diff options
Diffstat (limited to 'ChangeLogs/ChangeLog-03a')
-rw-r--r-- | ChangeLogs/ChangeLog-03a | 6070 |
1 files changed, 0 insertions, 6070 deletions
diff --git a/ChangeLogs/ChangeLog-03a b/ChangeLogs/ChangeLog-03a deleted file mode 100644 index bb91209b18f..00000000000 --- a/ChangeLogs/ChangeLog-03a +++ /dev/null @@ -1,6070 +0,0 @@ -Fri Feb 23 19:44:36 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * ACE version 5.1.13 released. - -Thu Feb 22 12:28:15 2001 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/TP_Reactor.cpp (handle_events): Here is the description of - how things were working and where the problem was: In the - TP_Reactor there is one leader thread waiting in select() and - the others waiting on the token. During deactivation(), the - <deactivation_> flag is set on the reactor, signal_all_threads() - is called on the token, and an empty notification is send to the - current leader thread. signal_all_threads() sets the - <signal_all_threads_> flag on the Token when there are waiters - on the Token (does nothing if there are no waiters). This - scheme worked fine when there were leader and follower threads - waiting. When a new leader is chosen, it returns the magic - number 2 from token.acquire_read() and hence exits gracefully - from the handle_events(). However, the problem was that when - some (or all) of the TP threads are busy running upcalls, - signal_all_threads() was a no-op and the magic number 2 was not - returned from token.acquire_read() and hence the exit from - handle_events() was flagged as an error. - - The fix is as described below: - - (1) ACE_Token::signal_all_threads() does not make sense. The - token is a mutex, not a condition variable. Therefore, it does - not make sense to have a signal_all_threads() methods. Plus I - believe that the method does not do as advertised. Therefore, I - removed this method from the token class. - - (2) There is no need to signal all the threads in the - TP_Reactor. Marking the Reactor as closed and signaling the - leader thread is enough. The leader thread will wake up to - handle the empty event, the next leader will see that the - Reactor is closed and exit gracefully. This will continue until - all the waiters are drained. - - (3) With the above change, there is no need to check for an - magic returns from token.acquire_read(). When a thread gets the - token, it checks the deactivation flag before proceeding. If - the Reactor is closed, it gracefully exits. Otherwise, it - continues as leader. - - (4) Since there is no need to call token.signal_all_threads(), - ACE_TP_Reactor::wakeup_all_threads() can be removed since it is - now the same as ACE_Select_Reactor_T::wakeup_all_threads(). - - Note that this change should help with making - Thread_Pool_Reactor_Test run without shutdown errors. - - * ace/Select_Reactor_T.i (deactivate): The setting of the - <deactivated_> flag and waking up of all the threads should be - atomic. I am being picky about this one but I think am I right. - Also added the guard to the accessor. - -Thu Feb 22 08:51:42 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS.i: - - Bleh. Changed an && to || to make sure we get a symbolname - variable for non-Unicode Win32. - - "He who lives without committing any folly is not so - wise as he thinks" - La Rochefoucauld - -Thu Feb 22 08:43:30 2001 Carlos O'Ryan <coryan@uci.edu> - - * include/makeinclude/rules.local.GNU: - Add SunWS_cache and ir.out to the list of files removed by - the realclean pseudo-target. This directory is used by Sun/CC - 5.0 (and Forte), if not removed the information there can become - stale, and break the builds. - This should fix the Sun/CC 5.0 nightly builds. - -Thu Feb 22 07:51:36 2001 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/OS.i: Reverted the change "Thu Feb 22 03:31:06 2001 - Krishnakumar B <kitty@ace.cs.wustl.edu> ". - -Thu Feb 22 03:31:06 2001 Krishnakumar B <kitty@ace.cs.wustl.edu> - - * ace/OS.i (dlsym): - Hopefully fixed the Win32 builds. Removed a extra ! in the - #ifdef. - -Thu Feb 22 02:58:36 2001 Krishnakumar B <kitty@ace.cs.wustl.edu> - - * VERSION: - * PROBLEM-REPORT-FORM: - * ace/Version.h: - Reverted the changes and restored the old versions of - these files as the beta build script was killed. Pain :-( - -Wed Feb 21 20:10:42 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS.i: - * tests/DLL_Test.cpp: - * tests/Thread_Pool_Reactor_Test.cpp: - - Fixed the remaining problems with the Win2000_Unicode tests. - [Bug 537] - -Tue Feb 20 21:19:39 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/Svc_Conf.h: - Add ACE_Export macro for ace_create_service_type() it is used - outside the library now. - -Tue Feb 20 06:34:26 2001 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * include/makeinclude/platform_sunos5_g++.GNU: Make sure to explicitly - set ACE_HAS_GNUG_PRE_2_8. Thanks to Christian von Mueffling - <cvm@aiss.de> for reporting this. - - * ace/Thread_Manager.cpp (cancel_grp): Fixed a typo in - the ACE_TRACE macro. Thanks to Mike Curtis for reporting this. - -Mon Feb 19 16:38:08 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * ace/Asynch_Acceptor.{h,cpp}: - Added method void handle (ACE_HANDLE) and made it call the base - class's handle(ACE_HANDLE) method to remove warning about - partial overloading. Fix suggested by Irfan. - -Sun Feb 18 21:55:16 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * ace/OS.h: - Added an extern "C" void ace_mutex_lock_cleanup_adapter before - the definition of ACE_PTHREAD_CLEANUP_PUSH to avoid warnings on - SunCC 5.x. - -Sun Feb 18 17:16:41 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * netsvcs/clients/Tokens/manual/manual.cpp (main), - * tests/Tokens_Test.cpp (main): Reordered the #ifdefs for the - ACE_HAS_TOKENS_LIBRARY stuff. Thanks to Kitty for reporting - this. - -Sun Feb 18 16:28:45 2001 Chris Gill <cdgill@cs.wustl.edu> - - * ace/RB_Tree.h: Changed order of method declarations to (hopefully) - dodge an internal compiler error on VxWorks. - -Sun Feb 18 09:59:18 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * netsvcs/clients/Tokens: - * netsvcs/servers/main.cpp (main): - * netsvcs/servers/svc.conf: - * netsvcs/lib/Token_Handler: Only compile this stuff if the - ACE_HAS_TOKENS_LIBRARY macro is enabled. Thanks to Carlos - for reporting this. - - * ace/Object_Manager.cpp: Added a check for ACE_HAS_TOKENS_LIBRARY - so we can skip initializing certain things if that macro is - undefined. - - * ace/Token_Collection.*: - * ace/Token_Invariants.*: - * ace/Token_Manager.*: - * ace/Local_Tokens.*: - * ace/Remote_Tokens.*: - * ace/Token_Request_Reply.*: - Only compile these files if ACE_HAS_TOKENS_LIBRARY is enabled - since this functionality does not appear to be widely used and - yet it takes up space. - - * tests/Tokens_Test.cpp (main): Only run this test if - ACE_HAS_TOKENS_LIBRARY is defined. - -Sat Feb 17 17:23:46 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/SOCK_Dgram_Mcast.cpp: - - Integrated a fix from Keith Nicewarner <knicewar@erg.sri.com> - In effect, this reverts the multicast code changed on - Fri Jan 26 11:18:15 2001 Douglas C. Schmidt ... - for all platforms except linux. - -Sat Feb 17 15:52:09 2001 Carlos O'Ryan <coryan@uci.edu> - - * bin/run_all_list.pm: - * tests/run_tests.pl: - Remove obsolete scripts. - -Sat Feb 17 13:26:17 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/POSIX_Asynch_IO.cpp (thread_function): Make sure to cast - -1 to void * since that's the signature of the method. Thanks - to Bala for reporting this. - - * ace/POSIX_Asynch_IO.cpp (thread_function): Zapped the unused - result local variable and also changed the logic to bailout if - handle_events() returns -1. This fixes a warning reported by - KCC. - -Fri Feb 16 17:39:16 2001 Mayur Deshpande <mayur@ics.uci.edu> - - * ace/config-qnx-rtp.h: Commented out ACE_HAS_NONSTATIC_OBJECT_MANAGER. - Thanks to Andrew Olson <andrew.olson@lmco.com> for the suggestion. - - * include/makeinclude/platform_qnx_rtp_gcc.GNU: Added -fexceptions to - CCFLAGS and 'exceptions=1' at end of file. - Thanks to Andrew Olson <andrew.olson@lmco.com> for the suggestion. - -Fri Feb 16 05:37:47 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Logging_Strategy.h: Updated the documentation to explain what - the various options mean. - - * tests/Log_Msg_Test.cpp (test_log_msg_features): Added - ACE_Log_Msg::PROCESS to a coupld or priority_mask() calls to - make the test work correctly. Thanks to Kitty for reporting - this. - -Thu Feb 15 17:53:31 2001 Frank Hunleth <fhunleth@cs.wustl.edu> - - * ace/Thread_Manager.i: Simplified casting in - ACE_Thread_Descriptor::get_next per suggestion from Carlos. - -Thu Feb 15 14:57:59 2001 Frank Hunleth <fhunleth@cs.wustl.edu> - - * ace/Thread_Manager.i: Fixed typo in ACE_Thread_Descriptor::get_next. - -Thu Feb 15 11:33:55 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * tests/Refcounted_Auto_Ptr_Test.cpp (main): Always call ACE_OS::exit(0) - to make sure that this test terminates... - - * ace/Thread_Manager.i (get_next): Added an ACE_const_cast() to - make certain compilers happy. - - * ace/Logging_Strategy.cpp: Fixed some Unicode problems on Win32. - -Thu Feb 15 08:49:44 2001 Carlos O'Ryan <coryan@uci.edu> - - * bin/nightlybuilds/builds.lst: - Rename WashU builds so all the Debian builds appear together. - -Thu Feb 15 07:15:39 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Logging_Strategy.{h,cpp}: Added support to generate multiple - log files, rather than just two. Thanks to Orlando Ribeiro - <oribeiro@inescporto.pt> for this fix. - -Thu Feb 15 03:36:46 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * bin/nightlybuilds/builds.lst (WEB): - Added a new build with TAO_MINIMUM linked statically. - -Wed Feb 14 05:28:59 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Thread_Manager.{h,i}: Added const to state(), grp_id(), - task(), get_next(), and self(). Thanks to Johnny Willemsen - <jwillemsen@remedy.nl> for reporting this. - -Tue Feb 13 17:51:47 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * tests/Makefile.bor (TESTS): Added Dirent_Test to list of TESTS. - Thanks to Mike Curtis for reporting this. - - * tests/run_tests.bat: Fixed a typo so that Dirent_Test will run. - Thanks to Mike Curtis for reporting this. - -Wed Feb 14 12:08:10 2001 Frank Hunleth <fhunleth@cs.wustl.edu> - - * ace/Read_Buffer.cpp: Fixed boundary condition in rec_read. - If record was a multiple of BUFSIZ and the record was - terminated with an EOF instead of a "term" character, - rec_read would return 0 instead of the actual data read in. - The problem was that the partial buffer fill size was being - checked instead of the total number of bytes read in. This - issue was noticed when trying to read in an IOR file in TAO - that was exactly 512 bytes (BUFSIZ * 2 on Win2k) long. - -Tue Feb 13 14:44:28 2001 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/ace_wchar.inl: Changed several variable from type <size_t> - to type <int> as on 64-bit whistler, size_t and int are not the - same size which confuses 64-bit compiler. (Per Kitty's - request.) - -Tue Feb 13 10:23:31 2001 Steve Huston <shuston@riverace.com> - - * apps/JAWS/server/IO.{h cpp}: (Side-affect necessitated by addition of - handle () method and handle_ member to ACE_Handler). Moved the - handle_ member to the JAWS_Synch_IO class - it needs it, and - JAWS_Asynch_IO class has it by inheritance from ACE_Handler. Left - the handle () methods in JAWS_IO, but made them pure virtual, - implemented in JAWS_Synch_IO (implementation moved from JAWS_IO) - and in JAWS_Asynch_IO, where they forward up to ACE_Handler's - handle () methods. - -Mon Feb 12 18:18:05 2001 Frank Hunleth <fhunleth@cs.wustl.edu> - - * tests/Thread_Mutex_Test.cpp: - Modified test to use process mutexes for Win32 platforms - so that timed mutexes could be tested. - -Sun Feb 11 15:33:39 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/QoS/ACE_QoS_Export.h: - - Added ACE_AS_STATIC_LIBS support. - - * ace/QoS/ACE_QoS.dsp: - * ace/QoS/ACE_QoS_Lib.dsp: - - Fixed to comply with ACE standards. - -Sun Feb 11 10:50:10 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * include/makeinclude/platform_sunos5_g++.GNU (exceptions): - Only set the -fno-exceptions option if we're *not* using GCC - 2.7.2! Thanks to Pradeep for explaining how to do this. - -Sun Feb 11 10:15:08 2001 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Makefile.bor: - * ace/QoS/Makefile.bor: - * include/makeinclude/ace_flags.bor: - Added build support for QoS library with Borland compiler. - -Sat Feb 10 15:29:08 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/Message_Block.h: - Improved documentation for the lenth(), size() and space() - methods. - -Sat Feb 10 07:19:37 2001 Christopher Kohlhoff <chris@kohlhoff.com> - - * include/makeinclude/outputdir.bor: - Removed the now out-of-date comment about pascal/non-pascal - build directories. - - * ace/config-win32-borland.h: - Added support for new ACE_HAS_VCL define to enable inclusion - of VCL header files. - -Fri Feb 9 16:47:44 2001 Pradeep Gore <pradeep@cs.wustl.edu> - - * ace/Makefile.am: - Removed references to the AQoSA files. - -Fri Feb 9 16:42:10 2001 Pradeep Gore <pradeep@cs.wustl.edu> - - * examples/QOS/client.dsp: - * examples/QOS/server.dsp: - Updated the corrected project files. - Thanks to Darrell for pointing this out. - -Thu Feb 8 19:25:03 2001 Carlos O'Ryan <coryan@uci.edu> - - * performance-tests/RPC/run_test.pl: - * performance-tests/TCP/run_test.pl: - * performance-tests/UDP/run_test.pl: - Add driver strings - -Thu Feb 08 19:18:36 2001 Angelo Corsaro <corsaro@cs.wustl.edu> - - * bin/auto_run_tests.lst: - - Removed comment from Exposed Policies test. The test has been fixed. - -Thu Feb 8 14:11:36 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/ws2tcpip.h: [Removed] - - I don't see any reason to have this file in ace; it is a system - header file. And from [Bug 722], our version is out of date anyway. - - * ace/OS.h: - - Changed to include system version of ws2tcpip.h - - * ace/Makefile.am: - * ace/ace-dll.icc: - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - - Removed references to ws2tcpip.h. - - * ace/SOCK_Connect.cpp: - - Changed get_ip_interfaces to remove a memory error. Thanks to - Viatcheslav Batenine <slavikb@infra2000.com> for helping out with - this. - - * ace/SOCK_Dgram_Mcast.cpp: - - It looks like the change made on: - - Fri Jan 26 11:18:15 2001 Douglas C. Schmidt ... - - doesn't work on Win32. Reverted back to the old scheme for Win32 - only. - -Thu Feb 8 00:34:57 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * include/makeinclude/platform_sunos5_g++.GNU: Added support for - -fno-exception when exceptions=0 is given to the Makefile. - Thanks to Christian v. Mueffling <cvm@aiss.de> for reporting - this. - -Wed Feb 7 23:02:46 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/SUN_Proactor.cpp: Fixed the cancel_aiocb() method and added a - new overridden cancel_aio() method. Thanks to Alexander Libman - <Alibman@baltimore.com> for contributing this. - -Thu Feb 8 02:09:05 2001 Pradeep Gore <pradeep@cs.wustl.edu> - - * examples/QOS/Makefile: - * examples/QOS/client.dsp: - * examples/QOS/server.dsp: - Modified to use the ACE_QoS lib. - * Receiver_QoS_Event_Handler.cpp: - * Sender_QoS_Event_Handler.cpp: - Added missing Log_Msg.h for ACE_DEBUG. - Other files were modified to reflect changed directory location - of AQosA imclude files. - -Thu Feb 8 00:17:43 2001 Pradeep Gore <pradeep@cs.wustl.edu> - - * ace/QoS_Decorator.cpp: - * ace/QoS_Session_Factory.cpp: - * ace/SOCK_Dgram_Mcast_QoS.cpp: - * ace/QoS_Decorator.h: - * ace/QoS_Session_Factory.h: - * ace/SOCK_Dgram_Mcast_QoS.h: - * ace/QoS_Manager.cpp: - * ace/QoS_Session_Impl.cpp: - * ace/SOCK_Dgram_Mcast_QoS.i: - * ace/QoS_Manager.h: - * ace/QoS_Session_Impl.h: - * ace/QoS_Session_Impl.i: - Moved the AQoSA classes to ace/QoS directory. - - * ace/Makefile: - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - Removed the QoS related files, added QoS directory to build. - - * ace/Qos/Makefile - * ace/QoS/ACE_QoS.dsw: - * ace/QoS/ACE_QoS.dsp: - * ace/QoS/ACE_QoS_Lib.dsp: - * ace/QoS/ACE_QoS_Export.h: - Added Makefile and MSVC project files. - Thanks to Craig Rodrigues <crodrigu@bbn.com> for suggesting this - change and providing the patch. - -Wed Feb 7 11:34:56 2001 Carlos O'Ryan <coryan@uci.edu> - - * tests/Message_Queue_Test_Ex.cpp: - Fix build problems for single threaded environments. - -Wed Feb 7 09:03:22 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * apps/Gateway/Gateway/Event_Channel.cpp (initiate_acceptors): Fixed - incorrect if/else branching. Thanks to Lu Yunhai - <luyunhai@huawei.com> for reporting this. - -Tue Feb 6 23:43:49 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * apps/Gateway/Peer/Peer.cpp (await_events): Fixed the use of - %s to be %*s since the string may not be nul terminated! Thanks - to Lu Yunhai <luyunhai@huawei.com> for reporting this. - -Tue Feb 6 20:11:28 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * include/makeinclude/platform_sunos5_sunc++.GNU: - - Remove double inclusion of $(CFLAGS) for Sun CC 5.0. - -Tue Feb 6 19:30:26 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * ace/Malloc_T.cpp (shared_malloc): - Added a #ifdef to remove the warnings from Digital Tru64 cxx. This - is a temporary workaround. TODO: Change ACE_SEH_* macros so that such - hacks are unnecessary. - -Tue Feb 6 19:23:16 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * include/makeinclude/platform_sunos5_sunc++.GNU: - Complete reverting the changes made by Irfan on Sat Jan 27 - 01:06:09 2001. This file was missed in the last revert done by me. - -Tue Feb 06 19:03:35 2001 Steve Huston <shuston@riverace.com> - - * include/makeinclude/platform_sunos5_sunc++.GNU: For C++ 5, make sure - to add CFLAGS into CCFLAGS, even if no other options are added. Else - things get built without -mt. - -Tue Feb 06 18:37:41 2001 Steve Huston <shuston@riverace.com> - - * tests/Thread_Mutex_Test.cpp: Changed wording of the message for - "timed acquire" not supported so as not to trip the error catcher - in run_tests.pl. - -Tue Feb 06 17:55:28 2001 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/post_completions.cpp: Allow this to - run with AIOCB_Proactor. - - * ace/Asynch_IO.{h cpp}: Added ACE_Handler::handle (ACE_HANDLE) - function to set the handle; changed ACE_HANDLE ACE_Handler::handle - to return the actual value. - - * ace/Asynch_Acceptor.cpp (handle_accept): Call ACE_Handler::handle - to set the handle value for the new handler before calling its - open() method. - -Tue Feb 06 17:16:43 2001 Steve Huston <shuston@riverace.com> - - * tests/MEM_Stream_Test.cpp: - * tests/MEM_Stream_Test.h: New file; moved definition of Echo_Handler - from MEM_Stream_Test.cpp to this file to make it visible to AIX - C++ 3.6.6 at template instantiation time. - -Mon Feb 5 16:19:20 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/ACE.cpp: ACE_TRACE ("ACE::basename") should be ACE_TRACE ("ACE::diname"). - Thanks to Lu Yunhai <luyunhai@huawei.com> for reporting this. - - * ace/Basic_Types.h: Changed the ACE_BIG_ENDIAN and ACE_LITTLE_ENDIAN - macros to use the form 0x0123 rather than 0123X. Thanks to - Craig Rodrigues <crodrigu@bbn.com> for helping find this - problem. - -Mon Feb 5 18:20:11 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * ace/config-sunos5.5.h (ACE_USES_OLD_IOSTREAMS): - - Reverted the change to config-sunos5.5.h and turned on - ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION, - ACE_TEMPLLATES_REQUIRE_SOURCE and ACE_USES_OLD_IOSTREAMS for the - SunCC 5.0 compiler. - -Mon Feb 5 16:19:20 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * examples/RMCast/Send_File/Receiver.cpp (init): Replaced - a hard-coded file permission of 0644 with - ACE_DEFAULT_FILE_PERMS. Thanks to Chris Kohlhoff - <chris@kohlhoff.com> for reporting this. - -Mon Feb 5 13:00:17 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * tests/Thread_Mutex_Test.cpp: (reported_notsup) - Removed a warning about defined but not used. - -Sat Feb 3 14:41:17 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS.h: - * ace/config-win32-msvc.h: - - Fixed problems with LLSEEK and ACE_stat on Windows CE. - -Sat Feb 3 14:12:36 2001 Darrell Brunsch <brunsch@uci.edu> - - * tests/Config_Test.cpp: - * tests/Config_Test.h: - * tests/Log_Msg_Test.cpp: - - Fixed ACE_USES_WCHAR problems. - -Fri Feb 02 18:36:42 2001 Steve Huston <shuston@riverace.com> - - * ace/config-linux-common.h: Added #define ACE_POSIX_AIOCB_PROACTOR - unless the user #defines ACE_POSIX_SIG_PROACTOR before including - config-linux.h. (At least prior to 2.4 kernels) queueing signals - across threads does not work - it fails silently, which breaks - basic functionality of the async model - like asynchronously - accepting TCP connections - and pretty much prevents multi-threaded - work with timers and probably other async operations. - - * examples/Reactor/Proactor/test_proactor.cpp: Disabled the ACE_DEBUG - calls that display message buffers - they're often too large and - overrun the ACE_Log_Msg buffers and crash the program. - -Fri Feb 2 08:38:37 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/WFMO_Reactor.cpp (handler): Set the existing_masks variable to 0 - to avoid warnings from static analysis tools. Thanks to Ivan - Murphy for reporting this. - -Thu Feb 01 22:00:17 2001 Sharath R. Cholleti <sharath@cs.wustl.edu> - - * tests/Message_Queue_Test_Ex.cpp - - Fixed the bugs which were giving runtime error on - Win32 - removed "queue_type" from single_thread_performance_test - and performance_test functions. Now these funtions dont - have args. - - * tests/tests.dsw - - Just an update of the workspace. - -Thu Feb 1 21:00:57 2001 Frank Hunleth <fhunleth@cs.wustl.edu> - - * bin/auto_compile_win32.pl: - * bin/run_all_list.pm: - - Added TAO RT Mutex unit tests to auto build and test - lists. - -Thu Feb 01 18:52:27 2001 Angelo Corsaro <corsaro@cs.wustl.edu> - - * RTJava: - Added Directory for RTJava related research. Check the - ChangeLog included in the newly directory for more info. - -Thu Feb 01 18:23:17 2001 Sharath R. Cholleti <sharath@cs.wustl.edu> - - * tests/Log_Msg_Test.dsp: - * tests/Config_Test.dsp: - - Added these files. - * tests/tests.dsw: - - Added the above two projects to the workspace. - -Thu Feb 1 18:01:26 2001 Luther J Baker <luther@cs.wustl.edu> - - * ace/OS.i (mutex_init): - - Removed call into PACE. - -Thu Feb 01 18:21:15 2001 Steve Huston <shuston@riverace.com> - - * tests/Thread_Mutex_Test.cpp: Only report "operation not supported" - once for a failed ACE_Thread_Mutex::acquire (ACE_Time_Value &). - Also, if it reports ENOTSUP while ACE_HAS_MUTEX_TIMEOUT is - defined, warns the user that it's probably the wrong setting. - -Thu Feb 01 17:36:28 2001 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/test_proactor3.cpp: Fixed a non-constant - array dimension and some output formatting. - -Thu Feb 01 17:07:35 2001 