diff options
Diffstat (limited to 'ACE/ChangeLogs/ChangeLog-02a')
-rw-r--r-- | ACE/ChangeLogs/ChangeLog-02a | 9018 |
1 files changed, 9018 insertions, 0 deletions
diff --git a/ACE/ChangeLogs/ChangeLog-02a b/ACE/ChangeLogs/ChangeLog-02a new file mode 100644 index 00000000000..870ec350d6e --- /dev/null +++ b/ACE/ChangeLogs/ChangeLog-02a @@ -0,0 +1,9018 @@ +Fri Jul 5 10:30:44 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Array_Base.cpp: Swapped the order of includes so that + Array_Base.h comes after Malloc_Base.h. Thanks to Al Pariante + <Albert.Pariante@axcelis.com> for reporting this. + +Thu Jul 4 11:27:05 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * Happy 226th Birthday USA! + +Sat Jul 06 19:32:40 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Time_Value.h (ACE_Time_Value): Fix for [BUGID 1240]. Some of + the operators were missing explicit declarations in the header + file (they were anyway defined in the inlined files) which + created problems in builds where inlining was diabled. Please + see + + http://deuce.doc.wustl.edu/bugzilla/show_bug.cgi?id=1240 for + more details. Thanks to Marina Spivak <marina@atdesk.com> for + reporting this. + +Fri Jul 5 18:44:39 2002 Steve Huston <shuston@riverace.com> + + * ace/ACE.cpp: Corrected undeclared 'select_width' errors and + signed/unsigned mismatch warnings. + +Fri Jul 5 16:14:51 2002 Steve Huston <shuston@riverace.com> + + * ace/OS.{h i} (send, sendto, recv, recvfrom): Change 'len' argument + from int to size_t. + Removed conditional typedef of ssize_t; it's done in Basic_Types.h. + + * ace/ACE.h (recv): Filled in documentation for varargs variant. + * ace/ACE.cpp (send, recv): Properly convert size_t 'count' arg to int + iovec count for call to ACE_OS::sendv/recvv(). + (handle_timed_accept, handle_timed_complete, handle_ready): Use + an int select_width for ACE_OS::select() calls to get the types + correct. This arg is ignored on Windows anyway. + + * ace/Basic_Types.h: If ACE_HAS_SSIZE_T not defined, and Win64, + typedef SSIZE_T ssize_t; else use the existing int typedef. + + * tests/SSL/Thread_Pool_Reactor_SSL_Test.cpp: + Change thread function to return ACE_THR_FUNC_RETURN, not void *. + +Tue Jul 02 21:43:53 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/common/XMLFilterImpl.h: One more pre/post mismatch. + +Tue Jul 2 16:58:55 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/XML_Types.h: + + Removed unnecessary includes and fixed a typo. + +Tue Jul 02 16:35:19 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/common/Attributes.h: + * ACEXML/common/Locator.h: Fixed mismatched pre/post inclusions. + + * ACEXML/parser/parser/Parser.cpp: Fixed compilation errors. + +Tue Jul 2 11:53:29 2002 Steve Huston <shuston@riverace.com> + + * ace/config-all.h: Changed definition of ACE_THR_FUNC to incorporate + the platform's thread function return type (DWORD on Windows, + void * on most others). Uses a new typedef ACE_THR_FUNC_RETURN + to specify the platform's thread function return type. + + * ace/OS.{h i cpp} (thr_exit, thr_join): The status arg is now + a ACE_THR_FUNC_RETURN *. + Also moved remaining thread function definition type things to + config-all.h with the rest of them. + + * ace/Base_Thread_Adapter.{h cpp} (invoke(), ace_thread_adapter()): + * ace/OS_Thread_Adapter.{h cpp} + * ace/Thread_Adapter.{h cpp} (invoke(), invoke_i()): + Return type changed from void * to ACE_THR_FUNC_RETURN. + + * ace/Event_Handler.{cpp h} (read_adapter): Change return value from + void* to ACE_THR_FUNC_RETURN. + + * ace/Thread_Hook.{h cpp}: ACE_Thread_Hook::start () returns + ACE_THR_FUNC_RETURN, not void *. + + * ace/Thread.{h i} (exit, join): 'status' arg is now + ACE_THR_FUNC_RETURN *, not void *. + + * ace/Thread_Control.{h inl cpp} (status, exit): + * ace/Thread_Manager.{h cpp} (exit, join): Thread exit status + type changed from void * to ACE_THR_FUNC_RETURN. + + * ace/Task.{h cpp} (ACE_Task_Base): Return type from svc_run now + ACE_THR_FUNC_RETURN. + + * examples/C++NPv1/RT_Thread_Per_Connection_Logging_Server.cpp: + * examples/C++NPv1/Thread_Per_Connection_Logging_Server.{cpp, h}: + * examples/C++NPv2/Select_Reactor_Logging_Server.cpp: + * examples/C++NPv2/Server_Shutdown.cpp: + * examples/C++NPv2/TP_Reactor_Logging_Server.cpp: + * examples/C++NPv2/WFMO_Reactor_Logging_Server.cpp: + * examples/C++NPv2/Client_Logging_Daemon.cpp: + * examples/Connection/blocking/SPIPE-acceptor.cpp: + * examples/IPC_SAP/SOCK_SAP/CPP-inserver.cpp: + * examples/IPC_SAP/SSL_SAP/SSL-server.cpp: + * examples/Reactor/WFMO_Reactor/Abandoned.cpp: + * performance-tests/TCP/tcp_test.cpp: + * tests/ACE_Init_Test.cpp: + * tests/MEM_Stream_Test.cpp: + * tests/Service_Config_DLL.cpp: + * tests/Task_Test.cpp: + * tests/Thread_Pool_Reactor_Resume_Test.cpp: + * tests/Thread_Pool_Reactor_Test.cpp: + Change thread function to return ACE_THR_FUNC_RETURN, not void *. + + * examples/C++NPv2/TP_Logging_Server.cpp (handle_input, svc): + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp (open): + * examples/C++NPv2/display_logfile.cpp (svc): Change + ACE_reinterpret_cast to ACE_static_cast. Thanks to Doug Schmidt + for finding this. + + * examples/C++NPv2/TP_Logging_Server.h: + * examples/C++NPv2/Reactor_Logging_Server_Adapter.cpp: Corrected use + of ACE_NEW_RETURN. + +Mon Jul 1 23:16:47 2002 Jeff Parsons <parsons@cs.wustl.edu> + + * */Makefile: + + Updated dependencies. + +Mon Jul 1 21:59:44 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/Attributes.h: + * ACEXML/common/AttributesImpl.h: + * ACEXML/common/Attributes_Def_Builder.h: + * ACEXML/common/CharStream.h: + * ACEXML/common/ContentHandler.h: + * ACEXML/common/DTDHandler.h: + * ACEXML/common/DTD_Manager.h: + * ACEXML/common/DefaultHandler.h: + * ACEXML/common/Element_Def_Builder.h: + * ACEXML/common/EntityResolver.h: + * ACEXML/common/Env.h: + * ACEXML/common/ErrorHandler.h: + * ACEXML/common/Exception.h: + * ACEXML/common/FileCharStream.h: + * ACEXML/common/HttpCharStream.h: + * ACEXML/common/InputSource.h: + * ACEXML/common/Locator.h: + * ACEXML/common/LocatorImpl.h: + * ACEXML/common/Mem_Map_Stream.h: + * ACEXML/common/NamespaceSupport.h: + * ACEXML/common/SAXExceptions.h: + * ACEXML/common/StrCharStream.h: + * ACEXML/common/Transcode.h: + * ACEXML/common/URL_Addr.h: + * ACEXML/common/Validator.h: + * ACEXML/common/XMLFilter.h: + * ACEXML/common/XMLFilterImpl.h: + * ACEXML/common/XMLReader.h: + * ACEXML/common/XML_Codecs.h: + * ACEXML/common/XML_Types.h: + * ACEXML/parser/debug_validator/Debug_Attributes_Builder.cpp: + * ACEXML/parser/debug_validator/Debug_Attributes_Builder.h: + * ACEXML/parser/debug_validator/Debug_DTD_Manager.h: + * ACEXML/parser/debug_validator/Debug_Element_Builder.h: + * ACEXML/parser/debug_validator/Element_Tree.h: + * ACEXML/parser/parser/Entity_Manager.h: + * ACEXML/parser/parser/Parser.cpp: + * ACEXML/parser/parser/Parser.h: + + ACE_LACKS_PRAGMA_ONCE fixes. + +Mon Jul 1 06:21:00 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Sock_Connect.cpp (count_interfaces): Added a MacOSX-specific + value for OSIOCGIFCONF. Thanks to John Zorko <j.zorko@att.net> + for this fix. + +Sun Jun 30 22:22:05 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * ace/MEM_Addr.cpp (same_host): This method was unnecessarily + creating and copying both the local and remote addresses. I + changed it so that we only compare the IP addresses. + + * ace/MEM_Connector.cpp (connect): Changed debugging statement to + provide additional information. + +Sun Jun 30 15:32:47 2002 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> + + * ace/Based_Pointer_Repository.h: + * ace/Array_Base.h: Added #include "ace/OS.h" to get the + definition of size_t on MacOS. Thanks to John Zorko + <j.zorko@att.net> for this fix. + + * ace/Synch.cpp: Fixed the ACE_recursive_mutex_state::reset() method + so that it sets m.LockCount to 0. Thanks to Davide Pasetto + <dipo@sis.it> for reporting this. + + * ace/Synch.cpp: Fixed the implementation of + ACE_Condition<ACE_Recursive_Thread_Mutex> so that it works + properly on Windows, where the recursive mutex count is + incremented by the OS. Thanks to Davide Pasetto <dipo@sis.it> + for reporting this. + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: + * examples/C++NPv2/Client_Logging_Daemon.cpp: Changed "blocks" + parameter to "chunk" to be consistent with the terminology in + the book. + + * examples/C++NPv2/TP_Logging_Server.h: + * examples/C++NPv2/Reactor_Logging_Server_Adapter.cpp: Added some + new code to safely allocate an array. + +Sun Jun 30 08:55:18 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * tests/Enum_Interfaces_Test.cpp (ACE_TMAIN): Fixed this to use + ACE_Sock_Connect::get_ip_interfaces() rather than ACE::get_ip_interfaces(). + Thanks to <rjgebis@hotmailer.com> for reporting this. + +Sun Jun 30 12:22:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * etc/tao_portablegroup.doxygen: + Added this file that configures the doxygen configuration of the TAO + PortableGroup library. + + * html/index.html: + * bin/generate_doxygen.pl: + Updated these files so that the new PortableGroup doxygen + documentation is generated and listed in the index page. + +Fri Jun 28 08:07:12 2002 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> + + * ace/Logging_Strategy.cpp: Added the ability to toggle + the LM_STARTUP and LM_SHUTDOWN messages. Thanks to + Adee Ran <adeeran@yahoo.com> for the patch. + +Fri Jun 28 07:37:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ACEXML/common/URL_Addr.cpp: + Fixed BCB Unicode build error. + +Thu Jun 27 16:50:19 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/TP_Reactor.cpp: + * ace/TP_Reactor.h: + * ace/TP_Reactor.i: Moved common code in get_socket_event_info () + to a new method clear_handle_read_set () and simplified the code + in get_notify_handle () so that it improves readability. + +Thu Jun 27 20:26:33 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * tests/DLL_Test.cpp: Define retval before preprocessor checks to + fix Lynx build problem. + +Thu Jun 27 14:32:50 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * bin/nightlybuilds/builds.lst: Moved Fuzz, DLL_Core, DLL_Debug + builds to the new scoreboard also. + +Thu Jun 27 07:47:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ACEXML/tests/HttpCharStream_Test.cpp: + Fixed BCB Unicode build errors. + + * ACEXML/common/URL_Addr.h: + Documentation update. + +Wed Jun 26 09:56:38 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * bin/msvc_auto_compile.pl: Changed the list of XML related + directory name from XML to ACEXML. + +Wed Jun 26 13:55:01 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * examples/Reactor/Multicast/Log_Wrappers.cpp: + * examples/Reactor/Ntalker/Log_Wrappers.cpp: + Remove references to SOCK_Dgram_Mcast_T. + +Tue Jun 25 15:13:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Basic_Types.h: Added ACE_WIN64 versions of + ACE_SSIZE_T_FORMAT_SPECIFIER and ACE_SIZE_T_FORMAT_SPECIFIER. + +Tue Jun 25 23:22:09 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/SOCK_Dgram_Mcast.cpp: + * ace/SOCK_Dgram_Mcast.h: + * ace/SOCK_Dgram_Mcast.i: + * ace/SOCK_Dgram_Mcast_T.cpp: Removed. + * ace/SOCK_Dgram_Mcast_T.h: Removed. + * ace/SOCK_Dgram_Mcast_T.i: Removed. + * ace/QoS/SOCK_Dgram_Mcast_QoS.cpp: + * ace/QoS/SOCK_Dgram_Mcast_QoS.h: + Reverted these files to before June 9. Existing code in QoS and + AVStreams were broken do to the changes. + +Tue Jun 25 14:45:31 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: Comment out + declarations for info(), suspend(), and resume() - they're left + as exercises for the reader. + + * examples/C++NPv2/AC_CLD.dsp: Added SSL link libraries. + +Tue Jun 25 12:13:39 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/AC_CLD.dsp: + * examples/C++NPv2/examples.dsw: Added project for the + Acceptor-Connector Client Logging Daemon in chapter 7. + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: Use the ACE + traits macros instead of relying on traits class support to + help the aged MSVC 6 compiler along. + +Mon Jun 24 20:55:45 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * ace/ACE.cpp (select): Fixed typo. + +Mon Jun 24 17:39:31 2002 Steve Huston <shuston@riverace.com> + + * ace/ace_dll64.mak: + * ace/ace_lib64.mak: NMAKE files for building ACE on 64-bit Windows + XP with the Platform SDK compiler. At some point, hopefully, + MSVC will allow 64-bit builds via the regular project files, + but for now, must use NMAKE. + + * ace/config-win32-common.h: #define ACE_WIN64 if doing a 64-bit build. + + * ace/ACE.cpp (select): Don't sync() the handles after select() on + Win64. The cast is illegitmate, and the call is a no-op on Windows. + +Mon Jun 24 17:55:58 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/wrapper_macros.GNU (ACE_BETA_VERSION): + + Grab the version information from Version.h rather than VERSION + as VERSION is not likely to installed when doing a make install. + Thanks to Bala for the suggestion. + + * ACEXML/parser/parser/Parser.cpp: + * ACEXML/parser/parser/Parser.h: + + Get the prefix right in case of a default namespace being + present. Previously we were using a URI a get to the prefix, + which doesn't work when the URI is the same as URI for another + namespace prefix. + +Mon Jun 24 19:54:34 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/config-g++-common.h: Define + ACE_HAS_STD_TEMPLATE_METHOD_SPECIALIZATION for gcc 2.95 and higher. + +Mon Jun 24 07:26:41 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Timer_Queue_Adapters.cpp: + * ace/Timer_Queue_Adapters.h: Changed uses of ACE_Recursive_Thread_Mutex + to ACE_RECURSIVE_SYNCH_MUTEX so things will work properly on + single-threaded configurations. Thanks to Craig Rodrigues for + reporting this. + +Mon Jun 24 12:07:00 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/RMCast/RMCast_IO_UDP.cpp: Remove ACE_SOCK_Dgram_Ex template instantiation. + +Mon Jun 24 03:26:10 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/AttributesImpl.cpp(isDuplicate): + + New function isDuplicate() to check duplicate attributes. Merged + support for duplicate attribute checking in setAttribute(). + + * ACEXML/common/AttributesImpl.h: + + Fixed documentation for a bunch of functions which had set/get + roles reversed. + + * ACEXML/common/AttributesImpl.i: + + Fixed more memory leaks. + + * ACEXML/common/NamespaceSupport.cpp(processName): + + Added support for processing attributes. + + * ACEXML/parser/parser/Parser.cpp: + * ACEXML/parser/parser/Parser.h: + + Implement XML Namespace (http://www.w3.org/TR/REC-xml-names) + support. ACEXML parser supports the SAX 2.0 XML API. Parser has + two user tunable features available: namespaces (defaults to + true) and namespace_prefixes (defaults to false). + + * ACEXML/examples/SAXPrint/Print_Handler.cpp: + * ACEXML/examples/SAXPrint/SAXPrint_Handler.cpp: + + Changed to reflect more event callbacks available now due to + above added support. + + * ACEXML/examples/SAXPrint/ns.svc.conf.xml: + + Changed to include a default namespace to test more code paths. + Found a bug in the process :-) Will fix tomorrow. + + The only remaining feature in ACEXML parser needed for CCM is + validator. Next target.... + + * ace/Message_Block.h: + + Fixed typo in documentation (ceases to exist and not seizes to + exist). + +Sun Jun 23 21:06:28 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + Bill Fulton <Bill_A_Fulton@raytheon.com> + + * ace/SOCK_Dgram_Mcast.cpp: + * ace/SOCK_Dgram_Mcast.h: + * ace/SOCK_Dgram_Mcast_T.cpp: + * ace/SOCK_Dgram_Mcast_T.h: + * ace/SOCK_Dgram_Mcast_T.i: + More SOCK_Dgram fixes supplied by Bill Fulton with modifications + by Craig Rodrigues. + - Remove proprietary typedefs for ACE mutex classes + - Make default mutex strategy for SOCK_Dgram_Mcast ACE_SYNCH_MUTEX + - (subscribe_ifs) fix check when determining if subscribed address + is the same as the bound address + +Sun Jun 23 14:51:48 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Synch.h: Put the class ACE_recursive_mutex_state inside + the section ACE_HAS_THREADS. Thanks to Craig Rodrigues + for noticing the need for this. + +Sun Jun 23 14:05:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ACEXML/common/HttpCharStream.cpp: + Fixed BCB Unicode compile errors. + +Fri Jun 21 19:18:05 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move to IA64_Linux_GCC_3.1 and + Tru64_CXX builds to new scoreboard. + +Fri Jun 21 14:34:51 2002 Steve Huston <shuston@riverace.com> + + * ace/config-sunos5.5.h: Enable + ACE_HAS_STD_TEMPLATE_METHOD_SPECIALIZATION for Forte 6 Update 1. + +Fri Jun 21 14:06:33 2002 Steve Huston <shuston@riverace.com> + + * ace/Timer_Queue_Adapters.cpp (ACE_Thread_Timer_Queue_Adapter): + Changed ACE_GUARD_RETURN uses of command_mutex_ from + ACE_Recursive_Condition_Mutex to ACE_SYNCH_MUTEX to match + its definition. + +Fri Jun 21 06:21:33 2002 Christopher Kohlhoff <chris@kohlhoff.com> + + * ace/config-win32-borland.h: + Re-enable ACE_HAS_STD_TEMPLATE_METHOD_SPECIALIZATION for all + versions of Borland C++. + +Thu Jun 20 20:50:51 2002 Steve Huston <shuston@riverace.com> + + * ace/config-sunos5.5.h: Re-enable + ACE_HAS_STD_TEMPLATE_METHOD_SPECIALIZATION for Forte 6 Update 2. + + * ace/Functor.i: + * ace/Synch.cpp: Commented out all the uses of + ACE_TEMPLATE_METHOD_SPECIALIZATION. The class specialization was + already declared such, and doing it on the methods is not right. + But they're only commented out for now, just in case... + +Thu Jun 20 20:29:31 2002 Steve Huston <shuston@riverace.com> + + * ace/Codecs.h (ACE_Base64): Added "friend class ace_dewarn_gplusplus" + to silence g++ warnings about private ctors and no friends. + +Thu Jun 20 20:21:39 2002 Steve Huston <shuston@riverace.com> + + * tests/Recursive_Condition_Test.icc: Visual Age C++ configuration + for this new test. + * tests/tests.icp: Added Recusrive_Condition_Test.icc. + +Fri Jun 21 00:21:48 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * examples/Timer_Queue/Thread_Timer_Queue_Test.cpp (dump): + Switch from ACE_SYNCH_MUTEX to ACE_SYNCH_RECURSIVE_MUTEX. + +Thu Jun 20 14:32:03 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Synch.cpp (ACE_recursive_mutex_state): Replaced the + ACE_OS::mutex_lock/unlock calls with + ACE_OS::thread_mutex_lock/unlock so things will work on NT. + Thanks to Jeff Parsons for reporting this. + + * ace/Synch.{h,cpp}: Only compile the new + ACE_Condition<ACE_Recursive_Thread_Mutex> template + specialization if ACE_HAS_THREADS. Thanks to Craig Rodrigues + for motivating this. + + * include/makeinclude/platform_sunos5_sunc++.GNU (CC_VERSION): Added + additional cases for SunC++ 5.2 and 5.3. Thanks to Carsten + Tonsberg Nielsen <ctn@maerskdatadefence.dk> for reporting this. + + * ace/ATM_Stream.cpp (get_peer_name): + * ace/Log_Msg_UNIX_Syslog.cpp (log): + * ace/Logging_Strategy.cpp: Replaced the use of ACE_OS::strtok() with + ACE_OS::strtok_r(). Thanks to Eugene Alterman + <eugalt@myrealbox.com> for motivating this. + +Thu Jun 20 16:02:22 2002 Steve Huston <shuston@riverace.com> + + * ace/config-sunos5.5.h: Commented out the below addition for now. + +Thu Jun 20 15:49:39 2002 Steve Huston <shuston@riverace.com> + + * ace/config-sunos5.5.h: For Forte 6, Update 2 and up, add + #define ACE_HAS_STD_TEMPLATE_METHOD_SPECIALIZATION. + Earlier Forte 6 versions may also support this, but I don't have + one to test on. + Fixes compile warning on Filecache.cpp. + +Thu Jun 20 13:37:55 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * examples/C++NPv2/README: Updated the README file to map the + filenames to the chapters in C++NPv2. + +Wed Jun 19 21:30:05 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + Bill Fulton <Bill_A_Fulton@raytheon.com> + + * ace/SOCK_Dgram_Mcast_T.cpp: Fix Win32 wchar build problems. + Thanks to Bill Fulton for supplying these fixes. + +Wed Jun 19 16:16:32 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * tests/tests.dsw: + * tests/Recursive_Condition_Test.dsp: Added a new project file. + +Wed Jun 19 06:50:54 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * tests/run_test.lst: + * tests/Makefile.bor: + * tests/Makefile: Added the new Recursive_Condition_Test. + + * ACE/Timer_Queue_Adapters.{h,cpp}: Enhanced + ACE_Thread_Timer_Queue_Adapter so it uses the new + ACE_Condition<ACE_Recursive_Thread_Mutex>. Thanks to + Stephen Howard <stephen.e.howard@lmco.com> for + suggesting this. + + * ace/Synch_T.h: Added a new RECURSIVE_CONDITION trait in + ACE_NULL_SYNCH and ACE_MT_SYNCH. + + * ace/Synch_T.h: Added a #define for ACE_SYNCH_RECURSIVE_CONDITION + so that we can parameterize it based on threads vs. non-threads. + + * docs/exceptions.html: Updated document to use ACE_RE_THROW + instead of the deprecated ACE_RETHROW macro. Thanks to + Andy Ling <Andy.Ling@Quantel.Com> for reporting this. + + * ace/Timer_Queue_Adapters.h: Changed the order in which + the condition variable and mutex are defined to ensure the + mutex is initialized before we pass it to condition_'s + constructor. + + * ace/Filecache.cpp: + * ace/Synch.cpp: + * ace/Atomic_Op.i: Replaced ACE_TEMPLATE_SPECIALIZATION with + ACE_TEMPLATE_METHOD_SPECIALIZATION so this will compile correctly + on certain (broken) compilers, such as SunC++ 6.1. + +Wed Jun 19 14:25:52 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/TP_Reactor.cpp (handle_socket_events): Couple of small + fixes. The fixes are for + + 1. Bug http://deuce.doc.wustl.edu/bugzilla/show_bug.cgi?id=1231 + and it is self-explanatory + + 2. The find () operation after dispatch was not holding the lock + and chances of race are high. This is documented as #2 in + bug http://deuce.doc.wustl.edu/bugzilla/show_bug.cgi?id=1233 + + Added some comments in the code explaining what is going + on. Thanks to Carlos for identifying the problem. + + * ace/Event_Handler.h: Added a enum to indicate the two states + returnable by the call resume_handler (). + + * ace/Event_Handler.cpp (resume_handler): Instead of returning 0, + we return ACE_Event_Handler::ACE_REACTOR_RESUMES_HANDLER. + +Wed Jun 19 13:42:16 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * tests/tests.dsw (Project): Added TP_Reactor_Test.dsp to the + workspace. + +Wed Jun 19 11:57:02 2002 Carlos O'Ryan <coryan@atdesk.com> + + * bin/auto_run_tests.lst: + Disable TAO/tests/Nested_Upcall_Crash on minimum_corba builds. + +Tue Jun 18 14:29:26 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Synch.{h,i}: Added a new template specialization of ACE_Condition<> + that supports recursive mutexes used in conjunction with ACE + condition variables. Thanks to Michael Lindner + <mikel@attmail.com> and Mike Vitalo <mvitalo@sprynet.com> for + contributing this. + + * ace/SOCK_Dgram_Mcast.cpp: Reverted the file to the previous version + since somehow things got messed up... + +Tue Jun 18 12:44:54 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/OS.{h,i}: Added a new method called stopped() that + returns 1 if the timer has already been stopped. Thanks to + Giovanni Zito <g.zito@acsys.it> for this suggestion. + +Tue Jun 18 12:17:58 2002 Jaiganesh Balasubramanian <jai@doc.ece.uci.edu> + + * ace/QoS/SOCK_Dgram_Mcast_QoS.cpp: + Added the RCS ID symbol. + +Tue Jun 18 12:41:15 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/CDR_Stream.i (adjust): Added a check for pointer overflow + befor setting the write pointer. thanks to Rich Siebel and Byron + Harris for pointing this out. + +Tue Jun 18 07:50:14 2002 Ossama Othman <ossama@uci.edu> + + * ace/Env_Value_T.h: + + Include "ace/OS.h". It is required for some types and function + calls made in this header. Thanks to Michael Kircher for + pointing this out. + +Mon Jun 17 17:36:52 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/AIO_Client_Logging_Daemon.cpp: Changed call + to ACE_Asynch_Connector::open() to take advantage of new + defaults, made on: + Sat Jun 15 20:30:41 2002 Steve Huston <shuston@riverace.com> + +Mon Jun 17 17:21:38 2002 Steve Huston <shuston@riverace.com> + + * ace/SUN_Proactor.cpp (handle_events): Remove the now-deleted + success argument from call to application_specific_code(). + +Mon Jun 17 13:40:24 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * include/makeinclude/platform_osf1_4.x_cxx.GNU: Removed the hard + override of LDFLAGS. Thanks to Martin Krumpolec + <krumpo@pobox.sk>. + +Mon Jun 17 06:20:55 2002 Christopher Kohlhoff <chris@kohlhoff.com> + + * ace/Object_Manager.cpp: + * ace/Service_Config.h: + * ace/Service_Config.cpp: + + Some components registered with the framework repository have + dependencies on the ACE_Log_Msg or ACE_Thread_Manager singletons. + Reorganised the ACE_Object_Manager::fini function so that the + framework repository is cleaned up before these objects. Thanks to + Don Hinton <dhinton@ieee.org> for assistance in tracking this down. + + * ace/Reactor.cpp: + + Prevent the reactor from being erroneously registered with the + framework repository a second time, as happened when the reactor + instance pointer was explicitly set. + +Sun Jun 16 20:38:27 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/DLL_Manager.cpp (ACE_DLL_Handle): + * ace/DLL.cpp: + + More mismatched delete/delete[] errors. I am fixing the easier + ones. There may be more. + +Sun Jun 16 20:22:28 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/AttributesImpl.i (qName): + + Fixed memory leaks caused by mismatched delete/delete[]. Caught + by Valgrind. + + * ACEXML/examples/SAXPrint/main.cpp: + + Give useful usage messages. + + +Sat Jun 15 20:56:29 2002 Steve Huston <shuston@riverace.com> + + * tests/tests.dsw: Added Proactor_Timer_Test.dsp. + +Sat Jun 15 20:30:41 2002 Steve Huston <shuston@riverace.com> + + * ace/Asynch_Acceptor.{h cpp}: Added a new template method, + virtual int validate_connection + (const ACE_Asynch_Connect::Result& result, + const ACE_INET_Addr &remote, + const ACE_INET_Addr& local) + It replaces the old validate_new_connection() method, which is + now marked deprecated. Semantics are the same (return -1 to + ask framework to abort the connection before activating a + service). However, more information is available, including + the socket handle. + + * ace/Asynch_Connector.{h cpp}: Replaced validate_new_connection() + method with: + virtual int validate_connection + (const ACE_Asynch_Connect::Result& result, + const ACE_INET_Addr &remote, + const ACE_INET_Addr& local) + Examining the result argument is the only way to find out if + the asynchronous connect attempt succeeded or not. + The connect completion calls validate_connection() on either + success or failure as long as the validate_new_connection + flag is set on open(). The default value was changed from + 0 (don't call) to 1 (call) since this is the only way to learn + about connection success/failure. + + * ace/POSIX_Proactor.{h cpp} (application_specific_code): + * ace/WIN32_Proactor.{h cpp} (application_specific_code): Removed the + 'int success' argument. Success/fail is determined by the error + value now (0 == success). Simplifies the code, and makes the + result.success() call return 1 iff the operation succeeded. + + Thanks very much to Alex Libman for these ideas and for working + to improve these classes for ACE and C++NPv2! + +Sat Jun 15 12:01:38 2002 Steve Huston <shuston@riverace.com> + + * ace/Obstack_T.cpp (request): Don't lose an initial set of characters + if the requested length won't fit. Thanks to Christopher W. Midgley + <cmidgley@livevault.com> for this report and fix. + + * tests/Obstack_Test.{cpp dsp icc}: New test for the above, and + its MSVC and Visual Age C++ files. + + * tests/Makefile: + * tests/Makefile.bor: + * tests/run_test.lst: + * tests/tests.dsw: + * tests/tests.icp: Added new Obstack_Test. + + * THANKS: Added Christopher W. Midgley. + +Sat Jun 15 12:41:07 Frank Hunleth <fhunleth@cs.wustl.edu> + + * THANKS: Added Nick Cross to the hall of fame. + +Fri Jun 14 14:39:33 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/Filecache.cpp: Removed excessive ACE_DEBUG statements. + Thanks to Charles F. Frasch <cfrasch@spawar.navi.mil> for + reporting them. + +Fri Jun 14 12:04:17 2002 Jaiganesh Balasubramanian <jai@doc.ece.uci.edu> + + * ace/QoS/SOCK_Dgram_Mcast_QoS.cpp: + * ace/QoS/SOCK_Dgram_Mcast_QoS.h: + Removed build errors for QoS library, due to dependencies to the + old SOCK_Dgram_Mcast class. Thanks to Bill Fulton for this fix. + +Fri Jun 14 14:17:49 2002 Carlos O'Ryan <coryan@atdesk.com> + + * bin/auto_run_tests.pl: + Fixed wrong application name in Usage message. + + * bin/auto_run_tests.lst: + Add the Nested_Upcall_Crash test. + +Fri Jun 14 09:34:41 2002 Ossama Othman <ossama@uci.edu> + + * ace/OS.i (ctime): + + For the ACE_HAS_WINCE, pass a statically allocated buffer to + ctime_r() since ctime_r() will return that same buffer as the + return value of ctime(). Fixes memory access violations. + Thanks to Michael Searles <msearles@base16.com> for pointing out + the problem. + +Thu Jun 13 21:26:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Configuration_Import_Export.cpp (squish): Replaced use + of naked isspace() with ACE_OS_String::ace_isspace() to work + with wide chars also. + +Thu Jun 13 21:09:33 2002 Steve Huston <shuston@riverace.com> + + * tests/Config_Test.cpp (run_tests): Added a test for importing + a previously-existing ini file and verifying its contents. + + * tests/Config_Test_Import_1.ini: File for the above test. + +Thu Jun 13 19:05:57 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/rules.local.GNU: + + Added a rule to install a symlink to libfoo.so.3 under + $(ACE_ROOT)/ace. This should fix all the build errors. ld.so + needs libfoo.so.3 to be in LD_LIBRARY_PATH. libfoo.so doesn't + do. Thanks to Marvin Wolfthal <maw@weichi.com> for reporting the + problem. + + * bin/make_release: + + Changes to ensure that all major releases are named + ACE-[0-9]+.0.0 and all minor releases are named + ACE-[0-9]+.[0-9]+.0. PLEASE TAG THE REPOSITORY AS ACE_5_3_0 and + TAO_1_3_0 or ACE_6_0_0 and TAO_2_0_0 WHEN MAKING A MINOR OR A + MAJOR RELEASE. Pointers to add this to somewhere more prominent + are welcome. + +Thu Jun 13 14:45:13 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/SOCK_Dgram_Mcast_T.cpp: Added wchar support - remove + warnings in MSVC when #define ACE_USES_WCHAR used in config.h. + Thanks to Bill Fulton for this fix. + +Wed Jun 12 21:28:29 2002 Steve Huston <shuston@riverace.com> + + * tests/Framework_Component_DLL.cpp (Server_T): Add template + parameter name to class template arg to get aC++ to eat it. + +Wed Jun 12 18:29:11 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/rules.lib.GNU: + * include/makeinclude/rules.local.GNU: + * include/makeinclude/wrapper_macros.GNU: + * TAO/rules.tao.GNU: + * include/makeinclude/platform_g++_common.GNU: + + Added support for shared library versioning. Libraries built + with GNU ld will be versioned by default. + + * include/makeinclude/platform_aix4_cset++.GNU: + * include/makeinclude/platform_aix4_g++.GNU: + * include/makeinclude/platform_aix_g++.GNU: + * include/makeinclude/platform_aix_ibm.GNU: + * include/makeinclude/platform_hpux.GNU: + * include/makeinclude/platform_hpux_aCC.GNU: + * include/makeinclude/platform_hpux_gcc.GNU: + * include/makeinclude/platform_hpux_kcc.GNU: + + Added null SONAME and SOVERSION for disabling shared library + versioning. + + * include/makeinclude/platform_dgux4_epc.GNU: + * include/makeinclude/platform_freebsd.GNU: + * include/makeinclude/platform_gnuwin32_common.GNU: + * include/makeinclude/platform_irix5.3_g++.GNU: + * include/makeinclude/platform_irix5.3_sgic++.GNU: + * include/makeinclude/platform_irix6.x_g++.GNU: + * include/makeinclude/platform_irix6.x_kcc.GNU: + * include/makeinclude/platform_irix6.x_sgic++.GNU: + * include/makeinclude/platform_linux.GNU: + * include/makeinclude/platform_linux_cxx.GNU: + * include/makeinclude/platform_linux_icc.GNU: + * include/makeinclude/platform_linux_kcc.GNU: + * include/makeinclude/platform_m88k.GNU: + * include/makeinclude/platform_macosx.GNU: + * include/makeinclude/platform_mvs.GNU: + * include/makeinclude/platform_netbsd.GNU: + * include/makeinclude/platform_osf1_3.2_cxx.GNU: + * include/makeinclude/platform_osf1_4.x_cxx.GNU: + * include/makeinclude/platform_osf1_4.x_g++.GNU: + * include/makeinclude/platform_osf1_4.x_kcc.GNU: + * include/makeinclude/platform_osf1_4.x_rcc.GNU: + * include/makeinclude/platform_psosim_g++.GNU: + * include/makeinclude/platform_qnx_neutrino.GNU: + * include/makeinclude/platform_qnx_rtp_gcc.GNU: + * include/makeinclude/platform_sco-nothread.GNU: + * include/makeinclude/platform_sco5.0.0-CC-fsu-pthread.GNU: + * include/makeinclude/platform_sco5.0.0-mit-pthread.GNU: + * include/makeinclude/platform_sco5.0.0-nothread.GNU: + * include/makeinclude/platform_sunos4_g++.GNU: + * include/makeinclude/platform_sunos4_lucid.GNU: + * include/makeinclude/platform_sunos4_sunc++3.x.GNU: + * include/makeinclude/platform_sunos4_sunc++4.x.GNU: + * include/makeinclude/platform_sunos5_centerline.GNU: + * include/makeinclude/platform_sunos5_g++.GNU: + * include/makeinclude/platform_sunos5_ghs.GNU: + * include/makeinclude/platform_sunos5_kcc.GNU: + * include/makeinclude/platform_sunos5_sunc++.GNU: + * include/makeinclude/platform_tandem.GNU: + * include/makeinclude/platform_unixware_g++.GNU: + * include/makeinclude/platform_unixware_udk.GNU: + + Changed SOFLAGS = to SOFLAGS += to avoid dealing with order of + includes of various Makefile rules. + +Wed Jun 12 11:25:22 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * tests/Config_Test.cpp (test): Removed what appeared to be + an unnecessary and unused called to + const ACE_Configuration_Section_Key &root = config->root_section (); + + * ace/Configuration.cpp (open_section): Fixed a warning in the + loop expression. + +Wed Jun 12 11:49:33 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: Changed to + match book. + + * examples/C++NPv2/Reactor_Logging_Server_Adapter.cpp (init): + Use ACE_Auto_Array_Ptr to be cleaner about allocated char ptr + array. Coincides with book change. + +Tue Jun 11 10:23:03 2002 Stephen Torri <storri@cs.wustl.edu> + + * ace/String_Base.h: Updated comments/documentation on the String_Base + class. + + * ace/String_Base.{i,cpp}: Added brackets to if/else statements for + better structure/readability. Changed "(cond) ? true : false" to + better "if/else" structure. + +Tue Jun 11 19:05:39 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/SOCK_Dgram_Mcast.h: Added a default constructor for + ACE_SOCK_Dgram_Mcast. + + * tests/Config_Test.cpp (test): Added tests for the following capability. + Thanks to Eugene Alterman <eugalt@myrealbox.com> for + contributing these tests. + + * ace/Configuration.{h,cpp}: + ACE_Configuration_Heap::open_section() now accepts path + separators in section name, so that it is consistent with the + updated version of + ACE_Configuration_Win32Registry::open_section(). Thanks to + Eugene Alterman <eugalt@myrealbox.com> for contributing this + patch. + +Tue Jun 11 20:02:39 2002 Steve Huston <shuston@riverace.com> + + * ace/DLL_Manager.cpp (ACE_DLL_Handle::error()): Change to a + different form of auto_ptr initialization that Visual Age C++ + likes better. + +Tue Jun 11 13:17:29 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Cache_Map_Manager_T.h: + * ace/Caching_Utility_T.h: + * ace/Env_Value_T.h: Added #include "ace/Global_Macros.h". + Thanks to Ty Tenait <tytenait@yahoo.com> for reporting this. + + * ace/Dirent_Selector.h: Added #include "ace/ACE_export.h". + Thanks to Ty Tenait <tytenait@yahoo.com> for reporting this. + +Tue Jun 11 09:42:06 2002 Ossama Othman <ossama@uci.edu> + + * ace/SOCK_Dgram_Mcast_T.cpp (make_multicast_ifaddr): + + Use ACE_HTONL instead of htonl(). The latter causes the + assembler to choke on some Linux installations. + +Mon Jun 10 14:50:46 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * examples/DLL/Dll.dsw: + * examples/ASX/CCM_App/CCM_App.dsw: Updated project dependencies. + +Mon Jun 10 14:23:28 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Reactor.h (ACE_Reactor): Added a comment explaining that + the mask passed to notify() can only be READ_MASK, WRITE_MASK, + or EXCEPT_MASK. Thanks to Cary Steinmetz + <cary.s.steinmetz@boeing.com> for motivating this comment. + + * ace/OS_String.cpp: Make sure to check whether src == dst and + avoid doing any copies in this case. Thanks to Michael Searles + <msearles@base16.com> for reporting this. + +Sun Jun 9 20:01:00 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Makefile: Something is wrong with my ACE_ROOT environment + variable, which caused problems with the dependency generation. + This is now fixed. + +Sun Jun 9 19:31:04 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/SOCK_Dgram_Mcast.h: Added ACE_Export to ACE_SOCK_Dgram_Mcast. + +Sun Jun 9 18:57:07 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * examples/Reactor/Ntalker/ntalker.cpp: + * examples/Reactor/Multicast/Log_Wrapper.cpp: Added template + instantiations for the new ACE_SOCK_Dgram_Mcast_Ex<> template. + +Sun Jun 9 18:47:22 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * tests/Config_Test.cpp (run_tests): Added code to exercise the + new subkey path feature contributed by Eugene Alterman. + + * ace/Configuration.{h,cpp}: Added support for the path separator + ('\\') in key names. Thanks to Eugene Alterman + <eugalt@myrealbox.com> for contributing this patch. + + * ace/Date_Time.