TAO_IDL/be/be_visitor_attribute.cpp: + * TAO_IDL/be/be_visitor_component.cpp: + * TAO_IDL/be/be_visitor_decl.cpp: + * TAO_IDL/be/be_visitor_exception.cpp: + * TAO_IDL/be/be_visitor_field.cpp: + * TAO_IDL/be/be_visitor_interface_fwd.cpp: + * TAO_IDL/be/be_visitor_operation.cpp: + * TAO_IDL/be/be_visitor_scope.cpp: + * TAO_IDL/be/be_visitor_sequence.cpp: + * TAO_IDL/be/be_visitor_structure.cpp: + * TAO_IDL/be/be_visitor_typedef.cpp: + * TAO_IDL/be/be_visitor_union.cpp: + * TAO_IDL/be/be_visitor_union_branch.cpp: + * TAO_IDL/be/be_visitor_valuetype_fwd.cpp: #include'd + ace/Log_Msg.h to all the above files. This helps compilation to + go smooth on AIX. + + * TAO_IDL/include/ast_module.h: #include'd ace/Unbounded_Set.h. + + * tao/Condition.h: #include'd tao/orbconf.h and forward declared ACE_Time_Value - * tao/FILE_Parser.cpp: - * tao/Typecode.cpp: #include'd ace/Malloc_Base.h. + * tao/FILE_Parser.cpp: + * tao/Typecode.cpp: #include'd ace/Malloc_Base.h. - All the above changes help compilation on AIX. Thanks to Jon - Wackley <jonw@mountaincable.net> for providing these patches. + All the above changes help compilation on AIX. Thanks to Jon + Wackley <jonw@mountaincable.net> for providing these patches. - * tao/Muxed_TMS.cpp: - * tao/Muxed_TMS.h: Use recursive mutex instead of a simple - mutex. This prevents deadlock that might occur, when a thread - tries to send a request during reply dispatching. + * tao/Muxed_TMS.cpp: + * tao/Muxed_TMS.h: Use recursive mutex instead of a simple + mutex. This prevents deadlock that might occur, when a thread + tries to send a request during reply dispatching. Sat Dec 28 13:09:53 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * tao/Strategies/SHMIOP_Transport.cpp (recv_i): Changed the recv_i - method to handle closed socket correctly. The original - implementation caused SHMIOP falling into an infinite loop when - the other end closed the connection. + * tao/Strategies/SHMIOP_Transport.cpp (recv_i): Changed the recv_i + method to handle closed socket correctly. The original + implementation caused SHMIOP falling into an infinite loop when + the other end closed the connection. Sat Dec 28 10:33:20 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tests/Bug_1361_Regression/Server_Thread_Pool.cpp: Fixed warnings - in Full builds. + * tests/Bug_1361_Regression/Server_Thread_Pool.cpp: Fixed warnings + in Full builds. Sat Dec 28 09:35:36 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tests/Makefile.bor: Removed ior_corbaloc test. + * tests/Makefile.bor: Removed ior_corbaloc test. Fri Dec 27 20:00:41 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tests/Bug_1361_Regression/run_test.pl: Added more threads in the - client to stress the ORB. + * tests/Bug_1361_Regression/run_test.pl: Added more threads in the + client to stress the ORB. Fri Dec 27 14:01:28 2002 Priyanka Gontla <pgontla@ece.uci.edu> - * orbsvcs/ImplRepo_Service/ImR_Locator_i.cpp: - Decrementing the reference count for object references - when unbinding it from the map. Thanks to Carlos for pointing - it out. Fixed bug 1403. + * orbsvcs/ImplRepo_Service/ImR_Locator_i.cpp: + Decrementing the reference count for object references + when unbinding it from the map. Thanks to Carlos for pointing + it out. Fixed bug 1403. - * orbsvcs/tests/ImplRepo/run_test.pl: - Fixed the typo in passing the name of the file to the ORBInitRef - option. Did some minor indentation modifications. + * orbsvcs/tests/ImplRepo/run_test.pl: + Fixed the typo in passing the name of the file to the ORBInitRef + option. Did some minor indentation modifications. Fri Dec 27 15:03:36 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tests/ior_corbaloc: Removed. + * tests/ior_corbaloc: Removed. Fri Dec 27 11:49:22 2002 Priyanka Gontla <pgontla@ece.uci.edu> - * tao/CORBALOC_Parser.h: - * tao/CORBALOC_Parser.cpp: - Modified the signature of parse_string_assign_helper. We dont - have to send the length of original string as an argument. - When invoking assign_ke_string, the length of the token needs - to be sent as opposed to the length of the whole address. + * tao/CORBALOC_Parser.h: + * tao/CORBALOC_Parser.cpp: + Modified the signature of parse_string_assign_helper. We