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/post_completions.cpp: Removed extraneous - 'd' from the 'count' display. - -Thu Feb 01 15:33:26 2001 Steve Huston <shuston@riverace.com> - - * ace/Reactor.h (purge_pending_notifications): - * ace/WFMO_Reactor.h (purge_pending_notifications): - * ace/Select_Reactor_Base.h (purge_pending_notifications): - Removed the default argument value - callers must now explicitly - specify a 0 ACE_Event_Handler* to remove all handler notifications. - Thanks to Edan Ayal <edana@bandwiz.com> for arguing this point. - -Thu Feb 1 09:57:41 2001 Jeff Parsons <parsons@cs.wustl.edu> - - * ace/SString.i: - - Changed behavior of ACE_WString::rep to be the same - as ACE_CString::rep, that is, to return an empty - (w)string rather than a null (w)string when the length - is 0. Thanks to Johnny Willemsen <johnny.willemsen@meco.nl> - for pointing out this discrepancy. - -Thu Feb 1 09:42:41 2001 Chad Elliott <elliott_c@ociweb.com> - - * include/makeinclude/platform_chorus4.x_g++.GNU: - - Modification to allow building for Chorus 4.x on x86. CPU will be - set to x86, but CLX_TYPE will be set to i486-elf-sysv4. - -Thu Feb 01 10:29:31 2001 Steve Huston <shuston@riverace.com> - - * ace/Synch.i (ACE_Null_Mutex::acquire (ACE_Time_Value&)): Removed - the arg name to fix "unused argument" warning from g++. - -Wed Jan 31 22:01:00 2001 Rick Stille <stille_r@ociweb.com> - * ace/OS.cpp(operator FILETIME): Corrected an error in the Jan 26 GHS - revision. The term "ACE_Time_Value::FILETIME_to_timval_skew" had been - omitted from the calculation. - -Wed Jan 31 18:43:26 2001 Krishnakumar B <kitty@cs.wustl.edu> - - * tests/Message_Queue_Test_Ex.cpp (single_thread_performance_test): - Removed the warning about queue_type using ACE_UNUSED_ARG. - -Wed Jan 31 17:27:35 2001 Steve Huston <shuston@riverace.com> - - * ace/Acceptor.cpp (ACE_Oneshot_Acceptor::handle_input): Unregister - the acceptor from the reactor before doing the shared_accept call - to guard against the use-case where the acceptor is destroyed during - the service handler activation sequence. - -Wed Jan 31 15:55:23 2001 Frank Hunleth <fhunleth@cs.wustl.edu> - - * ace/Synch.i: - Added ACE_Null_Mutex::acquire (ACE_Time_Value &tv) - to fix link error. - -Wed Jan 31 15:12:00 2001 Jeff Parsons <parsons@cs.wustl.edu> - - * ace/CDR_Stream.h: - - Made method do_byte_swap() public. If ACE_ENABLE_SWAP_ON_WRITE - is defined, TAO_GIOP_Message_Base::format_message() and - TAO_GIOP_Message_Lite::format_message() both call this method. - Thanks to Paul Rubel <prubel@bbn.com> for reporting this bug. - -Wed Jan 31 12:19:51 2001 Steve Huston <shuston@riverace.com> - - * docs/tutorials/Chap_2/ex03.html, ex04.html: Fixed these example - programs to 1) send the Nul terminators with the strings, - 2) allow enough space on the recv to get it all, 3) return a - value from main. - - * docs/tutorials/Chap_2/Chap_2.zip: Updated the above examples. - -Wed Jan 31 06:43:27 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Message_Queue_T.cpp: Make sure that dequeue_head() returns - the number of items still on the queue. Thanks to Johnny - Willemsen for reporting this. - -Tue Jan 30 21:39:15 2001 Sharath R. Cholleti <sharath@cs.wustl.edu> - - * tests/Log_Msg_Test.cpp: In funtion test_ostream(), removed the - if condition to check if ACE_OS::unlink() succeeds. Also set - the ostream to NULL before calling ACE_ERROR_RETURN(). Thanks - to Mike Curtis <mccurry@my-deja.com> for pointing out problem - caused when ACE_OS::unlink() fails. Thanks to Doug in helping - fix the problem. - -Tue Jan 30 18:04:08 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * tests/Message_Queue_Test_Ex.cpp (single_thread_performance_test): - Zap the Win32 and VxWorks-specific code from this function. - Thanks to Sharath for reporting this. - -Tue Jan 30 11:34:26 2001 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/test_aiosig_ace.cpp: Fixed a bunch - of mis-coded ACE_ERROR_RETURNs; added a explicit queued completion, - and disabled the display of buffers that may overrun ACE_Log_Msg's - internal buffers to prevent spurious crashes. - -Tue Jan 30 11:26:42 2001 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/post_completions.cpp: Added a count - to be sure that all expected completions are delivered, and - print error message if not. - -Tue Jan 30 10:47:52 2001 Steve Huston <shuston@riverace.com> - - * ace/Select_Reactor_Base.cpp - (ACE_Select_Reactor_Notify::purge_pending_notifications): - * ace/WFMO_Reactor.cpp - (ACE_WFMO_Reactor_Notify::purge_pending_notifications): Changed - to not purge notifications destined for the reactor only - (notification buffer with 0 handler value). Thanks to Edan - Ayal <edana@bandwiz.com> for this fix. - -Tue Jan 30 01:05:51 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * tests/Message_Queue_Test_Ex.cpp (single_thread_performance_test): - Add cases for VxWorks and NT. Thanks to Sharath for reporting - this. - - * ace/OS.h: #define _SC_AIO_MAX to 1 if it's not set by default. - Thanks to Jean-Francois Daune <Daune@IBA.be> for reporting this - fix for VxWorks. - -Mon Jan 20 18:36:41 2001 Steve Huston <shuston@riverace.com> - - * ace/Message_Queue_T.h (ACE_Message_Queue_Ex): Use _USE - rather than _DECL to define the ACE_Message_Queue - fixes - compile error. - -Mon Jan 29 17:29:31 2001 Steve Huston <shuston@riverace.com> - - * ace/Message_Queue_T.{h i cpp}: Corrected a bunch of references - to Message_Queue_Ex's queue_ member to use . rather than -> - and removed the ACE_NEW to allocate it (follow-ups from Doug's - Sun Jan 28 08:32:07 2001 change). Fixed other compile - errors to allow tests/Message_Queue_Test_Ex.cpp to build. - FYI... whomever added this, it probably should have been named - Message_Queue_Ex_Test to conform. - -Mon Jan 29 15:30:27 2001 Steve Huston <shuston@riverace.com> - - * ace/POSIX_Proactor.{h cpp} (ACE_POSIX_SIG_Proactor): Don't - mask/block all signals, just the one(s) being used for this - sig proactor instance. Blocking everything can break other - things, such as debuggers and the ability to kill a program. - Also, don't setup a handler for the signals being blocked/waited - as this is a combination whose behavior is undefined. - -Sun Jan 28 08:39:17 2001 Carlos O'Ryan <coryan@uci.edu> - - * tests/Thread_Mutex_Test.cpp: - Fixed use of ACE_TEXT and string concatenation. - -Sun Jan 28 08:03:08 2001 Carlos O'Ryan <coryan@uci.edu> - - * tests/tests.dsw: - * tests/Message_Queue_Test_Ex.dsp: - Add NT project files for the new test - -Sun Jan 28 08:32:07 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * netsvcs/lib/Client_Logging_Handler.cpp: Fixed the order in which - the iostream messages are logged so that we don't bomb out when - the VERBOSE option is enabled. Thanks to Craig Rodrigues - <crodrigu@bbn.com> for helping find this problem. - - * ace/Message_Queue_T.h: Make sure the lock() forwarding method - forwards to queue_.lock(). - - * ace/Message_Queue_T.{h,i,cpp}: There's no reason to dynamically - allocate the implementation of ACE_Message_Queue_Ex -- instead - it can be simply an instance of ACE_Message_Queue. - - * ace/Message_Queue_T.h: Make sure that the methods on - ACE_Message_Queue_Ex are properly virtual. Thanks to Kitty for - pointing this out. - -Sat Jan 27 19:40:36 2001 Ossama Othman <ossama@uci.edu> - - * bin/auto_run_tests.lst: - - Added TAO's Portable Interceptor "Service_Context_Manipulation" - test to the list of tests to run. - -Sat Jan 27 19:34:20 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * tests/Config_Test.{h,cpp}: Fixed the use of "bool," which is - not supported on older compilers. Thanks to Kitty for reporting - this. - -Sat Jan 27 15:22:03 2001 Ossama Othman <ossama@uci.edu> - - * tests/Thread_Mutex_Test.cpp (test): - - Added code that tests mutex timeout functionality. - -Sat Jan 27 15:16:33 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/POSIX_Proactor.h: Replace ACE_Thread_Mutex with ACE_SYNCH_MUTEX - to avoid the need for conditional compilation. - - * ace/SUN_Proactor.h: Don't use ACE_Condition<ACE_Thread_Mutex> - but instead use the ACE_SYNCH_CONDITION, which also avoids the - need to conditionally compile for threaded platforms. - -Sat Jan 27 13:29:51 2001 Ossama Othman <ossama@uci.edu> - - * ace/config-linux.h: - - Declare the pthread_mutex_timedlock() prototype as an - 'extern "C"' function. This fixes an "unresolved symbol" link - error. - -Sat Jan 27 10:39:28 2001 Carlos O'Ryan <coryan@uci.edu> - - * bin/nightlybuilds/builds.lst: - Apparently the blank in 'RedHat 7.0' breaks the scoreboard. - -Sat Jan 27 01:06:09 2001 Irfan Pyarali <irfan@cs.wustl.edu> - - * include/makeinclude/platform_sunos5_sunc++.GNU: SPRO 5.0 - compiler should be able to support automatic template - instantiation. Also, standard C++ library should be ok. -g was - added to DCCFLAGS. - - * ace/config-sunos5.5.h: Turned off - ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION, - ACE_TEMPLATES_REQUIRE_SOURCE and ACE_USES_OLD_IOSTREAMS for the - SPRO 5.0 compiler. - - Thanks to J. Russell Noseworthy <rnosewor@objectsciences.com> - for contributing the above patches. - -Sat Jan 27 00:37:26 2001 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/OS.i: - - - (mutex_lock): Windows code must use relative time since that - is what WaitForSingleObject() expects. Also, fixed extra - "return" before ACE_NOTSUP_RETURN. - - - (thread_mutex_lock): Made thread_mutex_lock() with timeout - close to the thread_mutex_lock() without timeouts. - -Fri Jan 26 21:53:45 2001 Ossama Othman <ossama@uci.edu> - - * ace/OS.h: - * ace/OS.i (thread_mutex_lock): - - Added an overloaded version of this method that accepts a - timeout argument. It delegates to the standard - ACE_OS::mutex_lock() method on non-Windows platforms. - Otherwise, it returns with an ENOTSUP. - - The addition of this method was necessary to fix a compile - problem related to the ACE_Thread_Mutex class. - - * ace/Synch.i (acquire): - - In ACE_Thread_Mutex::acquire() that accepts a timeout argument, - call the corresponding ACE_OS::thread_mutex_lock() method. This - fixes compile-time error on MS Windows builds due to the fact - that ACE_mutex_t and ACE_thread_mutex_t are different types on - that platform. - -Fri Jan 26 17:31:26 2001 Ossama Othman <ossama@uci.edu> - - * ace/OS.i (mutex_lock): - - Fixed unused argument warnings that occured when compiling for - the single threaded case. - -Fri Jan 26 16:58:03 2001 Ossama Othman <ossama@uci.edu> - - * ace/config-linux.h: - - glibc 2.1.x doesn't have the (very useful) __GLIBC_PREREQ - macro. Use a preprocessor test that doesn't rely on it. - -Fri Jan 26 16:48:06 2001 Ossama Othman <ossama@uci.edu> - - * ace/config-qnx-neutrino.h (ACE_HAS_MUTEX_TIMEOUTS): - * ace/config-qnx-rtp.h (ACE_HAS_MUTEX_TIMEOUTS): - - QNX supports pthread_mutex_timedlock(). Yes! - -Fri Jan 26 17:50:38 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/config-win32-common.h: - ace/config-vxworks-5.x.h: Added ACE_HAS_MUTEX_TIMEOUTS macro. - - * ace/OS.i: Added timed mutex operations for Win32, VxWorks, and - pSoS. - -Fri Jan 26 16:31:15 2001 Ossama Othman <ossama@uci.edu> - - * ace/OS.h (mutex_lock): - - Updated documentation for the version of this method that - accepts a timeout value. - - * ace/OS.i (mutex_lock): - - If the platform does not support timed mutexes, then return with - an ENOTSUP. - - * ace/Process_Mutex.h: - * ace/Process_Mutex.inl (acquire): - - Added version of this method that accepts a timeout value. - -Fri Jan 26 17:50:38 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * examples/Bounded_Packet_Relay/Thread_Bounded_Packet_Relay.cpp (report_stats): - Fixed some unintentional tabs. Thanks to Mike Curtis for - reporting this! - -Fri Jan 26 15:34:01 2001 Ossama Othman <ossama@uci.edu> - - * ace/OS.h: - * ace/OS.i (mutex_timedlock, mutex_lock): - * ace/Synch.i (acquire): - - Renamed the ACE_OS::mutex_timedlock() to the overloaded - ACE_OS::mutex_lock(). - -Fri Jan 26 14:13:32 2001 Ossama Othman <ossama@uci.edu> - - * ace/OS.h: - * ace/OS.i (mutex_timedlock): - - Wrapped the pthread_mutex_timedlock() function. It times out if - a lock hasn't been acquired by the given absolute time. - - * ace/Synch.h: - * ace/Synch.i (acquire): - - Added versions of this method that accept an ACE_Time_Value. - They are used to prevent blocking definitely while waiting for a - given mutex to become available. - - * ace/config-linux.h (ACE_HAS_MUTEX_TIMEOUTS): - - glibc 2.2.x supports the pthread_mutex_timedlock() function. - - * ace/config-linux-common.h: - - Undefine ACE_LACKS_{LLSEEK,LSEEK64}_PROTOTYPE prior to include - the KCC configuration file. Thanks to Jeff Greif - <jgreif@befree.com> for reporting this. - - * ace/README: - - Added ACE_HAS_MUTEX_TIMEOUTS to the list of ACE macros. - -Fri Jan 26 11:18:15 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Log_Msg.cpp (dump): Fixed some unintentional tabs. Thanks - to Mike Curtis for reporting this! - - * ace/SOCK_Dgram_Mcast.cpp (open): Modified the call to - ACE_SOCK_Dgram::shared_open() to just pass in this->mcast_ - rather than making a local copy that just contained the port - number. Thanks to Keith Nicewarner <knicewar@erg.sri.com> for - reporting this. - - * tests/Makefile.bor (TESTS): Added Message_Queue_Test_Ex.cpp. - -Fri Jan 26 13:37:30 2001 Luther J Baker <luther@cs.wustl.edu> - - * ace/OS.h: - - If using PACE, ACE_stat typedefs to pace_stat_s. - -Fri Jan 26 11:18:15 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * tests/Message_Queue_Test_Ex.cpp: Fixed some ASYS_* macros with - ACE_* stuff. Thanks to Darrell for reporting this. - - * ace/Message_Queue_T.h, - * ace/Message_Queue_T.i, - * ace/Message_Queue_T.cpp: - Added the ACE_Message_Queue_Ex, which provides a message queue - abstraction whose type is templatized. Thanks to Michael Vitalo - <mvitalo@sprynet.com> for contributing this. - - * tests/run_tests.lst: - * tests/run_test.lst: - * tests: Added support for the new Message_Queue_Test_Ex.cpp file. - Thanks to Michael Vitalo <mvitalo@sprynet.com> for contributing - this. - -Fri Jan 26 09:33:20 2001 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_run_tests.lst: - Add TAO/tests/Object_Loader to the list - - * include/makeinclude/rules.lib.GNU: - Allow both .cpp and .c files in the LSRC macro. This should - let the PACE guys use the real name for their files, and - generate proper dependencies. - -Fri Jan 26 02:16:42 2001 Darrell Brunsch <brunsch@uci.edu> - - * tests/OS_Test.cpp: - - Changed a use of struct stat to ACE_stat. - -Thu Jan 25 19:16:48 2001 Ossama Othman <ossama@uci.edu> - - * ace/config-linux.h: - - Only define ACE_LACKS_NETDB_REENTRANT_FUNCTIONS and - ACE_LACKS_THREAD_STACK_SIZE for glibc < 2.1. glibc 2.1+ has the - netdb reentrant functions (albeit with funky signatures), and it - has pthread_attr_setstacksize(). - - * ace/OS.h: - - Removed trailing whitespace. - - * ace/OS.i (getprotobyname_r, getprotobynumber_r, gethostbyaddr_r, - gethostbyname_r, getservbyname_r): - - Added glibc-specific calls to these functions. This allows us - to avoid grabbing a lock! Thanks to Ben Strong - <bstrong@bluelark.com> for providing the patch. - - * THANKS: - - Added Ben Strong to list of contributors. - -Thu Jan 25 20:10:24 2001 Irfan Pyarali <irfan@cs.wustl.edu> - - * include/makeinclude/platform_irix6.x_sgic++.GNU: - * include/makeinclude/platform_irix5.3_sgic++.GNU: - - Removed -rpath "$(ACE_ROOT)/ace" from LDFLAGS. - - * ace/Message_Queue_T.{h,i}: Moved code from .i to .h to work - around a compiler bug. The Sun Forte 6 (CC 5.1) compiler is - only happy if this is in the header file. - - * ace/Message_Block_T.cpp: Fixed template parameter name mismatch - between declaration and definition. - - Thanks to J. Russell Noseworthy <rnosewor@objectsciences.com> - for contributing the above patches. - -Thu Jan 25 10:50:35 2001 Steve Huston <shuston@riverace.com> - - * ace/Reactor.h: Add comments to schedule_wakeup/cancel_wakeup that - they don't cause an immediate re-examination of the masks like - register_handler does. Thanks to Oscar Rodriquez - <Oscar.Rodriquez@eln.ericsson.se> for prompting this clarification. - - * ace/WFMO_Reactor.cpp (ACE_WFMO_Reactor_Notify::handle_signal): - The call to ACE_Message_Queue::dequeue_head may change the - ACE_Time_Value passed to it, so don't override ACE_Time_Value::zero's - const-ness, lest ACE_Time_Value::zero take on some strange new value. - Thanks to Edan Ayal <edana@bandwiz.com> for this fix. - -Thu Jan 25 05:34:58 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Connector.cpp (connector): Added an ACE_const_cast() to - get rid of the const for the accessor method! - - * ace/Acceptor.cpp (acceptor): Added an ACE_const_cast() to - get rid of the const for the accessor method! - - * ace/Log_Msg.{h,cpp}, - * ace/Trace.{h,cpp}: The Log_Msg -I flag now uses the current - ACE_Trace nesting level rather than always using 1. Thanks to - Mike Curtis for providing this. - -Wed Jan 24 19:43:24 2001 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/OS.i (lstat,stat): Separated win32 cases out. - -Wed Jan 24 18:15:59 2001 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/OS.i (fstat): Oops, wrong version of this function was - modified. - -Wed Jan 24 14:44:12 2001 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/FILE.cpp: - * ace/Filecache.cpp: - * ace/Filecache.h: - * ace/OS.h: - * ace/OS.i: - * ace/OS.cpp: Changed the signature of <stat>, <fstat>, and - <lstat> to use ACE_stat instead of "struct stat". Also changed - to use _fstat and _stat on win32 instead of <fstat> and <stat> - directly. Also replaced several use of "struct stat" to - ACE_stat. Thanks to Mike Anderson <manderson@sonusnet.com> for - reporting this. [Bug 762] - -Wed Jan 24 16:13:10 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/OS_Dirent.cpp: Fixed a bug in the readdir_emulation() method. - Thanks to Yung Trinh <ytrinh@email.com> for reporting this. - - * ace/Connector.{h,cpp}: Added a new connector() accessor to - return the underlying ACE_PEER_CONNECTOR. Thanks to Pedro - Brandao <pbrandao@inescn.pt> for suggesting this. - -Wed Jan 24 10:17:03 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile: - Add dependencies (again). They were lost is a previous commit. - - * */Makefile: - Update dependencies (yet another time). - -Wed Jan 24 09:41:54 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/Configuration.cpp: - Fixed warning under gcc. Also made the for() loop more - ACE-like, not better, but at least consistent. - -Wed Jan 24 08:56:16 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.h: - The changed to document the ACE_SVC_* macros were accidentally - lost, fixed. - - * ace/Service_Config.cpp: - New debug message when processing an explicit directive, useful - to find out when and why TAO invokes - ACE_Service_Config::process_directive() directly. - As with all other debug messages in this file it only shows up - if ACE::debug() returns non-zero. - -Wed Jan 24 01:44:47 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Configuration.cpp (expand_path): Fixed a bug with assignment to - temp in the loop. Thanks to Jonathan Reis for reporting this! - -Tue Jan 23 17:02:42 2001 Darrell Brunsch <brunsch@uci.edu> - - * bin/msvc_auto_compile.pl: - - Removed PACE until we figure out how to handle it in the - DLL builds. - -Tue Jan 23 16:50:35 2001 Darrell Brunsch <brunsch@uci.edu> - - * tests/Dirent_Test.dsp: - - Updated to make consistent with the rest of the dsps in - this directory in regards to output directories and non-use - of ACE_ROOT. - - * tests/Dirent_Test.cpp: - - Updated wide character compatibility and removed the use - of ACE_ROOT (now looks at ../tests", which should give - the same results). - - * tests/run_test.lst: - - Added the Dirent_Test to the list - - * tests/version_tests/Dirent_Test.dsp: (added) - * tests/version_tests/Refcounted_Auto_Ptr_Test.dsp: (added) - * tests/version_tests/version_tests.dsw: - - Used the create_vt_dsps.pl script to update this directory. - -Tue Jan 23 18:03:26 2001 Jeff Parsons <parsons@cs.wustl.edu> - - * ace/CDR_Stream.cpp: - - Added patches to read_8 and write_8 sent in by - Keith Nicewarner <knicewar@erg.sri.com> to deal - with the ARM byte format for doubles - -Tue Jan 23 15:06:18 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/Configuration.cpp: - * ace/OS_Dirent.cpp: - - Updated wide character compatibility. - - * ace/ace_dll.dsp: - - Removed unneeded verbose library option. - -Tue Jan 23 10:54:14 2001 Jeff Parsons <parsons@cs.wustl.edu> - - * ace/config-doxygen.h: - * bin/autoconf_compile: - - Removed all occurrences of 'smart_proxies' and 'interface_repo'. - These are no longer necessary. - - * include/makeinclude/ace_flags.bor: - - Added necessary variables to go along with the Borland - makefiles for the new libraries tao/Domain and - tao/IFR_Client. - -Mon Jan 22 17:01:17 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS_String.cpp: - * ace/OS_String.inl: - * ace/ace_wchar.inl: - * ace/config-qnx-rtp.h: - - ACE_HAS_WCHAR fixes. Support is really bad on QNX RTP - so we just end up emulating most everything. As far as - I could tell, there isn't even a way to convert between - the two types of strings. - - * ace/OS_Dirent.cpp: - - Rewrote a paragraph to remove the use of "/*" in a larger - /*...*/ comment, since gcc complains about it. - - * tests/Message_Queue_Test.cpp: - - Backed out some changes to it on Jan 21, since it had - problems and I couldn't find any ChangeLog entry for the - changes (mistakenly checked in?). - -Mon Jan 22 13:51:11 2001 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/OS_Dirent.cpp (opendir_emulation): Added code to make the - emmulation work properly on Win32. Thanks to Edwin Wrench - <edwin.wrench@itt.com> for finding this. - - * tests/Dirent_Test.cpp: - tests/Makefile: - tests/run_tests.bat: - tests/run_tests.lst: Added test to verify correctness of above - fix. This test attempts to scan $ACE_ROOT/tests and will fail if - the number of entries is 0 or 1. - -Mon Jan 22 10:21:20 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Map_Manager.h: Clarify that find() returns 0, rather than - saying "a non-negative number". Thanks to Albert Wijna - <albert.wijnja@meco.nl> for clarifying this. - -Mon Jan 22 15:38:12 2001 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Containers_T.h: - ace/Containers_T.cpp: - Added operator* to the ACE_Ordered_Multiset_Iterator class - - * tests/Ordmultiset_test.cpp - Expanded test for the new operator* of the ACE_Ordered_Multiset_ - Iterator class - -Sat Jan 20 14:53:07 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Configuration.{h,cpp}: Fixed some problems with const char * - that was busting the compilation. Thanks to Carlos for pointing - this out. - - * ace/Configuration.