{h,i}: Added support for weekday. Thanks to + Eyal Lubetzky <eyall@BANDWIZ.COM> for reporting this. + +Sun Jun 9 18:11:02 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/SOCK_Dgram_Mcast.cpp: Removed the explicit inclusion of + "ace/SOCK_Dgram_Mcast_T.cpp" since this is #included appropriately + by SOCK_Dgram_Mcast.h. + + * ace/SOCK_Dgram_Mcast.cpp: Removed the ';' from the end of the + #pragma instantiates. + + * ace/Makefile: Added SOCK_Dgram_Mcast_T. + + * ace/SOCK_Dgram_Mcast_T.cpp + ace/SOCK_Dgram_Mcast_T.i + ace/SOCK_Dgram_Mcast_T.h: + Minor changes to conform to ACE coding conventions. + +Sun Jun 9 17:50:37 2002 Bill Fulton <bill_a_fulton@raytheon.com> + + * ace/SOCK_Dgram_Mcast.cpp + ace/SOCK_Dgram_Mcast.h + ace/SOCK_Dgram_Mcast_T.cpp <new> + ace/SOCK_Dgram_Mcast_T.i <new> + ace/SOCK_Dgram_Mcast_T.h <new> + + Somewhat massive changes to the (already very useful) + ACE_SOCK_Dgram_Mcast class; some fixes to existing functionality, + some new functionality, and a different way of controlling optional + (per-instance) functionality. The intent was to preserve existing + interface and functionality (when correct), including environment- + specific functionality. + + Fixes to Existing Functionality: + + - For ::unsubscribe(void); unsubscribes from all active subscriptions. + Previous version unsubscribed only from the most recently + subscribed group/net_if. + * This could change execution behavior of existing code. + + - Enhanced support for a passed port# of '0'. + The ephemeral port# assigned by the system when the socket is + bound is used for all local ::send methods. + + - Validation of non-zero port# and, optionally, the address passed + to the ::subscribe method, if the socket has been bound. + On second and subsequent ::subscribe invocations, the previous + version allowed passed parameters that were inconsistent with the + bound socket parameters. This could result in a confusing + condition; you would not receive messages sent to the port and/or + address passed to ::subscribe. + * This could change execution behavior of existing code. + + Added Functionality: + + - Support specification of the (multicast) address and interface to + be used by all local ::send methods, independent of the receive + address and interface. The same port# is used for all base class + ::recv methods and all local class ::send methods. The send + interface applies to all local and all base class ::send methods. + + - Enhanced ::dump method, to show the dynamic list of subscriptions, + the bound address/port#, the send address/interface, and the state + of optional functionality. Also did some output reformatting. + + Optional Functionality via Template Parameters and Ctor Args: + + In the previous implementation, some aspects of class functionality + were transparently configured based on the OS it was compiled + under. This has been changed so that all optional functionality is + explicitly specified, using (a baroque set of) optional + constructor arguments. In addition, new optional functionality + was added and is specified with template parameters and optional + constructor arguments. For backward compatibility, there is an + explicit specialization called ACE_SOCK_Dgram_Mcast - this should + have the same functionality, including OS-specific functionality, + as the previous class implementation, and can still be used for + most normal (if there is such a thing with multicast) apps. + Overview of new template parameter: + ACE_SDM_LOCK_* + - Strategized locking (very limited MT-safety - see below). + Overview of new, optional ctor arguments (enumerations): + opt_bindaddr_* + - Whether to bind the address as well as the port#. + opt_nulliface_* + - Defines semantics of passing net_if parameter of NULL; + i.e. use all or only one iface. + opt_dtorunsub_* + - Whether destructor explicitly unsubscribes from all + groups. + See SOCK_Dgram_Mcast_T.h for a full description. Environment- + specific default values are explicitly declared, for use by app + code. + + Changes to Previous Class (public) Interface & Functionality: + + - Added a public (overloaded) ::open method. + This allows explicit specification of the port# and the multicast + address (if opt_bindaddr_yes is used) that is bound to this + socket. This also allows specification of the address and the + network interface to be used by local ::send methods (and the + network interface used by super-class and sub-class ::send + methods). There is only one address parameter; the same parameter + is used for both the send address/port# and the bound + address/port#. The port# in the address may be 0 (to use a + system-assigned ephemeral port# for sending and receiving). + + NOTE: Use of this method is optional and, if used, must be + invoked _before_ any ::subscribe method is invoked - otherwise it + is _ignored_, and the (first) ::subscribe parameters will be used + to bind the socket and set the send parameters (as in the previous + code). Note that ::subscribe need not be invoked at all, now; if + only ::open is used, you can send but not receive multicast + messages, and you can specify the interface used for sending. + + - Validation of ::subscribe parameters is now strict. + Once the port# is bound (by ::open or the first ::subscribe), the + port# passed to every ::subscribe call must either be '0' or match + the bound port#. Additionally; if opt_bindaddr_yes is used, the + address passed must match the bound address. If these conditions + are not met, an error is returned (with errno=ENXIO). Note that + such subscriptions would have no (useful) effect, since messages + could never be received for the given port#/address, due to IP + stack filtering. This condition was ignored in the previous + version; it is now an error condition. This may break some + existing code - but note that it flushes out a (silent) logical + error condition. + The port# in ::unsubscribe(addr,...) is _not_ checked, though; + this is not logically consistent, but ... a mistaken port# here + seems less potentially dangerous and/or confusing. + + - The previous ::subscribe and ::unsubscribe methods had a special + return code (1) to indicate that (previously Win32-specific) + multiple-interface subscription/unsubscription had been done. This + seemed to be useful only internally, but was visible to the + caller. The new methods will always return '0' for success (even + if opt_nulliface_all is used). + + - The dynamic subscription list housekeeping adds slight processing + overhead, and even more overhead is added if locking is enabled. + This should not be noticeable, except in MT thrashing conditions + and/or if doing very frequent subscribes/unsubscribes. + + Known Restrictions: + + - Interaction/Integration with other ACE multicast-oriented classes + and library code is _undefined_, i.e.: + + ACE_SOCK_Dgram_Mcast_QoS: Presumably works ok, but inherits + legacy class; can't specify optional functionality. + + ACE_RMCast: Presumably works ok, but uses legacy class; + can't specify optional functionality. + + ACE_XTI_ATM_Mcast: I presume there's no integration desirable. + I've re-built vers. 5.2 ACE under Solaris and Windows with the + modified files, and it builds ok. Since the original functionality + is (basically) unchanged, it is _assumed_ that any other ACE lib + code that references the ACE_SOCK_Dgram_Mcast class signature + still functions ok. But ... things like adding templated code + could introduce problems! Also note that these classes use (now) + deprecated interface methods; e.g. RMCast passes protocol options + to ::subscribe that are now ignored. + + - Environment-specific notes/restrictions on use of options: + + The opt_nulliface_all option should be used only in environments + for which the ACE_Sock_Connect::get_ip_interfaces call is + correctly implemented (currently only Windows). In other + environments, using this option will _always_ return an error + from ::subscribe/::unsubscribe. + + The opt_dtorunsub_yes option allows finer control of multicast + group "leave" effects for environments that do not automatically + unsubscribe when the socket is closed, or in situations where + the underlying handle has been duplicated (e.g. via fork()), + but will not be used. All in all, this option is useful only + under special, controlled circumstances. + + The effect of using opt_bindaddr_yes is very environment- + specific; it may not do what you expect (or may do what you + don't expect :-) + The default option values are defined with these notes in mind. + + - Options enumeration weaknesses. + + The options enumeration list was put in the global namespace. + It seems like a pain for user code to have to prefix the option + values with the templated class name. + + All option values use the same enumeration; this could allow + option values to be used for the "wrong" option in the ctor + options list. This was a trade-off; I didn't want to add too + many things to the global namespace. (Note that more options + might be added to the list, in the future.) All new code + compares against explicit enum values (instead of 0), for safety. + + - The ::open parameter list may be too "dense" - it tries to do too + many different things with the passed parameters. This is an + artifact of its heritage. (Note that ::subscribe can also invoke + ::open behavior). In particular, it would be nice to separate + bound addr/iface from send addr/iface. + + - Does not explicitly check for duplicate active subscriptions (i.e. + subscription to the same address and interface while a previous + subscription is still active). (Subscription to the same address + but different interface is not a duplicate and is fully + supported.) Note that, for most IP stacks, a subscription attempt + for a duplicate addr/iface will fail (and a duplicate entry will + not be created). If it is allowed by the IP stack and you create + duplicate internal subscription list entries, you must call + ::unsubscribe(addr) for _each_ duplicate subscription (or + ::unsubscribe(void)) - until this is done, you will continue to + recv messages for this subscription. + + - This implementation carries forward (for backward compatibility) + certain environment-specific hacks, i.e.: + + A Win32 hack such that the default opt_nulliface_defopt value + for Win32 is different from all other environments. + + A Linux hack such that the full address is unconditionally bound. + These would seem likely to lead to unexpected behavior for the + "casual user", but changing the default would break existing code. + + - The SDM_OPT_LOCK parameter only controls locking of the dynamic + subscription list; and the need/use for this is debatable. It was + added as a fail-safe, in case one class instance is shared by + multiple threads. Note that a given instance can _not_ (reliably) + be shared by multiple threads; there are logical race conditions + (e.g. calling ::open) - but the logical race conditions should + cause only unpredictable operation, not corruption. However; the + subscription list could be corrupted if one instance is shared by + multiple threads, which could lead to undefined behavior/crashes. + + - The class interface is probably too "busy"; there are a _lot_ of + side-effects that may not be evident. + + - Currently only built and tested under Solaris w/ Forte-C++ and + Windows NT and 2000 w/ MSVC 6. + + Implementation Changes: + + - Now implemented as a templated class (ACE_SOCK_Dgram_Ex), via the + new SOCK_Dgram_Mcast_Ex.* files. The (remaining) SOCK_Dgram_Mcast.* + files are vestigial; they remain due to the build rules and + documentation conventions, and because existing user code + references the ace/SOCK_Dgram_Mcast.h file. Note that this + differs from most other *_Ex.* template source files in ACE, where + there is still at least some function implemented in the + non-templated source files; here there is none. + An explicit specialization is declared that matches the old class + signature. + + - The constructor now has optional arguments (to specify per-instance + optional functionality. A default ctor can still be used; the + legacy default environment-specific functionality will be declared. + + - I've tried to generalize code that was previously overly + environment-dependent and/or macro-driven. The intent was to make + future enhancements easier (there's still much desirable work to + be done ... speaking of which ...) + + Suggested Future Work: + + - Expose the SOCK_Dgram::send methods; i.e. allow specification of an + arbitrary dest address/port# with the ::send call. (Must check for + port# of '0' in the send address and set port# to the bound port#). + This would allow the specification of the network interface for + arbitrary multicast transmissions from one instance, and should + work as expected if the address specified is unicast. This seems + useful; currently, if you want to send multicast messages with + interface control, you would either have to create one instance of + this class for each send address, or force access to the inherited + ACE_SOCK_Dgram instance (which works ... "but it would be wrong"). + I'm not sure why access to these was disabled in the original + design. (Is there a good reason not to do this?) + + - Mark the current ::subscribe and ::unsubscribe interfaces as + deprecated and add new interfaces that remove all unnecessary + parameters. + + Remove the (optional) protocol* parameters from the ::subscribe + and ::unsubscribe methods - they don't do anything. I (really) + wanted to do this now, but it could break existing app code (but + you can't really specify anything other than the default values, + anyway!) + + Remove the (optional) reuse_addr parameter from the ::subscribe + method - it is only useful if ::open has not been previously + invoked. Problem: This will require that the ::open method be + explicitly invoked before invoking ::subscribe, which makes the + interface more clumsy. + Problem: Hard to get a unique signature for the new methods, + especially for ::unsubscribe. + + - It seems that an ACE_SOCK_Dgram_Mcast_Manager class could be a + useful "layered" addition, e.g.: + - Trap ::subscribe invocations using different port numbers, and + create separate instances/sockets for each unique port#. + - Implement software loopback filtering for environments that do + not implement loopback disabling in the IP stack (e.g. older + Windows). Also implement per-interface filters for environments + (like Win/NT) that don't do this right. + - Implement multicast group "sets" to support easy and atomic + joins/leaves of a user-defined list of group/interface + combinations. Include enhanced failure condition handling, such + as subscription "roll-back"; all-or-none of the set's members + are joined. + Actually; all of these could be done in the current class - but + there are too many options in this class already, and the function + seems (to me) more appropriate to a higher-level "manager". + + - Allow the send interface (and default send addr) to be changed, + after initial setting. (OS-specific support needed - most stacks + allow the interface to be changed once set, but I presume that not + all do.) + In general; might be worth splitting out the send from the + ::open/::subscribe methods - e.g. ::send_parameters. Might add + optional TTL and loopback parameters to this, too. + + - Add explicit methods to set common multicast socket options, e.g. + TTL, loopback. (So you don't have to remember the macro names and + to make these macros/actions more portable!). + + - Tighten up the parameter semantics and/or validation, e.g.: + + Verify that all addresses passed are multicast addresses. + For example, ::open does not check the send address. (It should + still work as expected - sending unicast on the route-derived + interface - but that's not the intended use.) + + The reuse_addr parameter in ::open is used for both REUSEADDR + and REUSEPORT (if implemented) option setting; should these be + distinct parameters? (They do different things!) + + - Enhanced integration with other ACE multicast classes, as needed + or desirable (see Restrictions section). + + - (?) For ::subscribe and ::unsubscribe - if successful, return the + (positive) number of subscribes/unsubscribes done (instead of '0'). + Possibly useful, as a way for the app to sense multi-interface + subs/unsubs (e.g. code that is passed an instance and does not + know what options were used when the instance was created). + + - (?) Add method ::unsubscribe(int n) to unsubscribe from most + recent n subscription(s) (including opt_nulliface_all option + support.) The implementation of group sets might obviate the + usefulness of this. + + - (?) Enhanced MT-safety and exception handling. + Don't know if any particular problems; but haven't tested + extensively. Also; might be able to make it more efficient. + In particular; if one instance is intended to be shared by multiple + threads, there is _much_ work to be done. + + - (??) Track subscription list state; i.e. recognize the state when + any ::unsubscribe emptys the subscription list. This state might + be used to re-define some parameters (especially the network + interface for sends) set with ::open and for other changes ... but + utility vs. added complexity ratio seems low. (Note that closing + the socket and reopening another can cause nasty side-effects, so + probably can't change the bound parameters, anyway.) + +Sun Jun 9 16:26:00 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/ATM_Acceptor.cpp: + * ace/ATM_Addr.cpp: + * ace/ATM_Connector.i: + * ace/ATM_QoS.cpp: + * ace/ATM_Stream.cpp: + * ace/CE_Screen_Output.cpp: + * ace/Configuration.cpp: + * ace/Handle_Set.i: + * ace/IOStream.cpp: + * ace/Lib_Find.cpp: + * ace/Log_Msg.cpp: + * ace/Log_Record.cpp: + * ace/MEM_Addr.cpp: + * ace/Malloc_T.i: + * ace/Memory_Pool.cpp: + * ace/Message_Queue.cpp: + * ace/Msg_WFMO_Reactor.cpp: + * ace/OS.cpp: + * ace/OS.i: + * ace/OS_String.cpp: + * ace/Process.cpp: + * ace/SOCK_Connector.i: + * ace/SPIPE_Connector.cpp: + * ace/Sock_Connect.cpp: + * ace/Synch_T.cpp: + * ace/ace_wchar.inl: Changed improper use of NULL to + 0. + +Sun Jun 9 15:33:05 2002 Jeff Parsons <parsons@cs.wustl.edu> + + * ace/Date_Time.i: + + Fixed some typos that caused build errors. + +Sat Jun 8 14:46:19 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Array_Base.h: Replaced + + #include "Global_Macros.h" + + with + + #include "ace/Global_Macros.h" + + Thanks to Bill Fulton <Bill_A_Fulton@raytheon.com> for reporting + this. + +Fri Jun 7 19:00:31 2002 Steve Huston <shuston@riverace.com> + + * ace/OS_String.inl (strstr): Also see: + Thu May 30 16:45:10 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + HP-UX 11.00 doesn't supply wcsstr(), only wcswcs (), so + special-case that one with #if defined (HPUX). + +Fri Jun 7 16:42:29 2002 Steve Huston <shuston@riverace.com> + + * ace/Date_Time.i (update): Use ACE_OS::localtime_r rather than + ACE_OS::localtime to avoid MT time confusion. Thanks to Kelly + Hickel <kfh@mqsoftware.com> for reporting this. + +Fri Jun 7 16:29:21 2002 Steve Huston <shuston@riverace.com> + + * ace/Acceptor.{h cpp}: + (ACE_Strategy_Acceptor) Added an open () method with signature + that matches ACE_Acceptor::open to avoid hiding the one in + ACE_Acceptor, according to Forte 6. In the new open(), however, + the ACE_Reactor* arg is required, to make it distinguishable from + the existing open(). The new open forwards everything + to the previously-existing, more full-featured open(). + + Constructors and full-featured open(), added optional reuse_addr + arg - it's passed down to the ACE_Accept_Strategy::open method. + + Removed the service_port_ member - it's not used. + + (ACE_Acceptor::open): If the reactor registration fails, close + the peer_acceptor_ object. + + * netsvcs/TS_Server_Handler.cpp (ACE_TS_Server_Acceptor::parse_args): + * netsvcs/Name_Handler.cpp (ACE_Name_Acceptor::parse_args): + * netsvcs/Server_Logging_Handler_T.cpp (ACE_Server_Logging_Acceptor_T): + Use a local variable for the -p port option; server_port_ is gone. + +Fri Jun 7 14:30:22 2002 Steve Huston <shuston@riverace.com> + + * ace/Message_Queue.{h cpp} (ACE_Message_Queue_NT): Change + deactivate (int pulse) to deactive() and pulse() - forgot this + one in "Thu Jun 6 18:10:45 2002 Steve Huston" set of changes. + Also, corrected the way the completion routine checks state + of the queue. enqueue() will always pass ACTIVATED as the + queue state to be sure that blocks are seen when dequeued. + +Thu Jun 6 17:41:24 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/Mem_Map_Stream.h: + * ACEXML/common/Mem_Map_Stream.cpp: + * ACEXML/common/HttpCharStream.h: + * ACEXML/common/HttpCharStream.cpp: + * ACEXML/common/URL_Addr.cpp: + + Fixed the explicit template instantiation problems on Solaris. + +Thu Jun 6 18:35:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Acceptor.h (ACE_Strategy_Acceptor): Made open() virtual + since its parent class (ACE_Acceptor) ope() is also virtual. + Resolves warning from Forte 6 that open() hides its parent + class's open(). + +Thu Jun 6 18:10:45 2002 Steve Huston <shuston@riverace.com> + + * ace/README: Removed ACE_HAS_OPTIMIZED_MESSAGE_QUEUE. No configs + use it, and it's functionality doesn't really support the + message queue deactivation/pulse semantics. + + * ace/Message_Queue.h: + * ace/Message_Queue_T.{h cpp i}: Removed the "int pulse" argument + from deactivate() and add a pulse() method. States stay the + same. + Took out all the ACE_HAS_OPTIMIZED_MESSAGE_QUEUE stuff. It's + not used any longer. + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: Use the pulse() + method instead of deactivate (1). + + * examples/C++NPv2/AIO_Client_Logging_Daemon.cpp: Removed answered + questions. Removed reimplemented handle_connect() method + because the needed things are now available via the framework. + + * ace/Asynch_Acceptor.{h cpp}: + * ace/Asynch_Connector.{h cpp}: Added new hook method, + int validate_connection (const ACE_Asynch_Accept::Result&result, + const ACE_INET_Addr& remote_addr, const ACE_INET_Addr& local_addr) + that allows access to success/fail, handle, and addresses. The + validate_new_connection (const ACE_INET_Addr&) method is now + deprecated. + + * ace/WFMO_Reactor.{h cpp} (ACE_WFMO_Reactor_Notify): Add an + optional argument to ACE_WFMO_Reactor_Notify constructor. + size_t max_notifies is used to specify a limit for how many + notifications can be queued. The value is used to calculate + new high and low watermarks to the message_queue_. Default 1024. + + * ace/Activation_Queue.cpp: + * ace/Message_Block.cpp: + * ace/Message_Queue_T.cpp: Added #include "ace/Log_Msg.h" to get + the ACE_DEBUG, etc. definitions. + + * ace/Read_Buffer.cpp: Added #include for ace/Log_Msg.h and + ace/Malloc_Base.h (for ACE_Allocator). + + * ace/Log_Msg.cpp (log (ACE_Log_Record&)): If there's a + ACE_Log_Msg_Callback, do it before sending the log record + to any other logging sinks. Allows the callback to munge + the data. + +Thu Jun 06 10:50:37 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.cpp (cancel_timer): + + Fixed race condition by adding an ACE_GUARD. Basically the same + fix that Doug made to the Select_Reactor below. + +Thu Jun 6 12:28:20 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Select_Reactor_T.cpp: Moved the cancel_timer() methods + from the Select_Reactor_T.i file and added ACE_GUARD's around + them to ensure we're serialized correctly. Thanks to Johnny + Chen <jchen@informatica.com> for motivating this. + +Thu Jun 6 08:24:43 2002 Jeff Parsons <parsons@cs.wustl.edu> + + * ace/Malloc_T.h: + + Removed extra comment closer */. + +Thu Jun 6 07:11:15 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * THANKS: Just added our 1,500th ACE+TAO contributor!!!!! + + * ace/Malloc_T.cpp: Clarify that the memory backing store must + reside in a directory with the appropriate visibility and + permissions. Thanks to Frank O. Flemisch + <f.o.flemisch@larc.nasa.gov> for reporting this. + +Wed Jun 5 23:06:11 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL_Manager.h: Fixed compile error. Thanks to Jai + Balasubramanian for pointing this out. + +Wed Jun 5 16:58:41 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/Mem_Map_Stream.cpp: + + Instantiate the templates needed when explicit templates are + defined. Thanks to Carlos for pointing this out. + + * ACEXML/common/HttpCharStream.cpp: + + Peek means "get the next character", not "get the character + after next". The previous change to the test caught this bug. + +Wed Jun 5 16:42:43 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/examples/SAXPrint/main.cpp (ACE_TMAIN): + + Modified the example to test fetching and parsing an XML URL. + This actually tests most of the functions of the new stream + class and is better than the lame test that I wrote the other + day. + +Wed Jun 5 13:54:03 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL_Manager.{h,cpp}: + * ace/Framework_Component.cpp: + + Made ACE_DLL_Manager a true singleton instead of using the + ACE_*Singleton templates, thus avoiding a problem with multiple + instantiations by broken compilers. Thanks to Chris Kohlhoff + <chris@kohloff.com> for this suggestion. + +Wed Jun 5 08:33:25 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Timer_Queue_T.h (class ACE_Timer_Queue_T): Clarify that the + second calculate_timeout() method doesn't use a lock. Thanks to + Johnny Chen <jchen@informatica.com> for motivating this. + + * ace/Process_Manager.cpp (handle_signal): Added a check for + i == -1. Thanks to Dimitrije Jankovic <djankov99@yahoo.com> + for reporting this and providing a fix! + +Tue Jun 04 16:47:27 2002 Ossama Othman <ossama@uci.edu> + + * ace/Service_Config.h: + + No need to include "ace/DLL.h". Forward declaring ACE_DLL is + enough. + + * ace/Service_Config.cpp: + + Include "ace/DLL.h" to pull in ACE_DLL class declaration. + +Tue Jun 04 16:34:18 2002 Ossama Othman <ossama@uci.edu> + + * ace/Service_Config.cpp: + + Removed duplicate "ace/Auto_Ptr.h" include. + + * ace/Svc_Conf_Lexer_Guard.cpp: + + Moved Svc_Conf.h include within the ACE_USES_CLASSIC_SVC_CONF + preprocessor block. It isn't needed for the XML-based Service + Configurator. + +Tue Jun 04 16:02:11 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/HttpCharStream.h: + * ACEXML/common/HttpCharStream.cpp: + * ACEXML/common/Filecharstream.h: + * ACEXML/common/Filecharstream.cpp: + + Removed the constructor which accepted a string containing + either a filename or a URL from these files. Problem was it was + relying on calling open to do the work, but if open fails there + was no way to report the error without throwing exceptions. So + don't allow people to use this constructor and always insist on + using open. + + * ACEXML/common/Mem_Map_Stream.cpp: + * ACEXML/common/URL_Addr.h: + * ACEXML/common/URL_Addr.cpp: + * ACEXML/common/XML_Codecs.cpp: + * tests/Codecs_Test.cpp: + + Other miscellaneous fixes including using + ACE_Auto_Basic_Array_Ptr instead of ACE_Auto_Array_Ptr. + + * ACEXML/common/Makefile.bor: + * ACEXML/tests/Makefile.bor: + + Added new files. + + * ACEXML/tests/HttpCharStream_Test.cpp: + + Cosmetic fixes. + +Tue Jun 04 14:18:56 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/tests/HttpCharStream_Test.dsp: + * ACEXML/tests/Tests.dsw: Added a new test project. + +Tue Jun 04 13:45:39 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/common/Mem_Map_Stream.h: Fixed typedefs of template + classes. + + * ACEXML/common/XML_Common.dsp: Added new files to project files. + +Tue Jun 4 12:35:12 2002 Jeff Parsons <parsons@cs.wustl.edu> + + * ace/Message_Queue*: Moved the state_ member and the accessor method + to the non-template base class to get rid of the Win32 build + bustage, since ACE_Message_Queue_NT is not a template class, and does + not inherit from ACE_Message_Queue<>. + +Tue Jun 4 10:03:19 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: Reverted the thr_count() + stuff since we need to reactivate the queue now. + + * ace/Message_Queue*: Changed all the message queue implementations to + use the new ACTIVATED/DEACTIVATED/PULSED implementation of + deactivate()/activate(). + +Tue Jun 4 09:48:34 2002 Jeff Parsons <parsons@cs.wustl.edu> + + * ace/Message_Queue.h: + * ace/Message_Queue_T.h: + + Fixed the signatures of deactivate and deactivate_i to make the + appearance of the default parameter 'int pulse' consistent. + +Tue Jun 4 06:46:27 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: Added the + "1" parameter to the call to deactivate() method to simply + "pulse" the queue. Also, changed the AC_Output_Handler::open() + method to check if thr_count() == 0 when determining if to + activate the service handler. + + * ace/Message_Queue_T.h: Added new pulse parameter to + ACE_Message_Queue::deactivate() so that callers can decide + whether to change the state of the queue to be "deactivated" or + just to pulse waiting threads to wake up and continue their + processing. This simplies one of the examples in C++NPv2. + +Mon Jun 3 16:11:12 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/FileCharStream.cpp: + + Fixed a memory leak. + + * ACEXML/common/HttpCharStream.cpp: + * ACEXML/common/HttpCharStream.cpp: + * ACEXML/common/HttpCharStream.h: + * ACEXML/common/Makefile: + * ACEXML/common/Mem_Map_Stream.cpp: + * ACEXML/common/Mem_Map_Stream.h: + * ACEXML/common/URL_Addr.cpp: + * ACEXML/common/URL_Addr.h: + + Functionality needed by the parser to fetch URI's (DTDs etc) + from the web. This is a simple http_get implementation and + doesn't work for https URLs (which can be fixed easily provided + there is interest). + + * ACEXML/common/Parser.cpp: + + Fixed a typo. + + * ACEXML/tests/HttpCharStream_Test.cpp: + * ACEXML/tests/Makefile: + + New regression test to test fetching of URI's from the web by + the parser. + + +Mon Jun 3 06:21:57 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * examples/C++NPv2/AIO_Client_Logging_Daemon.cpp: Removed + unnecessary 'return;' calls. + +Mon Jun 3 07:23:56 UTC 2002 Don Hinton <dhinton@ieee.org> + + * Framework_Component_DLL.bor: Fixed typo. + +Sun Jun 2 22:52:23 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + [Bug 1208] + * ace/README: Document new ACE_HAS_3_PARAM_WCSTOK macro. + * ace/OS_String.inl (strtok_r): Check ACE_HAS_3_PARAM_WCSTOK macro. + + * ace/config-aix-4.x.h: + * ace/config-linux-common.h: + * ace/config-sunos5.7.h: + Define ACE_HAS_3_PARAM_WCSTOK on these platforms after doing + the appropriate checks. + +Sun Jun 2 15:34:41 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL.cpp: + + Test for null dll_name_ before calling open() in the copy ctor, since + open would flag it as an error. Improved error messages. + + * ace/DLL_Manager.{h,cpp}: + * ace/Framework_Component.cpp: + + Changed ACE_DLL_Manager to use ACE_Unmanaged_Singleton so that it can be + explicitely closed by ACE_Framework_Repository after ACE_Service_Config + is closed. Improved error messages. Thanks to Steve Huston for finding + this. + +Sun Jun 2 09:03:06 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Select_Reactor_Base.cpp: Delete temp buffers + if enqueue_head() fails in the ACE_Select_Reactor_Notify open() + and notify() methods. Thanks to Chris Uzdavinis + <chris@atdesk.com> for reporting this. + +Sat Jun 1 22:42:17 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/AC_CLD.mak: + * examples/C++NPv2/TPCLS.mak: Added "ssl=1" to properly include + the SSL stuff in the build; these examples use SSL. + + * examples/C++NPv2/AIO_Client_Logging_Daemon.cpp: + * examples/C++NPv2/AIO_CLD_export.h: + * examples/C++NPv2/AIO_CLD.mak: Asynch I/O example from Chapter 8. + +Sat Jun 1 15:38:40 2002 Steve Huston <shuston@riverace.com> + + * ace/Asynch_Acceptor.cpp (open): bind_port is in ACE_Sock_Connect, + not ACE. + Added #include "ace/SOCK_Stream.h" that class is also used, and + #include "ace/Log_Msg.h" to get the ACE_ERROR macros. + + * ace/Asynch_Connector.cpp (handle_connect): clr_flags() is in + ACE_Flag_Manip, not ACE. + Added #include "ace/Log_Msg.h" to get ACE_ERROR macros. + +Fri May 31 13:35:41 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/TPC_Logging_Server.{h cpp}: + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: + Primarily formatting changes to match the book. + + * examples/C++NPv2/SLDex.mak: Removed Logging_Acceptor_Ex from + files list; it's only a header file. + +Fri May 31 11:07:25 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: Modified this + so use pointers rather than references to simplify things in + the book. + +Thu May 30 23:03:38 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/SR_Configurable_Logging_Server.{cpp mak}: + * examples/C++NPv2/Makefile: Added new ACE_Select_Reactor-based + configurable logging server. + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: Code changes + to match book and fix reconnect problems. + +Thu May 30 18:02:35 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/OS.h: Changed EACESS to EACCES to be POSIX compliant. Thanks + to Eric Eide <eeide@cs.utah.edu> for (re)reporting this! + +Thu May 30 10:01:12 2002 Ossama Othman <ossama@uci.edu> + + * ace/UPIPE_Stream.h (ACE_UPIPE_Stream): + + Added missing PEER_ADDR trait. + +Thu May 30 16:45:10 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + [Bug 1208] + * ace/OS_String.inl (strstr): Remove wcswcs(), use wcsstr() + instead. + +Thu May 30 9:15:19 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + Thanks to Edan Ayal <edanayal@yahoo.com> and Alex Libman + <AlexL@rumblegroup.com> for the following fixes. + + * ACE_WIN32_Asynch_Read_Stream_Result::complete () + ACE_WIN32_Asynch_Write_Stream_Result::complete () + ACE_WIN32_Asynch_Write_Dgram_Result::complete + ACE_WIN32_Asynch_Read_Dgram_Result::complete improved + readability and logic + + * ACE_WIN32_Asynch_Read_File_Result::complete () + ACE_WIN32_Asynch_Write_File_Result::complete () fixed + errors,improved readability and logic + + * ACE_WIN32_Asynch_Read_Stream::read () + ACE_WIN32_Asynch_Read_File::read () + ACE_POSIX_Asynch_Read_Stream::read () + ACE_POSIX_Asynch_Read_File::read () + Added check for enough space in message block + bytes_to_read <= mb->space() + + * ACE_WIN32_Asynch_Write_Stream::write () + ACE_WIN32_Asynch_Write_File::write () + ACE_POSIX_Asynch_Write_Stream::write () + ACE_POSIX_Asynch_Write_File::write () + Added check for correct length of message block + bytes_to_write <= mb->length() + + * ACE_WIN32_Asynch_Read_Stream::readv () + ACE_WIN32_Asynch_Read_File::readv () + ACE_WIN32_Asynch_Read_Dgram::recv Added checks for: never read + more bytes_to_read bytes; correct using of buffers - message + blocks + + * ACE_POSIX_Asynch_Write_Stream::writev () + ACE_POSIX_Asynch_Write_File::writev () + ACE_WIN32_Asynch_Write_Dgram::send Added checks for: never write + more bytes_to_write bytes; correct using of buffers - message + blocks improved readability and logic + +Thu May 30 08:48:19 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Configuration_Import_Export.cpp (squish): + Replaced (*cp == ' ') || (*cp == '\t') || (*cp == '\n') || (*cp + == '\r')) with isspace (*cp). Thanks to Gonzalo A. Diethelm + <gonzalo.diethelm@aditiva.com> and Dave Steele + <Dave_Steele@Mitel.COM> for this suggestion. + +Thu May 30 11:53:05 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL_Manager.cpp: Fixed Fuzz errors. Thanks to Johnny + Willemsen for pointing this out. + +Thu May 30 11:40:45 UTC 2002 Don Hinton <dhinton@ieee.org> + + * tests/Framework_Component_DLL.bor: + + Fixed cut-n-paste error that caused the symbols not to be + exported. Should fix the Borland builds. + +Thu May 30 10:18:27 UTC 2002 Don Hinton <dhinton@ieee.org> + + * tests/Framework_Component_Test.cpp: + + Modified test to use the new macro definitions and method names. + + * tests/Framework_Component_DLL.cpp: + + Added ACE_DLL_UNLOAD_POLICY declaration specifying LAZY unloading. + Changed an errant ACE_LIB_TEXT macro to ACE_TEXT. + + * ace/config-all.h: + + Added the dll unloading policies mentioned below as well as an + ACE_DLL_UNLOAD_POLICY macro dlls can use to set/export their own + policy. If no policy is exported, the current per-process policy + is used. + + * ace/DLL_Manager.{h,cpp}: + + Renamed unload_strategy to unload_policy and moved the policy + definitions to config-all.h so that they can be used without + including DLL_Manager.h (which includes OS.h, etc...). + + Refactored the unloading policies and implemented the per-dll + case by probing the dll for a "_get_dll_unload_policy" function + that returns the dll specific policy--if the dll chose to + implement it. + + Added missing template instantiations. + + * ace/DLL.{h,cpp} (symbol): + * ace/DLL_Manager.{h,cpp} (symbol): + + Added additional default parameter, ignore_errors, so the dll + can be probed for symbols without sending error messages to + the log. + +Thu May 30 08:53:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/Codecs_Test.cpp: + Removed not needed ; after the ACE_RCSID. This will fix a warning in + the Tru64 build + + * tests/Framework_Component_DLL.cpp: + Corrected playing of ;. This will fix a warning in the Tru64 build. + +Thu May 30 02:27:29 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/AC_Client_Logging_Daemon.cpp: + * examples/C++NPv2/AC_CLD_export.h: + * examples/C++NPv2/AC_CLD.mak: + * examples/C++NPv2/Makefile: Added remaining Chapter 7 example. + +Thu May 30 08:04:12 2002 Johnny Willemsen <jwillemsen@remedy.nl + + * ace/Asynch_IO.h: + * ace/Bound_Ptr.h: + Changed <p> in comment to \<p\> because the <p> is a commend + that is recognized by doxygen. + +Thu May 30 07:46:12 2002 Johnny Willemsen <jwillemsen@remedy.nl + + * ace/Makefile.bor: + Added Codecs + + * tests/Framework_Component_Test.cpp: + Fixed msvc wchar compile errors + + * etc/tao_costrader.doxygen: + Added this doxygen config file to generate TAO Cos Trader + documentation + + * bin/generate_doxygen.pl: + Added tao_costrader to generate TAO Cos Trader documentation + + * html/index.html: + Added link for CosTrader + +Wed May 29 22:33:53 2002 Ossama Othman <ossama@uci.edu> + + * ace/Acceptor.h (ACE_Acceptor, ACE_Oneshot_Acceptor): + * ace/Connector.h (ACE_Connector): + + Use ACE_TYPENAME where appropriate in newly added traits to + correct build problems in g++ 2.95.x or better builds. + +Wed May 29 22:13:19 2002 Ossama Othman <ossama@uci.edu> + + * ace/Acceptor.h (ACE_Acceptor, ACE_Oneshot_Acceptor): + * ace/Connector.h (ACE_Connector): + * ace/Svc_Handler.h (ACE_Svc_Handler): + + Added STL-style traits that may be useful when subclassing these + classes. + +Wed May 29 21:55:49 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/TPC_Logging_Server.{h cpp}: + * examples/C++NPv2/TPCLS_export.h: + * examples/C++NPv2/TPCLS.mak: + * examples/C++NPv2/Makefile: Added Chapter 7 examples. + + * ace/Acceptor.h (ACE_Acceptor::open): Made open() virtual. + +Thu May 30 01:38:51 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/OS_String.inl: Revert: + Wed May 29 09:38:53 2002 Ossama Othman <ossama@uci.edu> + The intended fix is satisfactorily addressed by: + Tue May 28 18:35:46 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + +Wed May 29 16:36:42 2002 Ossama Othman <ossama@uci.edu> + + * ace/CDR_Stream.cpp (write_wchar, read_wchar, skip_wchar): + * ace/CDR_Stream.i (write_wchar_array, read_wchar_array): + + Take into account the actual size of the ACE_CDR::WChar type + instead of assuming that it is always 16 bits long. + +Wed May 29 20:16:48 UTC 2002 Don Hinton <dhinton@ieee.org> + + * tests/Framework_Component_Test.cpp: + + Changed ACE_Service_Config::open() to call the + ignore_default_svc_conf_file flag since all the directives + are passed directly without using a config file anyway. + + * ACEXML/apps/svcconf/Svcconf_Handler.{h,cpp}: + + Removed handle_ member variable and pass the local svc_dll + object instead of handle_ to ACE_Service_Config::create_service_type() + in accordance with changes below to ACE_DLL. + +Wed May 29 12:35:21 2002 Ossama Othman <ossama@uci.edu> + + * ace/OS.h: + + If ACE_HAS_WCHAR is defined define the ACE_OS::WChar type to be + of type wchar_t. Otherwise, define it to be ACE_UINT16 (the + previous default). Addresses some wide character CDR and CORBA + C++ mapping issues. Thanks to David Smith + <dts@prismtechnologies.com> for reporting this. + +Wed May 29 09:46:56 2002 Ossama Othman <ossama@uci.edu> + + * ace/Message_Queue.h: + + Include "ace/Synch.h" to pull in ACE_Thread_Mutex class + declaration. Thanks to Marek Maleta <marek_maleta@agilent.com> + for reporting the problem. + + * THANKS: + + Added Marek to the Hall of Fame. + +Wed May 29 09:38:53 2002 Ossama Othman <ossama@uci.edu> + + * ace/OS_String.inl (strstr): + + Only use wcswcs() if _XOPEN_SOURCE is defined. In any case, the + Single UNIX Specification version 2 strongly encourages use of + wcsstr(). Fixes a compile-time error on glibc 2.x platforms. + + (strtok_r): + + Glibc has the 3 argument version of wcstok(). Use it instead. + Fixes a compile-time error on glibc 2.x platforms. + +Wed May 29 15:52:52 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL_Manager.{h,cpp}: + + Cleaned some problems with auto_ptr usage and added symbol + name to the debug message if the symbol isn't found--not + all systems seem to give you the name of the unfound symbol. + +Wed May 29 15:06:32 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move Debian_NoInterceptors to new + scoreboard. + +Wed May 29 10:29:10 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Client_Logging_Daemon.cpp: Use a safer scheme + to clean up sent blocks; compile the INET_Addr stuff clean. + +Wed May 29 12:54:23 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/DLL_Manager.cpp: + * ace/Framework_Component.cpp: + Fixed msvc wchar compile errors + +Wed May 29 08:41:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Codecs.h: + Removed the ; after the ACE_UNIMPLEMENTED_FUNC, the BCB compiler has + problems with it. + + * tests/Makefile.bor: + Added the new Codecs_Test. + +Tue May 28 19:40:44 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Client_Logging_Daemon.cpp: + * examples/C++NPv2/CLD_export.h: + * examples/C++NPv2/CLD.mak: + * examples/C++NPv2/TP_Logging_Server.