dont + have to send the length of original string as an argument. + When invoking assign_ke_string, the length of the token needs + to be sent as opposed to the length of the whole address. Thanks to Bertin Colpron <BColpron@interstarinc.com> - for pointing out. + for pointing out. Fri Dec 27 14:57:50 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * examples/Makefile: - * examples/Makefile.bor: Added ior_corbaloc to the list of - directories. + * examples/Makefile: + * examples/Makefile.bor: Added ior_corbaloc to the list of + directories. Fri Dec 27 14:41:44 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * examples/ior_corbaloc: Moved the test from tests/ior_corbaloc to - examples. Moving the test to examples directory would help us to - run the test in our daily builds. + * examples/ior_corbaloc: Moved the test from tests/ior_corbaloc to + examples. Moving the test to examples directory would help us to + run the test in our daily builds. Fri Dec 27 07:56:31 2002 Jeff Parsons <j.parsons@vanderbilt.edu> * TAO_IDL/be/be_generator.cpp (create_module): - Added a line to propagate the prefix of a module to a reopening of - that module in another IDL file, overlooked in the changes of + Added a line to propagate the prefix of a module to a reopening of + that module in another IDL file, overlooked in the changes of - Wed Dec 11 18:26:09 2002 Jeff Parsons <j.parsons@vanderbilt.edu> + Wed Dec 11 18:26:09 2002 Jeff Parsons <j.parsons@vanderbilt.edu> - which was causing a reference (in a third IDL file) to a member of - the reopened module to lack the correct prefix, if it had been set. - Thanks to Nanbor Wang <nanbor@cse.wustl.edu> for reporting the - problem. + which was causing a reference (in a third IDL file) to a member of + the reopened module to lack the correct prefix, if it had been set. + Thanks to Nanbor Wang <nanbor@cse.wustl.edu> for reporting the + problem. Fri Dec 27 07:15:34 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * examples/Simple/Simple_Util.h: Somehow missed this in my checkin. + * examples/Simple/Simple_Util.h: Somehow missed this in my checkin. Thu Dec 26 09:07:21 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * examples/Simple/Simple_Util.h: Added a method to return the - client side ORB. - * examples/Simple/grid/Grid_Client_i.cpp: Hold the Grid_Factory's - object reference in a var instead of a pointer. This prevents - memory leaks in the application. + * examples/Simple/Simple_Util.h: Added a method to return the + client side ORB. + * examples/Simple/grid/Grid_Client_i.cpp: Hold the Grid_Factory's + object reference in a var instead of a pointer. This prevents + memory leaks in the application. Wed Dec 25 17:47:47 2002 Bala <bala@cs.wustl.edu> @@ -302,264 +309,264 @@ Wed Dec 25 17:47:47 2002 Bala <bala@cs.wustl.edu> Wed Dec 25 06:24:01 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> - * docs/performance.html: Added some tips about how to reduce the size of - the compile-time and run-time memory footprint. + * docs/performance.html: Added some tips about how to reduce the size of + the compile-time and run-time memory footprint. Tue Dec 24 12:36:11 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * orbsvcs/tests/Security/Crash_Test/server.cpp: Fixed a compile - error in g++ builds. + * orbsvcs/tests/Security/Crash_Test/server.cpp: Fixed a compile + error in g++ builds. Mon Dec 23 22:48:58 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * orbsvcs/tests/Notify/Basic/AdminProperties.h: - * orbsvcs/tests/Notify/Basic/AdminProperties.cpp: Fixed compile - errors in g++ builds that were caused by improper signatures. + * orbsvcs/tests/Notify/Basic/AdminProperties.h: + * orbsvcs/tests/Notify/Basic/AdminProperties.cpp: Fixed compile + errors in g++ builds that were caused by improper signatures. Mon Dec 23 22:33:56 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - A slight redesign of the changes applied to fix BUG 1276. The + A slight redesign of the changes applied to fix BUG 1276. The redesign was motivated due to a drop in our performance. - * tao/Reply_Dispatcher.cpp: - * tao/Reply_Dispatcher.h: - * tao/Reply_Dispatcher.i: Did a slight redesign of how timeouts are - handled. This