{h,cpp}: Added several fixes for - mistakes with memory management and other minor problems. - Thanks to Jonathan Reis <reis@stentorsoft.com> for contributing - these fixes! - -Sat Jan 20 10:49:05 2001 Carlos O'Ryan <coryan@uci.edu> - - * docs/ACE-guidelines.html: - Update the guidelines to point people to the Doxygen - documentation, instead of the now abandoned OSE tools. - -Sat Jan 20 10:21:08 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.cpp: - Another pass fixing problems with ACE_BEGINTHREADEX() under - Win32, this time Win32 is working fine. - -Sat Jan 20 10:20:50 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * doc/tutorials: Integrated the new Service Configurator tutorial - into the main tutorials page. - -Fri Jan 19 16:57:43 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Svc_Handler.h: Updated the comments to explain why - operator new is overloaded. Thanks to Kirill Kuolechov - <Kirill.Koulechov@gmx.net> for reporting this. - -Fri Jan 19 15:39:43 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.cpp: - QNX/RTP has the same problem that LynxOS has, after fork() is - invoked trying to destroy the preallocated objects in - ACE_OS_Object_Manager fails. - Factor out this variation in a new macro: - ACE_HAS_BROKEN_PREALLOCATED_OBJECTS_AFTER_FORK - - * ace/config-lynxos.h: - * ace/config-qnx-rtp.h: - Define the new macro for the two relevant platforms. - - * ace/README: - Document the macro. - -Fri Jan 19 15:08:37 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.cpp: - I disabled too much in the commit at 10:50am... Now it seems to - be OK. - - * ace/Makefile: - Remove extra -v - -Fri Jan 19 13:16:49 2001 Priyanka Gontla <pgontla@ece.uci.edu> - - * include/makeinclude/platform_hpux_aCC.GNU: - Added new options to the CCFLAGS so that the known future errors - arising because of HPUX are suppressed. - -Fri Jan 19 13:01:13 2001 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/SSL/Makefile: - Put a check for defining ACE_SHLIBS to get it compiled on HPUX. - - * ace/RMCast/Makefile: - * ace/Makefile : - Updated dependencies - - * ace/config-all.h - Fix a build problem on HP-UX 11.00 with aCC A.03.25. The macro - used to determine the use of the std namespace changed from - RWSTD_NO_NAMESPACE to _NAMESPACE_STD. NOTE: I did preserve the - check for RWSTD_NO_NAMESPACE with aCC versions less than A.03.25. - - Thanks to Chad Elliott <elliott_c@ociweb.com> for making this fix. - -Fri Jan 19 11:22:42 2001 Priyanka Gontla <pgontla@ece.uci.edu> - - * docs/tutorials/022/client_handler.cpp : - * docs/tutorials/022/client_hander.h : - * docs/tutorials/022/Acceptor_Service.cpp : - * docs/tutorials/022/Acceptor_Service.h : - * docs/tutorials/022/server.cpp : - * docs/tutorials/022/Makefile : - - * docs/tutorials/022/client/Makefile: - * docs/tutorials/022/client/client.cpp: - - * docs/tutorials/022/page01.html: - * docs/tutorials/022/page02.html: - * docs/tutorials/022/page03.html: - * docs/tutorials/022/page04.html: - * docs/tutorials/022/page05.html: - - Developed a new tutorial for the ACE_Service_Configurator. - Based on the ACE tutorial 005. - -Fri Jan 19 10:50:11 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.cpp: - Use the Green Hills specific tricks *only* if we are compiling - under Green Hills, and only for the platform where they make - sense (in this case Win32). - -Thu Jan 18 23:21:53 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.h: - Document the infamous ACE_STATIC_SVC macros. - -Thu Jan 18 23:50:35 2001 Rick Stille <stille_r@ociweb.com> - - The Green Hills(GHS) Native x86 compiler does not provide support - for 64-bit integers, structural exceptions, or DLL's. Thus the - keywords __int64, __try, __dllimport, etc., are not defined. In - addition, exceptions do not work in a multi-threaded environment. - The following changes were made to address these shortcomings. - - * ace/Basic_Types.h: Because the GHS compiler has both - ACE_LACKS_LONGLONG_T and ACE_WIN32 defined, the - ACE_LACKS_LONGLONG_T macro must appear first, so the order of - these two macros has been reversed in the defines of both - ACE_UINT64_LITERAL(n) and ACE_UINT64_DBLCAST_ADAPTER(n). - - * ace/config-win32-common.h: The typedef of ACE_UINT64 as an - unsigned __int64 and the define of ACE_HAS_LLSEEK are blocked - out for ghs. - - * ace/config-win32-ghs.h: This is the new configuration file for - Green Hills/Win32. - - * ace/Configuration.cpp(resolve_key): The "return 0" at the end of - this function has been blocked out for ghs to eliminate an - "unreachable code" warning. - - * ace/Event_Handler_T.h(handle_signal): The type of the "signum" - parameter has been changed to "int" so that it matches the type - in the Event_Handler class. - - * ace/Event_Handler_T.i(handle_signal): The type of the "signum" - parameter has been changed to "int" so that it matches the type - in the Event_Handler class. - - * ace/High_Res_Timer.cpp(global_scale_factor): The Native x86 - compiler defines "QuadPart" in the LARGE_INTEGER union as a - "double," so it cannot be used in the calculation of - "global_scale_factor." Instead, for GHS, the low and high parts - of the LARGE_INTEGER structure are copied into an ACE_UINT64 - structure (which is actually an ACE_U_LongLong) and the - ACE_U_LongLong operators are used. The code segment if - (ACE_High_Res_Timer::global_scale_factor_ == 1u) // Failed to - retrieve CPU speed from system, so calculate it. - ACE_High_Res_Timer::calibrate (); is unreachable hen ACE_WIN32 - is defined, so it has been blocked out for Win32 platforms to - silence the Green Hills "unreachable code" warning. - - * ace/High_Res_Timer.i(hrtime_to_tv): There is no assignment - operator that converts an ACE_U_LongLong to a "long," so an - intermediate ACE_U_LongLong variable "usec64" has been - introduced. The type of "msec" and "usec" have been changed from - "long" to ACE_UINT32, as have been the casts. The conversion is - then carried out by using the lower part of "usec64." - - * ace/Local_Name_Space_T.cpp(remap): The "DWORD" cast has been - removed from both EXCEPTION_CONTINUE_SEARCH and - EXCEPTION_CONTINUE_EXECUTION beause the function return value is - of type "int" and the cast was causing an "integer conversion - resulted in a change of sign" warning. - - * ace/Memory_Pool.cpp(map): Removed references to the "mem_offset" - variable because it was not being used and the compiler was - issuing a warning to this effect. - - * ace/NT_Service.cpp(startup): This function was returning "-1" - when svc_sc_handle() failed, and returning a "0" when - QueryServiceConfig() failed. Since the return type of this - function is "DWORD," the compiler was issuing an "integer - conversion resulted in a change of sign" warning for the "-1" - value. Upon investigation of this issue, it was discovered that - "0" is a valid service type value, and thus the use of "0" to - indicate failure is a bug. The return value for svc_sc_handle() - failure has been changed to "MAXDWORD - 1," which is the DWORD - equivalent of "-2," and the return value for failure of - QueryServiceConfig() has been changed to "MAXDWORD," which is - the DWORD equivalent of -1. - - * ace/OS.cpp: The compiler cannot handle a 64-bit literal, so the - literal passed to FILETIME_to_timval_skew has been created as - ACE_U_LongLong(0xd53e8000, 0x19db1de). - - * ace/OS.cpp(set): The Native x86 compiler defines "QuadPart" in - the LARGE_INTEGER union as a "double," so it cannot be used in - the calculations of tv_sec and tv_usec. Instead, the low and - high parts of the LARGE_INTEGER structure are copied into an - ACE_U_LongLong structure, and the ACE_U_LongLong operators are - used. - - * ace/OS.cpp(operator FILETIME): The Native x86 compiler defines - "QuadPart" in the LARGE_INTEGER union as a "double," so it - cannot be used in the calculation of file_time. Instead, tv_sec - and tv_usec are converted iinto ACE_U_LongLong structures and - the ACE_U_LongLong operators are used. - - * ace/OS.cpp(uname): The Native x86 compiler names both the union - and the structure within the SYSTEM_INFO structure, so the - "wProcessorArchitecture" member within the "s" structure, which - is itself within the "u" union, must be accessed by using - ".u.s." - - * ace/OS.cpp(ACE_BEGINTHREADEX): Green Hills compiler gets - confused when __stdcall is imbedded in parameter list, so the - type ACE_WIN32THRFUNC_T has been defined, then used within the - parameter list of ::_beginthreadex(). - - * ace/OS.h(ACE_Time_Value): The Native x86 compiler defines - "DWORDLONG" as type "double," so FILETIME_to_timval_skew is - instead defined as type ACE_U_LongLong. - - * ace/OS.h: The pragmas "#pragma warning( disable : ... " are - unknown to the Native x86 compiler, so it produces warnings to - this effect. These pragma have been blocked out for ghs. - - * ace/OS.h: Because the Native x86 compiler does not support the - __int64 keyword, the typedef of ACE_hrtime_t has been changed - from "unsigned __int64" to "ACE_UINT64" for ghs. Because no - usage of "ACE_QWORD" could be foound, its definition has been - blocked out for ghs. - - * ace/OS.i(gettimeofday):The code at the end of this function is - not reachable if either ACE_HAS_WINCE or ACE_WIN32 is - defined. In order to quiet the Green Hills compiler warnings - about unreachable code, this code has been blocked out for - ACE_HAS_WINCE and ACE_WIN32. With this code blocked out, - however, the variables "tv" and "result" are not used, and the - compiler warns about about variables defined, but never used. In - order to quiet these warnings, the definitions of these two - variables have been blocked out for ACE_HAS_WINCE and ACE_WIN32. - - * ace/OS.i(gethrtime): The ACE_gethrtime() function is not meant - to be used by a Win32 platform, but because both "ghs" and - "ACE_HAS_PENTIUM" are defined for the Native x86 platform, a - call to ACE_gethrtime() was included in the code. This problem - has been solved by adding "&& !defined(ACE_WIN32)" to the two - "... defined (ghs) && defined (ACE_HAS_PENTIUM)" statements - which block out references to ACE_gethrtime(). - - The Native x86 compiler defines "QuadPart" in the LARGE_INTEGER - union as a "double," so it cannot be used as the return - value. Instead, because the return type is "ACE_hrtime_t," which - is type "ACE_UINT64," the low and high parts of the - LARGE_INTEGER structure are copied into an ACE_UINT64, which is - then returned. - - * ace/Process_Manager.cpp(wait): The "result >= WAIT_OBJECT_0" - test within the ACE_ASSERT statement causes the Native x86 - compiler to produce a ôpointless comparison of unsigned - integer with zeroö warning because "result" is of type - "DWORD," which is defined by the Native x86 compiler as type - "unsigned int." This test has been blocked out for ghs. - - * ace/Profile_Timer.cpp(elapsed_time): Because the "__int64" - keyword is not defined by the Native x86 compiler, it cannot be - used to cast the value of "delta_t" in the calculation of - "real_time." But "delta_t" is of type "ACE_hrtime_t," and - ACE_hrtime_t is typedefed as ACE_U_LongLong, so the division - operator defined for in the ACE_U_LongLong class can be used - directly with a divisor of type "double." Thus the "__int64" - cast has been removed for ghs. - - * ace/Sock_Connect.cpp(get_reg_value): This function is used - within the function get_ip_interfaces() function only when - either ACE_HAS_WINSOCK2 is not defined or when ACE_HAS_WINSOCK2 - is defined as "0." Because ACE_HAS_WINSOCK2 is defined on the - Native x86 platform, the compiler was issuing a ôdefined, but - never usedö warning. This warning has been eliminated by - blocking out the definition of get_reg_value() when - ACE_HAS_WINSOCK2 is defined as a value other than "0." - - * ace/Stats.cpp(accumulate): The comparison "rhs.samples_count () - == 0" was producing a warning because rhs.samples_count() - returns an ACE_UINT32, but is being compared to the signed - literal "0." The "0" has been changed to "0u." - - * ace/WFMO_Reactor.cpp(ok_to_wait): The comparison of the "switch" - statement parameter "result" to "DWORD" values in the case - statements elicited an "integer conversion resulted in a change - of sign" warning from the compiler, so the type of "result" has - been changed to "DWORD." - - * ace/WFMO_Reactor.cpp(dispatch): The comparison of the "switch" - statement parameter "wait_status" to "DWORD" values in the case - statements elicited an "integer conversion resulted in a change - of sign" warning from the compiler, so the type of "wait_status" - has been cast to "DWORD." - - * ace/WFMO_Reactor.cpp(dispatch_handles): The "wait_status >= - WAIT_OBJECT_0" comparison was producing a ôpointless - comparison of unsigned integer with zeroö warning because - "wait_status" is an unsigned integer. This was a known problem - with Borland, so the blocking out of this comparison has been - extende to ghs. - - * tests/Basic_Types_Test.cpp(main): Because both the - ACE_LACKS_LONGLONG_T and ACE_WIN32 macros are defined, the - ACE_LACKS_LONGLONG_T macro must appear first, so the order of - these two macros has been reversed. - - * tests/Cached_Accept_Conn_Test.cpp: The template class - ACE_Node<ACE_HANDLE> must be explicitly instantiated for this - platform. - - * tests/Cached_Conn_Test.cpp: The template class - ACE_Node<ACE_HANDLE> must be explicitly instantiated for this - platform. - - * tests/Handle_Set_Test.cpp: The template class - ACE_Node<ACE_HANDLE> must be explicitly instantiated for this - platform. This was already being done for the _CRAYMPP, so it - has been extended to ghs. - - * tests/Malloc_Test.cpp: Because the Native x86 compiler does not - support structural exceptions, it cannot support the "remap" - function, so the blocking of the ACE_TEST_REMAP_ON_FAULT macro - definition, which was already in effect for linux, has been - extended to ghs. - -Thu Jan 18 12:49:53 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/SUN_Proactor.cpp: Added support for cancel_aiocb(). Thanks - to Alexander Libman <Alibman@baltimore.com> for contributing - this. - - * tests/Makefile.bor (TESTS): Added Config_Test so that this test - runs on Borland. - - * tests: Moved the config_test stuff (which tests ACE_Configuration* - classes) from the examples/Configuration/ directory to the - tests/ directory and made it a one-button test. - - * ace/Future.h: Changed the order of the thread mutex and the condition - variable to avoid confusion in the future regarding "order of - initialization." Thanks to Edan Ayal <edan@bandwiz.com> for - reporting this. - - * examples/Configuration: Updated this example to include the new - test for the ACE_Configuration_Heap methods. Thanks to Michael - Searles <msearles@base16.com> for contributing this. - -Thu Jan 18 18:55:37 2001 Carlos O'Ryan <coryan@uci.edu> - - * bin/nightlybuilds/builds.lst: - Add the RedHat_Implicit_Templates build. - -Thu Jan 18 18:48:33 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/Log_Msg.cpp: - Declare sys_nerr as const int under QNX/RTP (and NTO). - - * ace/Pipe.cpp: - Use pipe() under QNX because it lacks both socketpair() and it - cannot open a socket to itself. - - * ace/config-qnx-rtp.h: - Remove ACE_NTRACE hacks. - - * include/makeinclude/platform_qnx_rtp_gcc.GNU: - Do not enable the repo by default it breaks QNX. - Instead of removing the definition of DCFLAGS by default we - set debug=0 all the time, that seems like a cleaner way to - disable debugging. - -Thu Jan 18 09:34:35 2001 Carlos O'Ryan <coryan@uci.edu> - - * include/makeinclude/platform_linux.GNU: - The -fno-implicit-templates flag was hardcoded, but this gave - the users no recourse to remove should they want to use - -fimplicit-templates. - The macro IMPLICIT_TEMPLATES_FLAG is used to set the flag now, - unless the user provides a value for the macro the makefiles - continue to work as they did before this change, i.e., adding - -fno-implicit-templates to CCFLAGS. - Notice that this functionality was available in - platform_linux_lxpthreads.GNU, but dissapeared when everything - was rolled into platform_linux.GNU. - -Thu Jan 18 06:26:42 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Configuration.{h,cpp}: Added functionality to read in all - values in a config file without requiring a type prefix on each - value, skips whitespace and allows comments on the end of lines. - Thanks to Michael Searles <msearles@base16.com> for contributing - this. - -Wed Jan 17 11:48:42 2001 Darrell Brunsch <brunsch@uci.edu> - - * bin/msvc_auto_compile.pl: - - Added PACE to the list of ACE directories so PACE is - compiled in full auto builds. - -Wed Jan 17 10:00:42 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/Handle_Ops.cpp: (handle_timed_open) - - Removed the use of ACE_NONBLOCK on Win32. This use - didn't function as expected, since ACE_NONBLOCK would - be interpreted as O_WRONLY. Also, the CreateFile API - call doesn't support any nonblocking option, so - redefining ACE_NONBLOCK would not seem to fix anything. - - Thanks to Matthias Wittig <m.wittig@com-on.de> for reporting - a problem report form about this. - - * bin/PerlACE/Process_Unix.pm: - - Fixed up the use of the RUNNING state, so wait and kill - work properly. Also made Kill automatically use waitpid - (since I can't think of any time we'd kill without wanting to - clean up the resources). - -Wed Jan 17 07:44:29 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * apps/Gateway/Gateway/Concrete_Connection_Handlers.cpp (send): Set errno to - EWOULDBLOCK if n < len so that the caller will initiate a - handle_output()-driven reply. Thanks to Klaus H. Wolf - <hw@cyland.com> for reporting this. - - * ace/Log_Msg.h (ACE_Log_Msg): Clarified that only the current message - is stored in TSS. Thanks to Pierre Oberson - <oberson@nagra-kudelski.ch> for pointing out the need for this - clarification. - - * examples/Log_Msg/test_log_msg.cpp (main): Fixed the example by - adding ACE_Log_Msg::PROCESS so that the messages do not print - out (which is the correct behavior for this test). Thanks to - Mike Curtis <mccurry@my-deja.com> for reporting this. - * apps/Gateway/Gateway/Concrete_Connection_Handlers.cpp (send): Set errno to - EWOULDBLOCK if n < len so that the caller will initiate a - handle_output()-driven reply. Thanks to Klaus H. Wolf - <hw@cyland.com> for reporting this. - - * ace/Log_Msg.h (ACE_Log_Msg): Clarified that only the current message - is stored in TSS. Thanks to Pierre Oberson - <oberson@nagra-kudelski.ch> for pointing out the need for this - clarification. - - * examples/Log_Msg/test_log_msg.cpp (main): Fixed the example by - adding ACE_Log_Msg::PROCESS so that the messages do not print - out (which is the correct behavior for this test). Thanks to - Mike Curtis <mccurry@my-deja.com> for reporting this. - -Tue Jan 16 15:28:22 2001 Carlos O'Ryan <coryan@uci.edu> - - * netsvcs/lib/Server_Logging_Handler_T.cpp: - The formal parameter names must match in the .h and .cpp file - to make Sun/CC 6.0 happy. - -Tue Jan 16 11:25:32 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Log_Msg.cpp (log): Convert strerror() text string to a - TCHAR so that this call will work correctly with Unicode. - Thanks to Matthias Wittig <m.wittig@com-on.de> for reporting - this. - -Tue Jan 16 09:58:20 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/ace_wchar.inl: - - Updated ACE_Ascii_To_Wide and ACE_Wide_To_Ascii to - preserve null pointers. - - Thanks to Matthias Wittig <m.wittig@com-on.de> for - reporting this and suggesting a fix. - -Tue Jan 16 10:47:33 2001 Frank Hunleth <fhunleth@cs.wustl.edu> - - * ace/config-win32-common.h: - * ace/config-win32-msvc.h: - Moved #define ACE_HAS_LLSEEK from config-win32-common.h to - config-win32-msvc.h, since _lseeki64 isn't defined for - Borland. Also implicitly removed ACE_HAS_LLSEEK for the - VisualAge compiler - it might have worked (didn't know, - couldn't test) - -Mon Jan 15 15:22:42 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Asynch_IO.h: Changed the ACE_Asynch_Read_Stream::read() - method's "bytes_to_read" parameter to "num_bytes_to_read" to - workaround a weird problem with some compilers. Thanks to Chris - Kohlhoff <chris@kohlhoff.com> for this fix. - - * examples/Service_Configurator/IPC-tests/README: Updated this - file so that it's now up-to-date and also explains how to - reconfigure a server if SIGHUP isn't supported. Thanks to Mike - Curtis <mccurry@my-deja.com> for reporting this. - -Mon Jan 15 12:13:48 2001 Darrell Brunsch <brunsch@uci.edu> - - * bin/nightlybuilds/builds.lst: [Added] - * bin/nightlybuilds/scoreboard_update.pl: [Added] - - Moved the scoreboard stuff from the internal cvs repos - to this one, so it is easier for people to edit the - list. - -Mon Jan 15 12:29:53 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/config-macosx.h: Fixed typo so that ACE_HAS_CLOCK_GETTIME - is spelled correctly. Thanks to Mike Curtis - <mccurry@my-deja.com> for reporting this. - - * ace/Strategies_T.cpp (open): Replace the ACE_OS::strncpy() functions - with ACE_OS::strcpy() functions. Thanks to Ron Hashimshony - <Ron@bandwiz.com> for reporting this. - -Mon Jan 15 09:44:16 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/RMCast/Makefile: - Fixed Makefile to link the ACE library, otherwise we get - undefined symbols under certain platforms (AIX). Thanks to - Scott Bolin <scott.bolin@pgs.com> for pointing this out. - -Mon Jan 15 09:11:12 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/config-win32-common.h: Added support for ACE_HAS_LLSEEK, though - this should probably be changed to be compiled only on some - Win32 platforms. - - * ace/OS.{h,i}: Added support for ACE_OS::llseek() on Win32. Thanks - to Nick Lin <nicklin@vantech.com.tw> for contributing this. - - * ace/Log_Priority.h: Clarified the fact that we only use 12 bits - for ACE_Log_Priority. Thus, the remaining bits are available - for application use. Thanks to David Singer - <David_P_Singer@res.raytheon.com> for motivating this comment. - -Sun Jan 14 16:53:11 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Configuration.{h,cpp}: Fixed a memory leak that occurred - with transient heaps. Also added a list of current limitations - to import/export routine. Thanks to Chris Hafey - <chafey@stentor.com> for contributing this stuff. - -Fri Jan 12 20:46:05 2001 Angelo Corsaro <corsaro@cs.wustl.edu> - - * include/makeinclude/platform_qnx_rtp_gcc.GNU: - Platform macros needed to run ACE on QNX RTP. - - * ace/config-qnx-rtp.h: - Configuration file needed to run ACE on QNX RTP. - - * ACE-INSTALL.html: - Added documentation on how to install ACE on QNX RTP, - and the issue with the current version of QNX RTP. - - -Fri Jan 12 14:59:15 2001 Carlos O'Ryan <coryan@uci.edu> - - * tests/RMCast/RMCast_Fragment_Test.cpp: - Fixed conversion from 64-bit ulong to 32-bit ulong under some - VxWorks compilers where 64 bits are emulated. Thanks to Edwin - McKay <mckayea@betalasermike.com> for pointing this out. - -Fri Jan 12 07:18:16 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/OS.