{h cpp}: + * examples/C++NPv2/TPLS_export.h: + * examples/C++NPv2/TPLS.mak: + * examples/C++NPv2/Makefile: Added new client logging daemon + and thread pool logging server examples from chapter 6. + +Wed May 29 02:09:32 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move RedHat_7.1_Full and + RedHat_7.1_NO_AMI_Messaging builds to new scoreboard. + +Tue May 28 19:20:58 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/Read_Buffer.cpp (rec_read): + + It seems that my changes trash the files written by test + programs. I need to write a test to check out what's happening + before changing the CVS head. Reverting my change of Tue May 28 + 17:00:21 2002 Krishnakumar B <kitty@cs.wustl.edu>. + +Tue May 28 23:13:45 UTC 2002 Don Hinton <dhinton@ieee.org> + + * tests/Framework_Component_DLL.cpp: + + Added missing ";" to end of template instantiation statements. + +Tue May 28 22:34:03 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL.cpp: Added missing ACE_const_cast(). + +Tue May 28 17:00:21 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/Read_Buffer.cpp (rec_read): + + Fixed stupid thinko when decrementing slot. Tell myself to think + twice to about using ?:. My previous change also ate the + character when only one character is in a file. ACE_Read_Buffer + is crying for a test under $ACE_ROOT/tests which tests variable + sized buffers with sizes of 1, < BUFSIZ, BUFSIZ, > BUFSIZ + instead of under examples/misc/test_tead_buffer.cpp. Will try to + get in one. + + * docs/ACE-guidelines.html: + + Update guidelines to be careful about using ?: in some + situations. Thanks to Doug for suggesting this. + +Tue May 28 21:54:44 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move Debian_NoInline and + Debian_GCC_3.0.4 builds to new scoreboard. + +Tue May 28 21:32:42 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL.cpp: + * ace/DLL_Manager.cpp: + * ace/Framework_Component.cpp: + * ace/Framework_Component_T.cpp: + + Fixed Fuzz errors. + +Tue May 28 18:35:46 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + [Bug 1208] + * ace/OS_String.cpp (strok_r_emulation): Call ACE_OS_String::strtok + instead of ::wcstok. + * ace/OS_String.inl (strtok): If ACE_HAS_XPG4_MULTIBYTE_CHAR is defined, + call 3 parameter version of wcstok, else call 2 parameter version. + * ace/config-linux-common.h: Only define ACE_LACKS_WCSTOK and + ACE_LACKS_WCSDUP_PROTOTYPE for glibc versions less than 2. Define + ACE_HAS_XPG4_MULTIBYTE_CHAR for glibc 2 and higher. + * tests/OS_Test.cpp: Add test for wchar_t version of strtok. + +Tue May 28 15:02:01 UTC 2002 Don Hinton <dhinton@ieee.org> + + * tests/tests.dsw: + * tests/Framework_Component_DLL.dsp: Added new dsp. + +Tue May 28 09:35:17 2002 Jeff Parsons <parsons@cs.wustl.edu> + + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: + + Added DLL_Manager.* to these projects. + +Tue May 28 10:47:50 UTC 2002 Don Hinton <dhinton@ieee.org> + + * tests/Logging_Strategy_Test.cpp: + + Fixed typo, ACE_AS_* > ACE_HAS_*. + + * tests/Makefile.Framework_Component_DLL: + * tests/Makefile: + * tests/Makefile.bor: + + Added/modified Makefiles to add new test file, and updated + dependancies. + + * tests/Framework_Component_Test.{h,cpp}: + * tests/Framework_Component_DLL.{h,cpp}: Added. + * tests/Framework_Component_DLL_export.h: Added. + + Modified to test to exercise the new ACE_DLL_Singleton and + ACE_DLL* classed in concert with the Service Config framework + and new Framework Component mechanisms. + + * tests/DLL_Test.{h,cpp}: + + Modified test to fully test all the features of the new + ACE_DLL, ACE_DLL_Handle, and ACE_DLL_Manager classes. + + * ace/Makefile: + * ace/Makefile.bor: + + Added DLL_Manager and updated dependancies. + + * bin/generate_export_file.pl: + + Added generation of library specific TRACE macros in + terms of the new ACE_TRACE_IMPL macro below. + + * ace/Global_Macros.h: + + Added new macro, ACE_TRACE_IMPL and implemented ACE_TRACE + in terms of it if tracing is enabled. This allows you + to turn tracing on for a particular cpp by just undef'ing + ACE_TRACE and redefining it as ACE_TRACE_IMPL. + + * ace/Parse_Node.{h,cpp}: + + Removed handle() method and added a dll() method. Also, + since ACE_DLL_Handle::symbol() calls ACE::ldname() internally, + changed the ACE::ldname() calls in the ctors for + ACE_Object_node and ACE_Function_Node to ACE::strnew(). + + * ace/Svc_Conf_y.cpp: + + Changed call to ACE_Location_Node::handle() to ACE_Location_Node::dll() + in ACE_Service_Type::ctor. + + * ace/Service_Object.{h,cpp,i}: + + Modified ACE_Service_Type to maintain an ACE_DLL member variable + instead of an ACE_SHLIB_HANDLE, and changed ctor accordingly. + Removed call to ACE_OS::dlclose() from dtor. Removed handle() + methods. + + * ace/Object_Manager.cpp: + * ace/Service_Config.cpp: + + The Object Manager is now always responsible for cleaning up + the ACE_Framework_Repository, but only after the Service + Config framework has been cleanup. ACE_Service_Config no longer + knows anything about ACE_Framework_Repository. + + * ace/Service_Config.cpp: + + Changed create_service_type() to take an ACE_DLL& instead of + an ACE_SHLIB_HANDLE. Removed unneeded call to ACE::ldname(). + + * ace/DLL_Manager.{h,cpp}: Added. + * ace/DLL.{h,cpp}: + + Added new classes, ACE_DLL_Manager_Ex and ACE_DLL_Handle, and + and moved all functionalility from ACE_DLL to the new classes. + ACE_DLL maintains the same interface and semantics but defers + all calls to the new classes. ACE_DLL_Manager is a singleton + that manages the lifetimes of instances of refcounted + ACE_DLL_Handle objects that represent the actual dlls. The + following unloading strategies are available: per-process/dll, + eager/lazy. + + Added copy ctor to ACE_DLL so that the get/set_handle() methods + aren't needed--there use in the Service Config framework can + complicate orderly destruction of singletons based on + ACE_DLL_Singleton. + + Added call to ACE::ldname() to the symbol() method so that + the caller no longer needs to do it themselves. + + * ace/Singleton.{h,cpp,i}: + + Added new singleton template, ACE_DLL_Singleton, that registers + itself with the ACE_Framework_Repository so that it can be + destroyed when the dll with which it is associated is unloaded. + + + * ace/Proactor.{h,cpp}: + * ace/Reactor.{h,cpp}: + + Added name() method as needed by changes below, and implemented + dll_name() method to return "ACE". + + * ace/Framework_Component.{h,cpp,inl}: + * ace/Framework_Component_T.{h,cpp}: + + Enhanced the Framework Component "framework" ;-) to allow the + removal of registered components, either individually by name or + as a group, based on the dll that registered them. + +Mon May 27 15:59:17 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/common/XML_Codecs.h: + * ACEXML/common/XML_Codecs.cpp: Fixed VC warnings. + + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: Added Codecs.*. + + * tests/tests.dsw: + * tests/Codecs_Test.dsp: Added a new test. + + * ACEXML/common/XML_Common.dsp: Added XML_Codecs.*. + +Mon May 27 15:47:44 2002 Steve Huston <shuston@riverace.com> + + * ace/OS.h: Add new service-related macro, ACE_Local_Service_Export. + This is intended for use in the ACE_FACTORY_DEFINE macro, for + static services whose factory functions need not be exported. + To use this when defining a factory function, do + ACE_FACTORY_DEFINE(ACE_Local_Service, StaticSvcClass) + + * examples/C++NPv2/Service_Reporter.cpp: Use ACE_Local_Service + in the ACE_FACTORY_DEFINE for the Service_Reporter service. + +Mon May 27 15:21:57 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/Codecs.h: + * ace/Codecs.cpp: + + New files meant to hold all encoding/decoding classes/functions. + Currently implements Base 64 encoding. + + * ACEXML/common/XML_Codecs.h: + * ACEXML/common/XML_Codecs.cpp: + * ACEXML/common/Makeile: + + Same guy under a different hood. + + * ace/Basic_Types.h: + + Added definition of ACE_Byte. + + * tests/Codecs_Test.cpp: + * tests/run_test.lst: + + Regression test which tests the encoding/decoding functionality + added above. + + * tests/Makefile: + * tests/RMCast/Makefile: + * ace/Makefile: + * ace/RMCast/Makefile: + + Updated dependencies. + + * ace/Read_Buffer.cpp: + + Decrement the extra increment of slot. This is not completely + correct, as we should abstract the handling of end-of-line across + Unix and Windows and then use it here and all the places which + read stuff from files. See Fri May 24 10:42:25 2002 Douglas C. + Schmidt <schmidt@tango.doc.wustl.edu> for another manifestation + of the same bug. + + * include/makeinclude/rules.local.GNU: + + Remove files named core.[0-9]+. Newer kernels dump a core file + associated with each process separately. + +Mon May 27 15:04:31 2002 Steve Huston <shuston@riverace.com> + + * ace/OS.h (ACE_FACTORY_DEFINE): Add the export declaration to + the generated factory, removing the need to also use the + ACE_FACTORY_DECLARE in simple (most common) situations. + + * examples/C++NPv2/SLD.cpp: + * examples/C++NPv2/SLDex.cpp: + * examples/C++NPv2/Server_Shutdown.cpp: Remove ACE_FACTORY_DECLARE. + It's not needed in these situations. + +Mon May 27 07:39:41 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Configuration_Import_Export.cpp (squish): Added a + check for '\r' to ensure that empty lines are handled properly. + Thanks to Vladimir Chovanec <Vladimir.CHOVANEC@asset.sk> for + reporting this. + +Mon May 27 09:20:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * bin/generate_doxygen.pl: + Added tao_costime to the list of doxygen files that must be run + + * etc/tao_costime.doxygen: + Created doxygen config file for this TAO service + + * html/index.html: + Added link for the TAO CosTime service doxygen documentation + +Mon May 27 08:18:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ACEXML/common/Attributes_Def_Builder.h: + * ACEXML/common/DTD_Manager.h: + * ACEXML/common/Element_Def_Builder.h: + * ACEXML/common/Validator.h: + * ACEXML/parser/debug_validator/Debug_Attributes_Builder.h: + * ACEXML/parser/debug_validator/Debug_Element_Builder.h: + * ACEXML/parser/debug_validator/Element_Tree.h: + Corrected doxygen commands, no space between @ and the command is + allowed. + + * ace/Active_Map_Manager.h: + * ace/Arg_Shifter.h: + * ace/Based_Pointer_Repository.cpp: + * ace/CDR_Stream.h: + * ace/Caching_Utility_T.h: + * ace/Configuration.h: + * ace/DEV_IO.h: + * ace/Dev_Poll_Reactor.h: + * ace/Dirent_Selector.h: + * ace/Env_Value_T.h: + * ace/FILE.h: + * ace/Get_Opt.h: + * ace/Hashable.h: + * ace/IO_Cntl_Msg.cpp: + * ace/Init_ACE.h: + * ace/Local_Tokens.h: + * ace/Log_Msg.cpp: + * ace/Log_Msg.h: + * ace/Memory_Pool.h: + * ace/Message_Block.h: + * ace/OS.cpp: + * ace/OS.h: + * ace/Object_Manager.cpp: + * ace/POSIX_Asynch_IO.cpp: + * ace/POSIX_Asynch_IO.h: + * ace/POSIX_Proactor.cpp: + * ace/POSIX_Proactor.h: + * ace/Pair_T.h: + * ace/Proactor.cpp: + * ace/Process.h: + * ace/Reactor_Impl.h: + * ace/Remote_Tokens.h: + * ace/SV_Message.h: + * ace/Select_Reactor_T.h: + * ace/Svc_Conf.h: + * ace/System_Time.h: + * ace/TLI_Stream.h: + * ace/TP_Reactor.cpp: + * ace/TP_Reactor.h: + * ace/Task.h: + * ace/Task_T.h: + * ace/Thread.h: + * ace/Timer_Wheel_T.cpp: + * ace/Token.h: + * ace/Token_Manager.h: + * ace/WIN32_Proactor.cpp: + Improved the doxygen documenation by converting old style documentation + to doxygen style. Also corrected small typing errors, start sentences with + capital letters and corrected incorrected doxygen commands. + + * ace/POSIX_Proactor.cpp: + Corrected the signature of handle_events from unsigned long to u_long to + match the declaration of this method. This was found using the warning file + of the doxygen generation. + + * ace/Future.h: + Corrected brief documentation. Brief documentation ends when a dot is found + followed by a space of a newline. We had "i.e. " in the brief documenation + and now this documentation ended after the second dot, changed it to + "i.e.,". This fixed [Bug 892]. + + * ace/config-doxygen.h: + Added ACE_HAS_TOKENS_LIBRARY to generate ACE token library documentation + + * bin/generate_doxygen.pl: + Added tao_cosnaming to the list of doxygen files that must be run + + * etc/tao_cosnaming.doxygen: + Created doxygen config file for this TAO service + + * etc/tao_cosnotification.doxygen: + Corrected tagfile name and directory that is generated + + * html/index.html: + Added link for the TAO Naming service doxygen documentation + +Sun May 26 21:48:49 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Reactor_Logging_Server_Adapter.cpp: Added + include guard. + + * examples/C++NPv2/SLD.dsp: Define SLD_BUILD_DLL, not ACE_BUILD_DLL, + to set up the export stuff correctly. + + * examples/C++NPv2/SLDex.dsp: + * examples/C++NPV2/examples.dsw: Add new project for second part + of chap 5 example. + + * examples/C++NPv2/SLD.cpp: + * examples/C++NPv2/SLDex.cpp: + * examples/C++NPv2/Server_Shutdown.cpp: Add ACE_FACTORY_DECLARE to + get correct export declaration for factory. + +Sun May 26 13:20:15 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/config-macosx.h: Added #define ACE_LACKS_SYSTIME_H. + Thanks to Dennis C. De Mars <demars@mminternet.com> for + recommending this. + +Sat May 25 17:45:21 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * tests/Proactor_Scatter_Gather_Test.dsp: + * tests/tests.dsw: Added Proactor_Scatter_Gather_Test. + +Sat May 25 16:42:46 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * tests/run_test.lst, + * tests/Makefile.bor, + * tests/Makefile: Added Proactor_Scatter_Gather_Test. + +Sat May 25 16:30:00 2002 Edan Ayal <edanayal@yahoo.com> + + * tests/Proactor_Scatter_Gather_Test.cpp: + Added a new test for the asynch scatter/gather I/O + functionality. It is currently supported (as the feature itself) + only under Win32 - actually NT4 SP2 and above. The test runs in + a single thread, and involves a single Sender, two Receivers and + a single Writer. The Sender async-reads (scattered) from a file + into chunks of <page size>. It async-sends (gathered) the odd + chunks to the first receiver over a stream, and the even chunks + to the second receiver over a different stream. The receivers + async-read (scattered) from the socket streams into chunks in + size of <page size>, and convey the data to the Writer. The + Writer reconstructs the file using async-write + (gathered). Finally, the reconstructed file is compared to the + original file to determine test success. The test therefore + covers both async scatter/gather stream I/O and async + scatter/gather file I/O. + +Sat May 25 14:52:06 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Strategies_T.cpp (make_svc_handler): Fixed a typo where + dll_name_ was misspelled as shared_library_. Thanks to Nathan + Krasney <natan-k@actcom.co.il> for reporting this. + +Fri May 24 10:42:25 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Configuration_Import_Export.cpp (squish): Added a + check for '\n' to ensure that empty lines are handled properly. + Thanks to Vladimir Chovanec <Vladimir.CHOVANEC@asset.sk> for + reporting this. + +Sat May 25 15:09:42 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/SLDex.{cpp mak}: + * examples/C++NPv2/Server_Shutdown.cpp: + * examples/C++NPv2/SLDEX_export.h: + * examples/C++NPv2/Makefile: Add new dll for second part of + Chapter 5 dynamic (re)config example. + +Sat May 25 11:47:33 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Reactor_Logging_Server_Adapter.cpp (init): + Add wide->narrow char conversion to get ACE_TCHAR into + a char-based method. + +Sat May 25 01:15:02 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move Debian_Minimum bulid to + new scoreboard. + +Fri May 24 23:52:39 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Configurable_Logging_Server.dsp: + * examples/C++NPv2/SLD.dsp: + * examples/C++NPv2/examples.dsw: Added new projects for chapter + 5's configurable logging server. + +Fri May 24 20:51:25 2002 Christopher Kohlhoff <chris@kohlhoff.com> + + * ace/Log_Msg_NT_Event_Log.cpp (open): + Need to use logger_key, when it is supplied, to register the + event source with the system. + + * ace/Log_Msg_NT_Event_Log.cpp (log): + Replace all newlines in the message with the CR-LF character + sequence so that line breaks are correctly displayed in the + event viewer. + +Fri May 24 20:18:41 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Reactor_Logging_Server.cpp: + * examples/C++NPv2/Reactor_Logging_Server_T.cpp: Move all the + method implementations for Reactor_Logging_Server<> into + Reactor_Logging_Server_T.cpp so they're reuseable in a + shared library. + + * examples/C++NPv2/Reactor_Logging_Server.h: Add template + implementation reference plumbing for compilers. + + * examples/C++NPv2/Select_Reactor_Logging_Server.cpp: + * examples/C++NPv2/TP_Reactor_Logging_Server.cpp: + * examples/C++NPv2/WFMO_Reactor_Logging_Server.cpp: Remove the + (duplicate) definitions of Reactor_Logging_Server<>; the + compiler now generates this using the code in + Reactor_Logging_Server_T.cpp. + + * examples/C++NPv2/svc.conf: + * examples/C++NPv2/SLD.cpp: + * examples/C++NPv2/SLD_export.h: + * examples/C++NPv2/SLD.mak: Code and Makefile for the dynamic + logging service in chapter 5. + + * examples/C++NPv2/Makefile: Added SLD.mak. + + * examples/C++NPv2/Logging_Acceptor.h: Added a PEER_ADDR trait. + The Reactor_Logging_Server_Adapter's use of Logging_Acceptor + needs the trait. + + * examples/C++NPv2/Service_Reporter.h: Make ctor public so + the service factory can instantiate this class. + +Sat May 25 00:20:50 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move Debian_WChar to new scoreboard. + +Fri May 24 10:18:39 2002 Steve Huston <shuston@riverace.com> + + * bin/auto_compile: Don't bail out if the only CVS conflict is + in ChangeLog or THANKS. + + * ace/CDR_Stream.h (ACE_InputCDR): Added comments clarifying + the behavior of the ctor that takes a const ACE_Message_Block *. + Thanks to Alexander Maack for motivating this. + + * examples/C++NPv2/display_logfile.cpp (svc): Fix memory leak + (thanks to Alexander Maack for reporting this) and don't lose + the host name in the data if demarshaling has to be aborted to + go read more data from the file. + + * examples/C++NPv2/Logging_Acceptor_Ex.h: + * examples/C++NPv2/Logging_Event_Handler_Ex.h: Remove trailing ';' + on inlined methods; Changed to match book. + + * examples/C++NPv2/Select_Reactor_Logging_Server.cpp: Changed to + match book. + + * examples/C++NPv2/Reactor_Logging_Server_Adapter.{h cpp}: + * examples/C++NPv2/Service_Reporter.{h cpp}: + * examples/C++NPv2/Configurable_Logging_Server.{cpp mak}: + Example code from Chapter 5 of C++NPv2 (Service Configurator). + + * examples/C++NPv2/Makefile: Added new examples. + + * THANKS: Added Alexander Maack to the Hall of Fame. + + * ace/Reactor.h: Mark run[_alertable]_event_loop, end_event_loop, + event_loop_done, and reset_event_loop static methods deprecated + in favor of calling the non-static methods on + ACE_Reactor::instance(). Thanks to Don Hinton for motivating this. + + * ace/String_Base.h (rep): Added note that directs the caller to free + the returned memory. + +Mon May 13 19:15:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Service_Config.h: Doxygen-ized comments for open(), and + added a note on -b that files are evaluated relative to "/". + + * netsvcs/servers/main.cpp: Move the reactor/signal initialization + to be after the service loading. If the user specifies -b (be a + daemon) it closes all handles, including the reactor + notification pipe. Also added ACE_TEXT decorator to the string + literals. + +Fri May 24 13:29:05 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move Debian_Full and + Debian_Minimum_Static builds to new scoreboard. + +Thu May 23 20:35:22 2002 Steve Huston <shuston@riverace.com> + + * ace/RB_Tree.h: Clarified that iterator is valid after changes to + container only if the iterated-over element is not removed. Thanks + to Craig Ching <cching@mqsoftware.com> for noticing this. + + * ace/Module.cpp (close): Fixed to only use the passed-in flags if + the task delete policy was not set explicitly. Thanks to Doug + Schmidt for this fix! + + * examples/C++NPv2/WFMO_Reactor_Logging_Server.cpp: Use new technique + for detecting "quit" command. Unified with book content. + +Thu May 23 20:36:46 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move Lynx_PPC build to new scoreboard. + +Thu May 23 14:04:52 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/common/LocatorImpl.cpp: + * ACEXML/common/LocatorImpl.h: + * ACEXML/common/LocatorImpl.i: + * ACEXML/parser/parser/Parser.cpp: + * ACEXML/parser/parser/Parser.h: + * ACEXML/parser/parser/Parser.i: + + Implemented the Locator for the XML parser. + + * ACEXML/common/CharStream.h: + * ACEXML/common/DTDHandler.h: + * ACEXML/common/Env.h: + * ACEXML/common/Exception.h: + * ACEXML/common/FileCharStream.h: + * ACEXML/common/InputSource.cpp: + * ACEXML/common/SAXExceptions.cpp: + * ACEXML/common/SAXExceptions.h: + * ACEXML/examples/SAXPrint/SAXPrint_Handler.cpp: + * ACEXML/examples/SAXPrint/SAXPrint_Handler.h: + + Cosmetic fixes including some fixing some memory leaks. + + * ACEXML/examples/SAXPrint/Print_Handler.cpp: + * ACEXML/examples/SAXPrint/Print_Handler.h: + * ACEXML/examples/SAXPrint/main.cpp: + + Add support for printing out errors which uses the Locator. This + will print out errors only if there is an error in the XML file. + + +Thu May 23 07:53:37 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * examples/Misc/test_read_buffer.cpp (ACE_TMAIN): Use ACE_STDIN + rather than 0. Thanks to Soeren Gerlach <soeren.gerlach@gmx.de> + for motivating this. + +Tue May 21 06:13:50 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Hash_Map_Manager_T.h (class ACE_Hash_Map_Manager_Ex): Made + shared_find() protected rather than private. Thanks to + Ganesh Pai <GPai@sonusnet.com> for reporting this. + +Wed May 22 23:18:26 2002 Priyanka <oci@cs.wustl.edu> + + * ACE version 5.2.3 released. + +Sat May 18 09:32:58 2002 Ossama Othman <ossama@uci.edu> + + * ace/SSL/SSL_SOCK_Stream.cpp (get_remote_addr): + + Set errno to ENOTCONN if the SSL connection hasn't been + completed yet, and errno to EBADF if the file descriptor hasn't + been set yet. This emulates the getpeername() errno semantics. + +Sat May 18 08:19:48 2002 Ossama Othman <ossama@uci.edu> + + * ace/SSL/SSL_Connector.h: + * ace/SSL/SSL_Connector.cpp: + + Temporarily removed these files since there are still some + thread-safefy issues to be addressed in the implementation. + Better to wait until after the forthcoming beta to ship them. + +Fri May 17 07:32:49 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Log_Msg.cpp: Replaced ACE_NEW_RETURN_I with ACE_NEW_RETURN. + Thanks to Johnny Willemsen for reporting this. + +Thu May 16 20:15:36 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/OS.i: Added ACE_HAS_ISASTREAM_PROTO guard around the + extern "C" int isastream() method. Thanks to Oleg Terletsky + <oterletsky@comcast.net> for reporting this. + + * ace/ARGV.h: Added a comment that points out the "add()" method + doesn't handle memory management. Thanks to Kyle Brost + <kyle.brost@quest.com> and Don Hinton for clarifying this. + +Thu May 16 17:05:19 2002 Ossama Othman <ossama@uci.edu> + + * ace/SSL/SSL_Connector.cpp (handle_input, handle_output): + + Qualify the AST type with its appropriate scope. g++ requires + it. + +Thu May 16 15:28:24 2002 Ossama Othman <ossama@uci.edu> + + * ace/SSL/SSL_Connector.h: + * ace/SSL/SSL_Connector.cpp: + + New ACE_SSL-specific subclasses of the standard + ACE_{Strategy_}Connector classes. The standard ACE_Connector + class assumes that a connection is established once the stream + handle becomes writable. However, that only holds for protocols + with a single handshake, such as TCP. Protocols with multiple + handshakes, such as SSL, require special handling to prevent the + connection from being considered completed once the stream + handle becomes writable. The subclasses simply override the + handle_input() and handle_output() methods for each of base + classes that were subclassed. If possible, these changes will + be generalized and merged into ACE proper. When that occurs, + these classes will be deprecated. As such, these classes are + only intended to be an interim solution. + + * ace/SSL/SSL_SOCK_Connector.cpp (connect): + + Fixed bug where the handle in the SSL_SOCK_Stream was not set if + a connection is in progress. In that case, the handle must also + be set in the SSL_SOCK_Stream so that the correct handle is + returned from SSL_SOCK_Stream::get_handle() when performing + non-blocking connect()s via the SSL_SOCK_Connector. + + (complete): + + No need to set the handle in SSL_SOCK_Stream since it is + already set in the connect() method. + + Only attempt to complete the TCP connection if that hasn't + already been done. + + * ace/SSL/SSL_SOCK_Stream.h (peer): + + Made this method public so that it can be called by the new + SSL_{Strategy_}Connector classes. + + * ace/SSL/SSL_SOCK_Stream.cpp (get_remote_addr): + + Overrode the base class implementations of this method. Some + applications use get_remote_addr() as a way of determining + whether or not a connection has been established. In SSL's + case, the remote addr will be available once the TCP handshake + has been complete. Despite that fact, the SSL connection may + not have been completed. In such a case, a successful return + from get_remote_addr() would be misleading. + +Thu May 16 13:34:19 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/examples/SAXPrint/main.cpp: Added constness to + <test_string>. This fixed a GCC 3.0 warning. + +Wed May 15 17:47:24 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL.cpp: Fixed compile error in an AIX ifdef'd + section of code I missed before. + +Wed May 15 17:18:58 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL.cpp (open): + + Removed unreferenced return value. Thanks to Nanbor + for noticing this. + +Wed May 15 16:44:27 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/DLL.cpp (save_last_error): + + Use ACE::strnew to make a copy of the saved string. + +Wed May 15 15:58:07 UTC 2002 Don Hinton <dhinton@ieee.org> + + * tests/DLL_Test.cpp: + + Modified test to make calles to ACE_DLL::error () in order to test + changes below. + + * ace/DLL.{h,cpp}: + + Removed the return in open() after an ACE_Lib_Find error, so that + dlopen() will be called and the appropriate error stored. + + Added new static flag, open_called_, to keep track of whether or not + a library function (specifically dlopen) has been called. This is + used internally to make sure that dlerror() isn't called unless the + library has been initialized with a call to dlopen first. Otherwise + it seg-faults on Linux (and perhaps others). + + Also added a last_error_ variable and save_last_error() method and + placed a call to save_last_error() after each library call. Then + changed error() so it just returns last_erro_. This solves a few + problems, since open() made calls to error() which returned the + appropriate error but cleared the value for the next call. So for + instance, when the ctor failed and tried to call error() to report + the error, it didn't return anything since open() had already called + it. Also, if open() failed, and the called tried to call error() to + find out why, they got a blank error message since error() had + already been called by open(). + +Tue May 14 19:40:11 2002 Priyanka Gontla <pgontla@ece.uci.edu> + + * tests/Proactor_Test.cpp (initiate_write_stream): + Applied the Patch Johnny Willemsen sent to fix the test for the + unicode builds. + +Tue May 14 19:14:11 2002 Priyanka Gontla <pgontla@ece.uci.edu> + + * tests/TP_Reactor_Test.cpp (Connector): + Fixed the warnings by removing the check condition + 'size_t index_ >= 0' which is always going to be true. + + * tests/Proactor_Test.cpp: + Integrated the changes sent by Alex Libman + <AlexL@rumblegroup.com>to fix the test on + linux and other platforms that has a weak AIO implementation. + +Mon May 13 16:43:34 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/parser/parser/Parser.cpp (parse_xml_prolog): + + I meant to change "standalong" to "standalone" not "encoding" to + "encodine". Brain damage... + +Mon May 13 16:31:33 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/parser/parser/Parser.cpp (parse_xml_prolog): + + Fixed a small typo. We should parse "standalone" not + "standalong". + +Mon May 13 10:19:48 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/Message_Block.cpp: Fixed a fuzz warning. + +Mon May 13 08:36:38 2002 Alex Libman <AlexL@rumblegroup.com> + + * tests/Proactor_Test.cpp: + + Improved flow-control in full-duplex mode by increasing the + window size, total_send - total_recv to achieve max performance + from TCP/IP + + Improved static info which includes the number of bytes + sent/recv and the number of read/write operations. This allows + to see distribution of I/O operations between handlers. Would be + useful for Proactor's future improvements. + + Impoved scatter/gather branch under Win32. Instead of splitting + one message into 3 small blocks we now send them as a + chain. This helps to improve performance. + +Mon May 13 07:33:43 2002 Chad Elliott <elliott_c@ociweb.com> + + * ACEXML/examples/SAXPrint/main.cpp: + + Fix the Debian build by using ACEXML_Char instead of ACE_TCHAR. + +Sun May 12 23:20:06 2002 Alex Libman <AlexL@rumblegroup.com> + + * tests/TP_Reactor_Test.{h,cpp}: + Improved flow-control in full-duplex mode by increasing the + window size, total_send - total_recv to achieve max performance + from TCP/IP + + Improved static info which includes the number of bytes + sent/recv and the number of read/write operations. This allows + to see distribution of I/O operations between handlers. Would be + useful for TP_Reactor's future improvements. + +Sun May 12 10:11:07 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Message_Queue_T.cpp: Updated all the enqueue*() and + dequeue*() methods so that they use the new + total_size_and_length() method, which is more efficient. + + * ace/Message_Block.h (ACE_Message_Block): Added new + total_size_and_length() method to speed up calculation of these + values by putting them into one loop. + +Sun May 12 11:12:00 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * Makefile: Moved ACEXML directory up so it would be compile right + after ACE. We will need them when XML-based Service + Configurator becomes the default. Thank to Kitty for pointing + this out. + +Sat May 11 13:04:49 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Synch.h: Updated the documentation explaining + what an ACE_Barrier does. Thanks to Timothy Culp + <timothy.culp@computer.org> for motivating this. + +Sun May 12 09:45:37 2002 Edan Ayal <edanayal@yahoo.com> + + * ace/WIN32_Asynch_IO.cpp: Fixed check for ACE_IOV_MAX + limitation in the following methods: + - ACE_WIN32_Asynch_Read_Dgram::recv + - ACE_WIN32_Asynch_Write_Dgram::send + - ACE_WIN32_Asynch_Read_Stream::readv + - ACE_WIN32_Asynch_Write_Stream::writev + - ACE_WIN32_Asynch_Read_File::readv + - ACE_WIN32_Asynch_Write_File::writev. + +Fri May 10 18:29:48 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/config-icc-common.h (ACE_HAS_ANSI_CASTS): + * include/makeinclude/platform_linux_icc.GNU: + + Fixed the config file to ensure compilation with Intel C++ + Compiler version 6.0. + +Fri May 10 18:16:50 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * ace/Connector.cpp (handle_output): Removed unnecessary debug + statement. + +Thu May 9 20:02:10 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * docs/ACE-bug-process.html: Fixed a typo in the hostname of + the TAO FAQ. Thanks to Jeff Adams <Jeff.Adams@gd-ais.com> + for reporting this. + +Thu May 9 16:26:48 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/examples/SAXPrint/main.cpp: Added missing include file. + + * ace/Service_Config.cpp: + * ace/Service_Config.h: + * ace/Svc_Conf.y: + * ace/Svc_Conf_y.cpp: Consolidated ace_create_service_type into + ACE_Service_Config::create_service_type_impl. + + +Thu May 9 14:35:24 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/XML_Svc_Conf.h: Changed to include ACE_export.h instead of + config-all.h and added #pragma once macros. + +Thu May 09 12:12:59 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/apps/svcconf/Svcconf.h: + * ACEXML/apps/svcconf/Svcconf.cpp: + * ACEXML/apps/svcconf/Svcconf_Handler.h: + * ACEXML/apps/svcconf/Svcconf_Handler.cpp: Commented out the code + if XML-based Service Configurator is not used. + + * ACEXML/apps/svcconf/Makefile: + * ACEXML/apps/svcconf/XML_Svc_Conf_Parser.dsp: Removed ACEXML from + the list of include path to be consistent with other projects. + +Thu May 09 10:53:51 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/common/XML_Common.dsp: One more mis-merged project file. + +Thu May 09 10:39:17 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: Fixed link error on win32. I forgot to merge + these project files differently (they can not be merged.) + +Thu May 9 01:07:56 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/common/StrCharStream.h: + * ACEXML/common/StrCharStream.cpp: Fixed Linux compilation + errors. + +Thu May 09 00:38:22 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * ace/WFMO_Reactor.cpp (event_handling): Added a fix for bug + 1161. The problem was described aptly and the fix was applied + exactly as suggested by Lyn Headley <headley@wolve.com>: + + Problem: "The do-while loop in this function is problematic, + because active threads is incremented once prior to executing + the loop, and then decremented on every iteration of the loop + inside safe dispatch(). Thus on the second iteration of the + loop active threads is decremented to 2^32, which causes the + thread to block on waiting to change state inside update + state(), which is never signaled (because there are no other + active threads)." + + Solution: "Moving the do clause of the loop up a few statements + to include the increment of active threads (and the acquisition + of the lock) would probably fix the larger bug." + + Also, thanks to Lyn Headley <headley@wolve.com> and Paxton Mason + <pmason@wolve.com> at Wolverine, and Kobi Cohen Arazi + <kobi@mivzak.com>, Mike Winter (mwinter@sonic.net), + <sturtesm@hotmail.com> who have all suggested fimilar problems + and fixes. + +Thu May 09 00:06:47 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * examples/Reactor/WFMO_Reactor: Added the "Static Debug", "Static + Release", and "Release" build configurations to the "Debug" + configuration for all the 16 dsp files in this directory. + +Wed May 8 23:57:32 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuilds/builds.lst: Move Redhat 7.1 Core build to new scoreboard. + +Wed May 8 15:15:41 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + Merged in the XML-based Service Configurator framework. The + XML-based Service Configruator is *not* enabled by default and + the merge should not affect existing applications. We will make + the XML-based Service Configurator the defualt one at some point + in the future. So you might want to try it out before that + time. :) To switch current ACE to use XML Service Configurtor, + Add + + #define ACE_HAS_XML_SVC_CONF + + in your <config.h>. You can use the + $(ACE_ROOT)/bin/svcconf_convert.pl to convert your existing + svc.conf files to svc.conf.xml. + + * ace/OS.h: + * ace/Parse_Node.cpp: + * ace/Parse_Node.h: + * ace/Service_Config.cpp: + * ace/Service_Config.h: + * ace/Svc_Conf.h: + * ace/Svc_Conf_Lexer_Guard.cpp: + * ace/Svc_Conf_l.cpp: + * ace/Svc_Conf_y.cpp: + * tests/Service_Config_DLL.cpp: Simplified the macro check. + + + * ace/Service_Config.h: + * ACEXML/apps/svcconf/Svcconf.cpp: Fixed Linux compilation warnings. + + * ACEXML/apps/svcconf/Svcconf_Handler.cpp: + * ace/Service_Config.cpp: Added more comprehensive error messages. + + * ACEXML/Makefile: + * ACEXML/Makefile.bor: + * ACEXML/apps/Makefile: + * ACEXML/apps/Makefile.bor: + * ACEXML/apps/svcconf/Makefile: + * ACEXML/apps/svcconf/Makefile.bor: + * ACEXML/parser/parser/Makefile: + * ace/Makefile: + * ace/Makefile.am: + * ace/Makefile.bor: + * ace/ace.icc: + * examples/ASX/CCM_App/Makefile: + * examples/ASX/CCM_App/Makefile.CCM_App: Added or Updated Makefiles. + + + * bin/svcconf-convert.pl: Added more debugging info printout. + Fixed incorrect mapping on initializer path/path and + state/params. + + * ace/Service_Object.h: Added enum definition to differentiate + ACE_Service_Type_Impl. The enum values are taken from + "ace/Svc_Conf_Toekns.h" so that the same service can work with + both classic and XML service configurator frameworks. + + + * ace/Service_Config.h: + * ace/Service_Config.cpp: Added <create_service_type> and + <create_service_type_impl> methods to ensure objects are + allocated in consistent DLL. + + * ACEXML/parser/parser/Parser.cpp: Made the exception messages + more comprehensive so it's easier to determine the origins of + parse errors. + + * ACEXML/apps/svcconf/Svcconf_Handler.h: + * ACEXML/apps/svcconf/Svcconf_Handler.i: + * ACEXML/apps/svcconf/Svcconf_Handler.cpp: Added more implementation. + + * examples/ASX/CCM_App/svc.conf.xml: + * netsvcs/servers/svc.conf.xml: + * tests/Service_Config_Test.conf.xml: Added new XML-based svc.conf files. + + * tests/Service_Config_DLL.cpp: Added xml-based inline svc.conf directives. + + + * ACEXML/apps/XML_Apps.dsw: Added new workspace for XML + applications. This should contains all projects under this + subdirectory. + + * ACEXML/apps/svcconf/README: + * ACEXML/apps/svcconf/Svcconf.cpp: + * ACEXML/apps/svcconf/Svcconf.h: + * ACEXML/apps/svcconf/Svcconf_Handler.cpp: + * ACEXML/apps/svcconf/Svcconf_Handler.h: + * ACEXML/apps/svcconf/Svcconf_Handler.i: + * ACEXML/apps/svcconf/XML_Svc_Conf_Parser.dsp: Added new files + that implement the parser for XML-based Service Configurator. + + * bin/svcconf-convert.pl: Updated the help message. + + * ace/config-all.h: Added ACE_DEFAULT_SVC_CONF_EXT to define the + default file extention name for svc.conf files. It is ".conf" + when using classic Service Configurator and ".conf.xml" when + using XML Service Configurator. + + * tests/Service_Config_Test.cpp: Changed to use different file + extension with different Service Configurator. + + * tests/Service_Config_Test.conf.xml: Added the converted XML + based svc.conf file. + + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: + * ace/XML_Svc_Conf.h: + * ace/XML_Svc_Conf.cpp: Added abstract interface definition for + DLL based XML parser. + + * ace/OS.h: Defined ACE_DEFAULT_SVC_CONF_FILE to "svc.conf.xml" if + ACE_USES_CLASSIC_SVC_CONF is not defined. + + * ace/Service_Config.cpp: Adapt Service_Config to use XML parser + to process an svc.conf.xml file. + + + * bin/generate_export_file.pl: Added code to record the flags used + to generate an export file. + + + * XML/examples/svcconf/Svcconf_Handler.h: + * XML/examples/svcconf/Svcconf_Handler.i: + * XML/examples/svcconf/Svcconf_Handler.cpp: Added Svcconf handling + simulation. + + * XML/examples/test_docs/svcconf.dtd: Revised Svcconf DTD. + + + * XML/examples/test_docs/svcconf.dtd: Updated the DTD again. + The <streamdef> tag also needed an "id" attribute. + + * XML/examples/SAXPrint/main.cpp: + * XML/examples/SAXPrint/svc.conf.xml: Modified the sample XML to + conform to the updated svcconf.dtd. + + * XML/examples/svcconf/README: + * XML/examples/svcconf/Svcconf.dsp: + * XML/examples/svcconf/Svcconf.dsw: + * XML/examples/svcconf/Svcconf_Handler.cpp: + * XML/examples/svcconf/Svcconf_Handler.h: + * XML/examples/svcconf/Svcconf_Handler.i: + * XML/examples/svcconf/main.cpp: New example to test out + svc.conf.xml parsing. Not finished. + + + * XML/examples/test_docs/svcconf.dtd: Simplied the svc.conf.dtd by + removing the <params> tags and making it an attribute for + <initializer> and <static> tags. + + + * ace/Parse_Node.cpp: + * ace/Parse_Node.h: + * ace/Service_Config.cpp: + * ace/Service_Config.h: + * ace/Svc_Conf.h: + * ace/Svc_Conf.y: + * ace/Svc_Conf_Lexer_Guard.cpp: + * ace/Svc_Conf_l.cpp: + * ace/Svc_Conf_y.cpp: Decoupled old Service_Config parser from ACE. + + + * XML/examples/SAXPrint/main.cpp: Added a new test to try the new + StrCharStream and switch to use ACE_Get_Opt to support more + robust command line arguments. + + * XML/common/StrCharStream.h: + * XML/common/StrCharStream.cpp: Added new CharStream class that + take a null-terminated ACEXML_Char string as an input source. + + * XML/common/FileCharStream.h: Removed unnecessary inclusion of + "ace/streams.h". + + * XML/common/Makefile: + * XML/common/Makefile.bor: + * XML/common/XML_Common.dsp: Added StrCharStream.