was necessitated since the previous design made - things slow. Though none of the synchronization primitives were - used for normal cases ie. invocations without timeouts, we still - created them for every invocation. Creation of synchronizaton - primitives actually made things slow, and they have now been - removed. + * tao/Reply_Dispatcher.cpp: + * tao/Reply_Dispatcher.h: + * tao/Reply_Dispatcher.i: Did a slight redesign of how timeouts are + handled. This was necessitated since the previous design made + things slow. Though none of the synchronization primitives were + used for normal cases ie. invocations without timeouts, we still + created them for every invocation. Creation of synchronizaton + primitives actually made things slow, and they have now been + removed. - * tao/Muxed_TMS.cpp: Removed calls to start_dispatch () and - end_dispatch () on the Reply_Dispatcher. + * tao/Muxed_TMS.cpp: Removed calls to start_dispatch () and + end_dispatch () on the Reply_Dispatcher. - * tao/Invocation.cpp: We now do a synchronization on the internal - lock used by Muxed_TMS strategy so that no two threads are - active on the Reply_Handler at the same time. This, though not - very elegant, provides the same effect as the changes in "Sun - Dec 22 11:26:30 2002 Balachandran Natarajan - <bala@isis-server.isis.vanderbilt.edu>". The gist of the changes - mentioned above have still been retained. + * tao/Invocation.cpp: We now do a synchronization on the internal + lock used by Muxed_TMS strategy so that no two threads are + active on the Reply_Handler at the same time. This, though not + very elegant, provides the same effect as the changes in "Sun + Dec 22 11:26:30 2002 Balachandran Natarajan + <bala@isis-server.isis.vanderbilt.edu>". The gist of the changes + mentioned above have still been retained. - * tao/Exclusive_TMS.cpp: Removed calls to start_dispatch and - end_dispatch () on the Reply_Dispatcher. + * tao/Exclusive_TMS.cpp: Removed calls to start_dispatch and + end_dispatch () on the Reply_Dispatcher. - These changes should get our performance back on track. + These changes should get our performance back on track. Mon Dec 23 18:21:51 2002 Pradeep Gore <pradeep@oomworks.com> - * orbsvcs/tests/Notify/Blocking/Notify_Structured_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Discarding/Notify_Sequence_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Discarding/Notify_Structured_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Ordering/Notify_Sequence_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Ordering/Notify_Structured_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Sequence_Multi_ETCL_Filter/Notify_Sequence_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Sequence_Multi_Filter/Notify_Sequence_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Structured_Filter/Notify_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Structured_Filter/Structured_Consumer.cpp: - * orbsvcs/tests/Notify/Structured_Filter/Structured_Supplier.cpp: - * orbsvcs/tests/Notify/Structured_Multi_Filter/Notify_Push_Consumer.cpp: - * orbsvcs/tests/Notify/Structured_Multi_Filter/Structured_Consumer.cpp: - * orbsvcs/tests/Notify/Structured_Multi_Filter/Structured_Supplier.cpp: - * orbsvcs/tests/Notify/performance-tests/Filter/Sequence_Consumer.cpp: - * orbsvcs/tests/Notify/performance-tests/Filter/Structured_Consumer.cpp: - * orbsvcs/tests/Notify/performance-tests/RedGreen/RedGreen_Test.cpp: - Debug messages are printed only if the debug level is set. + * orbsvcs/tests/Notify/Blocking/Notify_Structured_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Discarding/Notify_Sequence_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Discarding/Notify_Structured_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Ordering/Notify_Sequence_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Ordering/Notify_Structured_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Sequence_Multi_ETCL_Filter/Notify_Sequence_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Sequence_Multi_Filter/Notify_Sequence_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Structured_Filter/Notify_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Structured_Filter/Structured_Consumer.cpp: + * orbsvcs/tests/Notify/Structured_Filter/Structured_Supplier.cpp: + * orbsvcs/tests/Notify/Structured_Multi_Filter/Notify_Push_Consumer.cpp: + * orbsvcs/tests/Notify/Structured_Multi_Filter/Structured_Consumer.cpp: + * orbsvcs/tests/Notify/Structured_Multi_Filter/Structured_Supplier.cpp: + * orbsvcs/tests/Notify/performance-tests/Filter/Sequence_Consumer.cpp: + * orbsvcs/tests/Notify/performance-tests/Filter/Structured_Consumer.cpp: + * orbsvcs/tests/Notify/performance-tests/RedGreen/RedGreen_Test.cpp: + Debug messages are printed only if the debug level is set. Mon Dec 23 16:25:08 2002 Boris S Kolpackov <boris@isis-server.isis.vanderbilt.edu> - * orbsvcs/tests/Security/Makefile: Added - orbsvcs/tests/Security/Crash_Test/Crash_Test directory. + * orbsvcs/tests/Security/Makefile: Added + orbsvcs/tests/Security/Crash_Test/Crash_Test directory. Mon Dec 23 15:50:54 2002 Boris S Kolpackov <boris@isis-server.isis.vanderbilt.edu> - * orbsvcs/tests/Security/Crash_Test/Crash_Test.dsw: - * orbsvcs/tests/Security/Crash_Test/Crash_Test_Client.dsp: - * orbsvcs/tests/Security/Crash_Test/Crash_Test_Server.dsp: - * orbsvcs/tests/Security/Crash_Test/Makefile: - * orbsvcs/tests/Security/Crash_Test/README: - * orbsvcs/tests/Security/Crash_Test/cacert.pem: - * orbsvcs/tests/Security/Crash_Test/client.conf: - * orbsvcs/tests/Security/Crash_Test/client.cpp: - * orbsvcs/tests/Security/Crash_Test/client_cert.pem: - * orbsvcs/tests/Security/Crash_Test/client_key.pem: - * orbsvcs/tests/Security/Crash_Test/run_test.pl: - * orbsvcs/tests/Security/Crash_Test/server.conf: - * orbsvcs/tests/Security/Crash_Test/server.cpp: - * orbsvcs/tests/Security/Crash_Test/server.ior: - * orbsvcs/tests/Security/Crash_Test/server_cert.pem: - * orbsvcs/tests/Security/Crash_Test/server_key.pem: - * orbsvcs/tests/Security/Crash_Test/test.idl: - * orbsvcs/tests/Security/Crash_Test/test_i.cpp: - * orbsvcs/tests/Security/Crash_Test/test_i.h: - * orbsvcs/tests/Security/Crash_Test/test_i.i: - Added a new test to verifies that the ORB's connection - management mechanism is functioning properly in conjunction - with SSLIOP. For more information please refer to README. + * orbsvcs/tests/Security/Crash_Test/Crash_Test.dsw: + * orbsvcs/tests/Security/Crash_Test/Crash_Test_Client.dsp: + * orbsvcs/tests/Security/Crash_Test/Crash_Test_Server.dsp: + * orbsvcs/tests/Security/Crash_Test/Makefile: + * orbsvcs/tests/Security/Crash_Test/README: + * orbsvcs/tests/Security/Crash_Test/cacert.pem: + * orbsvcs/tests/Security/Crash_Test/client.conf: + * orbsvcs/tests/Security/Crash_Test/client.cpp: + * orbsvcs/tests/Security/Crash_Test/client_cert.pem: + * orbsvcs/tests/Security/Crash_Test/client_key.pem: + * orbsvcs/tests/Security/Crash_Test/run_test.pl: + * orbsvcs/tests/Security/Crash_Test/server.conf: + * orbsvcs/tests/Security/Crash_Test/server.cpp: + * orbsvcs/tests/Security/Crash_Test/server.ior: + * orbsvcs/tests/Security/Crash_Test/server_cert.pem: + * orbsvcs/tests/Security/Crash_Test/server_key.pem: + * orbsvcs/tests/Security/Crash_Test/test.idl: + * orbsvcs/tests/Security/Crash_Test/test_i.cpp: + * orbsvcs/tests/Security/Crash_Test/test_i.h: + * orbsvcs/tests/Security/Crash_Test/test_i.i: + Added a new test to verifies that the ORB's connection + management mechanism is functioning properly in conjunction + with SSLIOP. For more information please refer to README. Mon Dec 23 15:08:22 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * orbsvcs/tests/Notify/Basic/AdminProperties.cpp: - * orbsvcs/tests/Notify/Basic/AdminProperties.h: Fixed a warning - with SunCC builds. + * orbsvcs/tests/Notify/Basic/AdminProperties.cpp: + * orbsvcs/tests/Notify/Basic/AdminProperties.h: Fixed a warning + with SunCC builds. Mon Dec 23 15:16:21 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tao/RTPortableServer/RT_Object_Adapter_Factory.cpp: Moved the - setting of the <collocation_resolver_name_> from the constructor - to create (), since it is not required immediately on loading - the RT_PortableServer. This should fix a problem with GCC 2.95 - builds on Solaris where the static initializations are all - messed up! Surprising that other tests even