{h,i}: Moved some "using std::*" definitions from OS.i to - OS.h. Thanks to Mike Curtis <mccurry@my-deja.com> for reporting - this and to Christopher Kohlhoff <chris@kohlhoff.com> for - suggesting the fix. - -Fri Jan 12 15:34:28 2001 Steve Huston <shuston@riverace.com> - - * ace/Managed_Object.h: Added <TYPE> to the unimplemented operator= - declaration. Fixes compile w/ IBM C/C++ 3.6 on AIX. - -Fri Jan 12 07:59:19 2001 Joe Hoffert <joeh@cs.wustl.edu> - - * Makefile: - Add the PACE directory back into the build. PACE should - not generate any build failures for the platforms - currently supported. - -Wed Jan 10 23:30:21 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/FILE_Addr.cpp (set): Replace (sizeof this->filename_) - with (sizeof this->filename_ / sizeof (ACE_TCHAR)). Thanks to - Matthias Wittig <m.wittig@com-on.de> for reporting this. - -Tue Jan 9 07:27:19 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Containers_T.h: Clarified that the ACE_Ordered_Multiset class - allows duplicates. Thanks to Albert Wijna - <albert.wijnja@meco.nl> for clarifying this. - -Sat Jan 6 10:20:55 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/File_Lock.{h,inl}: Made the get_handle() method const. - Thanks to Johnny Willemsen <johnny.willemsen@meco.nl> for - reporting this. - - * ace/TP_Reactor.h: - * ace/Get_Opt.h: - * ace/Caching_Utility_T.h: Added "ACE_UNIMPLEMENTED_FUNC" macros - for all copy constructors and assignment operators. Thanks to - Johnny Willemsen <johnny.willemsen@meco.nl> for reporting this. - -Fri Jan 5 01:09:28 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * performance-tests/Misc/context_switch_time.cpp (get_options), - * performance-tests/Misc/preempt.cpp (open), - * examples/Reactor/Misc/pingpong.cpp (run_svc), - * examples/Reactor/Misc/notification.cpp (main), - * examples/IPC_SAP/FIFO_SAP/FIFO-test.cpp (main), - * examples/Shared_Malloc/Options.cpp (print_usage_and_die), - * examples/Service_Configurator/IPC-tests/client/: Updated all - these programs so that don't abort via %a, but instead call - ACE_OS::exit(). Thanks to Mike Curtis <mccurry@my-deja.com> for - suggesting this. - - * apps/Gateway/Peer/Options.cpp (print_usage_and_die): Make sure to - actually *die* when we're done with this function! Thanks to - Mike Curtis <mccurry@my-deja.com> for suggesting this. - -Thu Jan 4 00:10:26 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Select_Reactor_T.i: The cancel_timer() method should return the - number of events canceled, rather than 0 or 1. Thanks to Joe - Guan <jguan@mistwireless.com> for reporting this and providing - the fix. - - * ace/Synch.h: Added an acquire (ACE_Time_Value &) method to ACE_Null_Mutex - so that this will work properly when - ACE_HAS_OPTIMIZED_MESSAGE_QUEUE is enabled. Thanks to Yuriy - Zaporozhets <zaz@internet-zahav.net> for reporting this. - -Wed Jan 3 23:00:52 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Lib_Find.cpp (ldfind), - * ace/Makefile.bor (CFLAGS), - * include/makeinclude/outputdir.bor (OBJDIR), - * ace/config-win32-msvc.h: Added support for the new ACE_LD_DECORATOR_STR - macro which is used to work around the fact that Borland - attaches a so-called "decorator suffix" to all library names. - Thanks to Chris Kohlhoff <chris@kohlhoff.com> for reporting - this. - - * ace/Lib_Find.h: Fixed a typo where ACE_LIB_FIND_H was misspelled. - Thanks to Chris Kohlhoff <chris@kohlhoff.com> for reporting - this. - - * ace/SOCK_Stream.i, - ace/MEM_Stream.i: Updated the calls to ACE_OS::shutdown() to use - the new macros described below! - - * ace/OS.h: Added better support for the ACE_OS::shutdown() second - parmeter in the form of new macros ACE_SHUTDOWN_READ, - ACE_SHUTDOWN_WRITE, and ACE_SHUTDOWN_BOTH. Thanks to Mike - Curtis <mccurry@my-deja.com> for suggesting this. - - * ace/SOCK_Stream.h: Clarified the role of the timeout parameters in - the various methods. Thanks to Allen Broadman - <broadmana@plural.com> for motivating this. - -Thu Jan 11 17:24:30 2001 Carlos O'Ryan <coryan@uci.edu> - - * Makefile: - Add new target 'Core' to quickly compile the basic ACE+TAO - components. - -Thu Jan 11 09:56:43 2001 Frank <fhunleth@cs.wustl.edu> - - * ACE version 5.1.12 released. - -Thu Jan 11 05:38:24 2001 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/TP_Reactor.cpp: Reverted the change that was made for SPRO - 5.1 compilers. The change that was reverted is this - - Thu Jan 4 17:14:33 2001 Balachandran Natarajan - <bala@cs.wustl.edu> - - * ace/Handle_Set.cpp: Changed the macro - - ACE_DIV_BY_WORDSIZE(x) ((x) >> (ACE_Handle_Set::WORDSIZE)) to - - ACE_DIV_BY_WORDSIZE(x) ((x) >> ((int) - ACE_Handle_Set::WORDSIZE)). This seems to be a better - fix. Thanks to David Allen <dallen@storm.com> for suggesting - this. - -Tue Jan 9 09:53:53 2001 Carlos O'Ryan <coryan@uci.edu> - - * tests/MEM_Stream_Test.cpp: - Run the event loop for at most 1 minute. After that we know - that something is amiss with the client threads, and we should - exit. This fixes the deadlock problems we were seeing in the - nightly builds. - The real cause of the problem is something broken with the way - MEM_Addr objects compare local addresses, sometimes they fail to - compare equal. In the test this results in failed connections. - -Mon Jan 8 18:53:39 2001 Sharath R. Cholleti <sharath@cs.wustl.edu> - - * examples/Map_Manager/test_hash_map_manager.cpp: - Removed template instantiations for - ACE_Guard<ACE_SYNCH_RW_MUTEX>, - ACE_Read_Guard<ACE_SYNCH_RW_MUTEX>, and - ACE_Write_Guard<ACE_SYNCH_RW_MUTEX> to fix linker - errors under Lynx_I386. - -Mon Jan 8 10:44:57 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS.h: - - Added some more protection (!VXWORKS) around the cuserid - inline statement since VXWORKS doesn't have cuserid. - -Mon Jan 08 10:04:55 2001 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Thread_Manager.cpp (terminate): Reordered some statements so - we get the thread descriptor's log_msg pointer before we delete - the thread descriptor. Thanks to Tomer Amiaz - <Tomer@bandwiz.com> for reporting this. - -Sun Jan 7 16:51:42 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/ace_dll.dsp: - - Removed the PACE libraries from linking. We are going to come - up with a different solution for pace. - - * docs/run_test.txt: - - Changed the instructions a little to use PerlACE::LocalFile - for the server's IOR file. - -Sun Jan 7 08:59:09 2001 Balachandran Natarajan <bala@cs.wustl.edu> - - * tests/DLL_Test.cpp: Added a message in a ACE_ERROR statement. - * tests/run_test.lst: Disabled MEM_Stream_Test in Linux. - -Thu Jan 4 17:14:33 2001 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/TP_Reactor.cpp (dispatch_i): Eliminate a portion of the - code from SunCC5 builds. This piece of code was trying to set - things right when it sees all the masks clear and the <size_> - greater than zero. This seems to be creating problems with SunCC - 5 builds alone. Commenting out this piece for SunCC5 helps. This - should also be a work around for the bug #750 in bugzilla. - -Thu Jan 04 17:07:48 2001 Luther Baker <luther@cs.wustl.edu> - - * ace/ace_dll.dsp - - Consistent typography (win32 specific backslash). - -Thu Jan 4 15:34:14 2001 Luther J Baker <luther@cs.wustl.edu> - - * ace/ace_dll.dsp - - Added PACE library to the ACE DLL project file. - (also changed the library search path) - -Wed Jan 3 17:24:11 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS.h: - - Mistakenly changed the ace_timezone () method to use - timezone () instead of timezone. Changed back. - -Wed Jan 3 16:00:22 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/ACE.cpp - - Changed the types of day_of_week_name and month_name in - ACE::timestamp to ACE_TCHAR (since that was the data - being assigned to it and the type that is later expected). - This fixed errors in Win32's Unicode build. - -Wed Jan 3 15:44:31 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS.i: - - Removed ACE_TIMEZONE since we found another workaround for - platforms that define timezone () as a macro. - - * ace/OS.h: - * ace/OS.i: - - Moved much of the code from ACE_OS::timezone and - ACE_OS::difftime into ace_timezone and ace_difftime. Since - some platforms don't have timezone or difftime, we needed - to do something here so those methods aren't used. Instead - of figuring out what larger #if !... should be, it is easier - (and less error prone) to just move the code here. - -Wed Jan 3 13:10:39 2001 Darrell Brunsch <brunsch@uci.edu> - - * apps/JAWS/server/HTTP_Helpers.cpp: - - Added ACE_MT around an ACE_Guard to remove an unused var - warning in the single threaded builds. - -Wed Jan 3 12:56:13 2001 Darrell Brunsch <brunsch@uci.edu> - - * ace/Timer_Wheel_T.cpp: - * ace/Timer_Wheel_T.h: - - Doug noticed some redundant code in expire while making his - change on the 21st. I checked and he was right, it looks like - I was moving around the loop termination conditionals and left - some around. Removed them and also added some better comments - and made it more doxygen complient. - -Wed Jan 3 12:06:20 2001 Darrell Brunsch <brunsch@uci.edu> - - * tests/Refcounted_Auto_Ptr_Test.dsw: - - Projects are listed in tests.dsw, so this is unneeded. - - * tests/tests.dsw: - - Added Refcounted_Auto_Ptr_Test.dsp. - -Tue Jan 3 10:30:42 2001 Darrell Brunsch <brunsch@uci.edu> - - * docs/msvc_notes.txt: - - Added this file explaining some issues or explanations on how - ACE/TAO are set up for use with MSVC. Mainly I'm thinking of - this as a place to put random stuff about ACE and TAO on MSVC. - - Plus I think we need more nifty doxygen documents. ;-) A tip of - the hat goes to Letha Etzkorn <letzkorn@cs.uah.edu> and - Shivakumar Patil <shivakumar.patil@stdc.com> for bringing up - questions that led to the dump of info to this file. - - * ACE-INSTALL.html: - - Added a reference to msvc_notes.txt. - -Wed Jan 3 09:15:24 2001 Jeff Parsons <parsons@cs.wustl.edu> - - * ace/ACE.cpp: - - Replaced these lines - - SYSTEMTIME local; - ::GetLocalTime (&local); - - in the timestamp() method. - They were removed the last time this file was - modified, breaking the Win32 builds. - -Wed Jan 03 05:49:11 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.h: - The ace_cuserid() helper cannot be defined in Win32 - -Wed Jan 03 05:42:20 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.h: - Remove broken #include of unistd.h, it was a debug statement - that sneaked through the commit. - -Tue Jan 2 18:33:03 2001 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.h: - * ace/OS.i: - Workaround platforms that define timezone, difftime or cuserid - as macros. Instead of just dropping support for the ACE_OS::XXX - method we create an inline ace_XXX() function, that uses the - original macro expansion, then we #undef the offending macro, - and use the inline function in the code. - - * ace/config-tru64.h: - Enable cuserid() again, with the new workaround it should work - without problems. - This fix should remove some of the redness in the TRU64_CXX - builds. - -Tue Jan 2 14:22:21 2001 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_run_tests.lst: - Add the DII version of Param_Test to the nightly builds. - -Tue Jan 2 10:25:30 2001 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/Cached_Connect_Strategy_T.cpp: Added the patch provided by - Edan Ayal <edan@bandwiz.com> for - ACE_Bounded_Cached_Connect_Strategy::find_or_create_svc_handler_i (). - -Tue Jan 2 09:03:18 2001 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Asynch_Acceptor.{h,cpp}, - ace/POSIX_Async_IO.{h,cpp}, - ace/POSIX_Proactor.{h,cpp}, - ace/SUN_Proactor.{h,cpp}: Added enhanced versions of ACE_POSIX_Proactor - and ACE_SUN_Proactor to fix various problems uncovered and fixed by - Alexander Libman <Alibman@baltimore.com>. - - * ace/ACE.{h,cpp}: Fixed the timestamp() method so that it works - identically on Win32 and on UNIX. Also provided a new option - that'll make it possible to return a pointer to the beginning of - the time portion of "date and time." Thanks to Michael Searles - <msearles@base16.com> for contributing these fixes. - -Mon Jan 01 16:35:00 2000 Michael Kircher <Micahel.Kircher@mchp.siemens.de> - - * tests/Reader_Writer_Test.cpp: Added a #ifdef to check for native - RW lock support. If the platform supports native RW locks we should - not report a failure on upgrading, because native RW locks do not - support upgrading, only our RW lock emulation does this. - Thanks to Frank and Doug for pointing out this failure. - -Sun Dec 31 14:12:08 2000 Ossama Othman <ossama@uci.edu> - - * bin/auto_run_tests.lst: - - Added TAO's "ForwardRequest" PortableInterceptor test to the - list of tests to run. - -Sun Dec 31 13:59:02 2000 David L. Levine <levine@cs.wustl.edu> - - * tests/Refcounted_Auto_Ptr_Test.cpp: on DIGITAL_UNIX - only, added a call to ACE_OS::exit () at the end of - the test so that it will terminate. [Bug 758] - -Sun Dec 31 10:28:43 2000 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_osf1_4.0.GNU (LIBS): - wrapped /usr/lib/cmplrs/cxx/libcxxstd.a with - $(wildcard), so that is ignored on systems on which - it doesn't exist, such as Tru64 5.1. - -Sun Dec 31 09:23:09 2000 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_tru64_cxx.GNU: added - support for Tru64 5.1 (DIGITAL_UNIX 0x510). - - * ace/config-tru64.h: added ACE_LACKS_CUSERID with - DIGITAL_UNIX >= 0x510, because cuserid () is a macro - in Tru64 5.1. - - Thanks to Carlos for tracking down this source of compilation - failure. - -Sat Dec 30 16:21:49 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/Hash_Map_Manager_T.h: - * ace/Map_T.i: Fixed errors in g++ that were introduced while - applying the patches. - -Sat Dec 30 15:10:00 2000 Chris Gill <cdgill@cs.wustl.edu> - - * ace/RB_Tree.{cpp, h, i}: - tests/RB_Tree_Test.cpp: Fixed a bug in RB_delete_fixup that - occurs when the pointer to be fixed up is zero - solution is to - pass in a pointer to its parent, and maintain the parent pointer - and fixup pointer in tandem during the fixup function. Also - added a few functions to test the red-black invariant(s) at each - node of the tree, and added an invariant test call after each - bind and unbind in a new stress test in RB_Tree_Test.cpp. - - Thanks to Klaus H. Wolf <hw@cyland.com> for pointing out the - problem and for contributing the stress test (which in a - slightly modified form is now part of the RB_Tree_Test.cpp code) - that reproduced the problem. - -Sat Dec 30 11:12:08 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/Hash_Map_Manager_T.h: - * ace/Map_T.i: - * ace/Message_Block_T.cpp: Added patches provided by J. Russell - Noseworthy <rnosewor@objectsciences.com> for SPRO 5.1. The patch - to config-sunos-5.5.h has not been applied as it breaks builds - on SunCC 5.0. - -Fri Dec 29 17:08:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/auto_compile: - - Added a -sendreport option to enable the sending of email at all. - Now by default it will not. - -Fri Dec 29 16:53:22 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * performance-tests/TTCP/ACE-C++/wrapper-new-ttcp.cpp (main), - * performance-tests/TTCP/C/new-ttcp.cpp (main), - * tests/Conn_Test.cpp: Incorrect use of %r, which caused - problems... Thanks to Mike Curtis <mccurry@my-deja.com> for - pointing this out. - -Fri Dec 29 14:21:01 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/MEM_SAP.cpp: Fixed incorrect ACE_RCSID identifier. Thanks - to Jerry Odenwelder <jerryo@atl.fundtech.com> for noticing this. - -Thu Dec 28 16:02:51 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/sandbox.cpp: - This program executes a child process (and its children), in a - new session and process group. If the process tree does not - finish before a given timeout (the first argument) the whole - tree is terminated. - The program is useful for test scripts and programs, - occasionally they hang the whole auto build. - - * bin/auto_compile: - Add new options '-sandbox' and '-sandbox_timeout' to support the - sandbox program. - -Thu Dec 28 16:10:59 2000 Frank Hunleth <fhunleth@cs.wustl.edu> - - * performance-tests/UDP/udp_test.cpp: - * performance-tests/Server_Concurrency/Latency_Stats.h: - * performance-tests/Server_Concurrency/Queue_Based_Workers/workers.cpp - * performance-tests/Server_Concurrency/Leader_Follower/leader_follower.cpp - Fixed compile problems under VxWorks when handling - 64 bit integers. Basically, truncated the 64 bit integers - to 32 bit integers for operations not supported by - ACE_U_LongLong. - -Wed Dec 27 15:32:43 2000 Frank Hunleth <fhunleth@cs.wustl.edu> - - * examples/Reactor/WFMO_Reactor/test_network_events.cpp: - Added missing ; - -Wed Dec 27 11:11:02 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/Map_Manager/test_hash_map_manager.cpp: - * examples/Misc/test_trace.cpp: - * performance-tests/Misc/test_singleton.h: - * performance-tests/Misc/test_singleton.cpp: - Fixed problems in single threaded builds, basically I use the - ACE_SYNCH_* macros instead of using the class directly. - -Tue Dec 26 18:53:06 2000 Carlos O'Ryan <coryan@uci.edu> - - * performance-tests/Server_Concurrency/Makefile: - The tests make no sense if threads are disabled, so they are not - compiled under that configuration. - -Tue Dec 26 18:31:08 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/Naming/test_open.cpp: - Fixed compilation problems under Sun/CC 4.2, the Log_Msg.h file - was not #included - -Tue Dec 26 14:45:50 2000 Ossama Othman <ossama@uci.edu> - - * bin/auto_run_tests.lst: - - Added TAO's "Dynamic" and "IORInterceptor" PortableInterceptor - tests to the list of tests to run. - -Tue Dec 26 11:49:49 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * examples/IPC_SAP/SPIPE_SAP/server.cpp: Fixed a statement not - reachable warning on KCC. - -Sat Dec 23 18:11:02 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/SString.cpp: Enhanced the implementation so that it works - correctly when binary data (i.e., data that may contain NUL - characters) is stored in a string. Thanks to Michael Lindner - <mikel@att.net> for reporting this. - -Fri Dec 22 15:18:49 2000 Steve Huston <shuston@riverace.com> - - * ace/ace-dll.icc: Added Basic_Stats.{h cpp} to build clean. - -Fri Dec 22 14:00:41 2000 Steve Huston <shuston@riverace.com> - - * ace/config-all.h: On HP-UX 11, when figuring ACE_bad_alloc exception - class, account for older compilers that don't define __HP_aCC. - -Fri Dec 22 12:56:06 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/ACE.cpp (process_active): Call CloseHandle before return to - avoid handle leaks on Win32. Thanks to Arnaud Compan - <compan@ipanematech.com> for the fix. - -Fri Dec 22 13:04:58 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * bin/run_all_list.pm: Added the bidirectional test to the list. - -Fri Dec 22 12:43:54 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/OS.h: Fixed a typo. - -Fri Dec 22 09:12:37 2000 Steve Huston <shuston@riverace.com> - - * ace/OS.h: ACE_DEFAULT_FILE_PERMS - added FILE_SHARE_DELETE to the - bits if building on Win NT4 or greater. Thanks to Nanbor Wang - for calling attention to this. - -Fri Dec 22 08:06:10 2000 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> - - * tests/Refcounted_Auto_Ptr_Test.cpp: Fixed the problem with an - undefined ACE_Auto_Basic_Ptr<Scheduler> by adding an explicit - template instantiation. Thanks to Frank A. Hunleth - <fhunleth@cs.wustl.edu> for reporting this. - -Thu Dec 21 18:12:38 2000 Steve Huston <shuston@riverace.com> - - * ace/ace-dll.icc, ace-lib.icc: Added OS_Errno.(cpp h) to the files - list. Thanks to Jerry Odenwelder for pointing out the need for these. - -Thu Dec 21 16:45:28 2000 Steve Huston <shuston@riverace.com> - - * tests/Capabilities_Test.cpp: Added missing \n to a ACE_ERROR - statement, and added ACE_DEFAULT_FILE_PERMS to the ACE_OS::open - call to create the test file. It used to be 0600, but that was - removed to satisfy Win32 which now pays attention to that arg. - -Thu Dec 21 05:50:51 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Timer_Wheel_T.cpp (expire): Fixed a mistake where - next_earliest_time should have been next_earliest. Thanks to - Keith Brown <kalbrown@ix.netcom.com> for reporting this. - - * examples/IPC_SAP/SOCK_SAP/Makefile.bor: Reenable CPP-inserver-fancy.cpp - now that it compiles! Thanks to Johnny Willemsen - <johnny.willemsen@meco.nl> for confirming this. - - * examples/IPC_SAP/SOCK_SAP/CPP-inserver-fancy.cpp (handle_events): Added - casts for (SOCKET) when using FD_SET to work around problems - with Borland. Thanks to Johnny Willemsen - <johnny.willemsen@meco.nl> for reporting this. - - * ace/Message_Block.h (ACE_Dynamic_Message_Strategy): Fixed a bunch - of non-const accessor methods to be const. Thanks to Johnny - Willemsen <johnny.willemsen@meco.nl> for reporting this. - - * examples/Naming/Makefile (BIN2): Added the test_open.cpp file - to the Makefile. Thanks to Johnny Willemsen - <johnny.willemsen@meco.nl> for reporting this. - - * docs/tutorials/016/page02.html: Clarify that signal() or - broadcast() don't release the mutex. Thanks to Mike Curtis - <mccurry@my-deja.com> for pointing this out. - -Wed Dec 20 19:44:16 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * examples/Reactor/WFMO_Reactor/test_abandoned.cpp, - * examples/Reactor/WFMO_Reactor/test_directory_changes.cpp, - * examples/Reactor/WFMO_Reactor/test_network_events.cpp, - * examples/Reactor/WFMO_Reactor/test_prerun_state_changes.cpp, - * examples/Reactor/WFMO_Reactor/test_registration.cpp, - * examples/Reactor/WFMO_Reactor/test_registry_changes.cpp, - * examples/Reactor/WFMO_Reactor/test_talker.cpp, - * examples/Reactor/WFMO_Reactor/test_window_messages.cpp: - - * ace/Sock_Connect.cpp: The ACE 5.1.11 implementation of - ACE_Sock_Connect::get_ip_interfaces() fails when run on Windows - 2000 when ACE_HAS_WINSOCK2 = 0. This is because - get_ip_interfaces() checks the registry for the available - interfaces and fails when it encounters a new entry in windows - 2000 named "Device\NDisWanIP". The patch is to ignore invalid - devices and continue checking. Thanks to Chris Hafey - <chris@stentorsoft.com> for contributing this fix. - -Wed Dec 20 15:57:32 2000 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/RMCast/RMCast_IO_UDP.cpp - * ace/RMCast?RMCast_IO_UDP.h - * ace/RMCast/RMCast_IO_UDP.i : cleaned up tabs and trailing ws from - last submission... - -Wed Dec 20 16:18:49 2000 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/simple_test_proactor.cpp: Removed unused - 'initial_read_size' variable from file scope. - - * examples/Reactor/Proactor/test_aiosig_ace.cpp (setup_signal_handler): - * examples/Reactor/Proactor/post_completions.cpp: - * examples/Reactor/Proactor/test_end_event_loop.cpp: - * examples/Reactor/Proactor/test_cancel.cpp (Receiver::open): - Fixed unused variable warnings for gcc. - -Wed Dec 20 15:42:36 2000 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/RMCast/RMCast_IO_UDP.cpp - * ace/RMCast?RMCast_IO_UDP.h - * ace/RMCast/RMCast_IO_UDP.i : The g++ cross compiler 2.9-gnupro-98r2 - for target Lynx 3.0.1 host Solaris 2.6 was breaking with an internal - error at the end of handle_input(). Replacing a couple of lines of - code with the private, inline method allocate_and_bind_proxy() gets - around the problem. - -Wed Dec 20 15:32:44 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * examples/Threads/wfmo.cpp: Move the function calls etc. out of - the ACE_ASSERT() macros. Thanks to Mike Curtis - <mccurry@my-deja.com> for pointing this out. - -Tue Dec 19 15:08:40 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Logging_Strategy.cpp: If ACE_LACKS_IOSTREAM_TOTALLY we'll - use a "regular" file. Thanks to Sangwoo Jin for contributing - this fix. - - * ace/Mem_Map.cpp (map_it): In using ACE_Mem_Map, if the - ACE_Mam_Map::map's <offset> parameter is set to non-zero and the - <len> parameter is set to -1 then the right value to map should - be <mapped file size> - offset. Thanks to Sangwoo Jin for - reporting this and suggesting the fix! - - * ace/INET_Addr.