*. + +Wed May 8 16:29:05 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * bin/nightlybuilds/builds.lst (STATUS): + + Added more information about the build on toutatis. + +Wed May 8 13:57:58 2002 Chad Elliott <elliott_c@ociweb.com> + + * ace/OS_Dirent.h: + * ace/OS_Dirent.cpp: + + Fixed the Dirent_Test under wide character builds with Visual C++ + 6.0. The d_name member of the dirent struct is normally a one + element ACE_TCHAR array that is used as an ACE_TCHAR pointer. + This doesn't work properly under wide character builds with VC6. + Note: because of this change, memcpy on a 'struct dirent' no + longer works on Windows as it would on UNIX. + +Wed May 8 11:02:44 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * tests/TP_Reactor_Test.cpp: Turned off full traces, only the + error messages will goto the log. + +Wed May 8 10:58:15 2002 Alex Libman <AlexL@rumblegroup.com> + + * tests/Proactor_Test.cpp: Fixed the test to work right on + Linux. Turned off full traces, only the error messages and + statistics will go to the log. + +Wed May 8 06:14:08 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Proactor.h (ACE_Proactor): Fixed a comment that said + "ACE_Reactor" rather than "ACE_Proactor" (cut+paste error ;-)). + Thanks to Don Hinton for reporting this. + + * ace/Filecache.cpp (R_MASK): Allow file sharing by default. + Thanks to Peter Kullmann" <P.Kullmann@rog.de> for reporting + this. + +Tue May 07 21:03:01 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.cpp: + + Removed vestigial debugging code. + +Tue May 07 17:42:16 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * ace/OS.i: Added missing "\". + +Tue May 7 16:45:13 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/ACE.cpp (handle_timed_complete): + + Check for the return value from ACE_OS::getsockopt(). Return + ACE_INVALID_HANDLE if return value is -1. Thanks to Jonathan + Wackley <jwackley@legato.com> for reporting this. + + * ace/OS.i (ACE_LACKS_SETREUID_PROTOTYPE): + + Check if we are compiling with gcc in addition to checking for + _XOPEN_SOURCE. GCC unconditionally turns on _XOPEN_SOURCE, but + it is not defined in Solaris 2.5.1. So this hack to expose the + missing prototypes of setreuid() and setregid(). Thanks to + Jonathan Wackley <jwackley@legato.com> for reporting this. + + * include/makeinclude/platform_sunos5_sunc++.GNU: + + Cosmetic fixes. + +Tue May 07 15:44:41 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/auto_run_tests.lst: Reverted the change "Tue May 07 15:15:30 + 2002 Balachandran Natarajan <bala@cs.wustl.edu>". Got an idea + waht was going wrong with the test and fixed that. + +Tue May 07 15:15:30 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/auto_run_tests.lst: Removed No_Server_MT_Connect_Test from + the daily builds. The testing methodology needs to be + rethought. Thanks to Ru for ppointing out that the test is + failing on a few paltforms. + +Tue May 7 12:59:24 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * tests/Proactor_Test.cpp (handle_write_stream): Tried to fix + warnings and errors in Borland builds. Thanks to Ru for pointing + it out. + +Tue May 7 10:31:24 2002 Alex Libman <AlexL@rumblegroup.com> + + * tests/TP_Reactor_Test.cpp: + * tests/TP_Reactor_Test.h: Fixed the tests to work fine on all + platforms. + + * tests/run_test.lst (TP_Reactor_Test): Enabled the test for the + daily runs. + +Mon May 06 16:31:14 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * examples/Reactor/WFMO_Reactor/Abandoned.cpp (handle_timeout): + * examples/Reactor/WFMO_Reactor/APC.cpp (handle_signal): + + Make sure to remove handlers from the reactor if they are not + going to be around when the reactor closes down. + + * examples/Reactor/WFMO_Reactor/Handle_Close.cpp: Reactor creation + was buggy and the wrong reactor's event loop was being run. + + * examples/Reactor/WFMO_Reactor/Talker.cpp: STDIN handler was not + getting cleaned up properly. Also, Proactor was not getting + removed properly from the Reactor. + + * examples/Reactor/WFMO_Reactor/Network_Events.cpp: Network + listener was not getting cleaned up properly. + + * examples/Reactor/WFMO_Reactor/Multithreading.cpp (Task_Handler): + Fixed an "out of bounds" error where a "for" loop that was going + from 1 to n rather than from 0 to n-1. Also, added a debug + statement. + + * examples/Reactor/WFMO_Reactor/run_test.pl: Redirected some of + the output from STDERR to STDOUT. + +Mon May 06 15:45:35 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/parser/debug_validator/Debug_Validator.dsp: Fixed + rum-time library options to avoid compilation errors when MFC is + enabled. + +Mon May 6 14:26:52 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * tests/Proactor_Test.cpp: Fixed unused variable warnings in + non-win32 builds. + +Mon May 6 13:37:45 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/generate_performance_chart.sh: Changed the range of the y + range for the plots. + +Mon May 6 07:39:22 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Message_Queue_T.cpp (notify): Moved int queue_count definition + to outside the scope. Thanks to Chad Elliot for reporting this. + +Sun May 5 19:14:34 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Message_Queue_T.cpp: Modified all the enqueue*() methods so that + their calls to notify() occur *outside* of the monitor lock. + This change prevents deadlock from occurring when a reactor's + notification pipe is full. Thanks to Sasha Agranov + <sagranov@COMGATES.co.il> for reporting this. + +Mon May 6 10:24:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * apps/gperf/src/Makefile: + Removed INSBIN because this didn't work before my change of + Fri May 03 2002, but now it works and it causes error in all + builds. + +Sun May 5 22:23:00 2002 Edan Ayal <edanayal@yahoo.com> + + * tests/Proactor_Test.cpp (Sender): + - Totally removed the message content mem-copies by having the + message blocks assume ownership without copying, by + pre-allocating space for the ending '\0' (needed for the + printouts), and by using an additional complete message content + string for the non-scatter/gather case. + + - Added missing message blocks releases when the write + operations fail. + + - Added a missing message block rd_ptr adjustment at the message + content printout in the non-scatter/gather case. + +Fri May 03 20:29:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/wrapper_macros.GNU: + Only set INSBIN, INSLIB, INSMAN and INSINC when they are not set + yet. This way these can be overruled in a makefile. + +Fri May 03 08:38:27 2002 Ossama Othman <ossama@uci.edu> + + * ace/SSL/SSL_Context.cpp (ssl_library_init): + + Removed extraneous "SSL_library_init()" call. + SSLeay_add_ssl_algorithms(), which is an alias for + SSL_library_init(), is already called. + +Thu May 02 16:33:05 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/OS.i (mmap): On Win32, check for both ERROR_INVALID_NAME and + ERROR_FILE_NOT_FOUND to account for difference when running it + on WINNT but with ACE_HAS_WINNT4 = 0. Thanks to Ram Ben-Yakir + <ram@BANDWIZ.COM> for figuring this out. + +Fri May 3 07:21:20 2002 Ossama Othman <ossama@uci.edu> + + * tests/Dev_Poll_Reactor_Test.cpp: + + Uncommented ACE_START/END_TEST macros and removed extraneous + debugging statements. + +Fri May 3 08:00:55 2002 Chad Elliott <elliott_c@ociweb.com> + + * docs/tutorials/012/work.h: + + Include ace/Log_Msg.h to get the definition of ACE_DEBUG and + related macros. + +Fri May 3 07:07:31 2002 Chad Elliott <elliott_c@ociweb.com> + + * tests/Proactor_Test.cpp: + + Put ACE_TEXT around bare char strings. + +Fri May 3 06:57:13 2002 Chad Elliott <elliott_c@ociweb.com> + + * tests/Proactor_Test.cpp: + + Add a conditional to avoid building this test if threads are not + enabled. + +Fri May 3 06:47:23 2002 Chad Elliott <elliott_c@ociweb.com> + + * tests/Dev_Poll_Reactor_Test.cpp: + + Add a conditional main for when ACE_HAS_DEV_POLL and + ACE_HAS_EVENT_POLL are not defined. + +Fri May 3 06:41:10 2002 Chad Elliott <elliott_c@ociweb.com> + + * ace/Token.h: + + Remove parameter name to avoid a warning from g++. + +Thu May 02 15:54:48 2002 Ossama Othman <ossama@uci.edu> + + * ace/Select_Reactor_T.cpp (work_pending): + + Check if any timers are pending. If so, override any timeout + from the select() call. Corrects a problem where the decision + that no work was pending despite the fact that timers were + pending. + +Thu May 02 15:53:39 2002 Ossama Othman <ossama@uci.edu> + + * tests/Dev_Poll_Reactor_Test.cpp: + + Added some comments that briefly describe the test. + +Thu May 2 15:45:34 2002 Ossama Othman <ossama@uci.edu> + + * tests/Dev_Poll_Reactor_Test.cpp: + + New one-button test for the Dev_Poll_Reactor. Currently it is + fairly simple. Improvements to be added soon. + +Thu May 2 15:22:36 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.h (work_pending_i): + * ace/Dev_Poll_Reactor.cpp (work_pending_i): + + Accept a pointer to ACE_Time_Value instead of a reference. + + Corrected the logic for the test for pending timers. + +Thu May 02 14:38:53 2002 Ossama Othman <ossama@uci.edu> + + * examples/Reactor/Misc/pingpong.cpp: + + Include "ACE.h" to pull in methods in the ACE class/namespace. + +Thu May 2 14:02:20 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.cpp (work_pending_i): + + Take into account pending timers when deciding if there is work + to be done. + +Thu May 02 13:34:08 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.cpp: + + Include "ACE.h" to pull in methods in the ACE class/namespace. + +Thu May 2 13:28:01 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.h: + + Updated some documentation. + + * ace/Dev_Poll_Reactor.inl (upcall): + + New method containing common upcall operations for all types of + IO events. + + * ace/Dev_Poll_Reactor.cpp: + + Commented out "ready set" code since the upcall will simply loop + until the event handler no longer desires callbacks. This is + okay since multiple threads may dispatch event handlers (similar + to what the TP_Reactor does). + + (close): + + Cleaned up memory and resource leaks. + + (handle_events_i, work_pending_i): + + Moved "event polling" code to the new work_pending_i() method. + handle_events_i() now calls that method to determine if any + events must be dispatched. + + If the underlying ioctl() call was interrupted (e.g. via the + INTR signal), i.e. returns -1 with errno set to EINTR, then + automatically restart the event loop if so desired by the user. + + (work_pending): + + Implemented this method. It simply grabs the reactor lock and + calls the new work_pending_i() method. + + (dispatch_io_events): + + Call the remove_handler() method that acquires the lock instead + of the one doesn't. The lock must be reacquired since reactor + state will potentially change during an upcall and since the + lock was released prior to dispatching the upcall. Fixes a race + condition. + + Directly use pollfd pointers as the loop variables instead of an + integer. No need for the additional indirection. + + Since the underlying event demultiplexing mechansim (`/dev/poll' + or '/dev/epoll') is stateful, and since only one result buffer + is used, all pending events (i.e. those retrieved from a + previous poll) must be dispatched before any additional event + can be polled. As such, the Dev_Poll_Reactor keeps track of the + progress of events that have been dispatched. + + The semantics of the event loop in the presence of multiple + threads is non-trivial. The "start point" of the loop will be + incremented each time an event handler is dispatched, which may + be done across multiple threads. Multiple threads may change + the loop variables. Care must be taken to only change those + variables with the reactor lock held. + + (dispatch_timer_handlers): + + Release the reactor lock during the upcall. Fixes a potential + deadlock. + +Thu May 2 10:35:25 2002 Phil Mesnier <mesnier_p@ociweb.com> + + * ace/Log_Msg.cpp: ACE_Log_Msg::open() did not properly initialize a + custom backend due to improper evaluation of result code. + +Thu May 02 07:15:31 2002 Ossama Othman <ossama@uci.edu> + + * ace/IO_Cntl_Msg.h: + + Include OS.h to pull in size_t typedef. + +Thu May 2 07:50:29 2002 Chad Elliott <elliott_c@ociweb.com> + + * examples/Web_Crawler/URL_Addr.h: + + Include ace/ACE.h to avoid compilation error due to missing + method (ACE::hash_pjw) in inline method ACE_URL_Addr::hash(). + +Thu May 2 07:46:59 2002 Chad Elliott <elliott_c@ociweb.com> + + * websvcs/lib/URL_Addr.h: + + Include ace/ACE.h to avoid compilation error due to missing + method (ACE::hash_pjw) in inline method ACE_URL_Addr::hash(). + +Thu May 2 07:34:59 2002 Chad Elliott <elliott_c@ociweb.com> + + * ace/QoS/SOCK_Dgram_Mcast_QoS.cpp: + + Switch ACE::get_ip_interfaces() to + ACE_Sock_Connect::get_ip_interfaces() and include + ace/Sock_Connect.h. This only applies to Win32. + +Thu May 2 00:39:32 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * ace/Dynamic.i: Uncommented ACE_TRACE calls. + +Thu May 2 00:29:34 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * ace/TLI_Stream: Uninlined a methods that needed access to the + ACE class. + + * ace/Service_Object: Uninlined ACE_Service_Type::name(). + + * ace/Malloc_T.cpp: + * ace/Process_Mutex.cpp: + * ace/SOCK_Dgram_Bcast.cpp: + * ace/TLI_Acceptor.cpp: + * ace/TLI_Connector.cpp: + * ace/TLI_Stream.cpp: + + Added #include "ace/ACE.h". + +Wed May 1 21:07:23 2002 Ossama Othman <ossama@uci.edu> + + * ace/config-win32-msvc-6.h: + + Moved library "decorator" related macros to + config-win32-msvc.h. + + * ace/config-win32-msvc.h: + + Support "decorated" library names for MSVC 6 or better. Fixes + run-time library search problems in MS Visual Studio .NET + builds. + + * ace/ATM_Addr.h: + * ace/ATM_Params.h: + * ace/MEM_Addr.h: + * ace/Module.h: + * ace/Name_Space.h: + * ace/Naming_Context.h: + * ace/Remote_Name_Space.h: + + Include "ACE_export.h" instead of "ACE.h." The latter is + overkill. + + * ace/PI_Malloc.h: + + Include "ACE_export.h" instead of "ACE.h." The latter is + overkill. + + No need to include "ace/Malloc_T.h" and "ace/Memory_Pool.h" + + * ace/Date_Time.h: + * ace/Event_Handler.h: + * ace/High_Res_Timer.h: + * ace/Malloc.h: + * ace/Malloc_Allocator.h: + * ace/Malloc_T.h: + * ace/Mem_Map.h: + * ace/Pipe.h: + * ace/Profile_Timer.h: + * ace/Read_Buffer.h: + * ace/Shared_Memory.h: + * ace/Shared_Object.h: + * ace/SOCK.h: + * ace/SPIPE_Addr.h: + * ace/Stats.h: + * ace/SV_Message.h: + * ace/SV_Semaphore_Simple.h: + * ace/SV_Shared_Memory.h: + * ace/Synch.h: + * ace/Synch_Options.h: + * ace/Thread.h: + * ace/Typed_SV_Message.h: + + Include "OS.h" and "ACE_export.h" instead of "ACE.h." The + latter is overkill. + + * ace/Configuration.h: + * ace/DEV_Addr.h: + * ace/FIFO.h: + * ace/Get_Opt.h: + * ace/Vector_T.h: + + No need to include "ACE.h." + + * ace/Dynamic.h: + + No need to include "ACE.h", "Synch_T.h" and "Singleton.h." + "ACE_export.h" is enough. + + * ace/INET_Addr.h: + + Include "Sock_Connect.h" instead of "ACE.h". The latter is + overkill. + + * ace/Message_Block.h: + + Include "ace/OS.h" and "ace/ACE_export.h" instead "ace/ACE.h". + The latter is overkill. + + No need to include "Malloc.h" + + * ace/IO_Cntl_Msg.h: + * ace/Sock_Connect.h: + + Include "ACE_export.h" to pull in ACE_Export macro definition. + + * ace/Activation_Queue.cpp: + * ace/Message_Block.cpp: + + Include "ace/Malloc_Base.h" to pull in ACE_Allocator class + declaration. + + * ace/DEV_IO.h: + * ace/FIFO_Recv.h: + * ace/FIFO_Send.h: + * ace/NT_Service.h: + * ace/SOCK_Dgram.cpp: + * ace/SOCK_Dgram_Mcast.cpp: + * ace/SOCK_IO.h: + * ace/SPIPE_Stream.h: + * ace/SV_Semaphore_Simple.cpp: + * ace/Service_Types.h: + * tests/test_config.h: + + Include "ace/ACE.h" pull in methods in the "ACE" + class/namespace. + + * ace/Stream.h: + * ace/SV_Message_Queue.h: + + Include "config-all.h" instead of "ACE.h." The latter is + overkill. + + * ace/OS_Memory.h: + * ace/config-all.h: + + Moved default definition of the + ACE_HAS_POSITION_INDEPENDENT_POINTERS macro to + "ace/config-all.h". Allows for additional reductions in + inter-header dependencies. + + * ace/PI_Malloc.cpp: + + Moved non-essential includes within the + "ACE_HAS_POSITION_INDEPENDENT_POINTERS == 1" block. No need to + include them if the implementation won't be compiled. + + * ace/Select_Reactor_T.cpp: + + Include "ace/ACE.h" pull in methods in the "ACE" + class/namespace. + + (handle_events): + + Moved ACE_Countdown_Time declaration within the ACE_MT_SAFE + block. It is only used for the multi-threaded case. + + (work_pending): + + As part of the timeout take into account the time taken to + acquire the lock. + + Copy the timeout value. Do not ignore the interface by casting + away the const-ness. + + Do not bother continuing if the reactor has been deactivated. + +Wed May 1 15:49:33 2002 Steve Huston <shuston@riverace.com> + + * tests/DLL_Test_Impl.{h cpp}: Add a nothrow variant of operator new + for platforms with ACE_HAS_NEW_NOTHROW (like HP-UX, aC++). + +Tue Apr 30 14:54:02 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/auto_run_tests.lst: Added $TAO_ROOT/tests/Connection_Timeout + to the daily build list. + +Tue Apr 30 14:42:06 2002 Chad Elliott <elliott_c@ociweb.com> + + * ace/Synch.h: + * ace/Synch.i: + + Change the Null wait condition wait() method to take a const + ACE_Time_Value* like the Mutex wait condtion. + +Mon Apr 29 20:28:49 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * bin/nightlybuild/builds.lst: Move SunOS_SunCC51, SunOS_GCC_2_95, + Solaris8_FORTE_UP_1 to new scoreboard. + +Sun Apr 28 18:03:50 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/auto_run_tests.lst: Added the new tests to our daily list of + tests that needs to be run. + +Sun Apr 28 11:50:17 2002 Edan Ayal <edanayal@yahoo.com> + + * ace/WIN32_Asynch_IO.cpp: + ACE_WIN32_Asynch_Read/Write_Stream/File_Result::complete - + corrected the updates to the chained message blocks (rd_ptr and + wr_ptr) in case of scatter-gather IO. This was apparently a + cut-n-paste bug. + +Sun Apr 28 08:19:37 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * tests/Proactor_Test.cpp: Added code to test async scatter-gather + IO on Win32. Thanks Edan Ayal <edanayal@yahoo.com> for donating + this. + +Sun Apr 28 12:27:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/ace_flags.bor: + Added cflags and lib for new TAO IFRService library + +Sat Apr 27 20:00:27 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: Exclude template files Atomic_Op.cpp and + Framework_Component_T.cpp from compilation. + +Sat Apr 27 18:10:37 2002 Balachandran Natarajan <bala@guajira.cs.wustl.edu> + + * ace/Message_Queue_T.cpp (flush): Fixed fuzz error. + +Sat Apr 27 17:13:16 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Svc_Handler.cpp: Added a closing '}' for the method + ACE_Buffered_Svc_Handler::dump () const. This should fix the + compile errors with g++ that will start showing up shortly in + the Scoreboard. + + * ace/*makefile: Updated dependencies. + +Sat Apr 27 11:16:03 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + Thanks to Boris Temkin <borist@allcharge.com> for motivating the + following fixes: + + * ace/Svc_Handler.cpp: Updated the flush_i() method to call the + message queue's flush_i() method to avoid deadlocks on platforms + that lack recursive mutexes. + + * ace/Message_Queue_T.{h,cpp}: Added the flush() and flush_i() + methods to the ACE_Message_Queue in order to remove messages + without deactivating the queue. Also refactored the close() + method to use flush_i(). + +Sat Apr 27 09:26:43 2002 Edan Ayal <edanayal@yahoo.com> + + Added support (currently under win32 only) for asynchronous + scattered read and asynchronous gathered write with socket streams + and with files. + + * ace/Asynch_IO.{h cpp}: Added methods 'readv' to + ACE_Asynch_Read_Stream/File, and 'writev' to + ACE_Asynch_Write_Stream/File, for win32 only. The methods just + delegate to the implementation. + + * ace/Asynch_IO_Impl.h: Added pure virtual methods + 'readv' to ACE_Asynch_Read_Stream/File_Impl, and 'writev' to + ACE_Asynch_Write_Stream/File_Impl, for win32 only. + + * ace/WIN32_Asynch_IO.{h cpp}: Implemented the 'readv' + for ACE_WIN32_Asynch_Read_Stream/File and 'writev' for + ACE_WIN32_Asynch_Write_Stream/File. Modified the 'complete' + method and the ctor of ACE_WIN32_Asynch_Read_Stream/File_Result + and ACE_WIN32_Asynch_Write_Stream/File_Result. + +Sat Apr 27 16:07:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/Reactor/Proactor/test_udp_proactor.cpp: + Fixed BCB warning about hides virtual function by renaming open + to open_addr + + * ace/Asynch_IO.h: + Minor comment update + +Sat Apr 27 08:59:56 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Process.i: If getpid() is ACE_INVALID_HANDLE, don't call + ACE::terminate_process() or ACE_OS::kill() since + ACE_INVALID_HANDLE is -1 on UNIX, which causes all sorts of + problems due to the fact that -1 terminates the whole session + group! Thanks to Olivier Brunet <o.brunet@free.fr> for + reporting this. + +Sat Apr 27 15:28:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/Reactor/Proactor/test_proactor.cpp: + * examples/Reactor/WFMO_Reactor/Talker.cpp: + Fixed MSVC6 compile errors + +Fri Apr 26 17:00:23 2002 Dante J. Cannarozzi <djc2@cs.wustl.edu> + + * ace/Containers_T.h + * ace/Unbounded_Set.h + * ace/Hash_Map_Manager.h + * ace/RB_Tree.h + * ace/Vector_T.h + * ace/Unbounded_Queue.h + : with the help of Matt Hampton <mph2@cs.wustl.edu> updated + doxygen to contain a consistent html list that summarizes the + each container (like internal structure, duplicates allowed, etc.) + at the request of Steve Huston. + +Fri Apr 26 16:36:52 2002 Steve Huston <shuston@riverace.com> + + * netsvcs/lib/Server_Logging_Handler_T.cpp: Added #include + "ace/Log_Record.h" to pick up ACE_Log_Record declaration. + Fixes compile error on HP-UX. + +Fri Apr 26 15:10:21 2002 Steve Huston <shuston@riverace.com> + + * ace/ACE.cpp: + * ace/Sock_Connect.cpp: Moved ACE_Auto_Array_Ptr<> and + ACE_Auto_Basic_Array_Ptr<> instantiations from ACE.cpp to + Sock_Connect where they are now used. Also added MACOSX to the + platforms that get them instantiated. + Added include of OS.h in Sock_Connect.cpp to pick up the ifreq + stuff portably. + + * ace.icc: Added Time_Value, Copy_Disabled, Argv_Type_Converter, + Dev_Poll_Reactor, POSIX_CB_Proactor, WIN32_Proactor, + WIN32_Asynch_IO, Obstack, Vector_T. + Added the new group TIMER_FILES (and TIMER_INCLUDES) to parallel + the TIMER_FILES in ace/Makefile. + +Fri Apr 26 14:05:49 2002 Chad Elliott <elliott_c@ociweb.com> + + * ace/OS.i: + + Modify the ACE_OS::fflush() method to not call the system + implementation on VxWorks if the FILE* is 0. This causes + problems for the OS. Thanks to Michael Kirher + <Michael.Kircher@mchp.siemens.de> for reporting and providing + a fix for this. + + * ace/Svc_Conf.l: + * ace/Svc_Conf_l.cpp: + + Use ACE_OS::fflush() instead of fflush(). + +Fri Apr 26 13:22:18 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Svc_Handler.{h,cpp}: Applied the Thread-safe Interface pattern + by factoring out the flush logic into a flush_i() method (that + doesn't hold a lock) and call this method from the put() and + flush() methods (that do hold locks). Thanks to Boris Temkin + <borist@allcharge.com> for reporting this problem. + +Fri Apr 26 14:45:28 2002 Steve Huston <shuston@riverace.com> + + * tests/Proactor_Test.cpp: Replace CALLBACK with CB. Apparantly, + CALLBACK is a macro with MSVC. + +Fri Apr 26 14:39:37 2002 Steve Huston <shuston@riverace.com> + + * tests/Proactor_Test.cpp: Use default proactor type for platform + if no specific proactor type for platform is chosen. + +Fri Apr 26 11:39:22 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/examples.dsw: + * examples/C++NPv2/display_logfile.dsp: + * examples/C++NPv2/Reactor_Logging_Server.dsp: + * examples/C++NPv2/Select_Reactor_Logging_Server.dsp: + * examples/C++NPv2/TP_Reactor_Logging_Server.dsp: + * examples/C++NPv2/WFMO_Reactor_Logging_Server.dsp: + New project/workspace for the examples here so far. + + * examples/C++NPv2/WFMO_Reactor_Logging_Server.cpp: Make dtor + public - this object is created on the stack. + Added constructor that accepts a ACE_Reactor * and passes it up + the inheritance chain. + + * examples/C++NPv2/Select_Reactor_Logging_Server.cpp: + * examples/C++NPv2/TP_Reactor_Logging_Server.cpp: Add preprocessor + check to be sure that the correct C++ library is enabled on Windows. + Qualify "getline" call with "std::". + +Fri Apr 26 07:42:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/IOStream_Test.cpp: + Corrected the debug output string of server received so that + the formatting is the same as client received. Thanks to + Soeren Gerlach for reporting this. + + * examples/Reactor/WFMO_Reactor/Talker.cpp: + * examples/Reactor/Proactor/test_proactor.cpp: + * examples/Reactor/Proactor/test_proactor2.cpp: + Fixed BCB compile warnings + +Thu Apr 25 17:34:31 2002 Steve Huston <shuston@riverace.com> + + * bin/auto_compile: When doing 'realclean', do them in reverse + order of the builds, else ace/ACE_COMPONENTS.list gets wiped + out first, preventing things that examine the components list + (like netsvcs) from getting cleaned. + +Thu Apr 25 16:20:33 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/TP_Reactor_Logging_Server.{cpp mak}: + * examples/C++NPv2/WFMO_Reactor_Logging_Server.cpp: + New source and makefiles for C++NPv2 chapter 4 examples. + + * examples/C++NPv2/Makefile: Added TP_Reactor_Logging_Server. + +Thu Apr 25 15:46:39 2002 Steve Huston <shuston@riverace.com> + + POSIX asynch I/O improvements/corrections submitted by Alex + Libman <alibman@ihug.com.au>: + + * ace/POSIX_CB_Proactor.{h i cpp}: New POSIX Proactor implementation + that makes use of the AIO facility's callback feature. This has + only been tested on SGI Irix. + + * ace/Makefile: Added POSIX_CB_Proactor. + + * ace/POSIX_Proactor.{h cpp}: Refactored some code and added better + cancellation/cleanup handling. Added hooks for the new + ACE_POSIX_CB_Proactor class and refactored methods + close(), get_result_status(), create_result_aiocb_list(), and + delete_result_aiocb_list(). + + * ace/SUN_Proactor.{h cpp}: Add the new get_result_status() method + and refactored the old results code into the new method. Also, + make good use of the new aiocb create/delete methods. Improvements + to operation status detection based on input from Sun. + + * tests/Proactor_Test.cpp: Add support for new ACE_POSIX_CB_Proactor. + Added a log lock to be sure all messages from one transaction are + logged together. Also, sets up full duplex I/O for Windows and + Solaris; half duplex for all others due to general weakness in + AIO subsystems. If further testing reveals that more can be set + to full duplex, this can be expanded. + +Thu Apr 25 14:02:38 2002 Steve Huston <shuston@riverace.com> + + * tests/MT_Reactor_Upcall_Test.cpp: Fixed compile errors by using + sizeof (type) instead of sizeof type. + +Thu Apr 25 06:51:58 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/OS.h (ACE_FACTORY_DEFINE): Replaced an ACE_reinterpret_cast() + with an ACE_static_cast() since the latter works correctly with + multiple inheritance. Thanks to Chris Uzdavinis + <chris@atdesk.com> for reporting this. + + * ace/CDR_Base.h (ACE_CDR): Improved the documentation of the + consolidate() method. Thanks to Alain Decamps for helping with this. + +Thu Apr 25 03:34:43 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * tests/MT_Reactor_Upcall_Test.cpp: Changed the test such that the + end of the test is indicated by a shutdown message rather than + an explicit message count. + +Wed Apr 24 23:39:00 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * examples/Reactor/WFMO_Reactor: Updated file names to match + executable names. + +Thu Apr 25 07:34:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Log_Msg.h: + Doxygen-ized some of the comments + +Wed Apr 24 18:53:56 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * include/makeinclude/rules.local.GNU (clean.local): Reverted the + change "Wed Apr 24 16:08:30 2002 Balachandran Natarajan + <bala@cs.wustl.edu>" since it is bogus. + +Wed Apr 24 16:08:30 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * include/makeinclude/rules.local.GNU: Added a rule to clean up + Sunws_cache properly. Should help SunCC5_1 builds. + +Wed Apr 24 11:07:15 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/RMCast/Makefile (FILES): Removed a stray "\". Thanks to + Amir Kunst <amir@fats.co.il> for reporting this. Also reordered + the files so they would be in alphabetical order. + +Wed Apr 24 11:17:31 2002 Steve Huston <shuston@riverace.com> + + * PROBLEM-REPORT-FORM: Always ask for config.h and + platform_macros.GNU; not just on compilation issues. + +Wed Apr 24 07:56:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Future_Set.h: + Converted a comment to javadoc style because it is very long for + the brief documentation. + + * ace/DEV_IO.h: + * ace/TTY_IO.h: + Removed not needed includes + +Wed Apr 24 03:14:11 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * tests/MT_Reactor_Upcall_Test.cpp: Fix warnings on threads=0 + build. + +Tue Apr 23 18:45:31 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Logging_Event_Handler.h (get_handle): Cast + away const-ness of logging_handler_ member to call it's + peer() method, but use that in a const method call. Keeps + the interface contract intact. This example now works. + + * examples/C++NPv2/display_logfile.mak: Makefile for the + display_logfile example (chapter 9). + + * examples/C++NPv2/Select_Reactor_Logging_Server.{cpp mak}: New + example for the ACE_Select_Reactor section in chapter 4. + + * examples/C++NPv2/Makefile: Added Select_Reactor_Logging_Server. + +Tue Apr 23 16:01:47 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/auto_run_tests.lst: Prevented the Miop/McastHello test from + running in SUNCC51 builds. SUNCC51 is too broken. Time to remove + the build of the scoreboard. + +Tue Apr 23 11:31:43 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.inl (ACE_Dev_Poll_Handler_Guard): + (~ACE_Dev_Poll_Handler_Guard): + + Added "todo" comments about suspending and resuming the event + handler before and after the upcall, respectively. + +Tue Apr 23 11:38:40 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Makefile: Rearranged to build all the examples. + +Tue Apr 23 13:32:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/Reactor/Makefile.bor: + Added Proactor directory + + * examples/Reactor/Proactor/Makefile.bor: + Added this BCB makefile + + * examples/Reactor/Proactor/*.cpp + Made all files compiling when doing a wchar build. + +Tue Apr 23 10:03:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/Reactor/Makefile.bor: + Added WFMO_Reactor directory + + * examples/Reactor/WFMO_Reactor/Makefile.bor: + Added this BCB makefile + + * examples/Reactor/WFMO_Reactor/*.cpp + Fixed several unused argument warnings that appeared with bcb. + Made all files compiling when doing a wchar build. + +Tue Apr 23 07:35:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Base_Thread_Adapter.h: + * ace/Future.h: + * ace/Naming_Context.h: + * ace/Malloc.h: + * ace/Malloc_Base.h: + * ace/Malloc_T.h: + * ace/NT_Service.h: + * ace/Notification_Strategy.h: + * ace/Object_Manager.h: + * ace/OS_Thread_Adapter.h: + * ace/WIN32_Asynch_IO.h: + Doxygen-ized some of the comments + + * ace/OS.h: + Added ACE_ALLOCATOR_NORETURN and ACE_NEW_MALLOC_NORETURN, these + macro's set errno to ENOMEM incase of an memory error but don't + call return but guarantee that the pointer is 0. + + * ace/QoS/QoS_Manager.h: + * ace/QoS/QoS_Session_Impl.h: + Doxygen-ized some of the comments + + * ace/config-doxygen.h: + Added define ACE_YY_USES_PROTOS to parse ace/Svc_Conf_l.cpp correctly + + * tests/MT_Reactor_Upcall_Test.cpp: + Fixed unicode and single threaded build errors/warnings + +Mon Apr 22 20:40:41 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Logging_Event_Handler.{cpp h}: + Moved LOGFILE_SUFFIX to open() method. Changed get_handle() to be + const to match the signature inherited from ACE_Event_Handler. + + * examples/C++NPv2/Logging_Event_Handler_Ex.h: Moved MAX_CLIENT_TIMEOUT + into class declaration as an enum. + + * examples/Logging_Handler.h: Added terminating ';' to inlined + methods. + +Mon Apr 22 18:33:20 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * examples/Reactor/WFMO_Reactor/run_test.pl: This file will allow + us to run these tests nightly. Note that Console_Input, + Network_Events, and Talker are not included since they are + interactive tests. + + * bin/auto_run_tests.lst: Added new test script. + + * examples/Reactor/WFMO_Reactor/test_handle_close.cpp: + * examples/Reactor/WFMO_Reactor/test_multithreading.cpp: + * examples/Reactor/WFMO_Reactor/test_network_events.cpp: + Fixed several memory management and shutdown issues. + + * examples/Reactor/WFMO_Reactor/WFMO_Reactor.dsw: + Changed "Abondoned" to "Abandoned" + + * examples/Reactor/WFMO_Reactor/test_window_messages.cpp: + Fixed debug message formatting. + + * examples/Reactor/WFMO_Reactor/test_network_events.cpp: + Fixed documentation to says that ACE_DEFAULT_SERVER_PORT + defaults to 10002; correct value is 20002. + +Mon Apr 22 17:24:39 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Reactor_Logging_Server.mak: + * examples/C++NPv2/Logging_Acceptor.{cpp h}: + * examples/C++NPv2/Logging_Acceptor_Ex.h: + * examples/C++NPv2/Logging_Event_Handler.{cpp h}: + * examples/C++NPv2/Logging_Event_Handler_Ex.{cpp h}: + * examples/C++NPv2/Reactor_Logging_Server.{cpp h}: + New files for examples in C++NPv2, chapter 3. + +Mon Apr 22 15:43:15 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * examples/DLL/Today.h: + * examples/DLL/Today.cpp: + * examples/DLL/Newsweek.h: + * examples/DLL/Newsweek.cpp: Added overloaded new/delete opertors + and updated magazine messages. + +Mon Apr 22 15:20:02 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * tests/DLL_Test.h: + * tests/DLL_Test.cpp: Removed the destroy method as it didn't help + ensuring allocating/deallocating the DLL object with consistent + heap. Instead, we reverted back to use auto_ptr and relied on + the overloaded new/delete operators. + + * tests/DLL_Test_Impl.h: + * tests/DLL_Test_Impl.cpp: Added the aforementioned overloaded + new/delete operator implementations. + +Mon Apr 22 12:33:30 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * tests/MT_Reactor_Upcall_Test: This is a test that shows how to + handle upcalls from the TP_Reactor and the WFMO_Reactor when the + event loop is being run by multiple threads. + + The following files were updated to include the new test: + + - tests/Makefile + - tests/Makefile.am + - tests/Makefile.bor + - tests/Reactor_Dispatch_Order_Test.dsp + - tests/Reactor_Dispatch_Order_Test.icc + - tests/icc.bat + - tests/run_test.lst + - tests/run_tests.bat + - tests/run_tests.psosim + - tests/run_tests_remote.lst + - tests/tests.dsw + - tests/tests.icp + - tests/MT_Reactor_Upcall_Test_WinCE.vcp + - tests/tests_WinCE.vcw + +Mon Apr 22 15:27:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * docs/tutorials/021/page01.html + Corrected incorrect link + +Mon Apr 22 11:20:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Svc_Conf_y.cpp: + Removed some strange characters from this file + + * ace/config-doxygen.h: + Added define of __cplusplus so that the file Svc_Conf_y.cpp and + Svc_Conf_l.cpp are correctly parsed. We parsed the c-code and + we got the following defines which we don't want: + #define const + Also there was a global variable 'int size' which cluttered other + documentation. + +Mon Apr 22 08:55:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Timer_Queue_T.h: + Doxygen-ized some of the comments + + * etc/*.doxygen: + Added the following to all doxygen config files + DISTRIBUTE_GROUP_DOC = YES + When a doxygen group is used (starts with //@{ and ends with + //@}) and if there is only one description for all group members than + this description is used for all the members separately. If a member + has its own description than that description is used. + +Mon Apr 22 01:34:32 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/Task.h: Enhanced the comment for <wait> method to make it + clear that it will not wait for detached thread. Thanks to + Alain Decamps <Alain.Decamps@PIDPA.be> for suggesting this. + +Sun Apr 21 21:23:59 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_osf1_4.x_cxx.GNU (SOFLAGS): + + Re-enable -use_ld_input in SOFLAGS. Seems that this flag is used + to identify Digital Unix in some Makefiles which resort to all + kinds of hackery to get some files to link. + +Sun Apr 21 20:59:26 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_osf1_4.x_cxx.GNU (OCFLAGS): + * include/makeinclude/platform_osf1_4.x_g++.GNU (OCFLAGS): + + Set the optimization to level -O3. That way we will might just + get two builds in a day. This is also the recommended level in + the manual. + +Sun Apr 21 15:02:52 UTC 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/TP_Reactor_Test.cpp: Add checks for ACE_AS_THREADS. + +Sun Apr 21 09:42:05 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/config-win32-common.h: Changed the value of ACE_IOV_MAX to + be 64 rather than 1024. Thanks to Alain Decamps for motivating + this. + +Sat Apr 20 19:57:55 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/Asynch_Pseudo_Task.cpp (stop): Wrap ace_mon in ACE_MT() to eliminate + errors on builds which are configured single-threaded. + +Fri Apr 19 12:05:42 2002 Ossama Othman <ossama@uci.edu> + + * examples/Bounded_Packet_Relay/bpr_thread.cpp: + * examples/Timer_Queue/main_async.cpp: + * examples/Timer_Queue/main_reactor.cpp: + * examples/Timer_Queue/main_thread.cpp: + + The recent "ace/OS.h" inclusion reductions prevented a pragma + that disables a MSVC 6 warning about an overly long debugging + symbol from occurring. Explicitly disable that warning in this + file to silence the warning. + +Fri Apr 19 11:41:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Typed_SV_Message_Queue.h: + * ace/Strategies_T.h: + * ace/OS.h: + * ace/MEM_SAP.h: + * ace/Asynch_IO.h: + Doxygen-ized some of the comments + +Thu Apr 18 15:58:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Asynch_Connector.