work. + * tao/RTPortableServer/RT_Object_Adapter_Factory.cpp: Moved the + setting of the <collocation_resolver_name_> from the constructor + to create (), since it is not required immediately on loading + the RT_PortableServer. This should fix a problem with GCC 2.95 + builds on Solaris where the static initializations are all + messed up! Surprising that other tests even work. Mon Dec 23 10:44:19 2002 Jeff Parsons <j.parsons@vanderbilt.edu> * TAO_IDL/be/be_visitor_valuetype/valuetype_ch.cpp: - * TAO_IDL/be/be_visitor_ccm_pre_proc.cpp: + * TAO_IDL/be/be_visitor_ccm_pre_proc.cpp: - Fixed code generation errors for components and eventtypes. Thanks - to Nanbor Wang <nanbor@cse.wustl.edu> for his example IDL files - that uncovered the problem. + Fixed code generation errors for components and eventtypes. Thanks + to Nanbor Wang <nanbor@cse.wustl.edu> for his example IDL files + that uncovered the problem. Mon Dec 23 06:57:41 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * orbsvcs/orbsvcs/Notify/ThreadPool_Task.h: - * orbsvcs/orbsvcs/Notify/ThreadPool_Task.cpp: Fixed warnings in - SunCC builds. Yet another problem with SunCC compiler. This - should also hopefully fix a similar problem with Borland 6. + * orbsvcs/orbsvcs/Notify/ThreadPool_Task.h: + * orbsvcs/orbsvcs/Notify/ThreadPool_Task.cpp: Fixed warnings in + SunCC builds. Yet another problem with SunCC compiler. This + should also hopefully fix a similar problem with Borland 6. Sun Dec 22 22:46:50 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tao/Connection_Handler.cpp: Replaced the word "error" from one - of the debug statements. The test - tests/RTCORBA/MT_Client_Protocol_Priority uses a script that - parses the output from the test run to determine if the test was - right. The script searches for the word "error" in the output - and on seeing it barf's. Hence the change. + * tao/Connection_Handler.cpp: Replaced the word "error" from one + of the debug statements. The test + tests/RTCORBA/MT_Client_Protocol_Priority uses a script that + parses the output from the test run to determine if the test was + right. The script searches for the word "error" in the output + and on seeing it barf's. Hence the change. Sun Dec 22 23:18:58 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tao/Reply_Dispatcher.h: - * tao/Reply_Dispatcher.i: Added an accessor for <timeout_> + * tao/Reply_Dispatcher.h: + * tao/Reply_Dispatcher.i: Added an accessor for <timeout_> - * tao/Muxed_TMS.cpp: Call start_dispatch () and end_dispatch () - only if the Reply_Dispatcher has timout associated with it. This - should fix most of the problems that showed up in the latest - builds. + * tao/Muxed_TMS.cpp: Call start_dispatch () and end_dispatch () + only if the Reply_Dispatcher has timout associated with it. This + should fix most of the problems that showed up in the latest + builds. Sun Dec 22 20:22:50 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * performance-tests/Latency/Thread_Per_Connection/svc.conf: Used - the blocked connect strategy instead of the default - strategy. This is needed to work around the problem outlined in - BUG 1408. + * performance-tests/Latency/Thread_Per_Connection/svc.conf: Used + the blocked connect strategy instead of the default + strategy. This is needed to work around the problem outlined in + BUG 1408. Sun Dec 22 11:26:30 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - Fix for Bug 1276. Please see + Fix for Bug 1276. Please see - http://deuce.doc.wustl.edu/bugzilla/show_bug.cgi?id=1276 for more + http://deuce.doc.wustl.edu/bugzilla/show_bug.cgi?id=1276 for more details on this bug. The gist of the fix is this - - If a follower thread timesout, it first tries to unbind the - dispatcher from the table + - If a follower thread timesout, it first tries to unbind the + dispatcher from the table - - If unbinding is succesfull it throws a TIMEOUT exception to - the client. + - If unbinding is succesfull it throws a TIMEOUT exception to + the client. - - If unbinding is unsuccesful, it waits on a condition variable - in the reply dispatcher + - If unbinding is unsuccesful, it waits on a condition variable + in the reply dispatcher - - Once the reply is available it dispatches the