{h,i}: Make operator < const. Thanks to - Michael Lindner <mikel@att.net> for reporting this. - -Tue Dec 19 18:39:27 2000 Luther J Baker <luther@cs.wustl.edu> - - * ace/WFMO_Reactor.cpp (simple_dispatch_handler): - - When ACE_HAS_PACE on NT, ACE must use pace_siginfo_t instead - of creating and using its own. The ACE version of siginfo_t - happens to use constructors and so, the code that uses - this ACE siginfor_t instantiates the struct via a constructor. - When ACE_HAS_PACE, the pace_siginfo_t cannot be instantiated - via constructor. - - My edit just chooses the correct method of instantiation - based on ACE_HAS_PACE. - -Tue Dec 19 18:21:52 2000 Luther J Baker <luther@cs.wustl.edu> - - * ace/Thread.h (ACE_Thread): - - Removed the default value on the overriden join (...) - function that took three parameters. - -Tue Dec 19 14:50:38 2000 Steve Huston <shuston@riverace.com> - - * ace/OS.h: If on ACE_WIN32, set ACE_DEFAULT_FILE_PERMS to - (FILE_SHARE_READ | FILE_SHARE_WRITE), not 0666. - -Tue Dec 19 14:40:26 2000 Steve Huston <shuston@riverace.com> - - * ace/Filecache.cpp: Changed definition of R_MASK and W_MASK to - 0 for ACE_WIN32; used to be for __BORLANDC__. This is related - to the change below because the third arg to ACE_OS::open is - not ignored any longer for Win32. - - * tests/Capabilities_Test.cpp: On ACE_OS::open, don't supply a - <perms> arg - let it default. The file gets deleted at the end of - the test and has nothing to be protected. Related to the below - change as well. - - * tests/Mem_Map_Test.cpp: Changed all the 0666 args to ACE_OS::open - to ACE_DEFAULT_FILE_PERMS. - -Tue Dec 19 11:02:41 2000 Steve Huston <shuston@riverace.com> - - * ace/OS.{h cpp} (open(char), open(wchar)): Changed the <perms> - argument to be used in the CreateFile call on Win32. Made the - default value of the argument match the old behavior (defaults - to FILE_SHARE_READ | FILE_SHARE_WRITE (and FILE_SHARE_DELETE on - NT4 and Win2K)). Thanks to Edan Ayal <edana@bandwiz.com> for this - suggestion! - -Mon Dec 18 20:47:40 2000 Ossama Othman <ossama@uci.edu> - - * ace/POSIX_Asynch_IO.h: - - Reverted previous change. It introduced compile-time errors. - -Mon Dec 18 19:07:31 2000 Steve Huston <shuston@riverace.com> - - * ace/Asynch_IO.h, ace/POSIX_Asynch_IO.h: Clarified/corrected some - comments regarding the write() operation getting its data from - the message block's rd_ptr. - -Mon Dec 18 11:44:11 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * tests/MEM_Stream_Test.cpp (connect_client): Fixed an unused - argument warning on KCC compiler. - -Mon Dec 18 11:00:14 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * tests/MEM_Stream_Test.cpp: Fixed problems when threads are not - supported and when explicit template instantiation is required. - -Sun Dec 17 20:11:08 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * tests/MEM_Stream_Test.cpp: - * tests/MEM_Stream_Test.dsp: - * tests/Makefile: - * tests/Makefile.am: - * tests/Makefile.bor: - * tests/run_test.lst: - * tests/run_tests.bat: - * tests/run_tests.lst: - * tests/run_tests.psosim: - * tests/tests.dsw: - * tests/version_tests/MEM_Stream_Test.dsp: - * tests/version_tests/version_tests.dsw: Added a new - MEM_Stream_Test. - -Sun Dec 17 19:32:58 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Malloc_T.cpp (shared_malloc): After a win32 structural - exception occurs, we need to re-execution the expression that - causes the exception. Thanks to Roger Beck - <roger_beck@royalsun.com.au> for reporting this. - - * ace/MEM_IO.i (fetch_recv_buf, recv): Differentiate the - cases of end of message and errors. - - * ace/Malloc_T.h: - * ace/Malloc_T.i: Added a base_addr method to return the base - address of the ACE_Malloc so we don't have to depend on MMAP's - base_addr method. - - * ace/MEM_SAP.i (set_buf_len,get_buf_len): Changed to computer - buffer location using the base_addr method in ACE_Malloc as the - one in Memory_Pool only indicates the fixed based addr. - - * ace/MEM_Acceptor.i: - * ace/MEM_Acceptor.h: Added a method to set the prefix of MMAP - filename. - - * ace/MEM_Acceptor.cpp (ACE_MEM_Acceptor): - * ace/MEM_Connector.cpp (ACE_MEM_Connector): Changed to MMAP - Options to allow MMAP files to be mapped at any address. - -Sun Dec 17 18:42:13 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Refcounted_Auto_Ptr.h: Need to add "ace/Auto_Ptr.h" here to - get the definition of ACE_Basic_Auto_Ptr. - - * tests: Added the Refcounted_Auto_Ptr_Test.cpp, which tests - the new ACE_Refcounted_Auto_Ptr class. Thanks to Johnny Tucker - <JTucker@infoglide.com> for contributing this. - - * tests/Refcounted_Auto_Ptr_Test.dsp, - tests/Refcounted_Auto_Ptr_Test.dsw - tests/Refcounted_Auto_Ptr_Test.icc: Added the projects to - compile the Refcounted_Auto_Ptr_Test.cpp. - - * tests/Makefile, - * tests/run_test.lst, - * tests/run_tests.lst: Added the Refcounted_Auto_Ptr_Test. - -Sun Dec 17 11:00:47 2000 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Date_Time.i (update): Fixed a typo. - -Sun Dec 17 08:52:04 2000 Carlos O'Ryan <coryan@uci.edu> - - * etc/tao_smartproxies.doxygen: - This file was left out from a previous commit. - -Sat Dec 16 09:36:56 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Date_Time: Added a new method update() to reset the current - date/time and also added a new constructor that calls update() - to initialize an ACE_Date_Time implicitly, rather than having to - pass all the values into the constructor. Thanks to Jerry - Odenwelder <jerryo@atl.fundtech.com> for contributing this. - -Fri Dec 15 15:27:02 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_compile: - * bin/make_pretty.pl: - The DU/CXX compiler has three levels, 'Error', 'Warning' and - 'Info:', the latter was being ignored. - -Fri Dec 15 14:36:33 2000 Christopher Kohlhoff <chris@kohlhoff.com> - - * include/makeinclude/clean.bor: - * include/makeinclude/recurse.bor: - - Added new 'realclean' target to Borland makefiles for - removing all generated files (both intermediate and final). - Thanks to Johnny Willemsen <johnny.willemsen@meco.nl> for - suggesting this. - -Fri Dec 15 14:32:06 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Log_Msg.h (ACE_Log_Msg): Changed the documentation - to reflect the fact that the %a option causes the program to - abort() rather than exit(). Thanks to Mike Curtis - <mccurry@my-deja.com> for pointing this out. - -Fri Dec 15 14:29:44 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/generate_doxygen.pl: - Generate documentation for the new TAO_SmartProxies library. - -Fri Dec 15 10:22:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/auto_run_tests.pl: - - Included PerlACE::Run_Test instead of ACEUtils and - ConfigList. - -Fri Dec 15 09:22:05 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/Connection/non_blocking/Makefile: - Fixed dependencies between the binaries and the library - generated in this directory, otherwise parallel builds can - fail. - -Thu Dec 14 10:36:45 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/auto_run_tests.lst: - * tests/run_test.lst: - - Converted to the new PerlACE::ConfigList style of .lst - files (can now say stuff like "Test: MSVC !Borland") - - * bin/auto_run_tests.pl: - * bin/auto_compile: - * tests/run_test.pl: - - Changed to use PerlACE::ConfigList. - - [Bug 729] - - * bin/ACEutils.pm: - - Made sure it would eat the -Config option so it will not - be a problem if old run_test.pl's don't get updated - immediately. - - * bin/msvc_auto_compile.pl: - - Added -d flag for debugging (just prints out what would be - compiled) - - * bin/fuzz.pl: - - Added a rudimentary run_test.pl out-of-date test. - -Thu Dec 14 10:20:49 2000 Chad Elliott <elliott_c@ociweb.com> - - * include/makeinclude/platform_lynxos.GNU - Allow users to set the debug option without using - 'override debug='. debug=0 now works properly. - -Thu Dec 14 06:00:12 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace: Removed makefile-light since it's no longer needed. - Thanks to Johnny Willemsen <johnny.willemsen@meco.nl> for - reporting this. - -Wed Dec 13 18:33:19 2000 Carlos O'Ryan <coryan@uci.edu> - - * html/README: - Add README file explaining why the directory is empty and why - the index.html URLs don't work unless you download or generate - the Doxygen files. - - * html/index.html: - Add entry for the smart proxies library. - -Wed Dec 13 18:27:29 2000 Carlos O'Ryan <coryan@uci.edu> - - * THANKS: - Add 'Pedro Ferreira' to the hall of fame. - -Thu Dec 13 18:18:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/PerlACE/ConfigList.pm: (added) - * bin/PerlACE/Process.pm: (added) - * bin/PerlACE/Process_Unix.pm: (added) - * bin/PerlACE/Process_Win32.pm: (added) - * bin/PerlACE/Run_Test.pm: (added) - - These are a couple of Perl modules that will be used - to simplify current run_test.pl's and the auto_builds - and to allow us to run tests on Win32 for configurations - that output executables into subdirs (Win32 Release, - Win32 Static *, Borland). The conversion of scripts over - to the new style will occur in separate checkins. - - ConfigList provides a more powerful *.lst format for - specifying test lists. Process is a newer version of - the old bin/Process.pm that has a different interface - and a couple of extra methods for common tasks for - our scripts. And Run_Test.pm will replace the - bin/ACEutils.pm and be the main module for our - run_test.pl scripts. It contains some common subroutines - and automatically parses some arguments via ARGV. - - * docs/run_test.txt: (added) - - A bit of documentation on how new run_test.pl's will - look. - -Wed Dec 13 18:15:03 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/CDR_Stream.h: - Made some of the adjust() methods public, this is used to - optimize TAO's implementations of CORBA::Any. - -Wed Dec 13 12:59:49 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/PerlACE/Process_Unix.pm: - - Fixed PerlACE::Process on Unix to have the same interface as - the Win32 version. - -Tue Dec 12 09:52:00 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * Well, it looks like we finally have a new president in the - US, yow! - -Tue Dec 12 04:53:17 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Service_Config.{h,i}: Added an inequality operator. - Thanks to Mike Curtis <mccurry@my-deja.com> for pointing this - out. - - * ace/Service_Config.cpp (process_commandline_directives): Changed - a check for error values being -1 to being != 0 since process_directive() - can return - - 1) a positive number of parse errors - 2) and -1 on a certain ACE_NEW_RETURN call - 3) but always zero on success. - - Thanks to Mike Curtis <mccurry@my-deja.com> and - Munagala Ramanath <Munagala.Ramanath@PostX.com> - for pointing this out. - -Tue Dec 12 01:07:39 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Log_Msg.cpp: - - Changed a comment to remove a false warning by fuzz. - -Mon Dec 11 15:10:00 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Get_Opt.h: - In some platforms 'optind', 'optarg' and 'opterr' are defined as - macros, undefine them because we use the same identifiers as - variable names. - - * ace/OS.h: - If 'timezone' is defined as a macro in the system header files - then define ACE_TIMEZONE and #undef timezone. - The ACE_OS::timezone() method cannot be compiled otherwise, but - unfortunately this forces us to disable support for the method - under that sort of platform. - - * ace/OS.h: - Do not include sys/shm.h if ACE_LACKS_SYSV_SHMEM is defined, - there is no reason to #include the file if we are not using it - anyway. - - * ace/ACE.cpp: - Check the return value from getrlimit() before accepting its - results. - -Mon Dec 11 00:50:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/ace_wchar.h: - - Created an ACE_ANTI_TCHAR and ACE_TEXT_ANTI_TO_TCHAR. - ACE_ANTI_TCHAR has the opposite behavior of ACE_TCHAR. - - * ace/Process.cpp: - * ace/Process.h: - * ace/Process.i: - - Added wchar_t versions of: - - - ACE_Process_Options::working_directory (wd) - - ACE_Process_Options::command_line (format, ...) - - * ace/OS.cpp: - * ace/OS.h: - * ace/OS.i: - - Added wchar_t versions of: - - - ACE_OS::cond_init () - 2 versions - - ACE_OS::mutex_init () - - ACE_OS::thread_mutex_init () - - ACE_OS::event_init () - - ACE_OS::sema_init () - - On Win32, these should be implemented using the *W version - of the APIs. On other platforms, they call the char version - of the same ACE_OS call with a converted string. - - * ace/Log_Msg.h: - * ace/Log_Msg.cpp: - - Changed the first parameter to set and conditional_set from - being ACE_TCHAR to just char. We'll assume filenames are - always ansi strings (since __FILE__ seems to act this way). - There was actually a dangling pointer problem in the - conditional_set when ACE_USES_WCHAR was defined before. - - Added an ANTI_TCHAR version of log (priority, format, ...). - What this means is now we can use either narrow or wide - format strings with Log_Msg, regardless of whether ACE_USES_WCHAR - is defined or not. Hopefully this means we can get rid of a lot - of the ACE_TEXT and ACE_LIB_TEXT's in our code. - - Added a couple of more specifiers, %C and %w. So for characters - and strings we have: - - %c : print out an ansi character - %C : print out an ansi string - %s : print out an ACE_TCHAR * string - %w : print out a wide character - %W : print out a wide string - - An example, which will also function correctly even when - ACE_USES_WCHAR is defined: - - void print (char *a_str, wchar_t *w_str, ACE_TCHAR *t_str) - { - ACE_DEBUG ((LM_DEBUG, - "%C %s %W\n", - a_str, - t_str, - w_str)); - } - - Unfortunately this only really takes care of the string - specifiers. There is still work to be done with the other - implicit string specifiers (%p, %N, etc.). - - * ace/RMCast/RMCast_IO_UDP.cpp: - * ace/RMCast/RMCast_Reassembly.cpp: - * ace/SSL/SSL_Context.cpp: - * ace/SSL/SSL_SOCK_Acceptor.cpp: - * ace/SSL/SSL_SOCK_Connector.cpp: - * ace/SSL/SSL_SOCK_Stream.i: - - Removed unnecessary ACE_TEXT's to see how things go. - - * docs/wchar.txt: - - Updated the documentation. - -Fri Dec 8 13:47:58 2000 Jeff Parsons <parsons@cs.wustl.edu> - - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - - Added Refcounted_Auto_Ptr.{h,i} to these projects. - -Fri Dec 8 10:34:32 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Makefile: Added Refcounted_Auto_Ptr to the list of template - files. - - * ace/Auto_Ptr.{h,i}: - * ace/Refcounted_Auto_Ptr.{h,i}: Moved the new - ACE_Refcounted_Auto_Ptr into a separate file to reduce - unnecessarily coupling in ACE. Thanks to Darrell Brunsch for - pointing this out. - -Fri Dec 8 09:27:01 2000 Chad Elliott <elliott_c@ociweb.com> - - * include/makeinclude/platform_chorus4.x_g++.GNU: - - Modified to build only static libraries. Something changed - between the time of the port and now, which caused it to try - to build shared libraries (which Chorus 4.0 doesn't support). - -Fri Dec 8 06:47:12 2000 Christopher Kohlhoff <chris@kohlhoff.com> - - * tests/Makefile.bor: - - Use variable called TARGET_NAME instead of NAME to - prevent clash with nightly build scripts. - -Thu Dec 7 13:16:23 2000 Darrell Brunsch <brunsch@uci.edu> - - * examples/IPC_SAP/SPIPE_SAP/server.cpp: - - Removed PERMS global variable, since it wasn't being - used (which KCC noticed). - - * examples/System_V_IPC/SV_Message_Queues/MQ_Client.cpp: - * examples/System_V_IPC/SV_Message_Queues/MQ_Server.cpp: - * examples/System_V_IPC/SV_Message_Queues/TMQ_Client.cpp: - * examples/System_V_IPC/SV_Message_Queues/TMQ_Server.cpp: - - When calling ACE_OS::cuserid and passing a NULL pointer, it - should be cast to something like char * or wchar_t *. In - this case it should be a char *, so I added a - ACE_static_cast for it. - -Thu Dec 7 02:31:44 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/config-WinCE.h: - - Disabled inlining for SH3 Release configuration. The SH3 compiler - has problems with inlined functions (in this case, ACE_OS::dlerror's - static variable was staying unresolved). Thanks to Ricardo Chan - <ricchan@nortelnetworks.com> for pointing this out to me. - -Thu Dec 7 03:33:49 2000 Marina Spivak <marina@cs.wustl.edu> - - * bin/auto_run_tests.lst: - - Added RTCORBA MT test. - -Wed Dec 6 19:52:42 2000 Darrell Brunsch <brunsch@uci.edu> - - - * ace/Auto_Ptr.i: - - Added a couple of ACE_INLINEs and an include of Synch_T.h. - -Wed Dec 06 20:12:35 2000 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Containers_T.cpp (operator=): In case the existing array is - larger than the rhs (right hand side), we don't have to free and - reallocate the existing array. However, we do need to run the - destructor on the array elements that will be replaced. Hence, - I added a call to ACE_DES_ARRAY_NOFREE for the elements being - replaced. - - This fixes bug 704. Thanks to Martin Krumpolec - <krumpo@pobox.sk> for reporting the bug and for providing an - example illustrating the problem. - -Wed Dec 6 14:29:13 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Auto_Ptr.{h,i}: Added a new set of class (ACE_Refcounted_Auto_Ptr_Rep, - ACE_Refcounted_Auto_Ptr) that implements support for a reference - counted auto_ptr. Thanks to Johnny Tucker - <JTucker@infoglide.com> for contributing this. - - * ace/Managed_Object.h: Added ACE_UNIMPLEMENTED_FUNC macros for - copy constructor and assignment operator in class - ACE_Cleanup_Adapter. Thanks to Johnny Willemsen - <johnny.willemsen@meco.nl> for reporting this. - - * tests/run_tests.sh: Prevent null or zero-length string test name from being - passed into the run() function. Thanks to Jon Loeliger - <jloeliger@chiaro.com> for reporting this. - - * tests/Log_Msg_Test.cpp, - * examples/Log_Msg/test_log_msg.cpp (main): - Replaced int with u_long to make compilers happy. Thanks to - Mike Curtis <mccurry@my-deja.com> for pointing this out. - -Tue Dec 5 11:45:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/SSL/SSL_SOCK_Acceptor.cpp: - - Added ACE_TEXT to a debug message. - -Tue Dec 5 10:07:14 2000 Carlos O'Ryan <coryan@uci.edu> - - * performance-tests/Misc/childbirth_time.cpp: - Use &function instead of just function, hopefully that will make - Sun/CC 5.0 happy. - -Tue Dec 5 09:38:22 2000 Ossama Othman <ossama@uci.edu> - - * ace/SSL/SSL_SOCK_Acceptor.cpp (ssl_accept): - - Return "-1" instead of the result of the - ACE_SSL_SOCK_Stream::close() method. The latter may actually, - and most likely will return without error. - -Tue Dec 05 05:39:39 2000 Christopher Kohlhoff <chris@kohlhoff.com> - - * tests/Makefile.bor: - - Implemented 'clean' support in the ACE tests makefile. - -Tue Dec 05 02:02:51 2000 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/IPC_SAP.cpp (disable): Reverted this change: - - Tue Nov 21 17:42:36 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - The change seemed too low level where it might end up effecting - more code than was intended. However, as I discussed with Chris - Uzdavinis <chris@atdesk.com>, the correct solution is to - specialize the activation of the Svc_Handler rather than its - accepting. - -Mon Dec 04 23:54:10 2000 Irfan Pyarali <irfan@cs.wustl.edu> - - * examples/Reactor/Proactor/test_proactor.cpp: - - - Receiver::open(): Only duplicate the message block when we are - faking the result. Otherwise, initiate_read_stream will create - its own message block. - - - Receiver::open(): Delete the fake result. - - - Receiver::handle_read_stream(): When there is no more data to - read, release the message block and delete the receiver. - - - Sender::transmit_file(): No need to duplicate the trailer. - - * ace/Proactor.cpp (~ACE_Proactor_Timer_Handler): - - The following was changed: - - this->thr_mgr ()->wait (); - - to: - - this->thr_mgr ()->wait_grp (this->grp_id ()); - - since Thread_Manager::wait() if called while the - ACE_Object_Manager is shutting down (as a result of program - shutdown or ACE::fini), it will not wait for any threads to - complete. - - Also, the timer_handler thread was not created detached since - Thread_Manager::wait_grp() does not wait on detached threads. - -Mon Dec 04 16:32:11 2000 Ossama Othman <ossama@uci.edu> - - * ace/SSL/SSL_SOCK_Stream.i (recv): - - Shutdown the SSL connection if SSL_read() returns with an - SSL_ERROR_ZERO_RETURN OpenSSL error condition. - SSL_ERROR_ZERO_RETURN indicates that the peer is shutting down, - and that a "close_notify" message SSL message has been issued by - the peer. A "close_notify" message indicates that no further - data will be sent through the given connection, so there is no - need to keep the connection open. - - Removed SSL_ERROR_WANT_{WRITE,X509_LOOKUP} switch statement - cases. They are apparently not necessary when performing a SSL - read. - - (send): - - Removed SSL_ERROR_WANT_{READ,X509_LOOKUP} switch statement - cases. They are apparently not necessary when performing a SSL - write. - - * ace/SSL/SSL_SOCK_Acceptor.cpp (ssl_accept): - - Check that the X.509 verification didn't fail. If the - verification failed then do not continue. - - * ace/SSL/SSL_Context.cpp: - * ace/SSL/SSL_SOCK_Connector.cpp: - * ace/SSL/SSL_SOCK_Stream.cpp: - - Added missing ACE_RCSID macros. - - * ace/SSL/SSL_Context.h: - * ace/SSL/SSL_SOCK.h - * ace/SSL/SSL_SOCK_Acceptor.h: - * ace/SSL/SSL_SOCK_Connector.h: - * ace/SSL/SSL_SOCK_Stream.h: - - Doxygenated these headers. - - * ace/SSL/SSL_Context.i: - * ace/SSL/SSL_SOCK.cpp: - - Cosmetic updates. - -Mon Dec 04 12:45:46 2000 Steve Huston <shuston@riverace.com> - - * ace/High_Res_Timer.cpp: Corrected some problems with Linux on - Pentium that were introduced by this: - Fri Nov 10 10:42:31 2000 Steve Huston <shuston@riverace.com> - - * ace/High_Res_Timer.cpp: Replaced all occurrences of - "defined (ACE_HAS_PENTIUM)" with "defined (ACE_WIN32)" since - there is nothing Pentium-specific any longer in the Win32 - sections. - - Thanks to David Levine and Carlos O'Ryan for weeding this bug out. - -Mon Dec 4 07:39:16 2000 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Service_Configurator/IPC-tests/clients: - Added Borland C++ Builder specific makefile. - -Sun Dec 3 20:03:16 2000 Ossama Othman <ossama@uci.edu> - - * examples/Web_Crawler/Command_Processor.cpp: - - Removed duplicate template instantiations. Thanks to Paul Rubel - <prubel@bbn.com> for pointing out this