{h cpp}: Removed static member inet_addr_any_. + It doesn't work for all compilers (static in a template). Replaced + default arg for ACE_Asynch_Connector::connect with + ACE_INET_Addr ((u_short)0) to get the equivalent effect, i.e., let + the OS pick an unused port. + + * ace/Asynch_Acceptor.cpp (parse_address): Use the ACE_SOCK support + for dealing with addresses, else the ability to deal with both + IPv4 and IPv6 is lost. + Also get rid of a lot of debugging output; use ACE_TRACE instead. + + * tests/Proactor_Test.cpp: Fixed log file name. Corrected formatting + per ACE guidelines. + + * tests/run_test.lst: Enabled Proactor_Test and Proactor_Timer_Test + everywhere except Chorus, VxWorks, LynxOS. + +Thu Apr 18 11:34:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Asynch_IO.h: + * ace/ATM_Params.h: + * ace/ATM_Stream.h: + * ace/Caching_Strategies_T.h: + * ace/Event_Handler.h: + * ace/MEM_Acceptor.h: + * ace/MEM_Connector.h: + * ace/MEM_SAP.h: + * ace/Memory_Pool.h: + * ace/Message_Block.h: + * ace/Message_Queue_T.h: + * ace/Pair_T.h: + * ace/Parse_Node.h: + * ace/Select_Reactor_Base.h: + * ace/Thread_Adapter.h: + * ace/Thread_Adapter.h: + * ace/Thread_Manager.h: + * ace/WIN32_Asynch_IO.h: + * ace/WIN32_Proactor.h: + Doxygen-ized some of the comments + +Tue Apr 18 07:39:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * performance-tests/Synch-Benchmarks/Base_Test/Baseline_Test.h: + * performance-tests/Synch-Benchmarks/Perf_Test/Benchmark_Performance.h: + * performance-tests/Synch-Benchmarks/Perf_Test/Performance_Test_Options.h: + Added include of svc_export.h. + +Wed Apr 17 18:41:30 2002 Steve Huston <shuston@riverace.com> + + * ace/Activation_Queue.h: + * ace/Method_Request.h: Doxygen-ize and improve some documentation. + +Wed Apr 17 15:24:36 2002 Ossama Othman <ossama@uci.edu> + + * bin/auto_run_tests.lst: + + Added ORT test to the list of tests to run. + +Wed Apr 17 14:30:29 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/display_logfile.cpp (Logrec_Module ctor): Change + name from ACE_TCHAR * to const ACE_TCHAR *. + +Wed Apr 17 10:39:24 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.cpp (open): + + Corrected a potential resource (file descriptor) leak for the + "ACE_HAS_DEV_POLL" case. + + Fixed a warning. Use "ACE_NEW_RETURN" instead of "ACE_NEW." + +Wed Apr 17 10:29:14 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.cpp (ACE_HAS_DEV_POLL): + + Added missing "#endif" for this configuration. + +Wed Apr 17 12:47:41 2002 Steve Huston <shuston@riverace.com> + + * ace/WIN32_Asynch_IO.cpp (set_handle, handle_close): Fixed + "unused argument" warnings from C++Builder. + +Wed Apr 17 14:30:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/config-doxygen.h: + Added define UNICODE to generate ACE_Registry_Name_Space + + * ace/Proactor.h: + * ace/Proactor_Impl.h: + * ace/RB_Tree.h: + * ace/Reactor_Notification_Strategy.h: + * ace/Recyclable.h: + * ace/Registry.h: + * ace/Service_Types.h: + * ace/Signal.h: + * ace/Token_Request_Reply.h: + * ace/Typed_SV_Message.h: + Doxygen-ized some of the comments + +Wed Apr 17 12:27:23 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * etc/ace.doxygen: + * etc/ace_man.doxygen: + Removed ACE_COMPILE_TIMEPROBES as PREDEFINED because it should be + defined in Config-doxygen.h + + * ace/Config-doxygen.h + Added ACE_COMPILE_TIMEPROBES as define to generate timeprobe + documentation. This config file is automatically included when doing + a doxygen generation. + + * ace/Intrusive_List.h: + * ace/Synch.h: + Doxygen-ized some of the comments + + * ace/Message_Queue.h: + Fixed typo + +Wed Apr 17 07:28:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/DLL/Newsweek.cpp: + * examples/DLL/Today.cpp: + Added include of svc_export.h. + + * tests/Proactor_Test.cpp: + Fixed fuzz and unicode error + + * ace/Asynch_Acceptor.h: + * ace/Functor.h: + * ace/Message_Queue.h: + * ace/Module.h: + * ace/Thread_Manager.h: + Doxygen-ized some of the comments + +Tue Apr 16 23:32:49 2002 Steve Huston <shuston@riverace.com> + + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: Added Asynch_Connector.{h cpp} and + Asynch_Pseudo_Task.{h cpp}. + +Tue Apr 16 16:39:07 2002 Ossama Othman <ossama@uci.edu> + + * ace/POSIX_Asynch_IO.cpp (set_handle, handle_close): + + Fixed unused argument warnings. + +Tue Apr 16 19:25:47 2002 Steve Huston <shuston@riverace.com> + + * ace/Makefile.bor: + * ace/ace.icc: Added new Asynch_Connector and Asynch_Pseudo_Task files. + +Tue Apr 16 19:07:22 2002 Steve Huston <shuston@riverace.com> + + * tests/Proactor_Test.cpp: Basically replaced; new version also + tests new ACE_Asynch_Connector facility. + +Tue Apr 16 18:42:39 2002 Steve Huston <shuston@riverace.com> + + New feature, ACE_Asynch_Connect, contributed by Alex Libman + <alibman@ihug.com.au>. Allows asynchronous connect using the + ACE Proactor framework. The new classes follow the same arrangement + as the existing ACE_Asynch_Accept framework. + + * ace/Asynch_Connector.{h cpp}: New files + + * ace/Asynch_IO.{h cpp}: Added new ACE_Asynch_Connect class and + its result. Added new method, ACE_Handler::handle_connect(), to + handle completion of asynch connect operations. + + * ace/Asynch_IO_Impl.{h i cpp}: Added new classes + ACE_Asynch_Connect_Impl and ACE_Asynch_Connect_Result_Impl. + + * ace/Asynch_Pseudo_Task.{h cpp}: Generalized task that handles + asynch emulation where needed, for example, asynch accept/connect. + Replaces the ACE_Asynch_Accept_Task and used for both accept/connect. + + * ace/POSIX_Asynch_IO.{h cpp}: Removed ACE_POSIX_Asynch_Accept_Task + (subsumed by ACE_Asynch_Pseudo_Task, above) and add the + ACE_POSIX_Asynch_Connect and its Result class. + + * ace/POSIX_Proactor.{h cpp}: Added asynch connect plumbing. + + * ace/Proactor.{h cpp}: Added asynch connect support methods. + + * ace/Proactor_Impl.h: Added create_asynch_connect[_result] methods. + + * ace/SUN_Proactor.cpp: Change from asynch_accept_task to + asynch_pseudo_task. + + * ace/WIN32_Asynch_IO.{h cpp}: Add new ACE_WIN32_Asynch_Connect and + _Result. + + * ace/WIN32_Proactor.{h cpp}: Added new create_asynch_connect() and + create_asynch_connect_result() methods. + + * ace/Makefile: Added Asynch_Connector, Asynch_Pseudo_Task + +Tue Apr 16 11:49:00 2002 Ossama Othman <ossama@uci.edu> + + * ace/Service_Templates.cpp: + * ace/Thread_Manager.cpp: + + Only instantiate ACE_Auto_Basic_Ptr templates if + ACE_LACKS_AUTO_PTR is defined or if ACE_HAS_STANDARD_CPP_LIBRARY + is not defined. In the above cases, they are only used as + base classes for ACE's implementation of the standard auto_ptr + template. When using the auto_ptr implementation provided by + the standard C++ library in use, the ACE_Auto_Basic_Ptr template + instances aren't needed. Reduces footprint in cases where the + standard C++ library implementation is used, and explicit + template instantation is required. + +Tue Apr 16 14:38:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * etc/ace.doxygen: + * etc/ace_man.doxygen: + Set ACE_COMPILE_TIMEPROBES as PREDEFINED so that timeprobe doxygen + documentation is generated + + * ace/Dump.h: + * ace/Message_Block.h: + * ace/Task.h: + * ace/Task_T.h: + * ace/Thread_Manager.h: + * ace/TLI_Stream.h: + Doxygen-ized some of the comments + +Tue Apr 16 11:04:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Event_Handler.h: + * ace/Log_Msg.h: + * ace/Message_Block.h: + * ace/Timer_Hash_T.h: + Doxygen-ized some of the comments + +Mon Apr 15 22:25:28 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * THANKS: Added Andrew Guy to the hallf of fame. + +Mon Apr 15 21:43:44 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/Log_Msg.cpp (log): + * ace/Log_Msg.h: + + Added new option '@' to ACE_Log_Msg to print out pointers in + hexadecimal format. This is equivalent to the "%p" option of the + standard C library printf. + +Mon Apr 15 14:04:28 2002 Ossama Othman <ossama@uci.edu> + + * ace/Dev_Poll_Reactor.h: + * ace/Dev_Poll_Reactor.inl: + * ace/Dev_Poll_Reactor.cpp: + + Experimental implementation of a /dev/poll (or Linux /dev/epoll) + based Reactor. Refinements will be committed to the + implementation very soon. + +Mon Apr 15 11:29:49 2002 Ossama Othman <ossama@uci.edu> + + * ace/config-win32-msvc-7.h (ACE_NEEDS_FUNC_DEFINITIONS): + + MSVC 7 requires "hidden" functions/methods to be defined. A + declaration is not enough. Fixes link-time "unresolved symbol" + errors when using the ACE_UNIMPLEMENTED_FUNC macro in exported + templates. + +Mon Apr 15 16:20:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Timer_Queue_Adapters.{h,i,cpp}: + Made several additions to the ACE_Thread_Timer_Queue_Adapter + - Make it possible to pass a timer queue instance to the + ACE_Thread_Timer_Queue_Adapter instance using the constructor + - Make it possible to get/set the timer queue based on pointers, the + get method with a & is still available, but is marked as deprecated + - Make the thr_id method const + - When the timer queue is created by the + ACE_Thread_Timer_Queue_Adapter then it is also deleted, if it is + passed or set afterwards it isn't deleted by + ACE_Thread_Timer_Queue_Adapter (just like in the ACE_Reactor). + +Mon Apr 15 10:22:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Containers_T.h: + Doxygen-ized some of the comments + +Mon Apr 15 08:49:23 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/Makefile.bor: + * examples/Timer_Queue/Makefile.bor: + * exmaples/Timer_Queue/Async_Timer_Queue_Test.bor: + * examples/Timer_Queue/Reactor_Timer_Queue_Test.bor: + * examples/Timer_Queue/Thread_Timer_Queue_Test.bor: + Added BCB makefiles + + * examples/Timer_Queue/main_async.cpp: + * examples/Timer_Queue/main_reactor.cpp: + * examples/Timer_Queue/main_thread.cpp: + Made the example compiling in an unicode build + +Mon Apr 15 07:39:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/IPC_SAP/SOCK_SAP/CPP-unclient.cpp: + * examples/IPC_SAP/SOCK_SAP/FD-unclient.cpp: + * examples/IPC_SAP/SOCK_SAP/FD-unserver.cpp: + Added include of OS.h to get ACE_TMAIN macro + + * apps/JAWS/server/HTTP_Server.h: + Added include of svc_export.h. + +Sun Apr 14 20:33:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Connector.h: + * ace/Process.h: + Doxygen-ized some of the comments + +Sun Apr 14 19:25:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Future.h: + Fixed small typing errors in comments + +Sun Apr 14 17:56:33 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/SOCK_Dgram_Bcast.h: + * ace/SV_Message.h: + * ace/SV_Message_Queue.h: + Doxygen-ized some of the comments + +Sun Apr 14 11:59:12 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/SPIPE_Addr.h: + * ace/Filecache.h: + * ace/Time_Request_Reply.h: + * ace/TLI.h: + Doxygen-ized some of the comments + +Sat Apr 13 19:28:40 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/Basic_Types.h: Added include of pthread.h to pick + up typedef of pthread_key_t. + +Sat Apr 13 17:16:10 UTC 2002 Don Hinton <dhinton@ieee.org> + + * examples/Service_Configurator/IPC-tests/server/Handle_L_SPIPE.h: + * examples/ASX/CCM_App/CCM_App.cpp: + + Added include of svc_export.h. + +Sat Apr 13 15:42:03 UTC 2002 Don Hinton <dhinton@ieee.org> + + * tests/DLL_Test.{h,cpp}: + * tests/DLL_Test_Impl.{h,cpp}: Added methods to test the + the malloc/free and strnew/strdelete methods below. Also + removed use auto_ptr and added a destroy method to delete + the object within the dll/heap it was allocated. + + * ace/OS_Memory.{inl,cpp}: Changed malloc(), calloc() + realloc(), and free() to be non-inlined to avoid the heap + problem on Windows. + + * ace/ACE.{h,i,cpp}: Changed strnew() to be non-inlined and + added strdelete() for the same reason. + +Sat Apr 13 14:33:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Configuration.h: + * ace/Containers_T.h: + * ace/High_Res_Timer.h: + * ace/Object_Manager.h: + * ace/Reactor.h: + * ace/Reactor_Impl.h: + * ace/Select_Reactor_T.h: + * ace/WFMO_Reactor.h: + Improved doxygen comments. Added @deprecated to methods that + are deprecated so that in the doxygen description this is also + clearly list. + +Sat Apr 13 03:32:52 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/Proactor.h: + + Moved OS.h and ACE_export.h outside the #ifdef ACE_WIN32. + ACE_export.h beacuse the #else part uses it and OS.h because + sig_atomic_t is used in one of the dummy class's signatures. If + others have other opinions, please fix the case for the #else + part also. Fixes Tru64 bustage. + +Fri Apr 12 19:15:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Asynch_Acceptor.cpp (parse_address): Set the entire address + (address and port) instead of just the IP address part. + Thanks to Alex Libman <alibman@ihug.com.au> for this fix. + +Fri Apr 12 18:00:41 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/Log_Msg.h: Added missing includes. + + * ace/Basic_Types.h: Removed erroneous ENOMEM definition. + +Fri Apr 12 10:02:58 2002 Priyanka Gontla <pgontla@ece.uci.edu> + + * ace/Based_Pointer_T.h: + Included ace/Trace.h to fix the compilation errors about + undeclared ACE_TRACE. + +Fri Apr 12 08:56:46 2002 Priyanka Gontla <pgontla@ece.uci.edu> + + * tests/DLL_Test_Impl.cpp: + Included OS_Errno.h to fix the compilation error about + undeclared errno. + +Fri Apr 12 15:03:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * etc/*.doxygen: + Set JAVADOC_AUTOBRIEF to YES. From now on, when there is only + javadoc style comments, there first line until the first '.' is + used as brief comment. In the brief class description a lof of more + methods will have a description. + + * Timer_Heap_T.h: + Change the comment style for method remove_first from doxygen style + to javadoc style + +Fri Apr 12 13:32:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/WFMO_Reactor.h: + * ace/Timer_List_T.h: + * ace/Timer_Heap_T.h: + * ace/Timer_Hash_T.h: + * ace/Timeprobe_T.h: + Minor improvements doxygen comments + + * ace/POSIX_Asynch_IO.h: + Doxygen-ized some of the comments + +Fri Apr 12 00:15:32 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_sunos5_sunc++.GNU: + * ace/config-sunos5.5.h: + + Reverted to status quo. If and when things go fine, I will + re-enable them. Too many things breaking currently. + +Thu Apr 11 20:37:53 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/OS.i (ftell): + * ace/OS.h: + + Added new wrappers for ftell, fgetpos & fsetpos. + + * ACEXML/common/FileCharStream.cpp: + * ACEXML/common/FileCharStream.h: + + Implemented the previously unavailable available() call. + +Thu Apr 11 20:06:41 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/display_logfile.cpp: Spacing/line length + changes so it fits on the book pages. + +Thu Apr 11 18:29:27 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/Makefile: New Makefile to build display_logfile. + + * examples/C++NPv2/display_logfile.cpp: Finished and debugged on Linux. + +Thu Apr 11 22:07:40 UTC 2002 Don Hinton <dhinton@ieee.org> + + * netsvcs/lib/Client_Logging_Handler.cpp: Added missing + Log_Record.h include. + + * tests/Svc_Handler_Test.cpp: + Changed ACE_Log_Record::MAXLOGMSGLEN to ACE_MAXLOGMSGLEN. + +Thu Apr 11 20:04:22 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/ARGV.h: Rolled back the change and readded ACE.h. + Too many error down the line to fix. + + * ace/Sample_History.cpp: Added include of OS.h. + + * apps/apps/gperf/src/Options.h: Added include of OS.h. + +Thu Apr 11 14:37:09 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/config-win32-common.h: Added ACE_LACKS_PARAM_H. + +Thu Apr 11 18:21:25 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/Basic_Types.h: Added include of <sys/param.h> + to pickup MAXPATHLEN if available and get rid of warning. + +Thu Apr 11 17:37:20 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/Basic_Types.h: Added missing typedefs for pid_t and + ssize-t for WIN32 builds needed for my Log_Msg changes + below. + +Thu Apr 11 14:42:32 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/ARGV.h: Replaced include of ACE.h with config-all.h. + + * ace/Addr.cpp: + * ace/OS_Dirent.cpp: + * ace/Trace.cpp: + + Added include of OS.h. + + * apps/gperf/src/Bool_Array.cpp: Changed ACE_OS::memset() to + ACE_OS_String::memset() and added include of OS_String.h. + + * tests/Log_Msg_Test.cpp: + * examples/Logger/Acceptor-server.cpp: + * examples/Logger/simple-server/Logging_Handler.cpp: + * examples/Log_Msg/test_callback.cpp: + + Added includes of Log_Record.h. + + * ace/Log_Msg.{h,cpp}: + + Added new method, last_error_adapter(), and changed the + ACE_DEBUG, et al, macros to use new method instead of + ACE_OS::last_error(). Removed includes of OS.h and + Log_Record. Changed ACE_Log_Record::MAXLOGMSGLEN to + ACE_MAXLOGMSGLEN. Removed ACE_OS::cleanup_tss friend. + + * ace/Basic_Types.h: Added definition of MAXPATHLEN and + typedef of ACE_thread_t to prevent need to include OS.h + in Log_Msg.h above. + + * ace/Default_Constants.h: + * ace/OS.h: + + Moved definitions of ACE_MAXLOGMSGLEN, ACE_MAXTOKENNAMELEN, + and ACE_MAXCLIENTIDLEN here from OS.h. + +Thu Apr 11 14:26:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * Timer_Hash_T.h: + * Timer_Heap_T.h: + * Timer_List_T.h: + * Timer_Wheel_T.cpp: + * Timer_Wheel_T.h: + The timer queue classes schedule timers using absolute time. The + describtion in the implementaton files correctly described this, + but the header file comment and the argument name where describing + that the queues work on delta timers, so corrected this. + + * Timer_Queue_Adapters.cpp: + * Timer_Queue_Adapters.h: + Updated the describtion that the schedule method expects an absolute + time. + +Thu Apr 11 13:29:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/String_Base.h: + Fixed some small typo in comment + +Thu Apr 11 08:30:11 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/OS_String.inl (strnlen): + Fixed compile error. + +Thu Apr 11 08:16:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Timer_Queue_T.cpp: + Reverted my change of yesterday. Already the copy constructor is called + + * ace/Task.h: + * ace/Timer_Queue_Adapters.h: + Fixed some small typos in comment + + * ace/Capabilities.h: + Placed right comment with right method + +Wed Apr 10 20:00:57 2002 Ossama Othman <ossama@uci.edu> + + * ace/Proactor.h: + + Include "ace/ACE_export.h" to pull in definition of ACE_Export + macro. + +Wed Apr 10 19:53:01 2002 Ossama Othman <ossama@uci.edu> + + * ace/Argv_Type_Converter.cpp: + + Include "ace/OS_Errno.h" to pull in ENOMEM definition. + + (align_char_with_wchar): + + Fixed remaining ACE_OS::strcmp() to be ACE_OS_String::strcmp(). + +Wed Apr 10 19:37:58 2002 Ossama Othman <ossama@uci.edu> + + * apps/gperf/src/Bool_Array.h: + * apps/gperf/src/Iterator.h: + * apps/gperf/src/List_Node.h: + + No need to include "ace/OS.h". "ace/config-all.h" is enough. + +Wed Apr 10 19:25:39 2002 Ossama Othman <ossama@uci.edu> + + * apps/Gateway/Gateway/Config_Files.cpp (read_entry): + + Use "unsigned short" instead of the "u_short" typedef to avoid + including "ace/OS.h". + + * apps/Gateway/Gateway/Gateway.h: + * apps/Gateway/Gateway/Event_Channel.h: + * apps/Gateway/Gateway/Options.h: + * apps/Gateway/Peer/Options.h: + * examples/Service_Configurator/IPC-tests/server/Handle_Broadcast.h: + * examples/Service_Configurator/IPC-tests/server/Handle_L_CODgram.h: + * examples/Service_Configurator/IPC-tests/server/Handle_L_Dgram.h: + * examples/Service_Configurator/IPC-tests/server/Handle_L_FIFO.h: + * examples/Service_Configurator/IPC-tests/server/Handle_L_Pipe.h: + * examples/Service_Configurator/IPC-tests/server/Handle_L_Stream.h: + * examples/Service_Configurator/IPC-tests/server/Handle_R_Dgram.h: + * examples/Service_Configurator/IPC-tests/server/Handle_R_Stream.h: + * examples/Service_Configurator/IPC-tests/server/Handle_Timeout.h: + * examples/Service_Configurator/Misc/Timer_Service.h: + * netsvcs/clients/Naming/Client/Client_Test.h: + * netsvcs/clients/Naming/Dump_Restore/Dump_Restore.h: + * tests/DLL_Test_Impl.cpp: + + Include "ace/svc_export.h" to pull in ACE_Svc_Export macro. + +Wed Apr 10 18:57:11 2002 Ossama Othman <ossama@uci.edu> + + * netsvcs/lib/Client_Logging_Handler.h: + * netsvcs/lib/Name_Handler.h: + * netsvcs/lib/Server_Logging_Handler.h: + * netsvcs/lib/TS_Clerk_Handler.h: + * netsvcs/lib/TS_Server_Handler.h: + + Include "ace/svc_export.h" to pull in ACE_Svc_Export macro. + +Wed Apr 10 19:37:12 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/config-g++-common.h: + + Fixed a missing #define. This should fix the Lynx PPC build. + +Wed Apr 10 17:23:09 2002 Ossama Othman <ossama@uci.edu> + + * ace/Auto_IncDec_T.h: + + Include "ace/Global_Macros.h" to pull ACE_UNIMPLEMENTED_FUNC + macro definition. + +Wed Apr 10 19:28:27 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv2/display_logfile.cpp: First cut at the C++NPv2 + Streams chapter code for displaying a logfile written from any + C++NPv1 or C++NPv2 logging daemon. + +Wed Apr 10 17:38:31 2002 Steve Huston <shuston@riverace.com> + + * ace/README: New config macros ACE_HAS_STRNLEN and ACE_HAS_WCSNLEN. + + * ace/OS_String.{h inl}: New method ACE_OS_String::strnlen finds the + length of a string with a specified maximum length. Mimicks the + GNU strnlen(3) and wcsnlen(3) functions. + + * ace/config-linux-common.h: Added ACE_HAS_STRNLEN and ACE_HAS_WCSNLEN + if _GNU_SOURCE is defined, else glibc headers doesn't declare them. + +Wed Apr 10 11:16:57 2002 Ossama Othman <ossama@uci.edu> + + * ace/Addr.h (hash): + * ace/Addr.i (hash): + + Use "unsigned long" instead of the "u_long" typedef to avoid + including "ace/OS.h". + +Wed Apr 10 11:07:42 2002 Ossama Othman <ossama@uci.edu> + + * ace/Method_Request.h: + + Fixed remaining "u_long" to "unsigned long" change. + +Wed Apr 10 11:04:56 2002 Ossama Othman <ossama@uci.edu> + + * ace/Array_Base.h: + + Include "ace/Global_Macros.h" to pull in the ACE_DES_* macros + that required by the inlined ACE_Array_Base destructor. + +Wed Apr 10 11:01:56 2002 Ossama Othman <ossama@uci.edu> + + * ace/Global_Macros.h: + * ace/OS.h: + + Moved ACE_DES_* macros to Global_Macros.h. This allows some + sources to avoid including ace/OS.h. + +Wed Apr 10 10:54:58 2002 Ossama Othman <ossama@uci.edu> + + * ace/OS.h: + * ace/Time_Value.h: + + Moved "time" related includes from OS.h to Time_Value.h. Fixes + build problems on Unix platforms. + +Wed Apr 10 10:40:36 2002 Ossama Othman <ossama@uci.edu> + + Inter-header dependency reductions: + + * ace/Auto_IncDec_T.h: + * ace/Auto_Ptr.h: + * ace/Based_Pointer_T.h: + * ace/Bound_Ptr.h: + * ace/Cached_Connect_Strategy_T.h: + * ace/Cache_Map_Manager_T.h: + * ace/Caching_Strategies_T.h: + * ace/Caching_Utility_T.h: + * ace/Capabilities.h: + * ace/Cleanup_Strategies_T.h: + * ace/Containers_T.h: + * ace/CORBA_macros.h: + * ace/Dynamic_Service.h: + * ace/Env_Value_T.h: + * ace/Managed_Object.h: + * ace/Map.h: + * ace/Map_Manager.h: + + No need to include "ace/OS.h". "ace/config-all.h" is enough. + + * ace/Asynch_Acceptor.h: + + No need to include "ace/OS.h". "ace/Default_Constants.h" is + enough. + + * ace/Addr.h: + * ace/Arg_Shifter.h: + * ace/Argv_Type_Converter.h: + * ace/Based_Pointer_Repository.h: + * ace/Containers.h: + * ace/Containers.cpp: + * ace/Init_ACE.h: + * ace/Trace.h: + + No need to include "ace/OS.h". "ace/ACE_export.h" is enough. + + * ace/RB_Tree.h: + + No need to include "ace/OS.h". "ace/Global_Macros.h" is + enough. + + * ace/Global_Macros.h: + + Moved ACE_GUARD macros to this file, meaning that it is no + longer necessary to include "ace/OS.h" just to get those + macros. This should save on pre-processing times for some + sources. + + Include "ace/OS_Export.h" to pull in the ACE_OS_Export macro + definitions. + + * ace/Arg_Shifter.cpp: + * ace/Argv_Type_Converter.cpp: + + Include "ace/OS_String.h" to pull in static string manipulation + methods. + + Changed all ACE_OS string method calls to ACE_OS_String. + + * ace/Trace.cpp: + + Minor include file cosmetic tweaks. + + * ace/Time_Value.h: + * ace/Time_Value.inl: + * ace/Time_Value.cpp: + + Moved ACE_Time_Value class and related macros/constants to these + files. Files that only need the ACE_Time_Value class + declaration need only include this header instead of "ace/OS.h", + thus reducing pre-processing times. + + * ace/OS.h: + + Moved ACE_Time_Value class and ACE_GUARD macros out of this + header. + + * ace/OS.i: + * ace/OS.cpp: + + Moved ACE_Time_Value methods out of the files into the new + Time_Value.* files. + + * ace/Init_ACE.cpp: + + Use "unsigned int" instead of the "u_int" typedef to avoid + including "ace/OS.h". + + * ace/Hashable.h: + * ace/Hashable.cpp: + * ace/Method_Request.h: + * ace/Method_Request.cpp: + + No need to include "ace/OS.h". "ace/ACE_export.h" is enough. + + Use "unsigned long" instead of the "u_long" typedef to avoid + including "ace/OS.h". + + * ace/Proactor.h: + + No need to include "ace/OS.h" to pull in ACE_Time_Value class + declaration. A forward declaration is enough. + + * ace/Argv_Type_Converter.inl: + * ace/CE_Screen_Output.h: + + Cosmetic changes to improve conformance to our coding + style/guidelines. + + * ace/config-all.h: + + No need to include "ace/ACE_export.h" and "ace/svc_export.h". + They should only be included by headers that need them. + + Do not include "ace/OS_Errno.h". Fixes a circular dependency. + + * ace/Copy_Disabled.h: + * ace/Dirent.h: + * ace/Log_Msg_Backend.h: + * ace/Refcountable.h: + * ace/Recyclable.h: + * ace/String_Base_Const.h: + * ace/Thread_Adapter.h: + + Include "ace/ACE_export.h" to pull in definition of ACE_Export + macro. + + * ace/OS_Export.h: + + Include "ace/config.h" instead of "ace/config-all.h" to fix a + circular include. + + * ace/OS_Errno.h: + + No need to include "ace/config.h". It is already included + indirectly by "ace/OS_Export.h". + + * ace/OS_Dirent.h: + + Include "ace/OS_Errno.h" to pull in errno definitions. + + * ace/Handle_Ops.h: + + Include "ace/ACE_export.h" to pull in ACE_Export macro + definition. + + Added missing "#pragma once". + + * ace/Makefile: + * ace/Makefile.bor: + + Added new Time_Value.* sources to these Makefiles. + +Wed Apr 10 13:32:15 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Token.h: + * ace/Timer_Queue_T.h: + Doxygen-ized some of the comments + + * ace/Select_Reactor_T.cpp: + Fixed typo in comment + +Wed Apr 10 07:44:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Timer_Queue_T.cpp: + Instead of creating a ACE_Time_Value with the default constructor + and immediatly filling it with the assignment operator, create the + ACE_Time_Value with the copy constructor + + * examples/Logger/simple-server/Logging_Acceptor.cpp: + * examples/Service_Configurator/IPC-tests/server/Handle_R_Stream.h: + * examples/Reactor/Misc/test_signals_2.cpp: + Added missing includes of Log_Msg.h. These where caused by the + compilation speedup of Don Hinton on April 5th. + + * ace/Name_Request_Reply.h: + +Tue Apr 9 22:59:20 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/config-sunos5.5.h: + + Added ACE_LACKS_STATIC_DATA_MEMBER_TEMPLATES to fix errors. + + * ace/config-sunos5.6.h: + + Removed definition of ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION as + we turn it on on the command line. + +Tue Apr 9 19:48:15 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_osf1_4.x_cxx.GNU (LDFLAGS): + + Moved the -hidden and -non_hidden from LIBS to LDFLAGS. The + linker doesn't like it seeing at the end. This fixes the + problems with Tru64. + +Tue Apr 9 12:43:04 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/auto_run_tests.lst: Added POA_BiDir to the list of tests to + be run and prevented MT_BiDir from running in minimum_corba + builds. + +Tue Apr 9 11:46:52 2002 Steve Huston <shuston@riverace.com> + + * ace/Message_Block.h: Doxygen tweaks to block type enum. + +Tue Apr 9 11:20:19 2002 Steve Huston <shuston@riverace.com> + + * netsvcs/Server_Logging_Handler_T.cpp (handle_logging_record): + Moved variable 'count' inside the #if 0 block so it's not flagged + as unused. + +Tue Apr 9 15:02:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Containers_T.h: + Fixed two small typos + + * ace/Thread_Manager.h: + Doxygen-ized some of the comments + +Tue Apr 9 03:04:19 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/rules.lib.GNU (VLIB): + + Ensure that templates in the repository get added to the static + library during creation. This build now uses a new variable + called TMPINCDIR which points to the template repository from + which the templates should be pulled in. + + * include/makeinclude/platform_osf1_4.x_cxx.GNU: + + Added support for different template instantiations. + + * include/makeinclude/platform_tru64_cxx.GNU: + + Removed some repeated flags from the command line. + + * ace/config-cxx-common.h: + + Added ACE_TEMPLATES_REQUIRE_SOURCE to pull in the template + definitions to ensure proper template instantiation. + +Tue Apr 9 00:52:42 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/config-sunos5.5.h (ACE_TEMPLATES_REQUIRE_SOURCE): + + Moved this from within a guard to enable it unconditionally. + + * ace/Obstack_T.cpp: + + Added guards to make it consistent with other usage of _T.cpp + files. + + * tests/test_config.h: + + Reverted the change Sun Apr 7 18:04:56 2002 Krishnakumar B + <kitty@cs.wustl.edu>. The original code was right. + + * include/makeinclude/platform_g++_common.GNU: + + Cosmetic fixes. Explicitly check for egcs. It dumps out version + information in a non-standard format confusing the make + conditionals. Should fix the RedHat Static build. + + * include/makeinclude/platform_sunos5_sunc++.GNU: + + Cosmetic fixes to fix build bustage. + +Tue Apr 9 00:05:37 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Logging_Strategy.h (ACE_Logging_Strategy): Changed "private" to + "protected" so subclasses can access the implementation. Thanks + to Martin Krumpolec <krumpolec@asset.sk> for reporting this. This + fixes bugid 1182. + +Mon Apr 8 13:18:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Activation_Queue.h: + Corrected location of doxygen comments so that the right comment + is placed below the right method + +Sun Apr 7 18:04:56 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * tests/test_config.h (randomize): + + Wrapped the initialization of singleton_ within a + ACE_LACKS_STATIC_DATA_MEMBER_TEMPLATES guard. This should fix + the problems on FreeBSD and under RedHat static. Thanks to Craig + Rodrigues <crodrigu@bbn.com> for reporting this. + +Sun Apr 7 06:02:26 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_lynxos.GNU: + * include/makeinclude/platform_macosx.GNU: + + Didn't know that these platforms used g++. Fixing them. + +Sun Apr 7 05:27:35 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_g++_common.GNU: + + Fixed a comment from the previous checkin. + + * include/makeinclude/platform_sunos5_sunc++.GNU: + * include/makeinclude/platform_sunos5_ghs.GNU: + * include/makeinclude/platform_sunos5_kcc.GNU: + + Changed these files to the new template instantiation mechanism. + + * tests/test_config.h: + + Added missing definition for the template member. + +Sun Apr 7 04:22:19 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/wrapper_macros.GNU(templates): + + Added option templates which can be used to specify the + instantiation policy desired. Allowed values are explicit, + automatic and used. Added flag TEMPLATE_FLAGS which should be + set to the proper value in each of the platform config files. + + * include/makeinclude/platform_g++_common.GNU: + + New file which holds the parsing and setting logic for GNU C++. + + * include/makeinclude/platform_aix4_g++.GNU(templates): + * include/makeinclude/platform_aix_g++.GNU: + * include/makeinclude/platform_chorus4.x_g++.GNU: + * include/makeinclude/platform_freebsd.GNU: + * include/makeinclude/platform_gnuwin32_common.GNU: + * include/makeinclude/platform_hpux_gcc.GNU: + * include/makeinclude/platform_irix5.3_g++.GNU: + * include/makeinclude/platform_irix6.x_g++.GNU: + * include/makeinclude/platform_linux.GNU: + * include/makeinclude/platform_osf1_4.x_g++.GNU: + * include/makeinclude/platform_psosim_g++.GNU: + * include/makeinclude/platform_rtems.x_g++.GNU: + * include/makeinclude/platform_sco5.0.0-mit-pthread.GNU: + * include/makeinclude/platform_sco5.0.0-nothread.GNU: + * include/makeinclude/platform_sunos4_g++.GNU: + * include/makeinclude/platform_sunos5_g++.GNU: + * include/makeinclude/platform_unixware_g++.GNU: + + For all of the platforms above, when using g++ versions 2.95.x, + 2.96, 3.0.x or 3.x or later, the instantiation policy is set to + automatic. + + * include/makeinclude/platform_vxworks5.x_g++.GNU: + + Only exception to the above. The kind of parsing that is done in + this file scares me. So I explicitly set it to "explicit". + + Thanks to Alex Libman <AlexL@rumblegroup.com> for motivating + this. + + If people with any of the above platforms can test if this works + for them, it would be great. If suddenly your files don't link, + just set templates="explicit" in platform_macros.GNU and all + will be fine. This one was easy :-) Next round of changes is for + the above platforms with different compilers. + +Sun Apr 7 04:02:23 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/config-g++-common.h (ACE_LACKS_STATIC_DATA_MEMBER_TEMPLATES): + + Moved common definitions for egcs into a single block. Wrapped + ACE_LACKS_STATIC_DATA_MEMBER_TEMPLATES inside appropriate + compiler checks. + + * ace/config-all.h: + + Added a negation before check for + ACE_HAS_GNUC_BROKEN_TEMPLATE_INLINE_FUNCTIONS. + + * ace/config-sunos5.5.h: + + Enabled ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION and + ACE_TEMPLATES_REQUIRE_SOURCE only if Sun CC version <= 5.0. + +Sun Apr 07 10:25:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/IPC_SAP/SOCK_SAP/FD-unserver.cpp: + * examples/IPC_SAP/SOCK_SAP/FD-unclient.cpp: + * examples/IPC_SAP/SOCK_SAP/CPP-unclient.cpp: + Added include of Log_Msg.h to fix build errors in BCB + +Fri Apr 05 21:17:44 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/CDR_Stream.cpp: Initialized a null wstring properly. + + * tests/CDR_Test.cpp (test_get): Added a test case for a + marshalling and unmarshalling a zero length wstring. + + The above changes fix [Bug 1169]. Thanks to Duane Binder + <duane.binder@veritas.com> for suggesting these fixes. + +Fri Apr 5 19:20:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Proactor.h: Add #include "ace/OS.h" in the "non-AIO" section + to get the ACE_Time_Value class needed for the stubbed-out class. + + * ace/config-aix-4.x.h: Add some commentary about ACE_HAS_AIO_CALLS. + + * ace/Service_Manager.cpp: Added #include "ace/Log_Msg.h" to fix a + compile problem on AIX. + +Fri Apr 5 11:30:00 2002 Justin Michel <michel_j@ociweb.com> + + * ace/OS.i: + * ace/config-win32-common.h: + + SO_REUSEADDR fix brought over from 1.2a + FD_SETSIZE fix brought over from 1.2a + Thanks to Juergen Pfreundt <Juergen.Pfreundt@gft.com> for + motivating this. + +Fri Apr 5 07:45:54 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Local_Tokens.cpp: Fixed a compile error. + +Fri Apr 5 12:51:11 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/ACE.i (strnew): Added null pointer check to + the wchar_t version so its behavior matches the char + version. + +Fri Apr 5 11:40:00 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/ATM_Acceptor.h + * ace/ATM_Addr.cpp + * ace/ATM_Addr.h + * ace/ATM_Connector.cpp + * ace/ATM_Connector.h + * ace/ATM_Params.h + * ace/ATM_QoS.h + * ace/ATM_Stream.h + * ace/Asynch_Acceptor.h + * ace/Asynch_IO.h + * ace/Asynch_IO_Impl.cpp + * ace/Asynch_IO_Impl.h + * ace/CE_Screen_Output.cpp + * ace/CE_Screen_Output.h + * ace/FlReactor.cpp + * ace/FlReactor.h + * ace/IOStream.h + * ace/LSOCK.cpp + * ace/LSOCK.h + * ace/LSOCK_Acceptor.cpp + * ace/LSOCK_Acceptor.h + * ace/LSOCK_CODgram.cpp + * ace/LSOCK_CODgram.h + * ace/LSOCK_Connector.cpp + * ace/LSOCK_Connector.h + * ace/LSOCK_Dgram.cpp + * ace/LSOCK_Dgram.h + * ace/LSOCK_Stream.cpp + * ace/LSOCK_Stream.h + * ace/Local_Tokens.cpp + * ace/Local_Tokens. + * ace/Msg_WFMO_Reactor.h + * ace/POSIX_Asynch_IO.h + * ace/POSIX_Proactor.h + * ace/Proactor.cpp + * ace/Proactor.h + * ace/Proactor_Impl.h + * ace/QtReactor.h + * ace/Reactor.cpp + * ace/SUN_Proactor.h + * ace/TkReactor.cpp + * ace/TkReactor.h + * ace/UNIX_Addr.cpp + * ace/UNIX_Addr.h + * ace/WFMO_Reactor.cpp + * ace/WFMO_Reactor.h + * ace/WIN32_Asynch_IO.h + * ace/WIN32_Proactor.h + * ace/XTI_ATM_Mcast.h + * ace/XtReactor.cpp + * ace/XtReactor.h + Moved all includes inside the platform- or feature-specific + macro guards and added config-all.h to the headers in order + to speed up compiles when the guard isn't defined. + +Fri Apr 5 08:51:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * apps/Gateway/Peer/Peer.h: + Added missing explicit template instantion exports for msvc7. + Thanks to Tom Phan <tomp@telhub.com> for reporting this + +Thu Apr 4 21:01:11 2002 Carlos O'Ryan <coryan@atdesk.com> + + * ace/SOCK_Dgram.cpp: + I accidentally left out some code in my last change, without it + platforms that lack IPV6 support will probably break. Whoopsie. + +Thu Apr 4 20:39:51 2002 Carlos O'Ryan <coryan@atdesk.com> + + * bin/g++dep: + Remove bogus path(s) to find gcc, they were site-specific, + host-specific and version-specific, and then outdated at + that. The developer better has a decent version of gcc in her + PATH already, or the 'make depend' commands simply won't work. + +Thu Apr 4 13:17:52 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/auto_run_tests.lst: Added a new test to the list of tests + that need to be run. + +Thu Apr 4 11:52:45 2002 Carlos O'Ryan <coryan@atdesk.com> + + * ace/SOCK_Dgram.cpp: + Add support for anonymous PF_LOCAL/SOCK_DGRAM (aka + PF_UNIX/SOCK_DGRAM) sockets. As the class stood before these + changes it was impossible to create such sockets, forcing people + to choose a binding address even for sockets that are used only + to send messages. For PF_INET this is not a big deal, because + the OS (or ACE::bind_port) can choose a port for the + application. But there is no such luck for PF_LOCAL sockets, + where the application has to choose a filename for the socket, + functions like ACE_OS::tempnam() or ACE_OS::mktemp() are more + trouble than they are worth, as they open a security can of + worms. + This fixes bug 1179. + +Thu Apr 4 10:30:54 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/RMCast/Makefile: + * ace/SSL/Makefile: + * ace/Makefile: Updated dependency. Other directories need this + too. Will get to them before the beta. + +Thu Apr 4 11:05:48 2002 Carlos O'Ryan <coryan@uci.edu> + + * ace/Select_Reactor_Base.cpp: + Apply proposed patch for bug 1175, i.e. a possible deadlock + condition when ACE_HAS_REACTOR_NOTIFICATION_QUEUE is defined. + +Thu Apr 4 08:32:26 UTC 2002 Don Hinton <dhinton@ieee.org> + + * ace/CDR_Stream.cpp (ctor): + Modified default ctor to always add ACE_CDR::MAX_ALIGNMENT to + the size parameter since it is always required since the + subsequent call to ACE_CDR::mb_align() might advance the + (rd|wr)_ptr's up to ACE_CDR::MAX_ALIGNMENT-1 bytes. + +Wed Apr 3 18:01:33 2002 Steve Huston <shuston@riverace.com> + + * ace/SOCK_Dgram.cpp (shared_open): Add PF_INET6 to the test for + need to bind an unused port if ACE_HAS_IPV6 is defined. + +Wed Apr 3 11:04:37 2002 Steve Huston <shuston@riverace.com> + + * ace/Handle_Ops.{h, cpp} (handle_timed_open): + * ace/SPIPE_Connector.{h cpp} (ctor, connect): Added optional + LPSECURITY_ATTRIBTUES arg, defaults to 0. Allows Windows users + to set security for the new file/pipe open. + + * ace/SPIPE_Acceptor.