reply. + - Once the reply is available it dispatches the reply. - - This helps avoid the race conditions that are documented in - 1276. + - This helps avoid the race conditions that are documented in + 1276. - None of the above code should be executed when timeouts are not + None of the above code should be executed when timeouts are not enabled in the ORB ie. for the default path. - * tao/Bind_Dispatcher_Guard.cpp: - * tao/Bind_Dispatcher_Guard.h: - * tao/Bind_Dispatcher_Guard.i: Added method unbind_dispatcher () - which unbinds the dispatcher from the Transport_Mux_Strategy and - status (), which allows to set the status of the dispatcher - guard. - - * tao/Transport_Mux_Strategy.h - * tao/Exclusive_TMS.cpp: - * tao/Exclusive_TMS.h: - * tao/Muxed_TMS.cpp: - * tao/Muxed_TMS.h: Changed the signature of the method - unbind_dispatcher (). It now returns an int. Further, we mark - the Reply_Dispatcher as busy when the dispatching thread is - ready to call Reply_Dispatcher::dispatch_reply (). - - * tao/Reply_Dispatcher.cpp: - * tao/Reply_Dispatcher.h: - * tao/Reply_Dispatcher.i: Added a mutex and a condition variable - for the follower to wait on when a leader thread is within the - Reply_Dispatcher. Methods to wait on the CV and to mark the - reply_dispatcher as being used and free have been added. These - methods are essentially no-ops if timeouts are not set. - - * tao/Invocation.cpp: - * tao/Invocation.h: If the invocation thread gets a timeout, it - first unbinds its reply dispatcher from the Mux_Strategy. If the - unbind fails (this means that another thread has aleardy - collected the reply from the transport) , it waits on the - reply_dispatcher to collect the reply instead of just throwing a - TIMEOUT exception. + * tao/Bind_Dispatcher_Guard.cpp: + * tao/Bind_Dispatcher_Guard.h: + * tao/Bind_Dispatcher_Guard.i: Added method unbind_dispatcher () + which unbinds the dispatcher from the Transport_Mux_Strategy and + status (), which allows to set the status of the dispatcher + guard. + + * tao/Transport_Mux_Strategy.h + * tao/Exclusive_TMS.cpp: + * tao/Exclusive_TMS.h: + * tao/Muxed_TMS.cpp: + * tao/Muxed_TMS.h: Changed the signature of the method + unbind_dispatcher (). It now returns an int. Further, we mark + the Reply_Dispatcher as busy when the dispatching thread is + ready to call Reply_Dispatcher::dispatch_reply (). + + * tao/Reply_Dispatcher.cpp: + * tao/Reply_Dispatcher.h: + * tao/Reply_Dispatcher.i: Added a mutex and a condition variable + for the follower to wait on when a leader thread is within the + Reply_Dispatcher. Methods to wait on the CV and to mark the + reply_dispatcher as being used and free have been added. These + methods are essentially no-ops if timeouts are not set. + + * tao/Invocation.cpp: + * tao/Invocation.h: If the invocation thread gets a timeout, it + first unbinds its reply dispatcher from the Mux_Strategy. If the + unbind fails (this means that another thread has aleardy + collected the reply from the transport) , it waits on the + reply_dispatcher to collect the reply instead of just throwing a + TIMEOUT exception. Fri Dec 20 13:47:33 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * orbsvcs/tests/Security/Big_Request/Makefile: - * orbsvcs/tests/Security/Callback/Makefile: - * orbsvcs/tests/Security/MT_SSLIOP/Makefile: - * orbsvcs/tests/Security/Secure_Invocation/Makefile: - * orbsvcs/tests/Security/SecurityLevel1/Makefile: Too much of cut - and paste. Fixed places where BIN was assigned something - arbitrary. + * orbsvcs/tests/Security/Big_Request/Makefile: + * orbsvcs/tests/Security/Callback/Makefile: + * orbsvcs/tests/Security/MT_SSLIOP/Makefile: + * orbsvcs/tests/Security/Secure_Invocation/Makefile: + * orbsvcs/tests/Security/SecurityLevel1/Makefile: Too much of cut + and paste. Fixed places where BIN was assigned something + arbitrary. Thu Dec 19 13:45:50 2002 Jeff Parsons <j.parsons@vanderbilt.edu> * TAO_IDL/util/utl_global.cpp: Added a patch sent in by Greg Mulyar <greg_mu@yahoo.com> to make - sure that the ACE version of gperf is used when both the ACE and - GNU versions of gperf are present. This fixes BUG 1406. + sure that the ACE version of gperf is used when both the ACE and + GNU versions