bug, and for providing a - fix. - - * THANKS: - - Added Paul to the Hall-of-Fame. - -Sun Dec 3 15:23:50 2000 Carlos O'Ryan <coryan@uci.edu> - - * performance-tests/Misc/childbirth_time.cpp: - Use 'extern "C"' for the thread entry point, this should make - Sun/CC 5.0 (and other compilers) happy. - -Sun Dec 3 08:50:15 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/OS.i: Fixed a cut and paste error. This was from the BUG# - 734 in bugzilla. Thanks to Martin Krumpolec <krumpolec@pobox.sk> - for reporting this. - -Sat Dec 2 05:58:37 2000 Christopher Kohlhoff <chris@kohlhoff.com> - - * ace/config-win32-borland.h: - - Automatically define WIN32 macro if we are able to - determine that it is the target platform from pre- - defined compiler macros. - - * ace/Makefile.bor: - - Added missing template source file for install target. - -Sat Dec 2 00:19:38 2000 Marina Spivak <marina@cs.wustl.edu> - - * bin/auto_run_tests.lst: - - Added RTCORBA/Explicit_Binding test. - -Fri Dec 01 21:17:35 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Thread_Manager.cpp: - More ACE_LEGACY_MODE fixes. - -Fri Dec 1 13:49:36 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Log_Msg.cpp: - Fixed problems on non-legacy Win32 builds. - -Fri Dec 1 09:46:15 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Log_Msg.h: - * ace/Log_Msg.cpp: - The structured exception handler methods should show up on - ACE_LEGACY_MODE builds. Unfortunately, the change from - ACE_LATEST_AND_GRATEST to LEGACY_MODE didn't work here because - the LATEST_AND_GREATEST macro was misspelled. - -Fri Dec 1 09:30:08 2000 Carlos O'Ryan <coryan@uci.edu> - - * include/makeinclude/platform_linux.GNU: - Add support for wxWindows under linux. - -Fri Dec 1 10:56:42 2000 Jeff Parsons <parsons@cs.wustl.edu> - - * include/makeinclude/ace_flags.bor: - - Added definitions necessary for the new smart proxies - library and its Borland makefile. - -Fri Dec 01 09:21:31 2000 Steve Huston <shuston@riverace.com> - - * ace/Reactor.h: Clarified some comments. - -Thu Nov 30 22:49:10 2000 pradeep@cs.wustl.edu <pradeep@cs.wustl.edu> - - * ACE version 5.1.11 released. - -Wed Nov 29 23:55:37 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * apps/drwho/server.cpp: - * apps/drwho/BS_Client.cpp: - Used casts to quiet VxWorks/LynxOS. - * bin/auto_run_tests.lst: - Commented out this test - its broken. - -Wed Nov 29 10:35:26 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * examples/Service_Configurator/IPC-tests/client/remote_service_directory_test.cpp (parse_args): - Removed the colon after the "r" in the parse-args list. Thanks - to Mike Curtis <mccurry@my-deja.com> for pointing this out. - - * ace/Method_Request.h: Added ACE_UNIMPLEMENTED_FUNC to the - assignment operator and copy constructor of ACE_Method_Request. - Thanks to Johnny Willemsen <johnny.willemsen@meco.nl> for - pointing this out. - -Wed Nov 29 09:41:51 2000 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Reactor/Multicast/*.bor - Added Borland C++ Builder specific makefiles. Thanks to - Albert Wijna <albert.wijnja@meco.nl> for creating these. - -Tue Nov 28 15:39:51 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_compile: - Print out an error message when the script returns a non-zero - code. The errors where reported by email, but they didn't show - up in the make_pretty output. - -Tue Nov 28 15:36:10 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/generate_doxygen.pl: - Don't generate inline code for releases, it bloats the - documentation, and does not make things any clearer. - - * html/index.html: - Fixed URLs - - * etc/ace.doxygen: - * etc/ace_man.doxygen: - * etc/ace_rmcast.doxygen: - * etc/ace_ssl.doxygen: - * etc/tao_dynamicany.doxygen: - * etc/tao_dynamicinterface.doxygen: - * etc/tao_esf.doxygen: - * etc/tao_implrepo.doxygen: - * etc/tao_iormanip.doxygen: - * etc/tao_iortable.doxygen: - * etc/tao_portableserver.doxygen: - * etc/tao_rtevent.doxygen: - * etc/tao_strategies.doxygen: - Make sure that ACE_THROW_SPEC is properly defined. - - * etc/tao.doxygen: - Also fix the relative path to the ace doxygen files. - -Tue Nov 28 13:31:47 2000 David L. Levine <levine@cs.wustl.edu> - - * tests/run_test.pl: prefix exec of run_tests.sh with - $EXEPREFIX, so that it works without . in the user's - PATH. Thanks to Carlos for pointing this out. - -Tue Nov 28 08:09:35 2000 David L. Levine <levine@cs.wustl.edu> - - * tests/run_test.pl: added some more diagnostics to try - to figure out why run_tests.sh can't be exec'd on - UCI machines. - -Mon Nov 27 17:02:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ACE-INSTALL.html: - While following up on Johnny Willemsen's - <johnny.willemsen@meco.nl> suggestion to make sure I removed all - references to .mdp files, I noticed that some of the Win32 - instructions could use some updating. - -Mon Nov 27 17:44:52 2000 David L. Levine <levine@cs.wustl.edu> - - * tests/run_test.pl: if run_tests.sh doesn't exist, print - a diagnostic message and continue with the run_test.pl - version of the tests. For unknown reasons, this change: - Mon Nov 27 11:58:43 2000 David L. Levine <levine@cs.wustl.edu> - doesn't work on the UCI Linux machines, run_tests.sh can't - be exec'd. - -Mon Nov 27 10:17:58 2000 Ossama Othman <ossama@uci.edu> - - * docs/tutorials/005/005.dsp: - - Corrected name of output binary. It should have been - `server.exe' not `client.exe'. Thanks to David Channon - <djc@uq.net.au> for reporting this problem. - -Mon Nov 27 09:56:09 2000 Ossama Othman <ossama@uci.edu> - - The following updates are based on feedback from David Channon - <djc@uq.net.au>. Thanks David! - - * docs/tutorials/003/client.cpp: - * docs/tutorials/003/page01.html: - * docs/tutorials/004/client.cpp: - * docs/tutorials/004/page01.html: - * docs/tutorials/008/broadcast_client.cpp: - * docs/tutorials/008/directed_client.cpp: - * docs/tutorials/008/page02.html: - * docs/tutorials/008/page03.html: - * docs/tutorials/008/page04.html: - * docs/tutorials/008/server.cpp: - * docs/tutorials/009/broadcast_client.cpp: - * docs/tutorials/009/directed_client.cpp: - * docs/tutorials/009/page02.html: - * docs/tutorials/009/page03.html: - * docs/tutorials/009/page04.html: - * docs/tutorials/009/server.cpp: - * docs/tutorials/019/client.cpp: - * docs/tutorials/019/client2.cpp: - * docs/tutorials/019/page02.html: - * docs/tutorials/019/page03.html: - * docs/tutorials/019/page04.html: - * docs/tutorials/019/page05.html: - * docs/tutorials/019/server.cpp: - * docs/tutorials/019/server2.cpp: - * docs/tutorials/019/shmem.cpp: - * docs/tutorials/020/client.cpp: - * docs/tutorials/020/client2.cpp: - * docs/tutorials/020/mmap.cpp: - * docs/tutorials/020/page02.html: - * docs/tutorials/020/page03.html: - * docs/tutorials/020/page04.html: - * docs/tutorials/020/page05.html: - * docs/tutorials/020/server.cpp: - * docs/tutorials/020/server2.cpp: - - Include "ace/Log_Msg.h" to pull in the ACE_{DEBUG,ERROR} - macros. The example code in the HTML files was also updated. - - * docs/tutorials/011/message_queue.cpp: - * docs/tutorials/011/page02.html: - - Removed unnecessary line of code that NULL terminated a string. - It didn't serve any purpose since the string is already - terminated. If it wasn't terminated then that point in the code - would never have been reached since the code prior to it assumes - that the string is NULL terminated. - -Mon Nov 27 11:58:43 2000 David L. Levine <levine@cs.wustl.edu> - - * tests/run_test.pl: on platforms other than Win32, simply - exec run_tests.sh. run_tests.sh properly handles all of - the current ACE tests. [Bug 729] - -Mon Nov 27 14:12:09 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Activation_Queue.h: Added ACE_UNIMPLEMENTED_FUNC to the - assignment operator and copy constructor of - ACE_Activation_Queue. Thanks to Johnny Willemsen - <johnny.willemsen@meco.nl> for pointing this out. - -Sun Nov 26 19:07:48 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * examples/Map_Manager/test_hash_map_manager.cpp: - * examples/Shared_Malloc/Malloc.cpp: - * examples/Shared_Malloc/test_position_independent_malloc.cpp: - - Removed extraneous explicit template instantiation declarations - causing "multiply defined" errors on VxWorks and LynxOS. - -Sat Nov 25 22:08:46 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * apps/drwho/Makefile: - Disabled drwho compilation on Vxworks - it doesn't have - protocols/rwhod.h. - -Fri Nov 24 16:58:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/IPC_SAP.cpp: - - WSAEventSelect is not available on Windows CE, so the #ifdefs - were changed so it calls ioctl on Windows CE also. - -Fri Nov 24 15:43:20 2000 Carlos O'Ryan <coryan@uci.edu> - - * docs/ACE-categories.html: - Fixed URLs to Doxygen documents, I used the absolute URL to our - Doxygen site, but in the future we may want to do that - configurable. - - * bin/auto_run_tests.lst: - Disable the Connection_Purging test in TAO, it never worked and - the features are disabled while we continue to improve the - Connection Cache. - -Fri Nov 24 11:42:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ACE-INSTALL.html: - - Updated sections about Windows CE and MSVC. CE stuff - is documented in CE-status.txt, and removed references to old - MSVC 4.2 projects. Thanks to Johnny Willemsen - <johnny.willemsen@meco.nl> for pointing this out. - -Fri Nov 24 05:48:26 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Signal.h: Enhanced the documentation to clarify the role of - sigkey. Thanks to Oscar Rodriquez <Oscar.Rodriquez@eln.ericsson.se> - for motivating this. - - * ace/SOCK_Connector.h: Updated the documentation to clarify the - parameters to the constructors and the connect() method. Thanks - to Pierre Oberson <oberson@nagra-kudelski.ch> for suggesting this. - -Thu Nov 23 21:17:16 2000 David L. Levine <levine@cs.wustl.edu> - - * tests/Reactor_Notify_Test.cpp (main): added - auto_ptr<ACE_Reactor> so that the global Reactor instance - will be destroyed at the termination of main (). The cleans - up memory leaks that we introduced in this change to the test: - Tue Oct 24 12:30:47 2000 Steve Huston <shuston@riverace.com> - -Thu Nov 23 18:38:54 2000 Carlos O'Ryan <coryan@cs.wustl.edu> - - * Makefile: - Add ACE_wrappers/html to the list of files tagged, even though - they are not included in the tar or zip files. - -Thu Nov 23 18:13:32 2000 Carlos O'Ryan <coryan@cs.wustl.edu> - - Don't use 'grep -q' because Solaris' grep does not support it. - -Thu Nov 23 18:27:52 2000 Carlos O'Ryan <coryan@cs.wustl.edu> - - * Makefile: - The ACE-INSTALL file is generated in the staging directory, - there is no need to generate it locally too. - -Thu Nov 23 17:46:56 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Method_Request: Made the priority() method const. Thanks - to Johnny Willemsen <jwillemsen@remedy.nl> - -Thu Nov 23 14:43:40 2000 Carlos O'Ryan <coryan@uci.edu> - - * Makefile: - Do not include the html directory in the release. The doxygen - files are too big for this. - - * ace/config-doxygen.h: - Improve doxygen documentation. - - * bin/generate_doxygen.pl: - Add options to change the path to find dot and perl on the fly. - - * bin/make_release: - Generate the HTML documentation in a separate .tar.gz (or .zip) - file. - -Wed Nov 22 17:35:55 2000 Angelo Corsaro <corsaro@cs.wustl.edu> - - * bin/auto_run_tests.lst: - - Added entry for Collocation test. Moved entry for Client - Exposed Policy test from the old list file to this one. - -Wed Nov 22 09:29:32 2000 Darrell Brunsch <brunsch@uci.edu> - - * etc/ace.doxygen: - - Added the docs directory and *.txt files to be looked - at by doxygen. - - * docs/CE-status.txt: - * docs/wchar.txt: - - Fixed up these files, so they show up correctly in the - doxygen html pages. - -Tue Nov 21 19:55:49 200 0 Pradeep Gore <pradeep@cs.wustl.edu> - - * ace/IPC_SAP.cpp: - Casting param1 to SOCKET did the trick. Thanks Doug! - -Tue Nov 21 19:21:10 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * ace/IPC_SAP.cpp: - - Reverted previous change applied to ACE_IPC_SAP::disable because - of compile error. - -Tue Nov 21 18:54:14 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * apps/drwho/CM_Client.cpp: - - Explicitly cast Comm_Manager::sokfd_ to int in calls to - ACE_OS::recvfrom and ACE_OS::closesocket to quiet LynxOS. - -Tue Nov 21 17:42:36 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * ace/IPC_SAP.cpp: - - Modified ACE_IPC_SAP::disable to call ::WSAEventSelect on Win32. - Thanks to Chris Uzdavinis <chris@atdesk.com> for contributing - these changes. - -Tue Nov 21 11:00:14 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/run_test.lst: - - Borland shouldn't be running ACE_Init_Test. - -Tue Nov 21 10:15:15 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS_String.inl: - - Changed ::towlower to towlower since it can be a macro. - -Tue Nov 21 09:57:53 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/Reactor_Exceptions_Test.cpp: - - Code style fix. - - * tests/run_test.lst: - - New_Fail_Test just eats up memory when run, and it only tests - the behavior of new. Not really something we need to run - everynight (and as David Levine reports, using up all the - memory can cause problems in other processes on the nightly - build machines). Commented out the New_Fail_Test. - - * tests/run_test.pl: - - Fixed to clean up some of the backing_store_* files in the - tests directory after running the tests. - - * bin/ACEutils.pm: - - Added the ability to pick -ExeSubDir <dir> from the command - line to alter $EXEPREFIX. This should allow us to start - running tests with different configurations (ie. Win32 Release) - and on Borland. - -Tue Nov 21 07:24:54 2000 Chris Kohlhoff <chris@kohlhoff.com> - - * ace/SSL/Makefile.bor: - * include/makeinclude/build_dll.bor: - * include/makeinclude/build_exe.bor: - * include/makeinclude/build_lib.bor: - - Prevent SSL libraries from being built unless SSL_ROOT is defined. - - * include/makeinclude/build_example.bor: - - Add missing libraries required for statically linking TAO examples. - -Mon Nov 20 21:52:57 2000 Ossama Othman <ossama@uci.edu> - - * ace/Makefile.am: - - Added more missing files to the source lists. - -Mon Nov 20 19:16:12 2000 Ossama Othman <ossama@uci.edu> - - * ace/OS.i (mutex_init): - - Fixed an unused argument warning that was showing up in - ACE+Autoconf builds. - - * ace/Makefile.am (libACE_Utils_la_SOURCES): - - Updated "Init.cpp" to "Init_ACE.cpp." For some reason this - file wasn't updated when the name change was made. - - (libACE_Utils_la_SOURCES): - - Moved `Sock_Connect.cpp' to this list of sources. Similar to - the above change, this files wasn't updated when the - corresponding to changes was made to the classical Makefile. - -Mon Nov 20 15:00:26 2000 Marina Spivak <marina@cs.wustl.edu> - - * bin/auto_run_tests.lst: - - Added RTCORBA/Private_Connection, updated options for other - RTCORBA tests. - -Sun Nov 19 15:26:56 2000 Ossama Othman <ossama@uci.edu> - - * ace/SSL/sslconf.h - (ACE_DEFAULT_SSL_CERT_FILE, ACE_DEFAULT_SSL_CERT_DIR): - - On Win32, do not use UNIX-specific directories in certificate - paths. - -Sun Nov 19 10:37:48 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - Add the new files to the MSVC projects. - - * ace/Dynamic_Service_Base.h: - Add missing file. - - * ace/Dynamic_Service_Base.cpp: - Removed template-like protections, this is a regular class. - - * ace/Dynamic_Service.i: - * ace/Dynamic_Service.cpp: - Cosmetic fixes. - -Sun Nov 19 08:05:08 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Makefile: - Updated. - - * ace/Dynamic_Service.cpp: - * ace/Dynamic_Service.h: - * ace/Dynamic_Service_Base.cpp: - * ace/Dynamic_Service_Base.h: - Not all methods in Dynamic_Service were based on the SERVICE - types. Moved those to the base class 'Dynamic_Service_Base' and - the Dynamic_Service inherits from the base class. - -Sat Nov 18 15:39:49 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * examples/Service_Configurator/IPC-tests/client/remote_service_directory_test.cpp: - Fixed a mistake in the "usage" string. Thanks to Mike Curtis - <mccurry@my-deja.com> for pointing this out. - -Sat Nov 18 11:35:53 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/OS_Test.cpp: - - Changed the signature of main to use ACE_TCHAR. Although - I don't consider this the "right" way of doing this, it - will make the builds happy with ACE_USES_WCHAR until we - implement something better. - -Sat Nov 18 09:58:17 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Basic_Types.h: - * ace/OS_Log_Msg_Attributes.h: - * ace/config-all.h: - Under LynxOS we must #include stdarg.h before stdio.h. - Thanks to David for letting us know about this. - -Fri Nov 17 14:24:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * include/makeinclude/clean.bor: - - Added the removal of *Cli.cpp, *Ser.cpp, etc. files which - is the output style of the TAO Param_Test. - - * tests/Reactor_Exceptions_Test.cpp: - - Put the handler in its own scope so it must be cleaned up - before the reactor. For whatever reason, it wasn't happening - in the right order on MSVC before. - - Also changed the exception error to a debug message, since - it really doesn't seem an error (we explicitly cause one). - -Fri Nov 17 11:22:08 2000 Carlos O'Ryan <coryan@uci.edu> - - * include/makeinclude/platform_sunos5_kcc.GNU: - Same fix as for platform_linux_kcc.GNU. - A little more detail has surfaced: apparently under this - compiler we attempted to link the static run-time and system - libraries if static_libs_only=1 was set. Not only this would be - unique (in that we don't do this for other compilers), but it - fails when only one directory is compiled with - static_libs_only=1, hile only the dynamic version of ACE (or any - other of our libraries) is compiled. - IMHO it is better to make it consistent with all the other - platforms, and leave the extra optimization of using static - system libraries for a future makefile option. - - * include/makeinclude/platform_linux_kcc.GNU: - Remove obsolete comment - -Fri Nov 17 03:37:37 2000 Marina Spivak <marina@cs.wustl.edu> - - * bin/auto_run_tests.lst: - - Added RTCORBA/Banded_Connections test. - -Fri Nov 17 01:16:25 2000 Ossama Othman <ossama@uci.edu> - - * include/makeinclude/wrapper_macros.GNU (ssl): - - Removed definition of the ACE_HAS_SSL macro. It is no longer - used. - -Thu Nov 16 16:14:03 2000 Carlos O'Ryan <coryan@uci.edu> - - * Makefile: - * html/doxygen.css: - * html/index.html: - New html directory. People normally don't expect html documents - to show up under the man directory. - - * configure.in: - * bin/bootstrap: - * man/Makefile.am: - * man/html/Makefile.am: - Remove the man/html directory. - - * bin/make_release: - Insert code to generate the man pages index. - - * docs/ACE-categories.html: - Change references to man/html so they point to the new doxygen - generated files. - - * bin/README.html: - * bin/class2hxxcxx: - * bin/class2hxxcxxsingle: - * bin/class2info: - * bin/class2info.awk: - * bin/class2man: - * bin/class2mml: - * bin/class2src: - * bin/classinfo.ps: - * bin/generate_html_windex: - * bin/generate_man_pages: - * bin/hiding.fmt: - * bin/html-windex: - * bin/info2doc.awk: - * bin/info2doc.fmt: - * bin/info2head: - * bin/info2head.fmt: - * bin/info2headsrc: - * bin/info2man: - * bin/info2mml: - * bin/info2src: - * bin/info2src.awk: - * bin/man2html: - * bin/man2html1.awk: - * bin/man2html2.awk: - * bin/vendor.fmt: - Remove old scripts used in man page and html document - generation. We do everything with doxygen now. - -Thu Nov 16 17:55:18 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * apps/JAWS/clients/Blobby/Blob.{h,cpp}: - * apps/JAWS/clients/Blobby/Blob_handler.{h,cpp}: - * apps/drwho/Rwho_DB_Manager.h: - * apps/drwho/Multicast_Manager.{.h,cpp}: - * apps/drwho/Options.{h,cpp}: - Follow const correctness to make suncc5 happy. - - * examples/IOStream/server/iostream_server.h - Use ACE_HAS_TEMPLATE_TYPEDEFS rather than ACE_HAS_TYPENAME_KEYWORD - Thanks to Steve Huston <shuston@riverace.com> for the fix. - -Thu Nov 16 15:31:20 2000 Carlos O'Ryan <coryan@uci.edu> - - * include/makeinclude/platform_linux_kcc.GNU: - Builds with static libraries under Linux/KCC were failing due to - some magical -Bstatic options in the configuration file. When - the options are removed the problem went away. - -Thu Nov 16 15:31:20 2000 Carlos O'Ryan <coryan@uci.edu> - - * include/makeinclude/platform_linux_kcc.GNU: - Builds with static libraries under Linux/KCC were failing due to - some magical -Bstatic options in the configuration file. When - the options are removed the problem went away. - -Thu Nov 16 15:09:25 2000 Carlos O'Ryan <coryan@uci.edu> - - * tests/TkReactor_Test.cpp: - Fixed 'char*' vs. 