{h cpp} (ctor, open): Added optional + LPSECURITY_ATTRIBUTES arg, defaults to 0, same as above. + Also added a LPSECURITY_ATTRIBUTES member for NT4 and up. + The LPSECURITY_ATTRIBUTES passed in is remembered across + named pipe instances, and is used to create each new instance. + Doxygen-ized the comments. + +Wed Apr 3 10:15:00 2002 Si Mong Park <spark@ociweb.com> + + * tests/Atomic_Op_Test_WinCE.vcp: + * tests/Auto_IncDec_Test_WinCE.vcp: + * tests/Barrier_Test_WinCE.vcp: + * tests/Basic_Types_Test_WinCE.vcp: + * tests/Bound_Ptr_Test_WinCE.vcp: + * tests/Buffer_Stream_Test_WinCE.vcp: + * tests/Cached_Accept_Conn_Test_WinCE.vcp: + * tests/Cached_Conn_Test_WinCE.vcp: + * tests/Cache_Map_Manager_Test_WinCE.vcp: + * tests/Capabilities_Test_WinCE.vcp: + * tests/CDR_Array_Test_WinCE.vcp: + * tests/CDR_File_Test_WinCE.vcp: + * tests/CDR_Test_WinCE.vcp: + * tests/Collection_Test_WinCE.vcp: + * tests/Conn_Test_WinCE.vcp: + * tests/DLList_Test_WinCE.vcp: + * tests/DLL_Test_WinCE.vcp: + * tests/Dynamic_Priority_Test_WinCE.vcp: + * tests/Enum_Interfaces_Test_WinCE.vcp: + * tests/Future_Set_Test_WinCE.vcp: + * tests/Future_Test_WinCE.vcp: + * tests/Handle_Set_Test_WinCE.vcp: + * tests/Hash_Map_Bucket_Iterator_Test_WinCE.vcp: + * tests/Hash_Map_Manager_Test_WinCE.vcp: + * tests/High_Res_Timer_Test_WinCE.vcp: + * tests/Lazy_Map_Manager_Test_WinCE.vcp: + * tests/Logging_Strategy_Test_WinCE.vcp: + * tests/Log_Msg_Test_WinCE.vcp: + * tests/Malloc_Test_WinCE.vcp: + * tests/Map_Manager_Test_WinCE.vcp: + * tests/Map_Test_WinCE.vcp: + * tests/Max_Default_Port_Test_WinCE.vcp: + * tests/Mem_Map_Test_WinCE.vcp: + * tests/MEM_Stream_Test_WinCE.vcp: + * tests/Message_Block_Test_WinCE.vcp: + * tests/Message_Queue_Notifications_Test_WinCE.vcp: + * tests/Message_Queue_Test_Ex_WinCE.vcp: + * tests/Message_Queue_Test_WinCE.vcp: + * tests/MM_Shared_Memory_Test_WinCE.vcp: + * tests/MT_Reactor_Timer_Test_WinCE.vcp: + * tests/MT_SOCK_Test_WinCE.vcp: + * tests/Naming_Test_WinCE.vcp: + * tests/New_Fail_Test_WinCE.vcp: + * tests/Notify_Performance_Test_WinCE.vcp: + * tests/Object_Manager_Test_WinCE.vcp: + * tests/OrdMultiSet_Test_WinCE.vcp: + * tests/OS_Test_WinCE.vcp: + * tests/Priority_Buffer_Test_WinCE.vcp: + * tests/Priority_Reactor_Test_WinCE.vcp: + * tests/Priority_Task_Test_WinCE.vcp: + * tests/Process_Manager_Test_WinCE.vcp: + * tests/RB_Tree_Test_WinCE.vcp: + * tests/Reactors_Test_WinCE.vcp: + * tests/Reactor_Exceptions_Test_WinCE.vcp: + * tests/Reactor_Notify_Test_WinCE.vcp: + * tests/Reactor_Performance_Test_WinCE.vcp: + * tests/Reactor_Timer_Test_WinCE.vcp: + * tests/Reader_Writer_Test_WinCE.vcp: + * tests/Recursive_Mutex_Test_WinCE.vcp: + * tests/Refcounted_Auto_Ptr_Test_WinCE.vcp: + * tests/Reverse_Lock_Test_WinCE.vcp: + * tests/Semaphore_Test_WinCE.vcp: + * tests/Service_Config_Test_WinCE.vcp: + * tests/Sigset_Ops_Test_WinCE.vcp: + * tests/Simple_Message_Block_Test_WinCE.vcp: + * tests/SOCK_Connector_Test_WinCE.vcp: + * tests/SOCK_Send_Recv_Test_WinCE.vcp: + * tests/SOCK_Test_WinCE.vcp: + * tests/SPIPE_Test_WinCE.vcp: + * tests/SString_Test_WinCE.vcp: + * tests/Svc_Handler_Test_WinCE.vcp: + * tests/Task_Test_WinCE.vcp: + * tests/Thread_Manager_Test_WinCE.vcp: + * tests/Thread_Mutex_Test_WinCE.vcp: + * tests/Thread_Pool_Reactor_Resume_Test_WinCE.vcp: + * tests/Thread_Pool_Reactor_Test_WinCE.vcp: + * tests/Thread_Pool_Test_WinCE.vcp: + * tests/Timeprobe_Test_WinCE.vcp: + * tests/Timer_Queue_Test_WinCE.vcp: + * tests/Time_Service_Test_WinCE.vcp: + * tests/Time_Value_Test_WinCE.vcp: + * tests/Tokens_Test_WinCE.vcp: + * tests/TSS_Test_WinCE.vcp: + * tests/Upgradable_RW_Test_WinCE.vcp: + Added aygshell.lib to the link option. + +Wed Apr 3 00:03:05 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_sunos5_g++.GNU (exceptions): + * include/makeinclude/platform_qnx_neutrino.GNU: + * include/makeinclude/platform_qnx_rtp_gcc.GNU: + + Removed redefinitions of ACE_HAS_GNUG_PRE_2_8 as they were + unnecessary. Some more files also seem to be abusing this flag. + But I don't have access to those exotic platforms. + +Tue Apr 2 19:36:31 2002 Steve Huston <shuston@riverace.com> + + * ace/OS_String.inl (strtok_r): Another variant of wcstok()... + Linux/glibc uses the 3-arg version of wcstok(), and says it's + from UNIX98 and ISO/ANSI C. + +Tue Apr 2 16:21:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Get_Opt.h: Doxygen-ized the comments. + +Tue Apr 02 15:12:10 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * bin/nightlybuilds/builds.lst: Removed KCC builds from the list. + +Tue Apr 2 14:02:06 2002 Chris Cleeland <cleeland_c@ociweb.com> + + * ace/SSL/SSL_Context.cpp (dh_params): Changed this to use the + OpenSSL 'BIO' abstraction for file i/o rather than file pointers. + Using file pointers caused problems on Win32 platforms, and BIOs + don't. So, despite the fact that it's ugly and pollutes ACE code + with an OpenSSL abstraction, at least it works. We'll have to + figure out a more palatable way of dealing with this eventually. + +Tue Apr 2 12:53:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/Argv_Type_Converter.cpp: + Fixed possible memory leak in the Dtor with incorrect counter of + argv. Thanks to Don Hinton for finding this bug. + +Mon Apr 1 18:27:46 2002 Steve Huston <shuston@riverace.com> + + * ace/OS_String.inl (strtok_r): Corrected decision to call + wcstok() or wcstok_r(). This fixes builds on HP-UX 11. +Mon Apr 1 16:35:29 2002 Steve Huston <shuston@riverace.com> + + * ace/Containers_T.h (ACE_Array): Added performance characteristics + and requirements table. + +Mon Apr 1 11:56:55 2002 Steve Totten <totten_s@ociweb.com> + + * ace/Process.h: + * ace/Process.cpp: + + Applied change from Rich Seibel <seibel_r@ociweb.com> to add a + reset for the command_line_calculated_ flag (three places) and + rewrote the description of command_line_buf(). + +Mon Apr 1 12:38:44 2002 Steve Huston <shuston@riverace.com> + + * ace/Asynch_Acceptor.h: Corrected @arg to be @a for Doxygen. + +Sun Mar 31 22:44:00 2002 Si Mong Park <spark@ociweb.com> + + * apps/FaCE/FaCE_OS.h: + * apps/FaCE/FaCE.cpp: + * apps/FaCE/Main.cpp: + Changed 'LPWSTR' to 'ACE_TCHAR*' and 'LPCTSTR' to 'const ACE_TCHAR*'. + Also removed fuzz no-checking header to enable fuzz again. + +Sun Mar 31 22:08:00 2002 Si Mong Park <spark@ociweb.com> + + * ChangeLog: + Fixed incorrect path for FaCE related files on prior change log + items. + +Sat Mar 30 08:58:57 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Asynch_Acceptor.cpp: (handle_accept): Removed a stray + ACE_DEBUG() call. + Thanks to Edan Ayal <edanayal@yahoo.com> for reporting this. + +Sun Mar 31 01:18:00 2002 Si Mong Park <spark@ociweb.com> + + * apps/FaCE/FaCE_OS.h: + * apps/FaCE/FaCE.cpp: + * apps/FaCE/Main.cpp: + Added "// FUZZ: disable check_for_tchar" as the code is only and + specifically for the WinCE. Thanks to Nanbor Wang for help. + + * apps/CE_ARGV.H: + * apps/CE_ARGV.CPP: + Minor lower/upper case change in the include statement. + +Sat Mar 30 22:19:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/OS.h: + Removed FaCE_MAIN definition originally added during WinCE port + and moved to FaCE_OS.h in the FaCE package since it is FaCE + specific macro. + + * apps/FaCE/Face_OS.h: + Contains FaCE_MAIN macro definition to set entry point on WinCE. + + * apps/FaCE/ReadMe.txt: + Updated installation part for FacE_OS.h. + +Fri Mar 29 17:35:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Proactor.h: Doxygen-ize some comments; correct the + close() comment (no I/O completion port is known at this level). + +Fri Mar 29 13:40:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/OS.h: + Minor fix for WinCE IDC name. + + * apps/FaCE/ACE.ico: + * apps/FaCE/ACE_Racer.bmp: + * apps/FaCE/CE_ARGV.CPP + * apps/FaCE/CE_ARGV.H + * apps/FaCE/CE_Screen_Output.cpp: + * apps/FaCE/CE_Screen_Output.h: + * apps/FaCE/FaCE.cpp: + * apps/FaCE/FaCE.h: + * apps/FaCE/FACE.ico: + * apps/FaCE/FaCE.rc: + * apps/FaCE/FaCE.vcp: + * apps/FaCE/FaCE.vcw: + * apps/FaCE/FaCENOACE.vcp: + * apps/FaCE/FaCENOACE.vcw: + * apps/FaCE/License.txt: + * apps/FaCE/Main.cpp: + * apps/FaCE/newres.h: + * apps/FaCE/ReadMe.txt: + * apps/FaCE/resource.h + * apps/FaCE/TAO.BMP: + A new front-end framework utility/plug-in for ACE on WinCE. + +Fri Mar 29 11:33:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/ace_dll.vcp: + * tests/Refcounted_Auto_Ptr_Test_WinCE.vcp: + * tests/Service_Config_DLL_WinCE.vcp: + Added few files into project. + + * tests/CE_fostream.h: + * tests/CE_fostream.cpp: + A class that simulates fostream on WinCE for CDR File test. + + * tests/Bound_Ptr_Test.cpp: + * tests/CDR_File_Test.cpp: + * tests/Log_Msg_Test.cpp: + * tests/Logging_Strategy_Test.cpp: + * tests/Malloc_Test.cpp: + * tests/MEM_Stream_Test.cpp: + * tests/MM_Shared_Memory_Test.cpp: + * tests/MT_SOCK_Test.cpp: + * tests/Priority_Task_Test.cpp: + * tests/Refcounted_Auto_Ptr_Test.cpp: + * tests/SOCK_Send_Recv_Test.cpp: + * tests/SOCK_Test.cpp: + * tests/Svc_Handler_Test.cpp: + * tests/Thread_Pool_Test.cpp: + Minor updates for WinCE build. + +Fri Mar 29 11:08:29 2002 Steve Huston <shuston@riverace.com> + + * ace/Asynch_Acceptor.h: Doxygen-ized the method comments. + +Fri Mar 29 02:50:33 2002 Craig Rodrigues <crodrigu@bbn.com> + + * examples/QOS/Diffserv/README: Update links to RFC's for + Expedited Forwarding. + +Fri Mar 29 08:32:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/Makefile.bor: + Added new Cached_Allocator_Test + + * tests/Cached_Allocator_Test.cpp: + Fixed unicode build errors + +Thu Mar 28 23:47:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/OS.h: + * ace/OS.cpp: + Fixed WinCE entry point definitions, and factored out Unicode format + checking parts as a separate function for both WinCE and Unicode builds. + + * ace/config-WinCE.h: + Commented out the part that 'forces' WinCE to use DLL only. The + 'commented out' part has been left for possible future reference. + + * tests/test_config.h: + Changed path settings for WinCE file system as CE only supports absolute- + path and does not have concept of './'. + + * tests/tests_WinCE.vcw: + * tests/ARGV_Test_WinCE.vcp: + * tests/Atomic_Op_Test_WinCE.vcp: + * tests/Auto_IncDec_Test_WinCE.vcp: + * tests/Barrier_Test_WinCE.vcp: + * tests/Basic_Types_Test_WinCE.vcp: + * tests/Bound_Ptr_Test_WinCE.vcp: + * tests/Buffer_Stream_Test_WinCE.vcp: + * tests/Cached_Accept_Conn_Test_WinCE.vcp: + * tests/Cached_Conn_Test_WinCE.vcp: + * tests/Cache_Map_Manager_Test_WinCE.vcp: + * tests/Capabilities_Test_WinCE.vcp: + * tests/CDR_Array_Test_WinCE.vcp: + * tests/CDR_File_Test_WinCE.vcp: + * tests/CDR_Test_WinCE.vcp: + * tests/Collection_Test_WinCE.vcp: + * tests/Conn_Test_WinCE.vcp: + * tests/DLList_Test_WinCE.vcp: + * tests/DLL_Test_DLL_WinCE.vcp: + * tests/DLL_Test_WinCE.vcp: + * tests/Dynamic_Priority_Test_WinCE.vcp: + * tests/Enum_Interfaces_Test_WinCE.vcp: + * tests/Future_Set_Test_WinCE.vcp: + * tests/Future_Test_WinCE.vcp: + * tests/Handle_Set_Test_WinCE.vcp: + * tests/Hash_Map_Bucket_Iterator_Test_WinCE.vcp: + * tests/Hash_Map_Manager_Test_WinCE.vcp: + * tests/High_Res_Timer_Test_WinCE.vcp: + * tests/Lazy_Map_Manager_Test_WinCE.vcp: + * tests/Logging_Strategy_Test_WinCE.vcp: + * tests/Log_Msg_Test_WinCE.vcp: + * tests/Malloc_Test_WinCE.vcp: + * tests/Map_Manager_Test_WinCE.vcp: + * tests/Map_Test_WinCE.vcp: + * tests/Max_Default_Port_Test_WinCE.vcp: + * tests/Mem_Map_Test_WinCE.vcp: + * tests/MEM_Stream_Test_WinCE.vcp: + * tests/Message_Block_Test_WinCE.vcp: + * tests/Message_Queue_Notifications_Test_WinCE.vcp: + * tests/Message_Queue_Test_Ex_WinCE.vcp: + * tests/Message_Queue_Test_WinCE.vcp: + * tests/MM_Shared_Memory_Test_WinCE.vcp: + * tests/MT_Reactor_Timer_Test_WinCE.vcp: + * tests/MT_SOCK_Test_WinCE.vcp: + * tests/Naming_Test_WinCE.vcp: + * tests/New_Fail_Test_WinCE.vcp: + * tests/Notify_Performance_Test_WinCE.vcp: + * tests/Object_Manager_Test_WinCE.vcp: + * tests/OrdMultiSet_Test_WinCE.vcp: + * tests/OS_Test_WinCE.vcp: + * tests/Priority_Buffer_Test_WinCE.vcp: + * tests/Priority_Reactor_Test_WinCE.vcp: + * tests/Priority_Task_Test_WinCE.vcp: + * tests/Process_Manager_Test_WinCE.vcp: + * tests/RB_Tree_Test_WinCE.vcp: + * tests/Reactors_Test_WinCE.vcp: + * tests/Reactor_Exceptions_Test_WinCE.vcp: + * tests/Reactor_Notify_Test_WinCE.vcp: + * tests/Reactor_Performance_Test_WinCE.vcp: + * tests/Reactor_Timer_Test_WinCE.vcp: + * tests/Reader_Writer_Test_WinCE.vcp: + * tests/Recursive_Mutex_Test_WinCE.vcp: + * tests/Refcounted_Auto_Ptr_Test_WinCE.vcp: + * tests/Reverse_Lock_Test_WinCE.vcp: + * tests/Semaphore_Test_WinCE.vcp: + * tests/Service_Config_DLL_WinCE.vcp: + * tests/Service_Config_Test_WinCE.vcp: + * tests/Sigset_Ops_Test_WinCE.vcp: + * tests/Simple_Message_Block_Test_WinCE.vcp: + * tests/SOCK_Connector_Test_WinCE.vcp: + * tests/SOCK_Send_Recv_Test_WinCE.vcp: + * tests/SOCK_Test_WinCE.vcp: + * tests/SPIPE_Test_WinCE.vcp: + * tests/SString_Test_WinCE.vcp: + * tests/Svc_Handler_Test_WinCE.vcp: + * tests/Task_Test_WinCE.vcp: + * tests/Thread_Manager_Test_WinCE.vcp: + * tests/Thread_Mutex_Test_WinCE.vcp: + * tests/Thread_Pool_Reactor_Resume_Test_WinCE.vcp: + * tests/Thread_Pool_Reactor_Test_WinCE.vcp: + * tests/Thread_Pool_Test_WinCE.vcp: + * tests/Timeprobe_Test_WinCE.vcp: + * tests/Timer_Queue_Test_WinCE.vcp: + * tests/Time_Service_Test_WinCE.vcp: + * tests/Time_Value_Test_WinCE.vcp: + * tests/Tokens_Test_WinCE.vcp: + * tests/TSS_Test_WinCE.vcp: + * tests/Upgradable_RW_Test_WinCE.vcp: + Project files for WinCE build on eMbedded Visual C++ 3.0. + Note that some tests run fine under emulator but not on the real + machine, and some tests run okay on WinCE 3.0 but not on Pocket PC 2002. + The missing tests are mostly not supported by WinCE 3.0/PPC 2002. + For example, WinCE does not have 'fork' or environment variables. + Also, WinCE supports memory mapped file; however, the method is so different + to other Windows platforms that it is really hard to make it work correctly + on current ACE mem-map function structure. + +Thu Mar 28 16:15:17 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/TTY_IO.h: Added the ACE_Export macro to the Serial_Params + struct. Thanks to Pavel Repin <pavel@repin.com> for reporting + this. + +Thu Mar 28 13:31:19 2002 Priyanka Gontla <pgontla@ece.uci.edu> + + * ace/POSIX_Asynch_IO.cpp (handle_close): + Fixed the warnings on Debian_Core build that were caused by the + changes earlier today. + +Thu Mar 28 16:14:39 2002 Steve Huston <shuston@riverace.com> + + * ace/OS.h (ACE_Time_Value::dump): Added comments that'll go to + the man page to explain why this is a no-op. + +Thu Mar 28 14:26:19 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * tests/Cached_Allocator_Test.cpp: Added a missing template + instantiation. + +Thu Mar 28 10:45:16 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * tests/tests.dsw: + * tests/Cached_Allocator_Test.dsp: Added this new MSVC project. + +Thu Mar 28 10:12:13 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Malloc_T.{h,i,cpp}: Added support for + ACE_Dynamic_Cached_Allocator. + Thanks to Jaroslaw Nozderko <jareknz@polbox.com> for + contributing this. + + * tests: Added Cached_Allocator_Test.cpp. Thanks to + Jaroslaw Nozderko <jareknz@polbox.com> for contributing this. + + * tests/Makefile (BIN): + * tests/run_test.lst: Added Cached_Allocator_Test. + +Thu Mar 28 06:15:22 2002 Alex Libman <AlexL@rumblegroup.com> + + * ace/POSIX_Asynch_IO.cpp, + ace/POSIX_Asynch_IO.h: + ACE_POSIX_AOICB_Asynch_Operation and ACE_POSIX_Asynch_Operation + merged in one class ACE_POSIX_Asynch_Operation + + Since POSIX_SIG_Proactor and SUN_Proactor are based on + POSIX_AIOCB_Proactor and both of them use + ACE_POSIX_AOICB_Asynch_Operation, there is no necessity to + support extra class tree. + + * ace/POSIX_Asynch_IO.cpp, + ace/POSIX_Asynch_IO.h: + ACE_POSIX_AOICB_Transmit_Handler and ACE_POSIX_Transmit_Handler + merged in ACE_POSIX_Transmit_Handler for same reason as previous + change. + + * ace/POSIX_Asynch_IO.cpp, + ace/POSIX_Asynch_IO.h: ACE_POSIX_Asynch_Accept merged with + ACE_POSIX_Asynch_Accept_Hanlder and redesigned and added + new class ACE_POSIX_Asynch_Accept_Task. + + POSIX_AIOCB_Proactor has new member + ACE_POSIX_Asynch_Accept_Task asynch_accept_task_. + + Task activation should be done from the most derived + constructors , when the final table of virtual functions is + built (simular case with notify_manager). + + // start asynch accept task + this->get_asynch_accept_task.start (); + + All POSIX_Proactors implementations ( AIOCB,SIG,SUN ) should + stop ACE_POSIX_Asynch_Accept_Task in their destructors to avoid + post_completions from based classes + + // stop asynch accept task + this->get_asynch_accept_task.stop (); + +Thu Mar 28 06:14:22 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/TTY_IO.cpp (Serial_Params): Zero out the values in the + Serial_Params constructor. Thanks to Pavel Repin + <pavel@repin.com> for reporting this. + +Wed Mar 27 20:00:31 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * include/makeinclude/platform_linux.GNU: Fixed a typo where + PLATFORM_XT_LDFLAGS was spelled incorrectly. Thanks to Eric + Eide and Ansgar Konermann <ansgar.konermann@inf.tu-dresden.de> + for reporting this. + +Wed Mar 27 18:36:14 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_linux.GNU (CXX_VERSION): + + Reorganised the switches a bit for handling the implicit + templates. + + * include/makeinclude/wrapper_macros.GNU (ACE_HAS_GNUG_PRE_2_8): + + If the compiler has g++ or any mutations of g++, test whether + the version is less than 2.8. If so set ACE_HAS_GNUG_PRE_2_8 + to 1 else set it to 0. Thanks to James Haiar <haiar@ll.mit.edu> + for reporting this. + +Wed Mar 27 16:32:55 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * tests/Reactor_Dispatch_Order_Test.cpp: Minor compilation fixes. + +Wed Mar 27 15:10:27 2002 Irfan Pyarali <irfan@cs.wustl.edu> + + * tests/Reactor_Dispatch_Order_Test: Added a new test to check the + order of dispatching of ACE Reactors. Order should be: timeout, + output, and then input. Currently, Select and WFMO Reactors are + tested. + + The following files we updated to include the new test: + + - tests/Makefile + - tests/Makefile.am + - tests/Makefile.bor + - tests/Reactor_Dispatch_Order_Test.dsp + - tests/Reactor_Dispatch_Order_Test.icc + - tests/icc.bat + - tests/run_test.lst + - tests/run_tests.bat + - tests/run_tests.psosim + - tests/run_tests_remote.lst + - tests/tests.dsw + - tests/tests.icp + + * ace/WFMO_Reactor.cpp (upcall): Changed the dispatch order to + match the Select_Reactor's dispatch order. The order now is: + + FD_WRITE + FD_CONNECT + FD_OOB + FD_READ + FD_CLOSE + FD_ACCEPT + FD_QOS + FD_GROUP_QOS + + Thanks to Steve Huston <shuston@riverace.com> for pointing this + out. + + * tests/icc.bat: Removed duplicates. + +Wed Mar 27 10:25:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/High_Res_Timer.{h,cpp}: + Fixed a few small typing errors in comments + +Tue Mar 26 13:55:19 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/OS.i (operator *=): Explicitly promote sec() to double before + multiplying it to prevent problems with overflow. Thanks to + Eric Page <Eric_S_Page@raytheon.com> for reporting this. This + fixes BugId [1174]. + +Tue Mar 26 06:50:58 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Synch.h (ACE_Null_Mutex): Added a dummy "lock_" variable + to make ACE_Condition<ACE_Null_Mutex> work properly. Thanks to + Ido Yellin <Ido.Yellin@Focusengine.com> for reporting this. + +Wed Mar 21 08:11:15 2002 Boris Kolpackov <bosk@ipmce.ru> + + * THANKS: Added Frank Rybak <rybak@ll.mit.edu> to the Hall of fame. + +Wed Mar 20 11:23:38 2002 Priyanka Gontla <pgontla@ece.uci.edu> + + * ace/Service_Manager.cpp (handle_input): + Moved the declaration of 'error' outside the do-while loop to + fix the 'error (undeclared)' error. + +Wed Mar 20 09:51:47 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Acceptor.cpp (handle_input): Updated the comments to point out that + svc_handler->close() is called in accept_svc_handler() and + activate_svc_handler() if a failure occurs. Thanks to Rainer + Lucas <rainer.lucas@fun.de> for motivating this. + +Wed Mar 20 06:44:02 2002 Douglas C. Schmidt <schmidt@siesta.cs.wustl.edu> + + * ace/Service_Manager.cpp (handle_input): Fixed a bug that prevents + an infinite loop. Thanks to Sandro Doro <sandro@dorogroup.com> + for reporting this. + + * include/makeinclude/platform_vxworks5.x_g++.GNU (HOST_DIR): Fixed + a typo where -mlongcal should be -mlongcall. Thanks to Erik + Johannes <erik_johannes@teseda.com> for reporting this. + +Wed Mar 20 14:26:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * docs/usage-bugzilla.html: + Updated the link to the Bugzilla project + +Tue Mar 19 15:17:25 2002 Dante J. Cannarozzi <djc2@cs.wustl.edu> + + * ace/Containers_T.h: with the help of Matt Hampton + <mph2@cs.wustl.edu> updated doxygen comments for container classes + with more details at the request of Steve Huston. + +Tue Mar 19 02:40:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/OS_String.h: + * ace/OS_String.inl: + * ace/OS_String.cpp: + Corrected proper ACE_HAS_REENTRANT_FUNCTIONS macro checking location + for the strtok_r_emulation functions. Debian compilation error was + because ACE_HAS_REENTRANT_FUNCTIONS was not defined in the configuration. + If ACE_HAS_REENTRANT_FUNCTIONS is defined, then emulation functions + should not be visible. + +Mon Mar 19 08:32:34 2002 Boris Kolpackov <bosk@ipmce.ru> + * THANKS: Added Chen Jian <jchen@huawei.com> to the Hall of fame. + +Sun Mar 17 18:22:12 2002 Craig Rodrigues <crodrigu@bbn.com> + + * include/makeinclude/platform_freebsd.GNU: Copy lines from + platform_linux.GNU. exceptions=1 is now the default, + unless overridden by the user, just like for Linux. + +Sun Mar 17 11:53:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/compiler.bor: + When doing a wchar build don't automatically define UNICODE and _UNICODE. + +Fri Mar 15 18:08:28 2002 Steve Huston <shuston@riverace.com> + + * ace/config-aix-4.x.h: Removed the optional setting of + ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION for Visual Age C++ + 5 without __TEMPINC__ set. This was an experiment to try and + rid the build of all the "duplicate symbol" warnings, and it + broke the incremental builds. + +Fri Mar 15 12:19:58 2002 Chad Elliott <elliott_c@ociweb.com> + + * include/makeinclude/platform_vxworks5.x_ghs.GNU: + + Added support for building with exceptions enabled. + +Fri Mar 15 05:59:45 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/SPIPE_Acceptor.cpp: Removed the ACE_ASSERT (result == FALSE) + macro call. Thanks to Oleg Burlachenko <ua_fireball@yahoo.com> + for reporting this. + +Fri Mar 15 00:16:06 2002 Craig Rodrigues <crodrigu@bbn.com> + + * include/makeinclude/platform_freebsd.GNU: + * include/makeinclude/platform_netbsd.GNU: + Make -fno-implicit-templates conditional, instead of default. + Thanks to Denis Otchenashko <oko@bank.gov.ua> for motivating + me to look into this and fix this. + +Thu Mar 14 12:50:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/OS_String.h: + * ace/OS_String.inl: + * ace/OS_String.cpp: + Fixed a skipping strtok_r_emulation definition when ACE_HAS_WCHAR is + defined. + +Thu Mar 14 11:26:41 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * performance-tests/Misc/context_switch_time.cpp: Renamed the + internally used macro, DEBUG, to ACE_DEBUG_CST to avoid clashing + with the compiler defined macro. Thanks to Allan S Iverson + <allaniverson@sprynet.com> for motivating the fix. + +Thu Mar 14 09:17:55 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Strategies_T.cpp (activate_svc_handler): Added a call to + destroy() the svc_handler if ACE::fork() fails. Thanks to + Rainer Lucas <rainer.lucas@fun.de> for reporting this. + + * ace/Configuration.cpp (remove_section): Made the code consistent + for all versions of Windows. Thanks to Jon Lambert + <jlsysinc@ix.netcom.com> for the fix. + + * examples/Service_Configurator/IPC-tests/client/remote_service_directory_test.cpp (ACE_TMAIN): Added + '\n' to the end of command strings so that the client won't hang. Thanks to + Marc M Adkins <Marc.M.Adkins@doorways.org> for this fix. + +Wed Mar 13 15:55:54 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/OS.h: EACCESS was misspelled! Thanks to Eric Eide + <eeide@cs.utah.edu> for reporting this. + +Wed Mar 13 20:50:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/OS.i: + Minor change to resolve signed/unsigned comparison warning. + + * ace/OS_String.h: + * ace/OS_String.inl: + * ace/OS_String.cpp: + Added wide-char version of strtok_r and strtok_r_emulation. + + * ace/Functor.h: + * ace/Functor.i: + Added hash related templates for ACE_ANTI_TCHAR type + when ACE_USES_WCHAR is defined. + + * ace/ace_dll.vcp: + * ace/ace_os_dll.vcp: Added missing files. + +Wed Mar 13 11:57:41 2002 Steve Huston <shuston@riverace.com> + + * ace/Configuration_Import_Export.cpp: Replace ACE_ASSERTs with + validity checks that set errno and return -1. ACE_ASSERT is + a little too drastic for a simple error, and the ACE_ASSERT + check is removed altogether when building with ACE_NDEBUG. + +Wed Mar 13 10:36:02 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/config-freebsd-pthread.h: Fix test for POSIX RT signals. + Thanks to Denis Otchenashko <oko@bank.gov.ua> for reporting + these problems. + +Wed Mar 13 10:21:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Configuration_Import_Export.cpp (ACE_Ini_ImpExp::import_config): + Remove the "else value = ACE_LIB_TEXT("")" for zero-length + values - value is already a zero-length string. + +Wed Mar 13 09:40:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Registry_Name_Space.cpp: + Fixed warnings about unused variables in BCB Unicode build + +Tue Mar 12 18:37:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Configuration_Import_Export.{h cpp} (ACE_Ini_ImpExp): + Changed behavior of ACE_Ini_ImpExp::import_config() to take + the entire string, with or without whitespace, without requiring + quotes around the string. This puts functionality on par with + regular Windows INI files. + + Also, on ACE_Ini_ImpExp::export_config(), don't add quotes to + string values that are exported. import_config() will still read + files exported previously (which have quotes around strings) + but when re-exported, the quotes will not be added. + +Tue Mar 12 17:43:32 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/CDR_Base.cpp (consolidate): If the incoming message block + does not have a chain of message blocks, there is no need to do + a deep copy, a refcount increment on the incoming datablock + would do. Thanks to Lothar Werzinger + <Werzinger.Lothar@krones.de> for the patches. + +Tue Mar 12 15:32:46 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * PROBLEM-REPORT-FORM (Subject): Added a request for the version + of winsock on Windows based OS's. + +Tue Mar 12 15:22:29 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/OS.i (sendv): When using winsock1, we transfer data by + sending one iovec at a time. If transfer of a buffer isnt + complete because the socket buffers got full, we need to drop + out of the loop that sends data. Thanks to Stephan Gudmundson + <stephang@netacquire.com> for providing this patch. + + * THANKS: Added Stephan Gudmundson to the Hall of fame. + +Mon Feb 25 14:06:45 2002 Chad Elliott <elliott_c@ociweb.com> + + * ace/config-hpux-11.00.h: + + Do not define __HP_aCC if using KCC. + + * include/makeinclude/platform_hpux_kcc.GNU: + + Add the --one-instantiation-per-object parameter to avoid + build problems with KCC. + + * tests/Dirent_Test.cpp: + + Modify this test to work with Chorus (as it does with VxWorks). + + * ace/config-aix-4.x.h: + + Defined ACE_HAS_USING_KEYWORD for AIX 4.x with Visual Age 5 or + later. This change came from Yan Dai <dai_y@ociweb.com> + +Tue Mar 12 11:53:53 2002 Chad Elliott <elliott_c@ociweb.com> + + * bin/perltest2cpp.pl: + + Corrected the code for array assignments. + + * bin/vxworks_modify.pl: + + Use ACE_ENV_* instead of TAO_ENV_*. + +Tue Mar 12 10:45:00 2002 Justin Michel <michel_j@ociweb.com> + + * ace/OS.i: + + Added call to ACE_OS::set_errno_to_last_error() in two places + where Winsock WSARecvFrom or WSASendTo were called. + +Tue Mar 12 09:06:00 2002 Si Mong Park <spark@ociweb.com> + + * etc/Svc_Conf_l.cpp.diff: + Incorrect file had been checked in on prior commit. + This is the correct file. + + * tests/Config_Test.cpp: + Fixed incorrect delete statement. + +Tue Mar 12 08:54:00 2002 Si Mong Park <spark@ociweb.com> + + * etc/Svc_Conf_l.cpp.diff: + Fixed CR/LF (DOS format) problem: no content changed. + +Tue Mar 12 13:52:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * THANKS : Added Oleg Kraynov <olegvkr@yahoo.com> + +Mon Mar 11 21:05:26 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ACEXML/parser/parser/Parser.cpp (parse): + + Added comment on the order of processing that should be done by + the parser. + +Mon Mar 11 20:21:38 2002 Steve Huston <shuston@riverace.com> + + * ace/config-all.h: Fixed ACE_NEW section to work with HP-UX 10.20, + aC++ A.01.27 with -AA option. + +Mon Mar 11 18:48:41 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv1/Logging_Client.cpp: Added a compile-time check + for Win32 to be sure ACE_HAS_STANDARD_CPP_LIBRARY is set, else + trying to use the new getline() method with an old cin won't work. + +Mon Mar 11 17:43:00 2002 Si Mong Park <spark@ociweb.com> + + * ace/Svc_Conf_l.cpp: + Minor comment change to make diff file. + + * etc/Svc_Conf_l.cpp.diff: + Updated diff file for Svc_Conf_l.cpp. + +Mon Mar 11 16:59:42 2002 Steve Huston <shuston@riverace.com> + + * tests/TP_Reactor_Test.h (new file): + * tests/TP_Reactor_Test.cpp: Moved declarations for Receiver, + Acceptor, Sender, and Connector classes to the new file + TP_Reactor_Test.h so Visual Age C++ can find them when + instantiating templates. + +Mon Mar 11 16:38:49 2002 Steve Huston <shuston@riverace.com> + + * tests/Reactor_Notify_Test.cpp: Added missing #include "ace/Reactor.h" + to fix Visual Age C++ compile error. + +Mon Mar 11 16:27:59 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv1/RT_Thread_Per_Connection_Logging_Server.cpp: + Added missing #include "ace/Auto_Ptr.h" to fix compile problem + for platforms w/o native auto_ptr (like MSVC). + +Mon Mar 11 14:23:00 2002 Si Mong Park <spark@ociweb.com> + + * tests/Config_Test.cpp: + Removed one of the delete statement that deletes twice and thus + causing an invalid access violation. + +Mon Mar 11 10:28:45 2002 Chad Elliott <elliott_c@ociweb.com> + + * bin/vxworks_modify.pl: + + Update this script to work correctly with the new CORBA + Environment style. + +Mon Mar 11 12:17:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/build_dll.bor: + * include/makeinclude/build_exe.bor: + * include/makeinclude/build_lib.bor: + * include/makeinclude/compiler.bor: + * include/makeinclude/decorator.bor: + * include/makeinclude/make_flags.bor: + * include/makeinclude/outputdir.bor: + Added support for a real unicode build with BCB builder. A real + unicode build can for example be done with: + make -f makefile.bor -DUNICODE + Doing a real unicode build means that the defines ACE_USES_WCHAR, + UNICODE and _UNICODE are set, diffent BCB object files are used + to link with and the compiler gets the extra compiler option -WU. + The dll names have a 'u' added so that unicode dll's are separated + from the non-unicode dll's. Also the output files are build in a + separate subdirectory. + + * ACE-INSTALL.html: + Explained the BCB make options that can be used with ACE + +Mon Mar 11 11:35:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Registry_Name_Space.cpp: + Changed ACE_USHORT16 to ACE_WSTRING_TYPE because the underlying type + of ACE_WString is ACE_USHORT16 with MSVC and with BCB it is + wchar_t. The define ACE_WSTRING_TYPE hides these differences. When + UNICODE and ACE_USES_WCHAR are defined BCB gave a compile error. + +Sun Mar 10 18:53:48 2002 Si Mong Park <spark@ociweb.com> + + * ace/Svc_Conf_l.cpp: + Commented out 'break' line to resolve unreachable statement warning. + +Sun Mar 10 05:54:30 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * THANKS : Added Val Dumitrescu <val.dumitrescu@am-beo.com> to the + Hal of Fame. + * bin/auto_run_tests.lst: Added a new test to the daily builds. + +Sat Mar 9 21:22:51 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/CDR_Stream.cpp: + * ace/CDR_Stream.i: + * ace/CDR_Stream.h: Added accessor methods for GIOP versions of + Input and Output CDR streams. + +Sat Mar 9 07:52:45 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/SPIPE_Addr.cpp (set): Changed a strcpy() to a strcat() to + fix the code. Thanks to Robert Laferriere + <robert.laferriere@med.ge.com> for reporting this. + +Fri Mar 8 22:29:36 2002 Si Mong Park <spark@ociweb.com> + + * ace/Argv_Type_Converter.cpp: + * ace/Argv_Type_Converter.h: + * ace/Argv_Type_Converter.inl: + new class to convert command line parameter type between char + and wchar_t. + + * ace/Basic_Types.h: + Set the endian type to 'little endian' for ARM processor - + Pocket PC 2002 platform. + + * ace/Log_Record.cpp: + * ace/Log_Record.h: + Removed print method that uses ACE_CE_Bridge since CE_Bridge is + no longer supported by ACE CE port. + + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: + Added Argv_Type_Converter class to the project. + + * ace/WFMO_Reactor.cpp: + * ace/WFMO_Reactor.h: + * ace/WFMO_Reactor.i: + Updated to support WinCE. + + * ace/Process.cpp: + Fixed incorrect call to CreateProcess for WinCE. + + * ace/config-WinCE.h: + Updated for WinCE Pocket PC 2002. + + * ace/ace.vcw: + * ace/ace_dll.vcp: + * ace/ace_os_dll.vcp: + New ACE project file for WinCE Pocket PC 2002 build on eMbedded + Visual C++ 3.0. + + * ace/Makefile: + * ace/Makefile.am: + * ace/Makefile.bor: + Added Argv_Type_Converter file. + + * ace/Configuration.cpp: + * ace/Memory_Pool.cpp: + Updated for the WinCE port. + + * ace/Read_Buffer.cpp: + * ace/Read_Buffer.h: + Disabled Ctor with ACE_HANDLE on WinCE since ACE_HANDLE is same + as FILE* on WinCE. + + * ace/OS.cpp: + * ace/OS.h: + * ace/OS.i: + Updated for WinCE port. Changes includes removal of CE Bridge + added CE argv process class, enabled many file IO functions + that formerly disabled on CE, bypassing QoS parts for CE, and + other CE specific declarations. + + * ace/Dynamic_Service.h: + * ace/Dynamic_Service.i: + Added 'instance' method uses ACE_ANTI_TCHAR type name for when + ACE_TCHAR is wchar_t. + + * ace/CE_Screen_Output.cpp: + * ace/CE_Screen_Output.h: + screen output helper class only for WinCE platform - should + not be used on any other platform. + + * ace/OS_Memory.h: + * ace/OS_String.h: + Changed to skip including stddef.h for WinCE. + + * ace/MEM_Acceptor.h: + * ace/MEM_Acceptor.i: + * ace/SOCK_Acceptor.cpp: + * ace/SOCK_Acceptor.h: + * ace/SOCK_Connector.cpp: + * ace/SOCK_Connector.h: + * ace/SOCK_Dgram_Mcast.h: + Updated to skip QoS part on WinCE. + + * ace/Svc_Conf.l: + * ace/Svc_Conf_l.cpp: + Fixed a problem giving a parse error on reading carriage return + token on Unicode formatted svc conf file. + + * ace/Registry.cpp: + Changed connect function for WinCE to return -1 because CE does + not allow direct registry connection. + + * ace/Sock_Connect.cpp: + Changed to use Iphlpapi library (standard on WinCE) on WinCE to + find IP address instead of searching registry since CE has so + many variations on registry settings. + + * ace/Log_Msg.cpp: + Minor change for WinCE on stderr. + +Fri Mar 8 19:45:31 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv1/Thread_Per_Connection_Logging_Server.cpp: + * examples/C++NPv1/RT_Thread_Per_Connection_Logging_Server.cpp: + Use auto_ptr<> to manage dynamically-allocated Thread_Args. + Thanks to Chris Uzdavinis <chris@atdesk.com> for this suggestion. + +Thu Mar 7 16:10:02 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * include/makeincludes/platform_linux_icc.GNU, + * ace/config-icc-common.h, + * ace/config-linux-common.h: Added support for the Intel C++ compiler + (icc). Thanks to Roger Tragins for contributing this. + +Thu Mar 7 11:17:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/RMCast/RMCast_Fragment_Test.cpp: + * tests/RMCast/RMCast_Membership_Test.cpp: + * tests/RMCast/RMCast_Reassembly_Test.cpp: + * tests/RMCast/RMCast_Reordering_Test.cpp: + * tests/RMCast/RMCast_Retransmission_Test.cpp: + * tests/RMCast/RMCast_UDP_Best_Effort_Test.cpp: + * tests/SSL/Thread_Pool_Reactor_SSL_Test.cpp: + Replaced main with ACE_TMAIN. + +Wed Mar 6 21:48:15 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_linux_kcc.GNU (CCFLAGS): + + Moved --one_instantiation_per_object flag as a common flag. + Otherwise building with debugging enabled doesn't work with + libraries built without debug information. + +Wed Mar 6 21:42:37 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * bin/ace-install: + + Make sure that the .inl files are also installed. Thanks to + Carsten Prescher<carsten.prescher@sysde.eads.net> for reporting this. + +Wed Mar 6 18:30:49 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv1/Thread_Per_Connection_Logging_Server.h: + Reimplemented the Logging_Server::run method. Because the + handle_connections() method spawns a new thread to run the + logging session, and it calls handle_data(), the call to + handle_data() from Logging_Server::run() is incorrect. Thanks + to Raghuram Shetty <Raghuram.Shetty@comverse.com> for reporting + this issue. + + * examples/C++NPv1/Reactive_Logging_Server.h: + Fix wait_for_multiple_events() to correctly check error return + from select(). Also, in handle_connections(), sync the code + with the book to clear acceptor handles from the active_handles_ + after accepting all ready connections. + + * examples/C++NPv1/Reactive_Logging_Server_Ex.h (handle_connections): + Clear acceptor's handle from active_read_handles_ to keep from + dispatching it as a data handle. See, this is why you should use + the Reactor framework - so be sure to buy vol 2 ;-) + + * THANKS: Added Raghuram Shetty to the Hall of Fame. + +Wed Mar 06 17:06:43 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * tests/Cached_Accept_Conn_Test.cpp: + * tests/Cached_Conn_Test.cpp: Replaced main with ACE_TMAIN. + +Wed Mar 6 14:43:34 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/OS.h: Reverted my previous "fix" to the argv argument type. + + Mon Jan 21 23:27:03 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + A program entry poing <code>main</code> can take any of the + three forms: + + int main (int argc, char *argv[]) + int wmain (int argc, wchar_t *argv[]) + int ACE_TMAIN (int argc, ACE_TCHAR *argv[]) + + Of them, the entry point <code>main</code> always gives you + the command line arguemnt in char strings form. The entry + point <code>wmain</code> currently can only be used under + Win32 and it returns the command line arguments in wchar + strings format. Defining the <code>ACE_TMAIN</code> as the + program entry point is the more portable form. The command + line arguments are given in char strings in most cases, + or wchar strings when <code>ACE_USES_WCHAR</code> is defined. + See <code>$ACE_ROOT/docs/wchar.txt</code> for more information + on ACE support on wchar. + + * docs/ACE-guidelines.html: Added a new guideline explaining which + main entry point to use, as above. + + * bin/main2TMAIN.pl: Script to replace entry points of the form + + main (int, ACE_TCHAR *[]) + + to + + ACE_TMAIN (int, ACE_TCHAR *[]) + + * *.cpp: Changed to use the new ACE_TMAIN. + +Tue Mar 06 14:30:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/MEM_Stream_Test.cpp: + Fixed the shutdown of this test. When _TEST_USES_THREADS is defined + we must do a wait() on the thread manager, else on the process + manager. + +Tue Mar 05 20:34:22 2002 Ossama Othman <ossama@uci.edu> + + * ace/OS_Thread_Adapter.cpp (invoke): + * ace/Thread_Adapter.cpp (invoke_i): + + Fixed "jump out of __finally block" warning emanating from + MSVC 7. + +Tue Mar 05 15:46:15 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/Memory_Pool.h: + * ace/Memory_Pool.i: Added a base_addr () method for all memroy + pool classes. For memory pools that don't have a base_addr and + will never remap the pool to a different area, this method + always return 0. + + * ace/Malloc_T.cpp: Reset this->cb_ptr_ after acquiring new memory + to memory_pool's base_addr (if base_addr != 0). This makes sure + the cb_ptr_ points to the right memory after a remap. Thanks to + Ariel Peltz <Arielp@bigbandnet.com> for pointing this out. + +Tue Mar 5 12:20:26 2002 Ossama Othman <ossama@uci.edu> + + * ace/INET_Addr.cpp (get_host_name_i, set): + + h_errno -> h_error. h_errno conflicts with a declaration in + Microsoft's Winsock headers. Thanks to Nanbor for pointing out + the problem. + +Tue Mar 05 09:26:46 2002 