of gperf are present. This fixes BUG 1406. Thu Dec 19 08:48:04 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * orbsvcs/tests/Security/Big_Request/Makefile: - * orbsvcs/tests/Security/Callback/Makefile: - * orbsvcs/tests/Security/MT_SSLIOP/Makefile: - * orbsvcs/tests/Security/Secure_Invocation/Makefile: - * orbsvcs/tests/Security/SecurityLevel1/Makefile: Added our usual - checks to make sure that these dont get built if Security - service or SSLIOP libraries arent built. This shold fix compile - errors in NO_AMI_MESSAGING and NO_INTERCEPTORS builds. + * orbsvcs/tests/Security/Big_Request/Makefile: + * orbsvcs/tests/Security/Callback/Makefile: + * orbsvcs/tests/Security/MT_SSLIOP/Makefile: + * orbsvcs/tests/Security/Secure_Invocation/Makefile: + * orbsvcs/tests/Security/SecurityLevel1/Makefile: Added our usual + checks to make sure that these dont get built if Security + service or SSLIOP libraries arent built. This shold fix compile + errors in NO_AMI_MESSAGING and NO_INTERCEPTORS builds. Thu Dec 19 08:25:04 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * tao/Transport_Cache_Manager.cpp: Call - TAO_Transport::close_connection_no_purge () in - TAO_Transport_Cache_Manager::purge () instead of - close_connection_i (). Surprising that Chris Cleeland missed - this after we agreed that this was the right things to do. Looks - like Chris C has been caught up in a different world altogether - :-). This should fix the errors in the daily builds. + * tao/Transport_Cache_Manager.cpp: Call + TAO_Transport::close_connection_no_purge () in + TAO_Transport_Cache_Manager::purge () instead of + close_connection_i (). Surprising that Chris Cleeland missed + this after we agreed that this was the right things to do. Looks + like Chris C has been caught up in a different world altogether + :-). This should fix the errors in the daily builds. Wed Dec 18 19:45:45 2002 Jeff Parsons <j.parsons@vanderbilt.edu> * TAO_IDL/fe/fe_declarator.cpp: - Added a check for 0 return from a narrow to AST_Type. This - check catches attempts to use constants, enum values, - attributes, and operations as IDL sequence elements, and - outputs an error instead of crashing. Thanks to Petr - Ferschmann <pferschmann@kerio.com> for pointing out the - problem with enum values. + Added a check for 0 return from a narrow to AST_Type. This + check catches attempts to use constants, enum values, + attributes, and operations as IDL sequence elements, and + outputs an error instead of crashing. Thanks to Petr + Ferschmann <pferschmann@kerio.com> for pointing out the + problem with enum values. Wed Dec 18 14:48:32 2002 Jeff Parsons <j.parsons@vanderbilt.edu> @@ -567,55 +574,55 @@ Wed Dec 18 14:48:32 2002 Jeff Parsons <j.parsons@vanderbilt.edu> * tao/Any.i: Added cast operator to Any& for class Any_var. Thanks to Torsten - Kuepper <torsten.kuepper@nokia.com> for pointing out the lack of - this spec-required operator. This fixes BUG 1404. + Kuepper <torsten.kuepper@nokia.com> for pointing out the lack of + this spec-required operator. This fixes BUG 1404. Wed Dec 18 14:46:53 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * orbsvcs/orbsvcs/Makefile (TAO_ORBSVCS): Build SSL only when - corba_messaging and interceptors are infact turned on. + * orbsvcs/orbsvcs/Makefile (TAO_ORBSVCS): Build SSL only when + corba_messaging and interceptors are infact turned on. Wed Dec 18 08:22:39 2002 Ossama Othman <ossama@uci.edu> - * orbsvcs/orbsvcs/LoadBalancing/LB_Random.cpp - (tao_lb_random_init_routine): + * orbsvcs/orbsvcs/LoadBalancing/LB_Random.cpp + (tao_lb_random_init_routine): - Fixed compile-time problem on platforms that do not support for - the Pthread standard. This function should always be visible. + Fixed compile-time problem on platforms that do not support for + the Pthread standard. This function should always be visible. Tue Dec 17 09:53:02 2002 Chris Cleeland <cleeland_c@ociweb.com> - * tao/Transport_Cache_Manager.*: - - Fixed deadlock that can occur when trying to gain reactor token - and transport cache manager lock while purging cache entries. - Unfortunately , the classic solution to the classic - deadlock-- which is typically caused by an ordering problem, and is - solved by re-ordering--doesn't apply. We simply don't have - control of the order due to the fact that we have two frameworks - interacting, and the offending bodies of code don't have - visibility to both locks. - - The key is not trying to remove the handler while the transport - cache lock is held. So, we break this into two loops. The first - loop does everything cache-releated with only the cache lock held, - then releases it. The second loop then does only the - reactor-related things. - - Key changes are: - - - Move purge() from .inl to .cpp. - - Fold close_entries() into purge() (purge was the only thing that - called close_entries anyway). - - Use a stack to accumulate the transports that have been purged - in the first loop, then, in the second loop, pop the stack and - call the remove_handler on each transport. - - Note that the transports whose pointers are stuck in the stack - are properly reference counted using - TAO_Transport::_duplicate/release. - - Thanks to Raytheon in San Diego for providing the environment to - find this! + * tao/Transport_Cache_Manager.*: + + Fixed deadlock that can occur when trying to gain reactor token + and transport cache manager lock while purging cache entries. + Unfortunately , the classic solution to the classic + deadlock-- which is typically caused by an ordering problem, and is + solved by re-ordering--doesn't apply. We simply don't have + control of the order due to the fact that we have two frameworks + interacting, and the offending bodies of code don't have + visibility to both locks. + + The key is not trying to remove the handler while the transport + cache lock is held. So, we break this into two loops. The first + loop does everything cache-releated with only the cache lock held, + then releases it. The second loop then does only the + reactor-related things. + + Key changes are: + + - Move purge() from .inl to .cpp. + - Fold close_entries() into purge() (purge was the only thing that + called close_entries anyway). + - Use a stack to accumulate the transports that have been purged + in the first loop, then, in the second loop, pop the stack and + call the remove_handler on each transport. + - Note that the transports whose pointers are stuck in the stack + are properly reference counted using + TAO_Transport::_duplicate/release. + + Thanks to Raytheon in San Diego for providing the environment to + find this! Tue Dec 17 08:14:35 2002 Chad Elliott <elliott_c@ociweb.com> @@ -979,9 +986,9 @@ Fri Dec 13 00:10:23 2002 Pradeep Gore <pradeep@oomworks.com> * orbsvcs/tests/Notify/Basic/Sequence.h: * orbsvcs/tests/Notify/Basic/Sequence.cpp: Added a test for sequences. In the default run, this test sends - 15 events in batches of 5 events via a sequence supplier. It - sets the consumer's batch size to 3 and sets a pacing interval - of 2 seconds. it checks to see if 15 events are indeed received. + 15 events in batches of 5 events via a sequence supplier. It + sets the consumer's batch size to 3 and sets a pacing interval + of 2 seconds. it checks to see if 15 events are indeed received. * orbsvcs/tests/Notify/Basic/adminproperties.conf: Svc Conf file for the AdminProperties test - This prevents the diff --git a/TAO/tao/DynamicAny/DynamicAnyC.h b/TAO/tao/DynamicAny/DynamicAnyC.h index d5cfe902931..2be2bf31127 100644 --- a/TAO/tao/DynamicAny/DynamicAnyC.h +++ b/TAO/tao/DynamicAny/DynamicAnyC.h @@ -23,7 +23,7 @@ #define _TAO_IDL_ORIG_DYNAMICANYC_H_ #if !defined(TAO_DYNAMICANY_SAFE_INCLUDE) && \ - !defined(TAO_INSTANTIATING_TEMPLATES) + !defined(ACE_VACPP_INSTANTIATING_TEMPLATES) #error "You should not include DynamicAnyC.h directly, use DynamicAny.h" #endif /* !TAO_DYNAMICANY_SAFE_INCLUDE */ diff --git a/TAO/tao/RTCORBA/RTCORBAC.h b/TAO/tao/RTCORBA/RTCORBAC.h index c5316b8b29f..bca83e97ec9 100644 --- a/TAO/tao/RTCORBA/RTCORBAC.h +++ b/TAO/tao/RTCORBA/RTCORBAC.h @@ -45,7 +45,7 @@ class TAO_Network_Priority_Mapping_Manager_out; class TAO_Network_Priority_Mapping; #if !defined(TAO_RTCORBA_SAFE_INCLUDE) && \ - !defined(TAO_INSTANTIATING_TEMPLATES) + !defined(ACE_VACPP_INSTANTIATING_TEMPLATES) #error "You should not include RTCORBAC.h directly, use RTCORBA.h" #endif /* !TAO_RTCORBA_SAFE_INCLUDE */ |