Ossama Othman <ossama@uci.edu> + + * ace/INET_Addr.cpp (get_host_name_i, set): + + Do not clobber errno with the h_errno value returned from + gethostbyaddr_r(). They are two distinct types of errors. + + Renamed "error" to "h_errno" where appropriate to make it + more obvious that errno should not be set to the value of + h_errno. Thanks to Felix Wyss <Felix.Wyss@inin.com> for + reporting this. + +Tue Mar 5 05:51:44 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Process.cpp (spawn): If fork()/exec() fail, call ACE_OS::_exit() + rather than ACE_OS::exit() to avoid destructors being called + that will yield hang problems. Thanks to Renjie Tang + <rtang@informatica.com>, Max V. Zinal <Zlat0@mail.ru>, and + Michael Hornok <Michael.Hornok@alcatel.com> for this suggestion. + This fixes BugID 1147. + +Tue Mar 5 12:24:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Auto_IncDec_T.h: + Corrected typing error in comment + +Mon Mar 4 19:59:31 2002 Steve Huston <shuston@riverace.com> + + * tests/Process_Strategy_Test.h: Added #include "ace/Strategies_T.h" + so Visual Age C++ can find ACE_Process_Strategy when instantiating + templates. + +Mon Mar 4 19:52:27 2002 Steve Huston <shuston@riverace.com> + + * tests/Message_Queue_Test_Ex.h (new file): + * tests/Message_Queue_Test_Ex.cpp: Moved declaration of + User_Class to .h file so Visual Age C++ can find it when + instantiating templates. + +Mon Mar 4 19:43:51 2002 Steve Huston <shuston@riverace.com> + + * tests/Framework_Component_Test.h (new file): + * tests/Framework_Component_Test.cpp: Moved declaration of + My_Singleton to .h file so Visual Age C++ can find it when + instantiating templates. + +Mon Mar 04 11:20:45 2002 Carlos O'Ryan <coryan@uci.edu> + + * ace/Copy_Disabled.h: + * ace/Copy_Disabled.cpp: + Add new helper class to disable copy constructors and assignment + operators. I simply got sick of writing this repetitive code: + + // private & undefined + Foo (const Foo &); + Foo &operator= (const Foo&); + + The new class makes life *much* easier, simply say: + + class Foo : private ACE_Copy_Disabled + + Isn't that cool? + + * ace/Makefile: + * ace/Makefile.am: + * ace/Makefile.bor: + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: + Add new file to the project files and Makefiles. + +Mon Mar 4 07:36:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/Token_Strategy_Test.cpp: + Fixed MSVC6 unicode build errors. + +Sun Mar 3 15:56:23 2002 Craig Rodrigues <crodrigu@bbn.com> + + * tests/Token_Strategy_Test.cpp: + Add template instantiations for ACE_Array, ACE_Array_Base, + ACE_Array_Iterator. + +Sun Mar 3 16:23:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/Token_Strategy_Test.cpp: + Fixed MSVC6 unicode build errors. + +Sun Mar 3 11:00:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/Token_Strategy_Test.dsp: + Added MSVC project for this test + + * tests/Token_Strategy_Test.cpp: + Corrected small typing errors in comment + + * tests/tests.dsw: + Added new Token_Strategy_Test.dsp + +Sat Mar 2 09:17:45 2002 Douglas C. Schmidt <schmidt@siesta.cs.wustl.edu> + + * ace/INET_Addr.cpp (set): Change errno = EINVAL to errno = error if + the call to ACE_OS::gethostbyname_r() fails since errno should + already have been set! Thanks to Felix Wyss <FelixW@inin.com> + for reporting this. + +Fri Mar 01 08:03:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Makefile.bor: + Added Swap.cpp to the list of installable files + + * ace/SSL/SSL_Context.cpp: + Added ACE_TEXT_CHAR_TO_TCHAR around first argument of ACE_OS::fopen + call to convert char to ACE_TCHAR to fix compile errors in unicode + build + + * include/makeinclude/ace_flags.bor: + Corrected typing error for new tao messaging library + +Wed Feb 27 13:50:20 2002 Jaiganesh Balasubramanian <jai@kelvar.ece.uci.edu> + + * ace/INET_Addr.cpp: + Pull back changes from last night. + +Wed Feb 27 06:00:37 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Synch.h: Clarified that ACE_Auto_Event, ACE_Manual_Event, and + ACE_Event all support process-scope locking, but only Win32 + supports global naming and system-scope locking. Thanks to + Kobi Cohen Arazi <kobi@mivzak.com> for motivating this change. + +Wed Feb 27 09:24:30 2002 Carlos O'Ryan <coryan@uci.edu> + + * include/makeinclude/platform_linux.GNU: + Pull back Jai's change from last night. It looks like an + accident to me, and it is breaking all the builds. + +Wed Feb 27 12:09:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ACE-INSTALL.html: + Added description about setting environment variable BCBVER to + the Borland building instructions + +Wed Feb 27 10:47:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/ace_flags.bor: + Added lines for new TAO_Messaging library + +Wed Feb 27 01:12:32 2002 Carlos O'Ryan <coryan@uci.edu> + + * ace/Swap.h: + * ace/Swap.inl: + * ace/Swap.cpp: + Add helper template to swap variables, very useful when + implementing exception neutral/safe classes. + +Wed Feb 26 00:38:50 2002 UTC Don Hinton <dhinton@ieee.org> + + * ace/Select_Reactor_Base.h: + * ace/Select_Reactor.cpp: + * ace/Select_Reactor_T.cpp: + + Removed ACE_SELECT_REACTOR_HAS_DEADLOCK_DETECTION since + it isn't used. + +Tue Feb 26 22:54:50 2002 UTC Don Hinton <dhinton@ieee.org> + + * ace/Synch.i (ACE_Noop_Token::queueing_strategy): + + Removed ACE_UNUSED_ARG and commented out the parameter + instead. Thanks to Ossama and Craig for pointing this out. + +Tue Feb 26 16:17:45 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/Thread_Manager.h (ACE_Thread_Manager): Fixed a typo in + doxygen document tag. Thanks to Brad Hoskins + <BHoskins@slo.newport.com> for reporting this. + +Tue Feb 26 10:22:39 2002 Steve Huston <shuston@riverace.com> + + * include/makeinclude/platform_sunos5_sunc++.GNU: Added the ability + to enable RTTI in the compat=4 case. Removed the explicit + inclusion of an installation-specific path to -L for compat=4. + Installation-specific adidtions/changes should go in the + installation's platform_macros.GNU file. + +Tue Feb 26 09:48:24 2002 Craig Rodrigues <crodrigu@bbn.com> + + * apps/drwho/Protocol_Manager.cpp: Replace #include "new.h" + with #include "ace/config.h" to eliminate gcc 3.1 warning + about deprecated header. + +Tue Feb 25 10:16:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/Token_Strategy_Test.cpp: + Fixed unicode compile error + +Mon Feb 25 19:30:54 2002 Steve Huston <shuston@riverace.com> + + * tests/Collection_Test.cpp: Moved the declaration of UglyThing + to a new file, Collection_Test.h. Visual Age C++ needs it in + a separate file to do auto template instantiation. + * tests/Collection_Test.h: New file. + +Mon Feb 25 19:27:41 2002 Steve Huston <shuston@riverace.com> + + * ace/Log_Msg.cpp (log): In handling for %t, fixed the feature + test that changes behavior for AIX 4.2 and earlier. This + section missed my long-time-ago change to the AIX OS + version constants, and ended up being used for all AIX versions + which is wrong. Thanks to Yan Dai <dai_y@ociweb.com> for + reporting this problem. + + * THANKS: Added Yan Dai to the Hall of Fame. + +Mon Feb 25 19:16:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * docs/tutorials/007/thread_pool.{h|cpp}: + * docs/tutorials/007/page07.html: + * docs/tutorials/007/page08.html: + * docs/tutorials/013/mld.h: + * docs/tutorials/013/page03.html: + * docs/tutorials/017/Barrier_i.h: + * docs/tutorials/017/page03.html: + * docs/tutorials/018/page03.html: + Added missing include of 'ace/Atomic_Op.h' + +Mon Feb 25 18:51:37 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/Makefile.bor: + Added new Token_Strategy_Test + + * tests/Token_Strategy_Test.cpp: + Fixed small compile error that appeared with the BCB compiler + +Mon Feb 25 17:17:37 2002 UTC Don Hinton <dhinton@ieee.org> + + * ace/Select_Reactor.h: + * ace/Select_Reactor_Base.h: + * ace/Select_Reactor_T.h: Added a new typedef, ACE_SELECT_TOKEN, + to Select_Reactor_Base.h so that the choice of TOKEN type can be + made prior to typedefing the class. This allows the use of the + TOKEN type within the paramaterized class, i.e., + ACE_Select_Reactor_Token_T and ACE_Select_Reactor_T. Otherwise + you get an error on some compilers, e.g., M$VC, that complain + about generic types. + +Mon Feb 25 15:53:56 2002 UTC Don Hinton <dhinton@ieee.org> + + * ace/Synch.i (queueing_strategy): Added ACE_UNUSED_ARG for unused + queueing_strategy parameter, thanks to Craig Rodrigues + <crodrigu@bbn.com> for pointing this out. + +Mon Feb 25 09:17:39 2002 Chris Cleeland <cleeland_c@ociweb.com> + + * ace/SSL/SSL_Context.cpp (dh_params): Wrapped the second argument + to ACE_OS::fopen with ACE_TEXT so that it behaves properly on + wide character platforms. Thanks to Craig Rodrigues for + pointing this out and Ossama Othman for explaining the + difference btw. ACE_TEXT and ACE_LIB_TEXT. + +Mon Feb 25 13:50:43 2002 UTC Don Hinton <dhinton@ieee.org> + + * ace/Token.{h|i|cpp}: Added the ability to chose the queueing + strategy, FIFO or LIFO, by using the queueing_strategy() + methods. The default is FIFO, which was the previous behavior. + Now ACE_Token_Queue::insert_entry() is always called with the + queueing_strategy in order to determine where the thread should + requeue itself. + + * ace/Synch.{h|i}: Added queueing strategy methods to + ACE_Noop_Token. + + * ace/Select_Reactor_T.{h|cpp}: + * ace/TP_Reactor.{h|cpp}: Added QUEUEING_STRATEGY parameter to + ACE_Select_Reactor_Token_T, ACE_Select_Reactor_T, and + ACE_TP_Reactor ctors with FIFO default. + + * tests/Token_Strategy_Test.cpp : + * tests/Makefile : + * tests/run_test.lst: Added new Token_Strategy_Test.cpp to test + the FIFO/LIFO strategies. + +Mon Feb 25 13:44:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * docs/tutorials/018/Test_T.h: + Added missing include of Atomic_Op.h + +Fri Feb 22 15:54:32 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/Refcounted_Auto_Ptr.h: Fix comment, thanks to + Serge Kolgan <skolgan@objectsciences.com>. + +Fri Feb 22 09:31:35 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/QoS/QoS_Session_Factory.h (ACE_QoS_Session_Type): + Change "const static" to "static const" to remove gcc 3.1 + warning. + +Fri Feb 22 08:07:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Makefile.bor: Added String_Base.cpp and Atomic_op.cpp to the + list of files that must be installed when doing a make install. + Thanks to Cyrille Chépélov <cyrille@softek.fr> for reporting that + these files missed. + +Thu Feb 21 16:33:11 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * docs/tutorials/007: Rename Thread_Pool::open() to + Thread_Pool::start() and Thread_Pool::close() to + Thread_Pool::stop() and fixed all the usages. Also removed the + unneeded parameter from stop() and removed the unneeded + re-definition of close() (since we are not shadowing it any + more). Thanks to Peter Heitman <pheitman@cisco.com> for + reporting this. + +Thu Feb 21 13:12:44 2002 Chris Cleeland <cleeland_c@ociweb.com> + + * ace/SSL/SSL_Context.* (ACE_SSL_Context): Added new methods to + specify Diffie-Hellman parameters. These parameters are + required when using DSA certificates/keys. The new methods are + dh_params, dh_params_file_name, and dh_params_file_type. + +Thu Feb 21 09:32:56 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/TTY_IO.cpp (control): Use the appropriate B* macros to set + all the baudrate cases. Thanks to Olli Savia <ops@iki.fi> for + contributing this. + + * ace/Signal.cpp (dispatch): Added a cast of (ACE_SignalHandler) + to the SIG_DFL parameter so things will work on IRIX 6.5 with + GCC 3.0.1. Thanks to Dan Green <dan.c.green@lmco.com> for + reporting this. + +Thu Feb 21 13:09:13 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/streams.h: strstream was deprecated in the 1998 ISO C++ + standard [D.7 depr.str.strstreams], and has been replaced by + sstream. Including <strstream> or <strstream.h> causes annoying + warnings with gcc 3.1. Since strstream and sstream are not used + internally within ACE or TAO, remove includes for strstream, and + let the developer include them in their own code. + +Wed Feb 20 15:26:43 2002 Phil Mesnier <mesnier_p@ociweb.com> + + * apps/soreduce/Library.cpp: + * apps/soreduce/Library.h: + * apps/soreduce/Makefile: + * apps/soreduce/Obj_Module.cpp: + * apps/soreduce/Obj_Module.h: + * apps/soreduce/README: + * apps/soreduce/SO_Group.cpp: + * apps/soreduce/SO_Group.h: + * apps/soreduce/Sig_List.cpp: + * apps/soreduce/Sig_List.h: + * apps/soreduce/Signature.cpp: + * apps/soreduce/Signature.h: + * apps/soreduce/soreduce.cpp: + New application to assist in production of reduced footprint + shared libraries for specific collections of applications. For + more details see the enclosed README. + +Wed Feb 20 14:18:14 2002 Phil Mesnier <mesnier_p@ociweb.com> + + * ace/Process.cpp: + * ace/Process.h: Added a new method + ACE_Process_Options::release_handles() to fix the trouble of + using a pipe as stdout, where the pipe closes completely when + the child terminates. This method replaces some functionality in + the Process options destructor. + +Wed Feb 20 13:01:25 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Atomic_Op.i: Fix the return value of the ACE_GUARD_RETURN + macros so that if the lock fails, the comparison also fails. + Thanks to Ivan Pia <pia@octet.spb.ru> for reporting this. + +Wed Feb 20 17:00:34 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * include/makeinclude/platform_sunos5_sunc++.GNU: Users who use + 5.2 and above could use native exceptions with inlining turned + on. Thanks to Ken O'Brien <kmobrien@fedex.com> for reporting + this problem. + + * THANKS: Added Ken O'Brien <kmobrien@fedex.com> to the hall of + fame. + +Mon Feb 18 23:17:08 2002 Christopher Kohlhoff <chris@kohlhoff.com> + + * ACEXML/parser/parser/Makefile.bor: + Changed library name for the Borland build to be consistent with the + UNIX build. + + * include/makeinclude/ace_flags.bor: + Added macros for the ACEXML Parser library. + + * ACEXML/examples/Makefile.bor: + * ACEXML/examples/SAXPrint/Makefile.bor: + Added Borland makefiles for the SAXPrint example. + +Mon Feb 18 20:16:27 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Acceptor.cpp (make_svc_handler): + * ace/Connector.cpp (make_svc_handler): There doesn't seem to be + anypoint in *conditionally* assigning the Acceptor/Connector's + reactor to the Svc_Handler, so we'll just assign it... + +Mon Feb 18 18:41:33 2002 Steve Huston <shuston@riverace.com> + + Integrated the following from the ACE 5.2 stream: + + Wed Feb 13 15:37:41 2002 Steve Huston <shuston@riverace.com> + + * ace/Hash_Cache_Map_Manager_T.h: Added #include "ace/Synch.h" to + pick up ACE_Null_Mutex. + + * tests/Process_Manager_Test.cpp: Added #include "ace/Thread.h" to + pick up ACE_Thread::self (). + + * tests/Refcounted_Auto_Ptr_Test.h (new file): + * tests/Refcounted_Auto_Ptr_Test.cpp: Moved definition of Printer + from .cpp to .h so Visual Age C++ can find it when instantiating + templates. Also removes the compile warning where the compiler + warns that the test won't work... + + Tue Feb 12 21:59:42 2002 Steve Huston <shuston@riverace.com> + + * include/makeinclude/platform_aix_ibm.GNU: Added -qstaticinline + to CCFLAGS to tell compiler to generate inlined functions with + static scope instead of extern scope (how stupid is this?). + Removed error suppression options which should be unnecessary now. + + Sat Feb 9 18:34:59 2002 Steve Huston <shuston@riverace.com> + + * examples/C++NPv1/Reactive_Logging_Server_Ex.h: Removed an extra + master_handle_set_.set_bit call. Thanks to Craig Perras + <craigp@iswnet.com> for reporting this. + + Sat Feb 9 13:49:44 2002 Steve Huston <shuston@riverace.com> + + * ace/Hash_Map_With_Allocator_T.h: Added #include "ace/Synch.h" to + see the definition of ACE_Null_Mutex. + + * ace/config-all.h (ACE_RCSID): Make generated function static to + keep Visual Age C++ from complaining about the multiple definitions. + + Fri Feb 8 16:20:51 2002 Steve Huston <shuston@riverace.com> + + * ace/config-aix-4.x.h: Removed #define ACE_TEMPLATES_REQUIRE_SOURCE + for Visual Age C++ 5. As it turns out, wherever the compiler sees + source and a template is referenced, it generates the template + class functions used. This is very bad for size as well as for + situations counting on only one such as ACE_Singleton. + + Also added support for explicit template instantiation with Visual + Age C++. If the preprocessor define __TEMPINC__ is not defined, + explicit instantiation is turned on. + + Mon Feb 4 17:21:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Singleton.h (ACE_TSS_Singleton): Added ACE_UNIMPLEMENTED_FUNCS + for assignment and copy ctor methods. This is necessary to allow + the *_SINGLETON_DECLARE macro, that explicitly instantiates a + template class on Win32, to compile clean when instantiating an + ACE_TSS_Singleton class. Thanks very much to Nanbor Wang for + direction on solving this problem. + + * ace/config-win32-msvc-6.h: Added ACE_NEEDS_FUNC_DEFINITIONS. This + avoids warnings when explicitly instantiating an entire class, as + with ACE_TSS_Singleton and *_SINGLETON_DECLARE, above. + + +Mon Feb 18 18:23:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Reactor.h: Clarified behavior with respect to remaining + queued notifications when end_reactor_event_loop() is called + or when the reactor instance is closed/deleted. + +Mon Feb 18 19:28:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/README: + * ace/Proactor.cpp: + * ace/SString.h: + * ace/Task_T.h: + * ace/Timer_Queue_Adapters.h: + * ace/config-win32-msvc-7.h: + * netsvcs/lib/Client_Logging_Handler.h: + * netsvcs/lib/Name_Handler.h: + * netsvcs/lib/TS_Clerk_Handler.h: + * netsvcs/lib/TS_Server_Handler.h: + + Made ACE compiling with the Microsoft Visual C++ 7 compiler. + Template classes cannot be exported when doing a dynamic build, so + removed some export macro's from template definitions. When a class + is derived from a class template then the class template must be + explicit instantiated and be exported. To make sure that we only do + this when a compiler supports this we introduced the new define + ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION_EXPORT. Also vc++ 7 generates + now a warning when a class is exported is derived from a class that + is not exported. + + The following info was given by Microsoft on this: + The reason that a template can't be exported anymore is that it is + unlikely that anyone wants to export all specializations of a + class template. When B is a template and D is a class, the + construction 'class ACE_Export D : public B<D>' should + give no problems, but unfornately there is a bug in the vc++ + compiler. + + That's why we now explicit export the template instantations in ace. + +Mon Feb 18 11:00:17 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * html/index.html: Added an entry for ACEXML document. + + * ace/OS.h: Removed the extern "C" link designator from the + typedef of ACE_Service_Object_Exterminator and the definition of + gobbler functions in ACE_FACTORY_DEFINE. They are not used + outside of a DLL so it's okay to have a mangled gobbler name. + Thanks to Doug for noticing this. + +Mon Feb 18 11:12:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Refcounted_Auto_Ptr.h: Improved Doxygenation of some comments. + +Mon Feb 18 10:45:00 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/QoS/QoS_Session_Factory.h: + Add new constant + ACE_QoS_Session_Factory::ACE_DEFAULT_QOS_SESSION. Give + ACE_QoS_Session_Factory::create_session() a default argument of + ACE_DEFAULT_QOS_SESSION. + + * ace/QoS/QoS_Session_Factory.cpp: + Set the value of ACE_DEFAULT_QOS_SESSION to ACE_RAPI_SESSION on + platforms with RAPI RSVP support. Set it to + ACE_GQOS_SESSION on Win32 platforms with GQoS support. + Otherwise, issue a compilation error, since these are the only + two QoS types supported currently. + + * examples/QOS/Change_Receiver_FlowSpec/receiver.cpp: + * examples/QOS/Change_Receiver_FlowSpec/sender.cpp: + * examples/QOS/Change_Sender_TSpec/receiver.cpp: + * examples/QOS/Change_Sender_TSpec/sender.cpp: + * examples/QOS/Simple/receiver.cpp: + * examples/QOS/Simple/sender.cpp: + Remove reference to ACE_RAPI_SESSION from invocations of + create_session(), leave it empty and choose default argument + instead. This will allow the examples to compile and run on + Win32. + +Mon Feb 18 08:07:59 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/String_Base.{h,i}: Added two new overloaded operators that + work on characters. Thanks to Martin Krumpolec + <krumpolec@asset.sk> for contributing these patches. + +Mon Feb 18 13:05:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/config-all.h: + Added ACE_NEW_NORETURN macro. This macro is the same as ACE_NEW + but doesn't do a return when an out of memory error occured so that + the caller can do extra handling. + + * tests/New_fail_test.cpp: + Extended this test to test the new ACE_NEW_NORETURN macro. + +Mon Feb 18 01:45:07 2002 Christopher Kohlhoff <chris@kohlhoff.com> + + * include/makeinclude/compiler.bor: + * include/makeinclude/make_flags.bor: + Added support for Borland C++Builder 6. + +Sun Feb 17 16:32:01 2002 Venkita <venkita@cs.wustl.edu> + + * ACE version 5.2.2 released. + +Sun Feb 17 16:03:03 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * etc/acexml.doxygen: Fixed the output directory name for ACEXML. + +Fri Feb 15 10:50:26 2002 Venkita Subramonian <venkita@cs.wustl.edu> + + * ace/Dynamic_Service.h: + Fixed compile error. Added forward declaration for + ACE_Service_Object. + +Thu Feb 14 19:10:04 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Refcounted_Auto_Ptr.h: Make the rep_ protected rather + than private. Rodney Morris <rodyland@hotmail.com> for + motivating this. + +Thu Feb 14 15:26:06 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/Dynamic_Service.i (instance): Fixed instance to use an + ACE_dynamic_cast() so that the vptr is set correctly. Thanks to + Bill Dyer <bill.dyer@visogent.com> for suggesting this. + +Thu Feb 14 16:15:50 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * COPYING: Updated copyright years. + +Thu Feb 14 11:25:39 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/docs/bugs.txt: + * ACEXML/docs/guidelines.txt: Updated document. + +Thu Feb 14 08:17:40 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/config-all.h: There was a subtle difference between the + ACE_NEW based on try/catch and a 0 pointer. The version based on + the fact that new can return 0 always sets the pointer to 0 when + a memory error occured. The version that is based on try/catch + the pointer wasn't set to 0. If the pointer had a different + value, the pointer stays at the old value and wasn't set to 0. + This is now fixed. Thanks to Peter van Merkerk + <Peter.van.Merkerk@meco.nl> for noticing this and to Johnny + Willemsen for reporting it. + + * ace/Strategies_T.h: + * ace/Strategies_T.i: Allow the reactor of the Svc Handler to be + set to the reactor passed to the Creation Strategy. Thanks to + David Smith <smithdav@tycoelectronics.com> for motivating this. + +Thu Feb 14 01:14:40 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * include/makeinclude/ace_flags.bor: Updated ACE_XML_CFLAGS. + Thanks to Johnny Willemsen for reminding this. + +Thu Feb 14 01:01:10 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ACEXML/ACEXML.dsw: + * ACEXML/common/XML_Common.dsp: + * ACEXML/examples/SAXPrint/SAXPrint.dsp: + * ACEXML/parser/debug_validator/Debug_Validator.dsp: + * ACEXML/parser/parser/Parser.dsp: + * ACEXML/tests/NamespaceSupport_Test.dsp: + * ACEXML/tests/Transcoder_Test.dsp: Updated base include directories. + +Thu Feb 14 00:20:39 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * Makefile: + * Makefile.bor: + * ACEXML/common/Attributes.h: + * ACEXML/common/AttributesImpl.cpp: + * ACEXML/common/AttributesImpl.h: + * ACEXML/common/Attributes_Def_Builder.cpp: + * ACEXML/common/Attributes_Def_Builder.h: + * ACEXML/common/CharStream.cpp: + * ACEXML/common/CharStream.h: + * ACEXML/common/ContentHandler.h: + * ACEXML/common/DTDHandler.h: + * ACEXML/common/DTD_Manager.cpp: + * ACEXML/common/DTD_Manager.h: + * ACEXML/common/DefaultHandler.cpp: + * ACEXML/common/DefaultHandler.h: + * ACEXML/common/Element_Def_Builder.cpp: + * ACEXML/common/Element_Def_Builder.h: + * ACEXML/common/EntityResolver.h: + * ACEXML/common/Env.cpp: + * ACEXML/common/Env.h: + * ACEXML/common/ErrorHandler.h: + * ACEXML/common/Exception.cpp: + * ACEXML/common/Exception.h: + * ACEXML/common/FileCharStream.cpp: + * ACEXML/common/FileCharStream.h: + * ACEXML/common/InputSource.cpp: + * ACEXML/common/InputSource.h: + * ACEXML/common/Locator.h: + * ACEXML/common/LocatorImpl.cpp: + * ACEXML/common/LocatorImpl.h: + * ACEXML/common/Makefile: + * ACEXML/common/NamespaceSupport.cpp: + * ACEXML/common/NamespaceSupport.h: + * ACEXML/common/SAXExceptions.cpp: + * ACEXML/common/SAXExceptions.h: + * ACEXML/common/Transcode.cpp: + * ACEXML/common/Transcode.h: + * ACEXML/common/Validator.cpp: + * ACEXML/common/Validator.h: + * ACEXML/common/XMLFilter.h: + * ACEXML/common/XMLFilterImpl.cpp: + * ACEXML/common/XMLFilterImpl.h: + * ACEXML/common/XMLReader.h: + * ACEXML/common/XML_Types.h: + * ACEXML/examples/SAXPrint/Makefile: + * ACEXML/examples/SAXPrint/Print_Handler.h: + * ACEXML/examples/SAXPrint/SAXPrint_Handler.h: + * ACEXML/examples/SAXPrint/main.cpp: + * ACEXML/parser/debug_validator/Debug_Attributes_Builder.cpp: + * ACEXML/parser/debug_validator/Debug_Attributes_Builder.h: + * ACEXML/parser/debug_validator/Debug_DTD_Manager.cpp: + * ACEXML/parser/debug_validator/Debug_DTD_Manager.h: + * ACEXML/parser/debug_validator/Debug_Element_Builder.cpp: + * ACEXML/parser/debug_validator/Debug_Element_Builder.h: + * ACEXML/parser/debug_validator/Element_Tree.cpp: + * ACEXML/parser/debug_validator/Element_Tree.h: + * ACEXML/parser/parser/Entity_Manager.cpp: + * ACEXML/parser/parser/Entity_Manager.h: + * ACEXML/parser/parser/Makefile: + * ACEXML/parser/parser/Parser.cpp: + * ACEXML/parser/parser/Parser.h: + * ACEXML/tests/Makefile: + * ACEXML/tests/NamespaceSupport_Test.cpp: + * ACEXML/tests/Transcoder_Test.cpp: + * etc/acexml.doxygen: Renamed directory XML to ACEXML and moved the + base directory to include XML related files to $(ACE_ROOT). + Thanks to Johnny Tucker <jtucker@magisnetworks.com> for the + suggestion. + +Wed Feb 13 17:42:32 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Configuration.cpp (operator=): Fixed a warning in g++ + builds. Stupid mistake on my part :(. + +Wed Feb 13 12:45:06 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Configuration.cpp: + * tests/Config_Test.cpp (iniCompare): Fixed memory leaks. Thanks + to Johnny willemson for providing the patches. + +Wed Feb 13 11:46:54 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * XML/parser/parser/Makefile: Added a library (-lACEXML) to link + to. Thanks to John Michael Zorko <j.zorko@att.net> for + reporting this. + +Tue Feb 12 20:30:53 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * ace/WIN32_Proactor.cpp (handle_events): When the proactor + was called by the reactor in handle_signal() this method should + loop till all events are done. But the loop never got executed + twice because handle_events returned 1 on success and the loop + exits. To catch more than one notifications handle_events + should be called again. Even if the loop is executed twice and + no more events are outstanding handle_events should return 0 and + not -1 when calling with timeout 0. Calling + GetQueuedCompletionStatus with timeout value 0 returns FALSE and + errno "ERROR_SUCCESS". This check has to be added to + handle_events and 0 has to be returned. Thanks to Hartmut Quast + <HartmutQuast@t-online.de> for reporting this. + +Tue Feb 12 16:18:59 2002 Ossama Othman <ossama@uci.edu> + + * tests/Proactor_Test.cpp (logflag): + * tests/TP_Reactor_Test.cpp (logflag): + + Removed these unused global variables. Fixes an unused variable + warning. + +Tue Feb 12 11:50:18 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * bin/pippen.pl: Applied a patch from "the source" to fix a + problem in determining project dependencies. + +Tue Feb 12 09:37:56 2002 Ossama Othman <ossama@uci.edu> + + * ACE-INSTALL.html: + + Corrected EGCS documentation. Native exception support is now + the default. [Bug 1149] + + G++ 2.7.x is no longer supported. Updated accordingly. + +Mon Feb 11 16:31:04 2002 Ossama Othman <ossama@uci.edu> + + * bin/make_pretty.pl (is_warning): + + Do not flag Fuzz's "#pragma warning(push)/(pop)" test title as a + warning. + +Mon Feb 11 13:49:35 2002 Ossama Othman <ossama@uci.edu> + + * bin/fuzz.pl (check_for_push_and_pop): + + New test that verifies the number of #pragma warning(push) + pragmas matches the number of #pragma warning(pop) pragmas. + + * examples/IPC_SAP/SSL_SAP/SSL-client.cpp (shared_client_test): + * examples/IPC_SAP/SSL_SAP/SSL-client-simple.cpp + (shared_client_test): + + Do not convert the buffer length to network byte order when + allocating the buffer. Fixes excessive memory allocation. This + was apparently a cut-n-paste bug. Thanks to M Schulze + <m2.schulze@gmx.net>. + + * THANKS: + + Added M Schulze to the Hall of Fame. + +Mon Feb 11 05:42:02 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Connector.h: Fixed a typo in the coments. Thanks to Miljenko + Norsic (ETK) <Miljenko.Norsic@etk.ericsson.se> for reporting + this. + +Sun Feb 10 16:28:30 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/config-macosx.h + * ace/config-freebsd.h + * ace/config-freebsd-pthread.h + * ace/TTY_IO.cpp: + * TODO: Removed the ACE_USES_HIGH_BAUD_RATES macro since it no longer + seems to be necessary. Thanks to Olli Savia <ops@iki.fi> for + reporting this. + + * ace/TTY_IO.cpp: Replaced the two strcmp() calls with one + strcasecmp(). Thanks to Olli Savia <ops@iki.fi> for reporting + this. + +Sat Feb 9 15:17:45 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/make_release: Changed the path of gv as a new version of GV + was installed on deuce.doc. The old version had less colors and + it started mapping them to a smaller range. The graphs looked + very ugly. The new version fixes the problem and hence a change + in path. + +Fri Feb 8 22:56:29 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Log_Msg.cpp (log): Fixed a warning in TRU 64 builds. + +Fri Feb 8 14:54:21 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * apps/JAWS2/Makefile (LDFLAGS): + * apps/JAWS2/HTTPU/Makefile (LDFLAGS): Fixed some makefile bugs so + that this stuff compiles on AIX. Thanks to Steve Ige + <steve.ige@reuters.com> for reporting this. + +Thu Feb 7 18:13:03 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * tests/ACE_Init_Test.cpp (wait_and_kill_dialog): Replaced the call + to EndDialog() with EndModalLoop() to fix a race condition. + Thanks to Petru Marginean <petrum@ilx.com> for reporting this. + +Fri Feb 8 14:02:06 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * THANKS: Added Marco Kranawetter + <Marco.Kranawetter@icn.siemens.de> to the hall of fame. + +Fri Feb 08 11:24:36 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/String_Base.h: + * ace/Task_T.h: Removed the ACE_Export decl from ACE_Task and + ACE_String_Base. They were added as work-aronds for a VC7's + internal compiler bug but didn't seem to solve the problem. + Thanks to Patrick Bennett <patrickb@inin.com>, Johnny, and + Christian Veleba <christian.veleba@porsche.co.at> for reporting + this. + +Thu Feb 7 16:19:39 2002 Steve Huston <shuston@riverace.com> + + * ace/config-all.h: Define new macros, ACE_nothrow and ACE_nothrow_t, + to decide which variety of nothrow is used in new (nothrow). At + this point, HP aC++ is the only platform defined to use this + feature, so that's the only section that defines it. + + * ace/Svc_Handler.(cpp h): + * examples/Shared_Malloc/test_persistence.cpp: Use the new + ACE_nothrow[_t] macros in overridden operator new. + +Thu Feb 7 14:11:31 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Singleton.cpp (close): Fixed the implementation so that the + ACE_Unmanaged_Singleton's internal singleton point is reset to 0 + after cleanup to avoid double-deletion. Thanks to Marc Walrave + <marc.walrave@meco.nl> for this fix. + +Thu Feb 7 07:52:47 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Activation_Queue.{h,i}: Added get/set methods to access/update + the underlying ACE_Message_Queue so users can call methods on + the queue directly if necessary. Thanks to Timothy Kilbourn + <kilbourn@sep.com> for reporting this. + +Tue Feb 5 07:25:49 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * tests/TP_Reactor_Test.cpp: Improved the comments to clarify the + differences between this test and the Thread_Pool_Reactor_Test.cpp. + Thanks to Alex Libman for explaining this. + +Thu Feb 7 08:16:24 2002 Oliver Kellogg <oliver.kellogg@sysde.eads.net> + + * ACE-INSTALL.html: Document the include_env=1 make switch. + + * docs/exceptions.html: Replaced the "Transition from TAO_TRY + to ACE_TRY" section with "Transition from ACE_TRY_ENV usage + to ACE_ENV_ARG". + +Wed Feb 6 06:57:35 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * tests/run_test.lst: Disabled TP_Reactor_Test as the test is + hanging. + +Tue Feb 5 11:59:00 2002 Craig Rodrigues <crodrigu@bbn.com> + + * tests/TP_Reactor_Test.cpp (disable_signal): Eliminate unused + arguments warning on Win32 platforms. + +Mon Feb 4 16:22:20 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/OS.h: Include <new> instead of <new.h> if + ACE_USES_STD_NAMESPACE_FOR_STDCPP_LIB is defined. + +Mon Feb 4 19:58:03 2002 Boris Kolpackov <bosk@ipmce.ru> + + * ace/Log_Msg.cpp: + + Fixed minor bug in what's just commited before. + Thanks to Craig Rodrigues <crodrigu@bbn.com> + for pointing it out. + +Mon Feb 4 14:11:14 2002 Boris Kolpackov <bosk@ipmce.ru> + + * ace/Log_Msg.h: + * ace/Log_Msg.cpp: + + Added ability to install custom backend which is a + per-process entity as opposite to callback which is + a per-thread not-inheritable entity. + +Sun Feb 3 17:59:36 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * ace/config-sunos5.5.h (ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION): + + Explicitly defined the above macro as this is needed for SunOS + gcc to work. This was inside a __SUNPRO_CC #ifdef. I missed that + in my previous change. This should fix the builds under SunOS + gcc. + +Sun Feb 3 18:32:29 2002 Craig Rodrigues <crodrigu@bbn.com> + + * tests/TP_Reactor_Test.cpp: Use size_t instead of long + and int for index_ and sessions_ in order to eliminate + more compiler warnings. + +Sun Feb 3 09:20:04 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * tests/TP_Reactor_Test.cpp: Fixed a bunch of warnings. Thanks + to Venkita for reporting this. + +Sun Feb 3 08:22:28 2002 Venkita Subramonian <venkita@cs.wustl.edu> + + * tests/Makefile: + Regenerated makefile to create dependencies for TP_Reactor_Test. + +Sun Feb 3 08:05:12 2002 Venkita Subramonian <venkita@cs.wustl.edu> + + * tests/TP_Reactor_Test.dsp (RSC): + Regenerated the file in MSVC++. + +Sun Feb 3 11:16:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * tests/TP_Reactor_Test.cpp: + Fixed compile error in BCB unicode build + +Sat Feb 2 07:45:51 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * tests/run_test.lst: + * tests/TP_Reactor_Test.dsp: + * tests/Makefile.bor: + * tests/Makefile: Added the TP_Reactor_Test. + + * tests/TP_Reactor_Test.cpp: Added another test of the ACE_TP_Reactor. + Thanks to Alex Libman for contributing this. + + * ace/config-irix6.x-common.h: IRIX 6.5 supports AIO, so we'll + enable these features. Thanks to Alex Libman for validating + this. + + * ace/Select_Reactor_T.cpp: Fixed work_pending() so that it takes + into account pending timers that need to be expired. Thanks to + Russ Noseworthy for reporting this. + + * ace/Select_Reactor_T.cpp: Simplified the logic for calculating + timeouts in wait_for_multiple_events(). + + * ace/Process.{h,i,cpp}: When using ACE_Process_Options with the + inherit_environment set to off, i.e., ACE_Process_Options opts + (0), ACE_Process::spawn() was improperly setting the environment + in the child's process after fork (), before exec (). Changed + ACE_Process::spawn to check for the inherit_environment flag, + and to use the execve () call instead of execvp () if + inherit_environment is false. Thanks to James Risinger + <jrisinger@SignalSoftCorp.com> for contributing this fix. + + * ace/Process.{h,i}: Added "const" to the various accessor methods. + +Sat Feb 2 00:01:36 2002 Venkita Subramonian <venkita@cs.wustl.edu> + + * ace/config-sunos5.6.h: + Added missing #endif. + +Fri Feb 1 23:42:03 2002 Venkita Subramonian <venkita@cs.wustl.edu> + + * ace/config-all.h: + Removed extra ). + +Fri Feb 1 21:08:37 2002 Steve Huston <shuston@riverace.com> + + * tests/Framework_Component_Test.icc: + * tests/Vector_Test.icc: New Visual Age C++ test configurations. + + * tests/tests.icp: Add new test configurations to the project. + +Fri Jan 1 19:33:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Vector.(h i cpp): Removed 'const' from the 2nd template + argument (size_T DEFAULT_SIZE). A size_t is always const, + and having const there causes errors from HP aC++. I'm not sure + if they're completely legit, but Stroustrup 3rd Ed says the + template argument is const anyway... if this is a problem, + please let me know. + +Fri Feb 1 18:53:44 2002 Steve Huston <shuston@riverace.com> + + * ace/ace.icc: Added Framework_Component.(h cpp) to the files list. + +Fri Feb 1 10:19:46 2002 Jeff Parsons <parsons@cs.wustl.edu> + + * Thread_Manager.h: + + Removed extra '*/'. + +Fri Feb 01 00:00:12 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/Task.h: + * ace/Thread_Manager.h: Added more explanation on how to use the + <task> argument. Thanks to Petr Shelomovsky + <pedrodon@trustworks.com> for motivating the change. + +Thu Jan 31 19:18:37 2002 Steve Huston <shuston@riverace.com> + + * ace/NT_Service.{h cpp}: To avoid race condition at shutdown time, + moved the call to report_status(SERVICE_STOPPED, 0) from the + open() method to a new override of the fini() method. Setting + status to SERVICE_STOPPED frees up Windows to do its own shutdown + for the service, and that can't be allowed to commence until all + ACE_NT_Service things are done. Thanks to Zoran Cetusic + <ZoranC@inter-intelli.com>, Patrick Bennett and Felix Wyss from + Interactive Intelligence, Inc. for diagnosing this problem and + sending in a fix. + + * THANKS: Added Zoran Cetusic, Patrick Bennett, and Felix Wyss to + the Hall of Fame. + +Thu Jan 31 17:00:52 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/config-all.h: Need to include <new> with all versions of + SunCC compiler and not just CC 5.0, when the compiler is using + compat mode 4. + * ace/config-sunos5.6.h: Need to define ACE_LACKS_ACE_IOSTREAM + when higher versions of CC are used with compat mode 4 and + such. + + Thanks to Tim Rydell <tim.rydell@gd-ais.com> for the fixes. + + * THANKS: Added Tim Rydell to the hall of fame. + +Thu Jan 31 17:21:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Trace.cpp (constructor and destructor): Do not attempt + trace output if ACE has not been initialized. There is too + much not set up yet to bother trying. If you are on a platform + with ACE_HAS_NONSTATIC_OBJECT_MANAGER (such as Windows) and you + really, really need tracing in static objects, you should + try #define ACE_HAS_NONSTATIC_OBJECT_MANAGER 0 in your config.h + along with #define ACE_NTRACE 0. Beware, though, there are + crocodiles lurking there - platforms defined to use non-static + object manager are that way for good reason. + + Thank you to Shmulik Regev <shmul@vself.com> for reporting this. + +Thu Jan 31 13:32:07 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * XML/common/XML_Common.dsp: Fixed the LIB path to use relative + path. + +Thu Jan 31 19:18:16 2002 Oliver Kellogg <oliver.kellogg@sysde.eads.net> + + * include/makeinclude/wrapper_macros.GNU: + Corrected placement of the include_env switch. + include_env=1 is only sensible in combination with + exceptions=1. NB: The include_env switch is only + intended to facilitate transition to the ACE_ENV_ARG + macros and should not be used for new applications. + There will be unused-variable warnings when using this + build configuration. + +Thu Jan 31 11:57:07 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * XML/parser/debug_validator/Debug_Attributes_Builder.cpp: + * XML/parser/debug_validator/Debug_Element_Builder.cpp: + Temporarily removed unused arguments. + + * XML/common/FileCharStream.cpp (get): Made sure the character + read from the input file was converted to ACEXML_Char type + correctly. Casted the read XML_Char before comparing it to + 'EOF'. + +Thu Jan 31 13:06:12 2002 Boris Kolpackov <bosk@ipmce.ru> + + * THANKS: + + Added Koushik Banerjee to the hall of fame. + +Wed Jan 30 22:41:39 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_linux.GNU (CXX_VERSION): + + Made it work when someone wants to turn off the implicit + template instantiation. Care should be taken to #define + ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION in config.h also. + Surprisingly the code compiles without that also... + +Wed Jan 30 17:22:49 2002 Steve Huston <shuston@riverace.com> + + * ace/Process.cpp (wait (const ACE_Time_Value&, ACE_exitcode *)): + * ace/Process_Manager.cpp (wait (pid_t, const ACE_Time_Value&, + ACE_exitcode *)): + The mechanism for waiting up to a specified time for a child + process to exit has been replaced. Replaces the fix from: + Fri Jan 25 19:58:41 2002 Steve Huston <shuston@riverace.com> + and makes unnecessary any further work from: + Sat Jan 26 21:41:39 2002 Steve Huston <shuston@riverace.com> + + Both classes now do a timed wait for a child by doing an + ACE_OS::sleep, counting on being interrupted if a SIGCHLD + is delivered. In ACE_Process_Manager when a reactor hasn't + been specified, and always in ACE_Process, a temporary + SIGCHLD handler is installed for the duration of the wait. + This is necessary because the default SIGCHLD action on + POSIX (and holds true for most non-Win32) is SIG_IGN, and + SIGCHLD is not generated when a child process exits. + Therefore, a handler is installed to force the SIGCHLD. + It's not needed in ACE_Process_Manager when a reactor is in + place because the reactor already has a handler for SIGCHLD. + +Wed Jan 30 15:11:49 2002 Krishnakumar B <kitty@cs.wustl.edu> + + * include/makeinclude/platform_linux.GNU (CXX_VERSION): + * ace/config-g++-common.h: + + Turned off explicit template instantiation with gcc under Linux. + The specific versions are 2.95.x, 2.96, 3.0.x (3.x). Added a + flag implicit_templates to tweak the behaviour from the + platform_macros.GNU file. + + The combination of the compiler and binutils seems to give a + nice reduction in the footprint. + +Wed Jan 30 16:00:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Process_Manager.cpp (register_handler): Replaced ECHILD with + EINVAL if the pid is not found. Probably a more accurate + assessment of the situation, and should compile clean on WinCE. + +Wed Jan 30 13:50:17 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * docs/index.html: Fixed the ACE-inheritance.pdf document so + it isn't gzipped. Thanks to Michael Searles + <msearles@base16.com> for reporting this. + +Wed Jan 30 09:28:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * XML/parser/debug_validator/Debug_Attributes_Builder.cpp: + * XML/parser/debug_validator/Debug_DTD_Manager.cpp: + * XML/parser/debug_validator/Debug_Element_Builder.cpp: + * XML/parser/debug_validator/Element_Tree.cpp: + Added missing ACE_LIB_TEXT. This fixes the BCB unicode build errors + +Tue Jan 29 20:11:21 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * etc/*.doxygen (EXPAND_AS_DEFINED): Added ACE_CACHE_MAP_MANAGER + to the list in EXPAND_AS_DEFINED. Thanks to Don Hinton + <dhinton@ieee.org> for the suggestion. + +Tue Jan 29 19:36:24 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * bin/make_release: The whole release process has been moved to + Linux box. This is because the sun machines at WashU were having + problems and they are not dependable. The following are the list + of changes made + + - All the path related stuff have been changed ie. instead of + using /pkg/gnu tools, we use native tools on Linux now. + + - The gnu suffixes to many of the tools have been removed. + + - Most of the path to the tools have been hardcoded in the PATH + environment variabe. + + - Tools that were missing have been loaded on 3 main Linux boxes + at WashU including Graphviz and doxygen. + + - A beta cannot be cut from a sun box. + + - The script will recommend cutting a beta from deuce.doc. + + The script has been tested with a dummy release. + +Tue Jan 29 16:01:54 2002 Ossama Othman <ossama@uci.edu> + + * bin/fuzz.pl (check_for_missing_rir_env): + + Check for ACE_ENV_ARG_PARAMETER instead of + TAO_ENV_ARG_PARAMETER. The latter is deprecated. + +Tue Jan 29 20:47:24 2002 Oliver Kellogg <oliver.kellogg@sysde.eads.net> + + * docs/exceptions.html: Document the new ACE_ENV_ macros. + + * bin/subst_env.pl: Transform to ACE_ENV_ instead of TAO_ENV_. + +Tue Jan 29 08:39:24 2002 Oliver Kellogg <oliver.kellogg@sysde.eads.net> + + * ace/CORBA_macros.h: + Added ACE_ENV_ARG macros to replace the TAO_ENV_ARG macros + defined in TAO/tao/orbconf.h. All exception related macros + are now defined in ace/CORBA_macros.h, and the TAO_ENV_ARG + macros will soon be deprecated. + + * include/makeinclude/wrapper_macros.GNU: + Added the include_env switch for compatibility with the + exception handling use before TAO 1.2.2. + +Tue Jan 29 08:17:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/ace_flags.bor: + Added compiler flags for new cosevent orbsvcs test library CECTEST + Added compiler flags for new notify orbsvcs test library NotifyTests + +Mon Jan 28 17:44:51 2002 Steve Huston <shuston@riverace.com> + + * ace/Process_Manager.cpp (wait): When waiting for a non-specific + process, specify -1 pid for waitpid(). This is necessary because + of Fri Jan 25 19:58:41 2002 Steve Huston <shuston@riverace.com> + change to not alter the process group ID when ACE_Process_Manager + spawns a process. + Timed waits for a process still don't work on non-Win32, but + this fix corrects the failed wait with errno == ECHILD. + +Mon Jan 28 12:16:28 2002 Carlos O'Ryan <coryan@uci.edu> + + * bin/auto_run_tests.lst: + Removed EC_Basic and Event_Latency tests from nightly builds, I + left them there by mistake when I took them out of the + repository (around December, 25th 2001) + +Mon Jan 28 13:20:32 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * XML/common/AttributesImpl.cpp: Removed a bunch of inline + designators. + +Sun Jan 27 22:18:50 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * bin/msvc_auto_compile.pl: Projects in XML subdirectory are + interdependent. List out the order they should be built + explicitly. + +Sun Jan 27 12:48:48 2002 Ossama Othman <ossama@uci.edu> + + * bin/auto_run_tests.lst + * bin/performance_stats.sh: + + Updated in accordance with the new TAO "Latency" performance + test organization. + +Sun Jan 27 12:08:45 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * XML/common/Transcode.h (ACEXML_Transcoder): Improved the + documentation. + + The followings fixed the Tru64 warnings/errors. + + * XML/common/AttributesImpl.i (operator): + * XML/common/Env.i: Reordered inline functions. + + * XML/tests/Transcoder_Test.cpp: Removed an unused argument. + + * XML/common/Attributes_Def_Builder.h: Added inclusion of + "ace/Auto_Ptr.h". + + * XML/examples/SAXPrint/SAXPrint_Handler.cpp: + * XML/examples/SAXPrint/Print_Handler.cpp: Added inclusion of + "ace/Log_Msg.h". + +Sun Jan 27 15:03:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * Makefile.bor: + Build the XML library with BCB + + * XML/parser/Makefile.bor: + Added debug_validator + + * XML/parser/debug_validator/Makefile.bor: + Added BCB makefile + +Sat Jan 26 19:02:49 2002 Ossama Othman <ossama@uci.edu> + + * ace/Process_Manager.cpp (register_handler): + + Corrected code that always returned -1. Code that should only + have been run on error was always run since it was outside of an + "if block." Curly braces are a good thing (they were missing). + +Sat Jan 26 21:41:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Process_Manager.cpp (wait): Fixed compiler warning on + Linux about unused wait_until. This fix removes the ability + to spin around the 'for' loop waiting for signals multiple + times with the timeout decreasing to account for wait time. + Will have to come back to restore this functionality later. + +Sat Jan 26 14:40:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/ace_flags.bor: + Added compiler and linker flags for new XML library + + * XML/common/Makefile.bor: + * XML/parser/parser.Makefile.bor: + * XML/Makefile.bor: + * XML/parser/Makefile.bor: + * XML/tests/Makefile.bor: + Added BCB makefiles for the new XML library. + +Fri Jan 25 19:58:41 2002 Steve Huston <shuston@riverace.com> + + * ace/Process_Manager.cpp (wait(pid_t, const ACE_Time_Value &, + ACE_exitcode *status)): If platform offers sigtimedwait, use it + instead of setting ualarm and then doing sigwait. Once ualarm + is set, it will fire, even if this method has returned. This + causes Solaris processes to die on SIGALRM. + + Also, do not play with the process group ID by default. It's + not needed for doing normal signal management by most processes. + If processes really have a need to change or set a new process + group, they need to do it explicitly by using ACE_OS::setpgid() + or by setting a process group ID in an ACE_Process_Options object + when spawning processes. + + * tests/Process_Manager_Test.cpp: Added a bit more diagnostic info. + +Fri Jan 25 18:29:37 2002 Steve Huston <shuston@riverace.com> + + * ace/config-aix-4.x.h: Removed ACE_HAS_SIGTIMEDWAIT. It compiles, + but returns ENOSYS at run time. + + * ace/config-aix5.1.h: Added ACE_HAS_SIGTIMEDWAIT. + +Fri Jan 25 15:36:40 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/ace_dll.dsp: Removed /version flags from the project since + they have been taken care of by ace.rc file. Thanks to Ossama + for pointing it out. + +Fri Jan 25 14:45:00 2002 Venkita Subramonian <venkita@cs.wustl.edu> + + * bin/auto_run_tests.lst: + Added Two_Objects test to the list. + +Fri Jan 25 14:40:15 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * XML/common/Exception.cpp: + * XML/common/NamespaceSupport.cpp: + * XML/common/SAXExceptions.cpp: + * XML/parser/parser/Parser.cpp: Moved the initialization of static + members before the inclusion of inline files to avoid + compilation erros on Borland compiler. Thanks to Johnny + Willemsen <jwillemsen@remedy.nl> for figuring this out. + +Fri Jan 25 14:31:06 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * XML/common/NamespaceSupport.cpp: + * XML/parser/parser/Parser.cpp: Fixed several KCC warnings. + +Fri Jan 25 12:01:14 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + The following changes fixed SunCC5.1 compilation errors. + + * XML/common/Makefile: + * XML/parser/parser/Makefile: + * XML/tests/Makefile: + * XML/examples/SAXPrint/Makefile: Removed extra spaces for -I + flags. + + * XML/common/Attributes_Def_Builder.h: Removed a redundant comma. + + * XML/common/NamespaceSupport.i: Changed + ACE_TEMPLATE_METHOD_SPECIALIZATION to + ACE_TEMPLATE_SPECIALIZATION. + + * XML/tests/NamespaceSupport_Test.cpp: String literals needed to + be assigned to const char *. + +Fri Jan 25 09:42:12 2002 Ossama Othman <ossama@uci.edu> + + * ace/ace_dll.dsp: + + Corrected inconsistency in the DLL minor version. The correct + minor version for the ACE 5.2 series is "2," not "1." + +Fri Jan 25 11:21:28 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * bin/msvc_auto_compile.pl: Added XML into the list of auto build + targets. + +Fri Jan 25 00:37:00 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * We now have 1,400 contributors to the ACE+TAO software. Yow! + +Thu Jan 24 17:49:46 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/CDR_Stream.i: Fixed the check for the length within + ACE_InputCDR::read_*_array (). The method was checking just for + length passed in, which happens to be the number of elements in + the array, instead of the number of bytes necessary for the + elements. Thanks to William R Volz <WRVO@chevrontexaco.com> for + reporting this. + + * THANKS: Added William Volz to the hall of fame. + +Thu Jan 24 18:31:49 2002 Steve Huston <shuston@riverace.com> + + * tests/Process_Manager_Test.cpp: Better diagnostics added. + +Thu Jan 24 15:14:52 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/Lib_Find.cpp (ldfind): Restored previously removed Win32 + code and re-organized macros so we wouldn't upset CE builds. + +Thu Jan 24 14:53:38 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * bin/generate_doxygen.pl: + * etc/acexml.doxygen: Added the doxygen config file for XML + subdirectory. + + * Makefile: Added XML subdirectory into the lists to be compiled + and be included in the release. + + * XML/*: Merged in the XML parser code. + +Thu Jan 24 10:14:47 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/ace_wchar.h: Added the definition for ACE_TEXT_SearchPath. + + * ace/Lib_Find.cpp (ldfind): Fixed UNICODE and Fuzz builds + errors. Thanks to Johnny Willemsen <jwillemsen@remedy.nl> for + the fix. + +Wed Jan 23 16:48:54 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/Lib_Find.h: + * ace/Lib_Find.cpp (ldfind): Change to use Win32 API SearchPath to + search for the target DLL and updated the document for ldfind in + header file. Thanks to Eugene Alterman + <Eugene.Alterman@bremer-inc.com> for submitting the patch. + +Wed Jan 23 14:01:32 2002 Ossama Othman <ossama@uci.edu> + + * ace/config-lynxos.h (ACE_LACKS_INET_ATON): + + LynxOS does not implement the inet_aton() function. + +Wed Jan 23 16:37:52 2002 Steve Huston <shuston@riverace.com> + + * ace/NT_Service.cpp (insert): If the CreateService call fails, + be sure to save the error value before making another Win32 call + that will smash it. Thanks to Kelly Hickel <kfh@mqsoftware.com> + for reporting this. + Also ACE-ified the source better. + + * examples/NT_Service/main.cpp: Added some ACE_ERROR output if + operations requested from the command line fail. + +Wed Jan 23 16:14:43 2002 Boris Kolpackov <bosk@ipmce.ru> + + * include/makeinclude/platform_sunos5_sunc++.GNU + + Added work around for famous Sun CC "pure virtual function called" + bug. Unfortunately this involves introduction of yet another #define. + See TAO/tao/ValueBase.h for more information. + +Tue Jan 22 21:27:25 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/ace_dll.vcp: Add Frameork_Component.* to the builds. Thanks + to Venkita for pointing it out. + +Tue Jan 22 17:42:39 2002 Steve Huston <shuston@riverace.com> + + * ace/NT_Service.(h cpp): Added two new methods: + void capture_log_msg_attributes (void): Grabs a copy of the + calling thread's ACE_OS_Log_Msg_Attributes to facilitate + inheritance of the logging attributes in the service thread. + void inherit_log_msg_attributes (void): Called in a service + thread, inherits the main thread's logging attributes. + Modified the ACE_NT_SERVICE_RUN macro to capture the main + thread's logging attributes before starting the service control + dispatcher. Modified the ACE_NT_SERVICE_DEFINE macro to call + inherit_log_msg_attributes if the ACE_NT_Service object for + the service was set up before the thread started. + Fixes Bugzilla # 82. + + * examples/NT_Service/main.cpp: + * examples/NT_Service/ntsvc.cpp: Now writes a log file in the current + working directory which should have messages from both main and + service threads in it. + +Tue Jan 22 15:19:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/DLL.cpp: Changed to invoke this->open() in the + constructor. Thanks to Eugene Alterman + <Alterman@bremer-inc.com> for motivating this. + +Mon Jan 21 23:27:03 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/OS.h: Reordered main redefinition macros so that it actually + passed wchar argv to main when UNICODE is defined. + +Mon Jan 21 10:01:34 2002 Frank Hunleth <fhunleth@cs.wustl.edu> + + * bin/auto_run_tests.lst: + Added MIOP unit tests. + +Mon Jan 21 03:00:14 2002 Ossama Othman <ossama@uci.edu> + + * ace/Framework_Component.cpp (register_component): + + Removed debugging statements that always printed text. + +Mon Jan 21 07:45:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Framework_Component.cpp: + Fixed fuzz error + +Mon Jan 21 00:13:42 2002 Christopher Kohlhoff <chris@kohlhoff.com> + + * ace/streams.h: + Workaround for Borland C++ 5.5.1 bug we have now just hit. + +Sun Jan 20 21:42:53 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Framework_Component.cpp: + * ace/Framework_Component_T.cpp: Fixed fuzz errors. + +2002-01-20 Oliver Kellogg <oliver.kellogg@sysde.eads.net> + + * bin/subst_env.pl: New script to ease the transition to the + TAO_ENV_ARG macros defined in TAO/tao/orbconf.h. + +Sun Jan 20 12:38:28 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * tests/Framework_Component_Test.dsp: New dsp file for the test. + * tests/tests.dsw: Added the above test to the workspace. + +Sun Jan 20 12:25:28 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Framework_Component.h: Removed the definition of the + default constructor (in ACE_UNIMPLEMENTED_FUNC definition). The + other private constructor with a default argument tends towards + a default constructor and VC++ signals a multiple definition + error. Not sure how g++ didnt signal this one. + + * ace/Framework_Component_T.h: #include'd Framework_Component.h + * ace/Framework_Component_T.cpp: Added a #ifndef around the file. + + * ace/ace_dll.dsp: + * ace/ace_lib.dsp: Added the Framwork_Component* files to the + project file. + +Sun Jan 20 10:40:28 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * tests/Vector_Test.dsp: + * tests/tests.dsw: Added a new project for Vector_Test. + +Sun Jan 20 16:25:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Makefile.bor: + Added new Framework_Component + + * tests/Makefile.bor: + Added new Framework_Component_Test + +Sun Jan 20 00:00:30 2002 UTC Don Hinton <dhinton@ieee.org> + + * ace/Vector_T.cpp (dump): Commented out the contents of this + function for the time being. It assumed that the element + was an object with a dump() method, which won't always be + the case. + + * tests/Vector_Test.cpp: Changed a few data types from signed + to unsigned, size_t, to get rid of compiler warnings. + +Sat Jan 19 17:29:50 2002 Douglas C. Schmidt <schmidt@siesta.cs.wustl.edu> + + * ace/Vector_T.cpp (dump): Fixed problems with this method. Thanks + to Don Hinton for reporting this. + + * tests/Vector_Test.cpp: Changed the typedef of DATA from int to + size_t to avoid "type mismatch" compiler warnings. Thanks to + Don Hinton for reporting this. + +Sat Jan 19 22:30:26 UTC 2002 Don Hinton <dhinton@ieee.org> + + * apps/JAWS2/JAWS/Hash_Bucket_T.h: + * apps/JAWS2/JAWS/Assoc_Array.h: Added missing keyword "class" to + friend declarations. + + * ace/SString.cpp: Removed unneeded include of Service_Config.h. + + * ace/Service_Config.{h|cpp}: + * examples/Connection/misc/Connection_Handler.cpp: + + Removed static methods from ACE_Service_Config that delegated to + ACE_Reactor::instance(), and fixed a few instances where they were + still called. + + * ace/Object_Manager.cpp: + * ace/Service_Config.cpp: + * ace/Proactor.cpp: + * ace/Reactor.cpp: + + Added call to instance() methods that registers the singleton with + the new ACE_Framework_Repository so it can handle destruction, and + replaced explicit references to ACE_Reactor and ACE_Proactor with + calls to ACE_Framework_Repository. + + * ace/Framework_Component.{h|inl|cpp}: + * ace/Framework_Component_T.{h|inl|cpp}: + * ace/Makefile: + * tests/Framework_Component_Test.cpp: + * tests/Makefile: + * tests/run_test.lst: + + Added ACE_Framework_Repository to manage ACE_Framework_Component's, + e.g., singletons like ACE_Reactor or ACE_Proactor. It uses + External Polymorphism obviating any interface changes. The + components register themselves with repository in their instance + methods. This allows the Object_Manager and Service_Config to + manage these components without having to know about them a priori. + + This was needed to reduce footprint for applications like TAO that + don't need to use all the available components, e.g., ACE_Proactor. + +Sat Jan 19 10:23:39 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Makefile (TEMPLATE_FILES): + * tests/Makefile.bor: + * tests/Makefile: + * tests/run_test.lst: + * ace/Vector_T.{h,i,cpp}: + * tests/Vector_Test.cpp: Added support for the new ACE_Vector to + the appropriate places. This vector behaves like the STL + vector. Thanks to Gonzo and Craig Ching for contributing this. + + * ace/Future_Set.h: Updated the documentation to explain how + various features work better. Thanks to Johnny Tucker for + contributing this. + +Fri Jan 18 19:09:41 2002 Steve Huston <shuston@riverace.com> + + * tests/run_test.lst: Re-enabled Process_Manager_Test for all but + Chorus and VxWorks. Could not find a reason it was disabled. + Also enabled Process_Mutex_Test on Win32. + +Fri Jan 18 16:44:29 2002 Steve Huston <shuston@riverace.com> + + * ace/ace.icc: Added Reactor_Notification_Strategy.(h cpp) sources. + + * ace/Reactor_Notification_Strategy.cpp: Fixed ACE_RCSID to refer to + Reactor_Notification_Strategy, not Strategies. + + * tests/Get_Opt_Test.icc: + * tests/INET_Addr_Test.icc: New Visual Age C++ configs for these tests. + + * tests/tests.icp: Added Get_Opt_Test.icc and INET_Addr_Test.icc + +Fri Jan 18 12:56:36 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Log_Msg.cpp (init_backend): Added support for SysLog on platforms + that don't lack it. Thanks to Alexei I. Adamovich + <lexa@adam.botik.ru> for reporting this fix. + +Fri Jan 18 10:29:06 2002 Ossama Othman <ossama@uci.edu> + + * ace/Service_Config.h (process_file): + * ace/Service_Config.cpp (process_directives, process_file): + + Factored out code that processes a svc.conf file into the new + static process_file() method. This allows svc.conf files to be + explicitly parsed by the application at any arbitrary point in + time instead of Service Configuration initialization time alone. + +Thu Jan 17 18:51:09 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Name_Space.cpp (operator =): Fixed a memory leak. Thanks + to Ian Cahoon <icahoon@cisco.com> for reporting this. + +Thu Jan 17 12:13:51 2002 Ossama Othman <ossama@uci.edu> + + * ace/SSL/SSL_Context.h (private_key, verify_private_key): + + Added new documentation. These methods should only be called + after a certificate has been set since key verification is + performed against the certificate, among other things. + +Thu Jan 17 13:11:27 2002 Chad Elliott <elliott_c@ociweb.com> + + * ace/Message_Queue.h: + * ace/Message_Queue.cpp: + * ace/Message_Queue_T.h: + * ace/Message_Queue_T.cpp: + + Provide the ability to enqueue based on the message deadline and + to dequeue based on priority, deadline and from the end. + +Wed Jan 16 11:24:52 2002 Priyanka Gontla <pgontla@ece.uci.edu> + + * THANKS: + Updated to add Gerhard Voss <Gerhard_Voss@t-online.de>. + +Wed Jan 16 06:19:01 2002 Douglas C. Schmidt <schmidt@siesta.cs.wustl.edu> + + * ace/OS.i: Replaced "set" with "sset" in sigtimedwait() and sigwait() + to avoid STL symbol clashes with MSVC++ 6.0. Thanks to Shmulik + Regev <shmul@vself.com> for reporting this. + +Wed Jan 16 09:01:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Containers_T.{h,cpp}: + Added ACE_Fixed_Set_Const_Iterator to make it possible to + iterate through a const ACE_Fixed_Set instance + +Wed Jan 16 07:53:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/ace_flags.bor: + Added new flags for the new TAO ETCL orbsvcs library + +Tue Jan 15 17:24:53 2002 Steve Huston <shuston@riverace.com> + + * ace/SSL/SSL_Context.cpp (report_error()): Set ACE_OS::last_error() + to ERR_get_error() so the caller can get the error code later. + + * ace/SSL/SSL_SOCK_Connector.cpp: If the SSL handshake phase of a + connection attempt fails, close the underlying socket. + +Tue Jan 15 15:35:41 2002 Steve Huston <shuston@riverace.com> + + * ace/SOCK_Connector.(h cpp): + * ace/LSOCK_Connector.(h i cpp): + * ace/MEM_Connector.(h cpp): + * ace/SSL/SSL_SOCK_Connector.(h cpp): + Improved the Doxygenation and removed the protocol_family and + protocol arguments from the ctors and connect() methods. The + protocol family is always taken from the ACE_Addr remote_sap + argument since it can now be either PF_INET or PF_INET6 (for + SOCK_Connector objects) and should be PF_UNIX for LSOCKs. + It is pointless to allow the user to request something that + is impossible to do correctly. + +Tue Jan 15 10:52:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/ace_flags.bor: + Added new flags for the new TAO PortableGroup library + +Mon Jan 14 14:40:25 2002 Carlos O'Ryan <coryan@uci.edu> + + * bin/g++dep: + Fixed small problems in the dependency generation: + - The script did not properly handle files with '+' in their + names. + - In some cases the script generated escaped blanks, i.e. lines + containing a blank preceded by a backslash. Such blanks are + interpreted as part of a dependency name and break havoc with + the builds. + +Mon Jan 14 16:49:37 2002 Steve Huston <shuston@riverace.com> + + * ace/OS.h (ACE_STATIC_SVC_DEFINE): Corrected the documentation to + say the service-implementing class must be derived from + ACE_Service_Object, not ACE_Service_Config. + +Mon Jan 14 07:40:16 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/OS.i (mmap): There was a typo that prevented the ACE Memory Map + stuff from working properly on Win9x. Thanks to Edan Ayal + <edanayal@yahoo.com> for reporting this. + +Sun Jan 13 18:59:37 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Memory_Pool.{h,cpp}: Added a new option that makes is possible + to control whether or not a fixed address will be used when + remapping a memory-mapped file. Thanks to Jonathan Reis + <reis@stentor.com> for this enhancement. + +Mon Jan 14 11:02:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * include/makeinclude/ace_flags.bor: + Added flags for new TAO FT_ORB library + +Sun Jan 13 08:20:05 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/config-all.h: Make sure that ACE_bad_alloc + is defined as std::bad_alloc if + ACE_USES_STD_NAMESPACE_FOR_STDCPP_LIB macro is set. + Fixes gcc 3.0.3 compilation problem. + +Fri Jan 11 22:54:22 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/config-lynxos.h: Added #define ACE_HAS_USING_KEYWORD to teh + file. The compiler supports namespaces. According to the new + rules at the doc_group, we dont use any compilers that dont + support namespaces. The above macro is itself a waste. But we + cannot remove it overnight as it has far reaching + consequences. Working around that for the timebeing. + +Thu Jan 10 18:35:41 2002 Steve Huston <shuston@riverace.com> + + * ace/Profile_Timer.h: Clarified that elapsed_time() calculates time + from start() to stop(). Improved Doxygenation. + +Thu Jan 10 16:53:41 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * examples/Service_Configurator/IPC-tests/server/server.dsp: The + Release version of the library needs to link in ADVAPI32.LIB as + GetUserName is used in ACE's inline code. + +Wed Jan 9 22:07:50 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Logging_Strategy.cpp (fini): Make sure to cancel the + timer if interval_ and max_size_ are > 0. Thanks to Yaniv Ben + Ari <yanivb@bis.co.il> for reporting this. + +Wed Jan 9 11:38:58 2002 Ossama Othman <ossama@uci.edu> + + * tests/SSL/Makefile (LDLIBS): + + Added missing SSL and crypto libraries. Fixed link errors. + Thanks to Marvin Wolfthal <maw@weichi.com> for reporting the + error and suggesting a fix. + +Wed Jan 9 12:24:39 2002 Steve Huston <shuston@riverace.com> + + * ace/Process.cpp (spawn): Don't attempt ACE_OS::setpgid if + ACE_LACKS_SETPGID is defined. Thanks to Victor Terber + <vterber@csksoftware.de> for reporting this. + +Wed Jan 09 11:19:07 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/OS.h: Updated the comment for ACE_CE_Bridge to indicate that + it's obsolete and will be removed in the future. + +Wed Jan 9 00:48:48 2002 Don Hinton <dhinton@gmx.net> + + * ace/Get_Opt.cpp: Make sure to cast away constness + before deleting an ACE_TCHAR array. Thanks to + Bala for reporting this. + +Tue Jan 8 17:29:33 2002 Steve Huston <shuston@riverace.com> + + * ace/SSL/SSL_SOCK_Connector.cpp: Don't try to dereference a 0 + timeout pointer. Gack. Thanks to Ossama for pointing this out. + +Tue Jan 8 15:51:06 2002 Don Hinton <dhinton@gmx.net> + + * ace/Get_Opt.cpp + * ace/Service_Config.cpp: + + Moved the template instantiations from Service_Config.cpp to + Get_Opt.cpp where they belong. + + * ace/Get_Opt.{h.cpp}: Replaced ACE_TString with ACE_TCHAR for + type of member variable ACE_Get_Opt_Long_Option since it + wasn't really needed and took up space. + +Tue Jan 8 10:43:48 2002 Ossama Othman <ossama@uci.edu> + + * ace/config-sunos5.5.h (ACE_LACKS_INET_ATON): + + Solaris does indeed implement the inet_aton() function, but it + is found in `libresolv.*'. It doesn't seem worth it to link + another library just for that function. Just use the emulation + in ACE that has been used for years. + +Tue Jan 8 11:31:22 2002 Steve Huston <shuston@riverace.com> + + * tests/Makefile: When doing realclean, use the DLL_Test and + Service_Config_DLL Makefiles to clean their files up. + +Tue Jan 8 08:36:33 2002 Steve Huston <shuston@riverace.com> + + * ace/SSL/SSL_SOCK_Acceptor.cpp (ssl_accept): + * ace/SSL/SSL_SOCK_Connector.cpp (ssl_connect): Corrected order + of operations checking EWOULDBLOCK, and fixed compile errors. + Thanks to Vlado Chovanec <Vladimir.CHOVANEC@asset.sk> for this fix. + +Mon Jan 7 19:55:39 2002 Steve Huston <shuston@riverace.com> + + * ace/SSL/SSL_SOCK_Acceptor.cpp (ssl_accept): + * ace/SSL/SSL_SOCK_Connector.cpp (ssl_connect): Added extra check + for SSL_accept/connect status failure to avoid looping on a bad + socket if the socket closes during handshake. Thanks to Vlado + Chovanec <Vladimir.CHOVANEC@asset.sk> for this fix. + + Also added timeout countdown support for SSL_SOCK_Connector, same + as in: Sun Jan 6 09:37:02 2002 Ossama Othman <ossama@uci.edu> + +Mon Jan 7 15:55:26 2002 Ossama Othman <ossama@uci.edu> + + * ace/INET_Addr.cpp (set): + + Pass a pointer to a "struct in_addr" to inet_aton(), i.e. the + proper type, instead of a forcibly casted ACE_UINT32. Also + updated existing code to use the in_addr::s_addr member instead + of the previous ACE_UINT32 variable. + +Mon Jan 7 15:13:09 2002 Mayur Deshpande <mayur@ics.uci.edu> + + * performance-tests/Misc/context_switch_time.cpp (main): + Since the Yield_test does seem to work on VxWorks now (see + ChangeLog below), the 'ifdefs' for bypassing VxWorks for the + Yield-Test have now been removed. + +Mon Jan 7 15:08:25 2002 Mayur Deshpande <mayur@ics.uci.edu> + + * ace/OS.i (thr_yield): + Changed ::taskDelay (1) to ::taskDelay (0) for VxWorks in + thr_yield (). The change with (0), now does seem to perform + the yield correctly as reflected in the Yield-Test of + context_switch_time. Thanks to Charlie Grames + <charlie.grames@windriver.com> for this tip. + +Mon Jan 7 15:16:10 2002 Ossama Othman <ossama@uci.edu> + + * ace/OS.h (INADDR_NONE): + + If the platform does not define this constant, then define it. + + * ace/OS.cpp (inet_aton): + + For some reason we were emulating inet_aton() on all platforms + using the now deprecated inet_addr() function. Use the native + inet_aton() function unless ACE_LACKS_INET_ATON is defined. + + Instead of performing a memcpy() of the IPv4 32-bit address into + the in_addr data structure, simply assign it to the s_addr field + of that data structure. It's not clear why we didn't do this in + the first place. + + (inet_ntoa): + + Fixed PSoS emulation of this method. The result is supposed to + be stored in a statically allocated string, not a dynamically + allocated one. Fixes a memory leak. Note that this change + makes the implementation non-reentrant. However, inet_ntoa() + was not designed to be reentrant to begin with. + + * ace/OS.i (inet_addr): + + On error, inet_addr() is supposed to return INADDR_NONE. + + The return value should be a 32 bit unsigned integer, not a + signed one. + + * ace/config-win32-common.h: + + MS Windows does not support the inet_aton() function. Define + ACE_LACKS_INET_ATON. + +Mon Jan 7 12:20:26 2002 Ossama Othman <ossama@uci.edu> + + * bin/auto_run_tests.lst: + + Added the MT_SSLIOP test to the regression test suite list. + +Sun Jan 6 21:19:10 2002 John Aughey <jha@cs.wustl.edu> + + * tests/run_test.lst: Uncommented out Conn_Test from daily builds. + +Sun Jan 6 21:09:10 2002 John Aughey <jha@cs.wustl.edu> + + * ace/INET_Addr.cpp: + * ace/INET_Addr.h: + + Reverted to January 1 version until I have time to put the + set_host_name() method in correctly. + +Sun Jan 6 20:01:10 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * tests/run_test.lst: Commented out Conn_Test from the daily + builds. This test seems to hang blocking build progress. Have + sent a mail to John Aughey on this. + +Sun Jan 6 09:37:02 2002 Ossama Othman <ossama@uci.edu> + + * ace/SSL/SSL_SOCK_Acceptor.cpp (accept, ssl_accept): + + Take into account the time to complete the basic TCP handshake + and the SSL handshake. Specifically, ACE_Countdown_Time is used + to reduce the timeout value after each IO operation + (e.g. accept(), SSL_accept()) used during SSL passive connection + establishment. [Bug 1110] + + Commented out debugging statements. + +Sat Jan 5 20:57:36 2002 Venkita Subramonian <venkita@cs.wustl.edu> + + * ace/Future.cpp (get): Added another ACE_const_cast in addition + to Doug's changes to fix compile errors. See below. + +Sat Jan 5 14:57:36 2002 Craig Rodrigues <crodrigu@bbn.com> + + * ace/OS_QoS.h: Fix comments, put in doxygen format. + +Sat Jan 5 08:59:41 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> + + * ace/Future.cpp (get): Added an ACE_const_cast() to silence certain + C++ compilers. Thanks to Venkita for reporting this. + +Fri Jan 5 15:17:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/OS.{h,cpp}: + Added ACE_TSS_Emulation::release_key() method to release a + thread_key within the TSS_Emulation when a thread is stopped. + Added ACE_TSS_Emulation::tss_keys_used_ member to administrate which + thread_keys are used and which not. + Added ACE_TSS_Keys::is_set() method to test whether a specific + thread_key is marked as used. + Changed ACE_TSS_Emulation::next_key() method to return a thread_key + that is not used yet, this key is then marked as used at the same + time. + Changed ACE_OS::thr_keyfree() method to release the key in the + TSS_Emulation when ACE_HAS_TSS_EMULATION is defined. + + These changes fix the bugzilla bugs 223 and 657. The ACE_TSS_Emulation + now recycles keys that are released earlier. + +Fri Jan 4 19:59:03 2002 John Aughey <jha@cs.wustl.edu> + + * ace/INET_Addr.cpp: Fixed the new set_host_name method + +Fri Jan 4 18:59:27 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ChangeLogs/ChangeLog-01b: Added a new file. Trimmed this file + to have entries only in 2002. + +Fri Jan 4 15:50:42 2002 Steve Huston <shuston@riverace.com> + + * ace/SSL/SSL_SOCK_Acceptor.cpp (ssl_accept): + * ace/SSL/SSL_SOCK_Connector.cpp (ssl_connect): + On ACE::select-reported timeout or failure, set status to return + a -1 to caller, not 0. Thanks to Vladimir Chovanec + <Vladimir.CHOVANEC@asset.sk> for reporting this and sending a fix. + +Fri Jan 4 08:31:49 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> + + * tests/Thread_Manager_Test.cpp (test_task_record_keeping): Fixed + a typo in an expression on line 226. Thanks to Harvinder + Sawhney <harvindersawhney@yahoo.com> for reporting this. + +Fri Jan 4 05:51:22 2002 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> + + * ace/Future.{h,cpp}: Made the get() and ready() methods const. + Thanks to Ran Kohavi <ran@kashya.con> for reporting this. + +Fri Jan 4 15:06:31 2002 Steve Huston <shuston@riverace.com> + + * ace/String_Base.h (operator=): Add <CHAR> to ACE_String_Base + return type. Fixes compile error on IBM C/C++. + + * ace/SSL/SSL_SOCK_Acceptor.cpp (ssl_accept): + If SSL_get_error() returns SSL_ERROR_SYSCALL and it's EWOULDBLOCK, + don't blindly set both read and write handles for select. Check + if SSL is indicating SSL_want_write() and set the proper handle. + Also, don't ACE_ASSERT SSL_pending before return... if there's + an SSL handshake screw-up (like someone trying to break in) + just report the failure, don't abort/crash. + +Wed Jan 02 13:27:09 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/INET_Addr.h: + * ace/INET_Addr.cpp: Removed tabs and trailing whitespaces. + +Wed Jan 2 08:19:18 2002 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> + + * ace/FILE_Connector.h, + * ace/OS.h (ACE_OS): Clarified the weak semantics of O_APPEND + on Win32. Thanks to Eugene Alterman <eugalt@myrealbox.com> for + reporting this. + +Wed Jan 2 12:43:00 2002 John Aughey <jha@aughey.com> + + * ace/INET_Addr.h + * ace/INET_Addr.cpp : Added set_host_name method and moved + relevant code into this method. Changed signature of + set_address method to take a void pointer rather than a + char *. + +Wed Jan 2 12:30:01 2002 Chris Gill <cdgill@cs.wustl.edu> + + * ace/RB_Tree.i + * tests/RB_Tree_Test.cpp : added check for valid current node to + forward_i and reverse_i methods of iterator base class. Thanks to + Craig L. Ching <cching@mqsoftware.com> for reporting this! + +Wed Jan 2 08:19:18 2002 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> + + * tests/README: Clarify that run_test.pl should be used rather + the run_tests.sh. + + * tests/run_tests.bat: Clarify that run_test.pl should be used + on Win9x. Thanks to Edward A Thompson <ed4becky_2000@yahoo.com> + for prompting this. + +Wed Jan 2 07:37:01 2002 Balachandran Natarajan <bala@cs.wustl.edu> + + * ace/Handle_Set.h: + * ace/Handle_Set.cpp: Added a method reset_state () to the + ACE_Handle_Set_Iterator class. + +Tue Jan 2 11:39:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/OS.i: + Added missing ACE_UNUSED_ARG in ACE_OS::event_timedwait + +Tue Jan 1 15:36:39 2002 Steve Huston <shuston@riverace.com> + + * include/makeinclude/platform_sunos5_sunc++.GNU: Added support + for the buildbits=64 make option. + +Tue Jan 1 20:05:12 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * ace/Name_Request_Reply.{h,cpp}: + Changed type of 3 constructor arguments from size_t to ACE_UINT32 + because the members in which these arguments are stored are also + of type ACE_UINT32 + + * ace/OS.i + In ACE_OS::umask method, changed the type in the ACE_OSCALL_RETURN + macro from int to mode_t because that is the return type of the + method + +Tue Jan 1 08:47:25 2002 Douglas C. Schmidt <schmidt@siesta.cs.wustl.edu> + + * ace/Thread.h: Clarify how the ACE_Thread_Adapter is deleted + when spawn() is called. Thanks to Preston Elder + <prez@srealm.net.au> for reporting this confusion. + +Tue Jan 1 14:09:26 2002 Johnny Willemsen <jwillemsen@remedy.nl> + + * examples/Map_Manager/test_hash_map_manager.cpp: + Made this example compiling when ACE_USES_WCHAR is set + + * Makefile.bor: + Added examples directory because all examples for which there are + BCB makefiles now build when ACE_USES_WCHAR is set + +Tue Jan 1 00:02:12 2002 Nanbor Wang <nanbor@cs.wustl.edu> + + * ace/ace_dll.vcp: Added String_Base_Const.*. |