'const char *' conflict. - -Wed Nov 15 14:23:54 2000 Ossama Othman <ossama@uci.edu> - - * configure.in: - - Fixed bug in the auto_ptr / C++ std namespace test where the - template argument for the auto_ptr being instantiated was - a non-existent type. - Reported by Kevin Marhsall <KCMarshall@att.net> - - * THANKS: - - Added Kevin to the hall of fame list. - -Wed Nov 15 10:19:45 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/Threads/thread_specific.cpp: - Fixed warning under linux - -Wed Nov 15 08:32:58 2000 Carlos O'Ryan <coryan@uci.edu> - - * apps/JAWS/server/Makefile: - * performance-tests/Misc/Makefile: - Change dependencies to ensure that the libraries are created - before we attempt to link the binaries, otherwise the - compilation fails with parallel builds. - -Tue Nov 14 17:55:52 2000 Carlos O'Ryan <coryan@uci.edu> - - * Makefile: - PACE is disabled by default, unless pace=1 is set in the command - line or the platform_macros.GNU file. - -Tue Nov 14 19:36:47 2000 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/test_siosig_ace.cpp: Don't log - the read/written file using ACE_DEBUG - it will usually overflow - the ACE_Log_Record max message length and do bad things. - -Tue Nov 14 19:19:16 2000 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/simple_test_proactor.cpp: Don't log - the read/written file using ACE_DEBUG - it will usually overflow - the ACE_Log_Record max message length and do bad things. - Thanks to John Buckman <john@lyris.com> for reporting this. - -Tue Nov 14 18:43:51 2000 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/test_proactor3.cpp: Replace use of - list<Sender *> with Sender * [] so it builds without inifintely - complicated explicit instantiations on gcc. KISS ;-) - -Tue Nov 14 17:43:26 2000 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/test_aiocb_ace.cpp: Fixed to work - correctly if one of the aio ops completes, but not the other, - during a pass through the Test_Aio::do_aio loop. Thanks to - John Buckman <john@lyris.com> for reporting this. - -Tue Nov 14 14:19:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Log_Record.cpp: - - In ACE_Log_Record::print, changed the call to ACE_OS::fprintf - to use ACE_TEXT("%s") as the second argument instead of just - "%s". This should make logging under ACE_USES_UNICODE readable. - Thanks to James Buck <jim@nowsol.com> and Nick Pratt - <npratt@microstrategy.com> for both reporting this problem and - submitting patches. - -Tue Nov 14 09:19:16 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/Threads/thread_specific.cpp: - Simply use %u to print an ACE_hthread_t. This is what we seem - to be doing in other places in the code. However, this seems to - be non-portable, the code in ACE_Log_Msg to print the thread ID - is a lot more complicated. - -Mon Nov 13 15:13:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Object_Manager.cpp: - - Protected the assert dialog box disable with _MSC_VER - since it is only for MSVC. - -Mon Nov 13 03:46:21 2000 Darrell Brunsch <brunsch@uci.edu> - - * docs/CE-status.txt: - * docs/wchar.txt: (added) - - Updated documentation == good thing. - -Mon Nov 13 03:39:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/config-win32-borland.h: - - It looks like Borland also uses _wcsdup, so added the - ACE_WCSDUP_EQUIVALENT here. - -Mon Nov 13 01:07:36 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS_String.cpp: - * ace/config-win32-borland.h: - * ace/config-win32-msvc.h: - - Christopher Kohlhoff <chris@kohlhoff.com> mentioned to me - that Borland did have a wcsdup, but it was named differently - than the _wcsdup in MSVC. I took his suggestion and added - a ACE_WCSDUP_EQUIVALENT for MSVC's version and enabled - wcsdup for Borland's config. - -Mon Nov 13 00:36:29 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Obstack.cpp: - * ace/Obstack.h: - - Nanbor mentioned to me the other day that the interfaces to - this class probably should remain as char *. After looking - at it, I agreed and changed it to that. But since Svc_Conf - uses this class in such a way that it is passing wchar_t - strings to it when ACE_USES_WCHAR is defined, I also added - a wchar_t version of copy to make things happy. - -Sun Nov 12 17:00:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/ace_ce.dsw: (removed) - * ace/ace_dll_ce.dsp: (removed) - * ace/ace.vcw: (added) - * ace/ace_dll.vcp: (added) - * ace/ace_os_dll.vcp: (added) - - Removed the old Windows CE project files, and replaced - them with ones created by the eMbedded Visual C++ from - MS. - - * ace/OS_Errno.cpp: (added) - * ace/OS_Errno.h: (added) - * ace/OS_Errno.inl: (added) - * ace/config-all.h: - * ace/OS.cpp: - * ace/OS.h: - * ace/OS.i: - - Moved the Errno stuff from ACE_OS to a separate class, - since it needed to be in its own layer for Windows CE. - - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - * ace/Makefile: - * ace/Makefile.am: - * ace/Makefile.bor: - - Added OS_Errno.* files. - - * ace/ACE.cpp: - * ace/Sock_Connect.cpp: - - Moved get_reg_value () to Sock_Connect since that is where - it is being used. - - * ace/ARGV.cpp: - * ace/OS.cpp: - * ace/OS.h: - * ace/OS.i: - - Added ACE_LACKS_ENV for platforms (like Windows CE) which - doesn't have an environment. Maybe in the future this can - be emulated, for now, we do nothing. - - * ace/Base_Thread_Adapter.h: - * ace/OS.h: - - Replace ACE_Export with ACE_OS_Export. - - * ace/Base_Thread_Adapter.inl: - * ace/Thread_Adapter.inl: - - Moved some of the ACE_OS_Thread_Descriptor methods to - Base_Thread_Adapter, since they were declared in - Base_Thread_Adapter.h. - - * ace/SUN_Proactor.cpp: - - Removed ACE_BUILD_DLL from this file, since it should only - be defined in the project file. - - * ace/OS_Dirent.h: - - Added ACE_WINCE to the #if block around some of the includes. - - * ace/OS_Memory.h: - * ace/OS_Dirent.inl: - * ace/OS_String.inl: - * ace/OS.h: - - Added /**/ to the #includes of pace, since they seem to - totally confuse EVC's automatic dependency generation. - - * ace/OS_String.cpp: - * ace/OS_String.h: - * ace/OS_String.inl: - - Added strerror. On machines without it, just returns - "Unknown Error" all the time. - - Added strspn_emulation, strpbrk_emulation, strtol_emulation, - and strtoul_emulation. - - New compile time defines: - - ACE_LACKS_STERROR - - ACE_LACKS_STRPBRK - - ACE_LACKS_STRSPN - - ACE_LACKS_STRTOL - - ACE_LACKS_STRTOUL - - ACE_LACKS_STRTOD - - * ace/Log_Msg.cpp: - - Since there is now strerror in ACE, removed the !ACE_HAS_WINCE - blocks and replaced them with direct calls to ACE_OS_String's - version. - - * ace/config-win32.h: - - Changed check for #include-ing config-wince.h to be based on - _WIN32_WCE because this definition has been automatically added - to the default projects by the misc WinCE editors. - - * ace/Service_Config.i: - - The char versions of the following fuctions (which are defined - only under WinCE) were recursive and would fail miserably if - ever used. I added ACE_CHAR_TO_TCHAR to resolve the infinate - recursion. - - ACE_Service_Config::initialize () - - ACE_Service_Config::resume () - - ACE_Service_Config::suspend () - - ACE_Service_Config::remove () - - * ace/config-win32-msvc.h: - - We were defining ACE_HAS_GNU_CSTRING_H. I doubt that MSVC has - GNU's cstring.h, so I removed it. - - * ace/OS.cpp: - * ace/OS.h: - * ace/OS.i: - - Added ACE_LACKS_CUSERID and ACE_LACKS_CHDIR. - - Added writev and readv emulation methods to ACE_OS, since they - were just being defined as global functions. This is more - consistent with what we are doing elsewhere. - - Added ACE_LACKS_SYS_NERR emulation. - - * ace/config-WinCE.h: - - Added all the necessary defined to get ACE compiling for Windows - CE 3.0. As for previous versions, they do not currently compile, - although that could change in the future - -Sun Nov 12 12:29:31 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Cache_Map_Manager_T.h: - * ace/Codeset_IBM1047.h: - * ace/Date_Time.h: - * ace/Dump_T.h: - * ace/Flag_Manip.h: - * ace/Hash_Cache_Map_Manager_T.h: - * ace/LSOCK_Acceptor.h: - * ace/Memory_Pool.h: - * ace/QoS_Decorator.h: - * ace/SV_Semaphore_Complex.h: - * ace/Strategies_T.h: - * ace/Synch_Options.h: - * ace/Time_Request_Reply.h: - * ace/Timer_Wheel_T.h: - * ace/WIN32_Proactor.h: - * ace/ace_wchar.h: - - Fixed @file comments. - -Sat Nov 11 18:55:04 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * ace/Profile_Timer.cpp: - Included Log_Msg.h to fix compile error. - -Fri Nov 10 19:41:38 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/fuzz.pl: - - Added a check for the @file comments used by doxygen. - They must be the same as the actual filename, or doxygen will - not pick up the comments for that file. - -Fri Nov 10 20:26:01 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Profile_Timer.cpp: Replaced ACE_OS::perror() with - ACE_ERROR macros so that the output can be redirected to the - same place that the ACE logging output is configured to go. - Thanks to Jeff Greif <jgreif@befree.com> for reporting this. - -Fri Nov 10 17:15:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/fuzz.pl: - - Made the output look a little closer to MSVC's. - -Fri Nov 10 17:08:52 2000 Carlos O'Ryan <coryan@uci.edu> - - * etc/tao_strategies.doxygen: - New doxygen configuration file for the TAO_Strategies library. - - * bin/generate_doxygen.pl: - Add tao_strategies.doxygen to the list. - - * etc/ace_man.doxygen: - Generate man pages in the man directory instead of doxygen_man. - -Fri Nov 10 12:58:53 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/generate_doxygen.pl: - * etc/tao_implrepo.doxygen: - - New Implementation Repository documentation. - -Fri Nov 10 12:53:28 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_compile: - Format test output so make_pretty.pl can recognize where each - test starts. - - * etc/ace_rmcast.doxygen: - * etc/ace_ssl.doxygen: - * etc/tao.doxygen: - * etc/tao_dynamicany.doxygen: - * etc/tao_dynamicinterface.doxygen: - * etc/tao_esf.doxygen: - * etc/tao_iormanip.doxygen: - * etc/tao_iortable.doxygen: - * etc/tao_portableserver.doxygen: - * etc/tao_rtevent.doxygen: - Don't define the macros explicitly, use the config-doxygen.h - file. - Must define ACE_RCSID and ACE_UNDEFINED_FUNC properly, so - doxygen does not get confused. - -Fri Nov 10 14:26:41 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/config-sunos5.6.h: The AIO definitions were missing the #if - !defined guards that are present in the solaris 2.7 config - header. Thanks to Alex Hornby <alex@anvil.co.uk> for reporting - this. - -Fri Nov 10 11:44:15 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/RMCast/Send_File/Receiver.cpp: - Fixed errors in Tru64/DUCXX builds. - -Fri Nov 10 10:42:31 2000 Steve Huston <shuston@riverace.com> - - * ace/OS.i (ACE_OS::gethrtime): Removed the ACE_HAS_PENTIUM check - from the #if defined (ACE_WIN32) line surrounding the use of - QueryPerformanceCounter - it works fine on 486 also, per - Bernd Annamaier <bernd.annamaier@de.adtranz.com>. Thanks to - Bernd for testing this out. - - * ace/High_Res_Timer.cpp: Replaced all occurrences of - "defined (ACE_HAS_PENTIUM)" with "defined (ACE_WIN32)" since there - is nothing Pentium-specific any longer in the Win32 sections. - - * THANKS: Added Bernd Annamier to the Hall of Fame. - -Fri Nov 10 12:28:42 2000 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/log_msg/makefile.bor: - * examples/map_manager/makefile.bor: - * examples/makefile.bor: - * performance-tests/makefile.bor: - * performance-tests/Server_Concurrency/makefile.bor: - New Borland makefiles for ACE - -Thu Nov 9 18:39:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/config-sunos5.5.h: - - Added ACE_LACKS_MEMORY_H for SunCC 5, since it seems there is - a bad interaction between memory.h and string.h on that config. - -Thu Nov 9 17:47:37 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/IPC_SAP/SPIPE_SAP/producer_msg.cpp: - Antoher missing Log_Msg.h - - * examples/Threads/thread_specific.cpp: - Tru64 does not like the convertion from ACE_hthread_t to - unsigned long. I'm trying with a reinterpret_cast, but I don't - have much hope for it. - - * examples/Service_Configurator/IPC-tests/server/Makefile: - * examples/Service_Configurator/Misc/Makefile: - * examples/Timer_Queue/Makefile: - * apps/Gateway/Gateway/Makefile: - * apps/Gateway/Peer/Makefile: - Avoid problems where the library is linked twice, once for the - required dependency in parallel builds an another for the LDLIBS - flag. - - * etc/ace.doxygen: - * etc/ace_man.doxygen: - Use a separate file to generate man pages. We don't want to - document .cpp files or include code in man pages, but that looks - great in HTML documents. - - * etc/tao.doxygen: - * etc/tao_dynamicany.doxygen: - * etc/tao_dynamicinterface.doxygen: - * etc/tao_esf.doxygen: - * etc/tao_iormanip.doxygen: - * etc/tao_iortable.doxygen: - * etc/tao_portableserver.doxygen: - * etc/tao_rtevent.doxygen: - Update the replacement path to some of the tag files, the - generated documents did not get the right cross references. - - * bin/make_pretty.pl: - Add yet another linker error. - - * bin/generate_doxygen.pl: - Parse the man page generation code to create its output path too. - -Thu Nov 9 15:10:28 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/make_pretty.pl: - In the HTML we replace < with < and > with > otherwise the - output for templates makes no sense. - -Thu Nov 9 12:11:52 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Makefile (UTILS_FILES): - Modified the Makefile to help when building subsetted shared - libraries. Specifically, moved Addr, INET_Addr, and Sock_Connect - into the Sockets component. That removes the dependancy of the - Utilities and Sockets components on the non-socket IPC component. - - Committing on behalf of Phil Mesnier <mesnier_p@ociweb.com> - -Thu Nov 9 10:38:33 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/Makefile: - ACE_HAS_GNUG_PRE_2_8 should be used only after - wrapper_macros.GNU is included. - -Thu Nov 9 10:26:01 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/make_pretty.pl: - Add yet another link error message. - -Thu Nov 9 08:51:04 2000 Carlos O'Ryan <coryan@uci.edu> - - * include/makeinclude/rules.local.GNU: - It seems like I cannot get the rules to create the .obj and - .shobj directories right. This last iteration seems to be - happier, it works fine with parallel builds, it works fine for - static and shared libraries, it does not recompile everything - just because the .obj or .shobj directories changed, it creates - the Templates.DB file when needed. - The only problem that I'm aware off is that it can potentially - produce a warning on parallel builds: two separate rules can try - to create the .obj directory simultaneously, adding a dependency - to have only one rule do it results in full-recompilations when - a .obj directory is modified for whatever reason. - I'm using ACE_NUL_STDERR to avoid such warnings, and it seems to - work correctly. - -Thu Nov 9 10:14:06 2000 Jeff Parsons <parsons@cs.wustl.edu> - - * include/makeinclude/ace_flags.bor: - - Added the line - - TAO_DYNAMICINTERFACE_LIB = - $(CORE_BINDIR)\TAO_DynamicInterface$(LIB_DECORATOR).lib - - that I had overlooked when doing the BCB makefile changes - related to the DynamicInterface library. Thanks to - Johnny Willemsen <Johnny.Willemsen@meco.nl> for sending - in the patch. - - -Wed Nov 8 16:11:11 2000 Carlos O'Ryan <coryan@uci.edu> - - * tests/RMCast/RMCast_Membership_Test.cpp: - Remove unused variables - - * bin/make_pretty.pl: - Add regular expression to detect more linker errors. - -Wed Nov 8 14:10:20 2000 Carlos O'Ryan <coryan@uci.edu> - - * examples/IPC_SAP/FIFO_SAP/FIFO-Msg-server.cpp: - * examples/IPC_SAP/SPIPE_SAP/client.cpp: - * examples/IPC_SAP/SPIPE_SAP/consumer_msg.cpp: - * examples/IPC_SAP/SPIPE_SAP/consumer_read.cpp: - * examples/IPC_SAP/SPIPE_SAP/server.cpp: - * examples/RMCast/Send_File/Sender.cpp: - * examples/Reactor/Misc/test_event_handler_t.cpp: - * examples/Service_Configurator/IPC-tests/server/Handle_L_SPIPE.h: - We must explicitly #include "ace/Log_Msg.h", unfortunately the - errors were only detected on platforms that do not define - ACE_TEMPLATES_REQUIRE_SOURCE. - - * include/makeinclude/rules.local.GNU: - More fine tuning, the Templates.DB directory must be created by - the Makefile because SunCC 4.2 generates a warning if it is not - present. - -Wed Nov 8 14:40:25 2000 Chad Elliott <elliott_c@ociweb.com> - - * ace/config-all.h - - Fix a build problem on HP-UX 11.00 with aCC A.03.25. The macro - used to determine the use of the std namespace changed from - RWSTD_NO_NAMESPACE to _NAMESPACE_STD. NOTE: I did preserve the - check for RWSTD_NO_NAMESPACE with aCC versions less than A.03.25. - -Wed Nov 8 10:27:00 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/generate_doxygen.pl: - Automates the invocation of doxygen. It modifies the doxygen - templates in $ACE_ROOT/etc and creates the right directory - hierarchy to receive the files. - - * ace/config-doxygen.h: - Configuration file used to generate the doxygen files. - Notice that this is not a valid configuration for any platform, - for example, it enables both the Win32 and the POSIX AIO calls. - - * etc/ace.doxygen: - * etc/ace_rmcast.doxygen: - * etc/ace_ssl.doxygen: - * etc/tao.doxygen: - * etc/tao_dynamicany.doxygen: - * etc/tao_dynamicinterface.doxygen: - * etc/tao_esf.doxygen: - * etc/tao_iormanip.doxygen: - * etc/tao_iortable.doxygen: - * etc/tao_portableserver.doxygen: - * etc/tao_rtevent.doxygen: - Minor mods to make the generate_doxygen.pl job easier. - -Wed Nov 8 09:02:10 2000 Carlos O'Ryan <coryan@uci.edu> - - * apps/Gateway/Gateway/Makefile: - * apps/Gateway/Peer/Makefile: - * examples/Timer_Queue/Makefile: - * examples/Service_Configurator/IPC-tests/server/Makefile: - * examples/Service_Configurator/Misc/Makefile: - When generating a library and a binary that uses it in the same - Makefile we must add explicit dependency between them, otherwise - parallel builds can fail. - - * apps/JAWS/server/HTTP_Helpers.cpp: - * examples/IPC_SAP/FILE_SAP/client.cpp: - * examples/Naming/test_multiple_contexts.cpp: - * examples/Threads/thread_specific.cpp: - * examples/Timer_Queue/Thread_Timer_Queue_Test.cpp: - Fixed warnings about mismatches between printf-like strings and - their arguments. - -Wed Nov 8 02:57:15 2000 Darrell Brunsch <brunsch@uci.edu> - - * Local_Name_Space.cpp: - * Local_Name_Space.h: - - Conversion between ACE_WString and ACE_NS_String depended on - ACE_WString storing two byte char's (because it was using - fast_rep ()) Changed to ushort_rep (), but this required - ACE_NS_String to clean up the string once it is done. Added - a destructor and a flag to ACE_NS_String for this purpose. - -Tue Nov 7 11:47:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/Cached_Accept_Conn_Test.cpp: - * tests/Cached_Conn_Test.cpp: - * tests/Collection_Test.cpp: - - Protected some of the explicit template instantiations with #if's - to make sure we don't get duplicate ones with ACE_INT32 == int - (since ACE_HANDLE and DATA are also int in some configurations). - -Tue Nov 7 09:04:15 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/RMCast/RMCast.h: - Undefine macro MT_DATA if defined already. - -Tue Nov 7 09:00:08 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_compile: - By default build everything under ACE_wrappers and - ACE_wrappers/TAO, we used to build just a subset of ACE. - We need to provide full coverage because this script is - replacing all nightly builds, if auto_compile does not test it, - nothing will. - - * bin/show_log_dir.pl: - The script was removing the second and third most recent logs - from the list, this used to be required to remove the brief and - html files, but now we filter those earlier on the script. - -Mon Nov 6 16:23:56 2000 Carlos O'Ryan <coryan@uci.edu> - - * tests/RMCast/RMCast_Fragment_Test.cpp: - * tests/RMCast/RMCast_Reassembly_Test.cpp: - * tests/RMCast/RMCast_Retransmission_Test.cpp: - Fixed warnings reported by DU/CXX - -Sun Nov 5 17:23:55 2000 Carlos O'Ryan <coryan@uci.edu> - - * etc/ace.doxygen: - * etc/ace_rmcast.doxygen: - * etc/ace_ssl.doxygen: - * etc/tao.doxygen: - * etc/tao_dynamicany.doxygen: - * etc/tao_dynamicinterface.doxygen: - * etc/tao_esf.doxygen: - * etc/tao_iormanip.doxygen: - * etc/tao_iortable.doxygen: - * etc/tao_portableserver.doxygen: - * etc/tao_rtevent.doxygen: - Commit doxygen templates, I'm still writing the script to use - them effectively. - -Sun Nov 5 14:37:13 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_compile: - Make log files world readable, under Solaris they are created - with 600 mode by default. - -Sun Nov 5 14:05:13 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/create_vt_dsps.pl: - * tests/version_tests/Atomic_Op_Test.dsp: - * tests/version_tests/Auto_IncDec_Test.dsp: - * tests/version_tests/Barrier_Test.dsp: - * tests/version_tests/Basic_Types_Test.dsp: - * tests/version_tests/Buffer_Stream_Test.dsp: - * tests/version_tests/CDR_Array_Test.dsp: - * tests/version_tests/CDR_File_Test.dsp: - * tests/version_tests/CDR_Test.dsp: - * tests/version_tests/Cache_Map_Manager_Test.dsp: - * tests/version_tests/Cached_Accept_Conn_Test.dsp: - * tests/version_tests/Cached_Conn_Test.dsp: - * tests/version_tests/Capabilities_Test.dsp: - * tests/version_tests/Collection_Test.dsp: - * tests/version_tests/Conn_Test.dsp: - * tests/version_tests/DLL_Test.dsp: - * tests/version_tests/DLList_Test.dsp: - * tests/version_tests/Dynamic_Priority_Test.dsp: - * tests/version_tests/Enum_Interfaces_Test.dsp: - * tests/version_tests/Env_Value_Test.dsp: - * tests/version_tests/Future_Set_Test.dsp: - * tests/version_tests/Future_Test.dsp: - * tests/version_tests/Handle_Set_Test.dsp: - * tests/version_tests/Hash_Map_Bucket_Iterator_Test.dsp: - * tests/version_tests/Hash_Map_Manager_Test.dsp: - * tests/version_tests/High_Res_Timer_Test.dsp: - * tests/version_tests/IOStream_Test.dsp: - * tests/version_tests/Lazy_Map_Manager_Test.dsp: - * tests/version_tests/MM_Shared_Memory_Test.dsp: - * tests/version_tests/MT_Reactor_Timer_Test.dsp: - * tests/version_tests/MT_SOCK_Test.dsp: - * tests/version_tests/Malloc_Test.dsp: - * tests/version_tests/Map_Manager_Test.dsp: - * tests/version_tests/Map_Test.dsp: - * tests/version_tests/Mem_Map_Test.dsp: - * tests/version_tests/Message_Block_Test.dsp: - * tests/version_tests/Message_Queue_Notifications_Test.dsp: - * tests/version_tests/Message_Queue_Test.dsp: - * tests/version_tests/Naming_Test.dsp: - * tests/version_tests/New_Fail_Test.dsp: - * tests/version_tests/Notify_Performance_Test.dsp: - * tests/version_tests/OS_Test.dsp: - * tests/version_tests/Object_Manager_Test.dsp: - * tests/version_tests/OrdMultiSet_Test.dsp: - * tests/version_tests/Pipe_Test.dsp: - * tests/version_tests/Priority_Buffer_Test.dsp: - * tests/version_tests/Priority_Reactor_Test.dsp: - * tests/version_tests/Priority_Task_Test.dsp: - * tests/version_tests/Process_Manager_Test.dsp: - * tests/version_tests/Process_Mutex_Test.dsp: - * tests/version_tests/Process_Strategy_Test.dsp: - * tests/version_tests/RB_Tree_Test.dsp: - * tests/version_tests/Reactor_Exceptions_Test.dsp: - * tests/version_tests/Reactor_Notify_Test.dsp: - * tests/version_tests/Reactor_Performance_Test.dsp: - * tests/version_tests/Reactor_Timer_Test.dsp: - * tests/version_tests/Reactors_Test.dsp: - * tests/version_tests/Reader_Writer_Test.dsp: - * tests/version_tests/Recursive_Mutex_Test.dsp: - * tests/version_tests/Reverse_Lock_Test.dsp: - * tests/version_tests/SOCK_Connector_Test.dsp: - * tests/version_tests/SOCK_Send_Recv_Test.dsp: - * tests/version_tests/SOCK_Test.dsp: - * tests/version_tests/SPIPE_Test.dsp: - * tests/version_tests/SString_Test.dsp: - * tests/version_tests/SV_Shared_Memory_Test.dsp: - * tests/version_tests/Semaphore_Test.dsp: - * tests/version_tests/Service_Config_Test.dsp: - * tests/version_tests/Sigset_Ops_Test.dsp: - * tests/version_tests/Simple_Message_Block_Test.dsp: - * tests/version_tests/Svc_Handler_Test.dsp: - * tests/version_tests/TSS_Test.dsp: - * tests/version_tests/Task_Test.dsp: - * tests/version_tests/Thread_Manager_Test.dsp: - * tests/version_tests/Thread_Mutex_Test.dsp: - * tests/version_tests/Thread_Pool_Reactor_Test.dsp: - * tests/version_tests/Thread_Pool_Test.dsp: - * tests/version_tests/Time_Service_Test.dsp: - * tests/version_tests/Time_Value_Test.dsp: - * tests/version_tests/Timeprobe_Test.dsp: - * tests/version_tests/Timer_Queue_Test.dsp: - * tests/version_tests/Tokens_Test.dsp: - * tests/version_tests/UPIPE_SAP_Test.dsp: - * tests/version_tests/Upgradable_RW_Test.dsp: - * tests/version_tests/version_tests.dsw: - - Updated to use the new ACE_AS_STATIC_LIBS macro. - - * ace/Sample_History.cpp: - * ace/SSL/SSL_Context.cpp: - * ace/SSL/SSL_SOCK_Stream.i: - - Fixed miscellaneous ACE_USES_UNICODE problems with ACE_LIB_TEXT. - -Sat Nov 4 19:33:39 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_compile: - Fixed the mail message, the path to the CGI script was garbled. - - * include/makeinclude/rules.local.GNU: - My last fixed did not work for static builds, the SOLINK rules - got in the way. It works correctly for both static and shared - builds now. - -Sat Nov 4 18:19:46 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/RMCast/RMCast_IO_UDP.cpp: - * ace/RMCast/RMCast_Reassembly.cpp: - * tests/Reactor_Notify_Test.cpp: - Add missing template instantiations, they only showed up in - builds with disabled inlining. - -Sat Nov 4 17:05:08 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/auto_compile: - New revision of the script. It can copy log files directly to - the destination directory, and it also runs the make_pretty - script to generate both the HTML and HTML summary of the log. - - * bin/show_log_dir.pl: - Corrections to the document and more fixes to make it Apache - friendly. - - * include/makeinclude/rules.local.GNU: - Add special target to force the creation of the .shobj and .obj - directories, even with parallel builds. The faster our - computers get the more often we get failures due to this - problem. - -Sat Nov 4 14:53:39 2000 Carlos O'Ryan <coryan@uci.edu> - - * bin/make_pretty.pl: - This is a modified version of Darrell's script to beautify our - auto_compile logs. This version works better with Unix builds - and is going to be invoked by bin/auto_compile to generate - summary logs. - - * bin/show_log_dir.pl: - A CGI script to present all the logs for a particular builds in - a single page. It is a modified version of Darrell's script, - but this time it works better on Unix web servers. - -Fri Nov 3 17:22:10 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/config-sunos5.5.h: Backed of changes given by Russ - Noseworthy through bug id #710. This was breaking SunCC5.0 - builds. - -Fri Nov 3 09:41:20 2000 Carlos O'Ryan <coryan@uci.edu> - - * tests/RMCast/RMCast_Membership_Test.cpp: - Fixed warning about variables possibly uninitialized before - their first use. In fact the variables are initialized before - their first use, but the code flow is too complex for the - compiler to figure that out. - -Fri Nov 3 09:34:39 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile: - * tests/Makefile: - * examples/Makefile: - Disable the RMCast builds for old versions of g++ - -Thu Nov 2 12:11:43 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Name_Request_Reply.h: - * ace/Timer_Hash_T.h: - * ace/Map_Manager.h: - * ace/Log_Record.h: - Minor adjustments to make doxygen happier. - -Thu Nov 2 11:57:37 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile: - * tests/Makefile: - * examples/Makefile: - Enable the RMCast library, tests and examples by default. - -Wed Nov 1 17:24:59 2000 Pradeep Gore <pradeep@cs.wustl.edu> - - * tests/RMCast/Makefile.bor: - Added Makefile, thanks to Christopher Kohlhoff - <chris@kohlhoff.com> and Johnny.Willemsen <Johnny.Willemsen@meco.nl> - for contributing. - -Wed Nov 1 14:11:48 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/*.h: - Change all the header files to generate better docs with - Doxygen. Thanks to Darrell for his wonderful script to automate - this process. - -Tue Oct 31 19:01:19 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * netsvcs/lib/Client_Logging_Handler.cpp (handle_input): Added - a comment explaining why we don't go to heroic lengths to recv() - all the data if it fails after the second recv(). Thanks to - Steve Sivier <Steven.Sivier@Eng.Sun.COM> for motivating this. - -Tue Oct 31 14:21:11 2000 David L. Levine <levine@cs.wustl.edu> - - * bin/check_build_logs: sort the output file names, so - that guajira's cxx build shows up first (locally). - Limit total number of lines to 2000 to keep below - majordomo's size limit. - -Tue Oct 31 10:54:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Process.i: - - Changed the CREATE_UNICODE_ENVIRONMENT section to be only - in ACE_WIN32 environments. - -Tue Oct 31 11:35:31 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/config-sunos5.5.h: - * ace/Log_Msg.cpp: Added patches provided by Russ. Bugzilla id 710. - -Tue Oct 31 08:14:33 2000 Chris Cleeland <cleeland_c@ociweb.com> - - * wrapper_macros.GNU: - - Set defaults for exceptions, rtti, and fast to all be zero - (0),i.e., off. Henceforth, if you're creating a new port and - you want exceptions to be on, you must explicitly have - "exceptions=1" in your platform_*.GNU file. - - * platform_irix6.x_common.GNU: - - Added an "exceptions=1" directive per the comment above. - -Mon Oct 30 18:56:33 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS.cpp: - * ace/OS.h: - * ace/OS.i: - - Changed several of the socket calls to use char instead - of ACE_TCHAR, seeing that no OS supports anything but char. - But I am keeping the ACE_TCHAR or dual char/wchar_t - interfaces on classes higher up the stream. - - Methods changed are: - - - gethostbyaddr - - gethostbyname - - gethostbyname2 - - gethostbyaddr_r - - gethostbyname_r - - getprotobyname - - getprotobyname_r - - getservbyname - - getservbyname_r - - inet_addr - - inet_ntoa - - inet_aton - - inet_ntop - - inet_pton - - This should take care of a couple of things. First we don't - have to worry about returning a wchar_t * static buffer - instead of a char * static buffer from something like - inet_ntoa and this should remove some slowdowns reported from - gethostbyname. Thanks to Nick Pratt <npratt@microstrategy.com> - for reporting both of these. - - * ace/INET_Addr.cpp: - * ace/INET_Addr.h: - * ace/INET_Addr.i: - - Fixed references to interface changes and added dual - char/wchar_t interfaces to replace some ACE_TCHAR ones. - - * ace/MEM_Addr.cpp: - * ace/MEM_Addr.h: - * ace/MEM_Addr.i: - - Changed get_host_name and get_host_name to return char * - - * ace/SOCK_Dgram_Mcast.cpp: - * ace/SOCK_Dgram_Mcast_QoS.cpp: - - Fixed all the references to the interface changes. - - * ace/Object_Manager.cpp: - - Added ERROR to the output for AV's and Unhandled Exceptions. - -Mon Oct 30 18:51:40 2000 Darrell Brunsch <brunsch@uci.edu> - - * examples/IPC_SAP/SSL_SAP/SSL-client.dsp: - * examples/IPC_SAP/SSL_SAP/SSL-server.dsp: - - Removed hardcoded OpenSSL include directories. - - * ace/ssl/SSL_Context.cpp: - * ace/ssl/SSL_Context.h: - * ace/ssl/SSL_SOCK.cpp: - * ace/ssl/SSL_SOCK_Acceptor.cpp: - * ace/ssl/SSL_SOCK_Acceptor.h: - * ace/ssl/SSL_SOCK_Connector.cpp: - * ace/ssl/SSL_SOCK_Connector.h: - * ace/ssl/SSL_SOCK_Connector.i: - * ace/ssl/SSL_SOCK_Stream.cpp: - * ace/ssl/SSL_SOCK_Stream.h: - * ace/ssl/SSL_SOCK_Stream.i: - - Removed the ACE_HAS_SSL #ifdefs, since ACE_SSL is a - separate library, so it isn't really needed. - -Mon Oct 30 16:34:33 2000 Chris Cleeland <cleeland_c@ociweb.com> - - * include/makeinclude/platform_vxworks5.x_g++.GNU: - - More exceptions/fast/rtti canonicalization fixes. Thanks again, - Carlos. - -Mon Oct 30 11:56:13 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile.am: - * ace/Makefile.bor: - Add missing files to the Borland and Automake makefiles. - -Mon Oct 30 11:38:07 2000 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/Basic_Stats.cpp: - * ace/Basic_Stats.inl: - * ace/Sample_History.cpp: - * ace/Stats.cpp: - Fixed problems on platforms that lack 64-bit integers. - -Mon Oct 30 11:10:26 2000 Chris Cleeland <cleeland_c@ociweb.com> - - * include/makeinclude/platform_linux.GNU: - * include/makeinclude/platform_chorus_ghs.GNU: - - The change to canonicalize the values used for - exceptions/fast/rtti caused the default for "exceptions" on these - two platforms to change from exceptions off to exceptions on. - There may be others, but without access to the compiler itself - it's hard to tell what the default behavior is for the compiler - and thus set the value for exceptions/rtti/fast accordingly. - - Thanks to Carlos O'Ryan <coryan@uci.edu> for pointing this out. - -Mon Oct 30 00:57:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/OS_Test.cpp: - - Was a bit overzealous on a previous [] to * change, - so switched one back to [] to remove a segfault - with the strncmp section. - -Sun Oct 29 23:46:54 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/ace_wchar.h: - - Removed the definitions for LPTSTR and LPCTSTR. They - will need to be defined outside of ace or in config.h - to be made available on other platforms. Here is a - sample block which can be added to a non-Win32 config.h: - - typedef char *LPSTR; - typedef const char *LPCSTR; - typedef wchar_t *LPWSTR; - typedef const wchar_t *LPCWSTR; - - #if defined (UNICODE) - typedef const wchar_t *LPCTSTR; - typedef wchar_t *LPTSTR; - typedef wchar_t TCHAR; - #define __TEXT(literal) L##literal - #else - typedef const char *LPCTSTR; - typedef char *LPTSTR; - typedef char TCHAR; - #define __TEXT(literal) literal - #endif - - #define TEXT(literal) __TEXT(literal) - - * ace/config-vxworks5.x.h: - - Added ACE_LACKS_WCSDUP. - -Sun Oct 29 01:59:11 2000 Darrell Brunsch <brunsch@uci.edu> - - * bin/doxygen-convert-h.pl: (added) - - New script to convert old header files to new doxygen - style commented files. Works well, but some things - do throw it for a loop, be sure to check output before - checking in. - -Sat Oct 28 15:02:40 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Stats.h: - Cosmetic fixes to make doxygen happier. - -Fri Oct 27 17:21:56 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/OS_Test.cpp: - - Darrell learned how to use strtok_r properly. 'nuff said. - -Fri Oct 27 17:04:02 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/OS_Test.cpp: - - Changed the arrays ([]) to array pointers (*) since the - pointer arithmatic I do doesn't seem to work with the - arrays. - -Fri Oct 27 15:45:57 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - Add the new files to the MSVC projects. - -Fri Oct 27 15:02:56 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile: - * ace/Makefile.am: - * ace/Makefile.bor: - * ace/Sample_History.h: - * ace/Sample_History.inl: - * ace/Sample_History.cpp: - Add new class to keep a full history of all the samples in a - performance test. - - * ace/Basic_Stats.h: - * ace/Basic_Stats.inl: - * ace/Basic_Stats.cpp: - * ace/Stats.h: - * ace/Stats.cpp: - Refactor code to compute minimum, maximum, average and jitter - into its own class. It used to be part of ACE_Througthput_Stats - - * performance-tests/TCP/Makefile: - * performance-tests/TCP/tcp_test.cpp: - Use the new ACE_Sample_History class. - -Fri Oct 27 12:04:17 2000 Chris Cleeland <cleeland_c@ociweb.com> - - * platform_chorus.GNU: - * platform_chorus4.x_g++.GNU: - * platform_chorus_ghs.GNU: - * platform_cray.GNU: - * platform_hpux_kcc.GNU: - * platform_linux_cxx.GNU: - * platform_osf1_4.0.GNU: - * platform_sunos5_ghs.GNU: - * platform_sunos5_kcc.GNU: - * platform_sunos5_sunc++.GNU: - * platform_vxworks5.x_diab.GNU: - * platform_vxworks5.x_g++.GNU: - * wrapper_macros.GNU: - - Changed the way that 'rtti', 'exceptions', and 'fast' makeflags - get processed in wrapper_macros.GNU. There was a discrepancy - between the "command line" value and the "platform_macros file" - value each needed in order to turn the feature off, e.g., zero - on the command line and "null" in the file. After - investigation, the discrepancy seemed to be without reason, so - I've canonicalized them. You can now put "exceptions=0" on - either the command line OR in the platform_macros.GNU file. - -Fri Oct 27 10:51:26 2000 Steve Huston <shuston@riverace.com> - - * tests/Reactor_Notify_Test.cpp: Moved definition of 'int status' - outside the auto_ptr-scoping block. - -Fri Oct 27 02:06:29 2000 Luther J Baker <luther@cs.wustl.edu> - - * ace/OS.cpp: - * ace/OS.i: - - Fixed some warnings due to int ace_result_ declared for NT. - -Thu Oct 26 22:18:42 2000 Alex Arulanthu <Alex.Arulanthu@sylantro.com> - - * ace/Log_Msg.h: - * ace/Log_Msg.cpp: - * ace/Service_Config.h: - * ace/Service_Config.cpp: - * ace/Service_Config.i: - - Setting the right default values for process and thread level - priority masks. I had already talk to Steve H about this - change. I just forgot to get this in. - -Thu Oct 26 14:04:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS.h: - - Reverted Doug's change which defined ACE_MAX_USERID to - L_cuserid on all platforms. It ended up that this - is only valid on Windows when _POSIX_ is defined. - -Thu Oct 26 11:20:43 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/RMCast/RMCast_IO_UDP.cpp: - * ace/RMCast/RMCast_Reassembly.cpp: - * tests/CDR_Array_Test.cpp: - * 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: - - The usual additions of ACE_TEXT to make ACE_USES_WCHAR builds - happy. - -Thu Oct 26 12:19:45 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Memory_Pool.cpp (unmap): Fixed a bug that was preventing - the handle from being closed. Thanks to Ivan Murphy and Dieter - Quehl for reporting this. - -Thu Oct 26 10:56:36 2000 Steve Huston <shuston@riverace.com> - - * ace/OS.i (ACE_OS::rename): Added && (ACE_HAS_WINNT4 == 1) to the - defined (ACE_HAS_WINNT4) to avoid this code section on Win95/98. - -Thu Oct 26 07:39:22 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - ace/OS.h: Added a new macro called ACE_MAX_USERID that is 32 on - Win32 and L_cuserid on other platforms. - - * examples/Service_Configurator/IPC-tests/client/local_dgram_client_test.cpp, - examples/Service_Configurator/IPC-tests/server/Handle_L_Stream.cpp, - examples/Service_Configurator/IPC-tests/server/Handle_Thr_Stream.cpp, - examples/Service_Configurator/IPC-tests/server/Handle_R_Stream: - Fixed the calls to cuserid() so they don't pass in 0, but - instead pass in a buffer of size ACE_MAX_USERID. This fixes a - problem that occurs on WinNT. Thanks to Mike Curtis - <mccurry@my-deja.com> for pointing this out. - - * tests/Reactor_Notify_Test.cpp: Fixed an unused args warning - and also added the use of auto_ptr to ensure that dynamic memory - is deleted no matter what return path is taken. - -Thu Oct 26 04:00:00 2000 Luther J Baker <luther@cs.wustl.edu> - - * ace/OS.h: - * ace/OS.i: - * ace/OS.cpp: - * ace/OS_Dirent.inl: - Fixing more ACE_HAS_PACE over NT. Most fixes just skip the - PACE call. NT doesn't implement much of posix and so - many ACE calls will skip pace until PACE/WINNT emulates - more posix functionality. - -Wed Oct 25 20:38:33 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/SSL/ACE_SSL.dsp: - - Removed the include/lib directories put here. Instead - we rely on the fact that OpenSSL should be in - DevStudio's global include and library directories. - -Wed Oct 25 13:20:24 2000 Irfan Pyarali <irfan@cs.wustl.edu> - - * tests/Purgable_Map_Manager_Test: Removed this deprecated test. - The following files were effected: - - - Makefile.am - - icc.bat - - t.icc - - These files were removed: - - - Purgable_Map_Manager_Test.cpp - - Purgable_Map_Manager_Test.icc - -Wed Oct 25 11:24:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/adapter/README: (removed) - * ace/adapter/ace/ACE.h: (removed) - * ace/adapter/ace/ARGV.h: (removed) - * ace/adapter/ace/ATM_Addr.h: (removed) - * ace/adapter/ace/Arg_Shifter.h: (removed) - * ace/adapter/ace/Capabilities.h: (removed) - * ace/adapter/ace/DEV_Addr.h: (removed) - * ace/adapter/ace/DLL.h: (removed) - * ace/adapter/ace/Dirent.h: (removed) - * ace/adapter/ace/Dynamic_Service.h: (removed) - * ace/adapter/ace/FIFO.h: (removed) - * ace/adapter/ace/FIFO_Recv.h: (removed) - * ace/adapter/ace/FIFO_Recv_Msg.h: (removed) - * ace/adapter/ace/FIFO_Send.h: (removed) - * ace/adapter/ace/FIFO_Send_Msg.h: (removed) - * ace/adapter/ace/FILE_Addr.h: (removed) - * ace/adapter/ace/Functor.h: (removed) - * ace/adapter/ace/High_Res_Timer.h: (removed) - * ace/adapter/ace/INET_Addr.h: (removed) - * ace/adapter/ace/MEM_Addr.h: (removed) - * ace/adapter/ace/Malloc_T.h: (removed) - * ace/adapter/ace/Memory_Pool.h: (removed) - * ace/adapter/ace/OS.h: (removed) - * ace/adapter/ace/OS_Dirent.h: (removed) - * ace/adapter/ace/OS_String.h: (removed) - * ace/adapter/ace/SOCK_Dgram_Bcast.h: (removed) - * ace/adapter/ace/SOCK_Dgram_Mcast.h: (removed) - * ace/adapter/ace/SOCK_Dgram_Mcast_QoS.h: (removed) - * ace/adapter/ace/SPIPE_Addr.h: (removed) - * ace/adapter/ace/Service_Config.h: (removed) - * ace/adapter/ace/Service_Object.h: (removed) - * ace/adapter/ace/Stats.h: (removed) - * ace/adapter/ace/System_Time.h: (removed) - * ace/adapter/ace/Task_T.h: (removed) - * ace/adapter/ace/Trace.h: (removed) - - The adapter classes were removed since they were - not ready for prime time and their main purpose was - invalidated. If need be, they can be brought back in - the future. - -Wed Oct 25 12:26:25 2000 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/RMCast/RMCast_UDP_Reliable_Sender.cpp: - Yet another missing #include - -Wed Oct 25 09:50:48 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/RMCast/RMCast_IO_UDP.h: - * ace/RMCast/RMCast_IO_UDP.cpp: - * ace/RMCast/RMCast_UDP_Reliable_Receiver.h: - * ace/RMCast/RMCast_UDP_Reliable_Receiver.i: - * ace/RMCast/RMCast_UDP_Reliable_Receiver.cpp: - * ace/RMCast/RMCast_UDP_Reliable_Sender.h: - * ace/RMCast/RMCast_UDP_Reliable_Sender.i: - * examples/RMCast/Send_File/Makefile: - * examples/RMCast/Send_File/Sender.cpp: - * examples/RMCast/Send_File/Receiver.cpp: - Change several open() methods to init() because the base class - already has a virtual open() method, but with a different - signature, some compilers generate warnings when this happens. - Add #include required in platforms that do not define - ACE_TEMPLATES_REQUIRE_SOURCE - -Wed Oct 25 09:32:10 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Object_Manager.cpp: - - Changed to only use the CRT_ERROR disabling under - ACE_DISABLE_WIN32_ERROR_WINDOWS when _DEBUG is defined, - since crtdbg.h is only included by ACE when _DEBUG is - defined. - -Wed Oct 25 10:51:26 2000 Steve Huston <shuston@riverace.com> - - * ace/Select_Reactor_Base.cpp (purge_pending_notifications): Added - ACE_UNUSED_ARG (eh) for the non-ACE_HAS_REACTOR_NOTIFICATION_QUEUE - case to quiet g++'s unused argument warning. - -Tue Oct 24 23:14:55 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS_String.cpp: - * ace/config-win32-borland.h: - - Added ACE_LACKS_WCSDUP to do something similar to - ACE_HAS_STRDUP_EMULATION. Added this to Borland's config, - - Also fixed several warnings in OS_String's emulation - functions that show up with g++. - - * ace/Object_Manager.cpp: - - Added code in the Object Manager to disable some of the - error windows that pops up in Win32 apps. If - ACE_DISABLE_WIN32_ERROR_WINDOWS is defined, then - ACE_ASSERT's and Access Violations will not pop up - error box's anymore. This was mainly added for auto - build purposes. - - * ace/Select_Reactor_Base.cpp: - - Changed ACE_TEXT's to ACE_LIB_TEXT. - - * ace/WFMO_Reactor.cpp: - - In ACE_WFMO_Reactor_Notify::purge_pending_notifications - there was an instance of this type of code: - - for (size_t index = 0; index < ....) - ... - for (index = 0; index ....) - - On Borland, the index is scoped within the for loop, so - the second use of it produces an error. I moved the - declaration outside of the loop to fix this. - -Tue Oct 24 23:40:33 2000 Irfan Pyarali <irfan@cs.wustl.edu> - - * tests/SString_Test.cpp (main): Added tests for string - assignments. - -Tue Oct 24 19:30:16 2000 Steve Huston <shuston@riverace.com> - - * ace/Select_Reactor_Base.cpp (purge_pending_notifications): Correct - erroneous use of 'index' as a variable when it's also a function. - -Tue Oct 24 13:54:49 2000 Ossama Othman <ossama@uci.edu> - - * ace/Makefile (TEMPLATE_FILES): - - Added missing Test_and_Set source file to the list. - - * ace/Makefile.am (libACE_Utils_la_SOURCES, HEADER_FILES, - INLINE_FILES, TEMPLATE_FILES): - - Updated these lists of sources with the newly added sources. - -Tue Oct 24 12:30:47 2000 Steve Huston <shuston@riverace.com> - - * ace/Reactor.h (ACE_Reactor): - * ace/Reactor_Impl.h (ACE_Reactor_Notify and ACE_Reactor_Impl): - * ace/Select_Reactor_Base.h (ACE_Select_Reactor_Notify and - ACE_Select_Reactor_Impl): - * ace/WFMO_Reactor.h (ACE_WFMO_Reactor_Notify and ACE_WFMO_Reactor): - Added new function, purge_pending_notifications (ACE_Event_Handler *) - - * ace/Reactor.i (ACE_Reactor): - * ace/Select_Reactor_Base.i (ACE_Select_Reactor_Impl): - Added implementation of purge_pending_notifications that forwards - the call on to the proper implementation/handler object. - - * ace/Select_Reactor_Base.cpp (ACE_Select_Reactor_Notify): Added - implementation of purge_pending_notifications. Works when - ACE_HAS_REACTOR_NOTIFICATION_QUEUE is defined. Else it does - ACE_NOTSUP_RETURN. - - * ace/WFMO_Reactor.cpp (ACE_WFMO_Reactor, ACE_WFMO_Reactor_Notify): - Added implementation of purge_pending_notifications. - - Thanks to Edan Ayal <edana@bandwiz.com> for supplying the code for - this new feature. - - * ace/Reactor.i (notify): Set the event handler's reactor pointer if - it doesn't already have one. Enables the event handler to purge - its notifications upon destruction. - - * ace/Event_Handler.cpp (~ACE_Event_Handler): Call reactor's - purge_pending_notifications function to avoid delivering notifies - to a deleted handler. - - * tests/Reactor_Notify_Test.cpp: Added the run_notify_purge_test - function to exercise the above new purge_pending_notifications. - -Tue Oct 24 12:05:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS_String.inl: - - Removed an extra } after an #endif that was producing - warnings on OSF. - -Tue Oct 24 09:27:04 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Test_and_Set.cpp: - Removed the repeated inclusion of the header file. - Thanks to Darrell. - -Tue Oct 24 03:24:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Basic_Types.h: - * ace/config-win32-common.h: - - Moved the code that defines ACE_SIZEOF_WCHAR to the config file, - so it is easier to add it to other platforms. - - * ace/config-linux-common.h: - - Added ACE_SIZEOF_WCHAR, ACE_LACKS_ITOW, ACE_LACKS_WCSICMP, and - ACE_LACKS_WCSNICMP definitions. - - * ace/Synch.h: - - Moved the ACE_LEGACY_MODE includes further down in the file. It - was producing errors the way it was, since Test_and_Set needed - the Guard classes defined before it was included. - -Mon Oct 23 15:31:52 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Malloc_Allocator.cpp: - Yet another calloc() method that was declared but not defined. - -Mon Oct 23 11:37:06 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Init_ACE.cpp: - Removed the semicolon at the end of ACE_RCSID. - -Mon Oct 23 11:20:53 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Malloc_T.i: - Add missing function, now things actually link. - -Mon Oct 23 10:02:36 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Malloc_T.h: - Add missing prototype, now the code actually compiles. - -Mon Oct 23 10:59:21 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/OS_String.cpp: Fixed a compile error in gcc 2.95.2. - - * ace/Cached_Connect_Strategy_T.h - * ace/Cached_Connect_Strategy_T.cpp: Added fixes provided by Edan - Ayal <edana@bandwiz.com>. - -Mon Oct 23 08:37:49 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Malloc_T.{h,cpp} - * ace/Malloc_Base.h, - * ace/Malloc_Allocator.{h,cpp}: Added a new calloc() method that - has the same (default) signature as the one in the C/C++ - standard library. Thanks to Shourya Sarcar - <Shourya.Sarcar@geind.ge.com> for reporting this. - - * examples/ASX/Event_Server/Event_Server/Peer_Router.cpp (handle_input): - Fixed a mistake where we were calling release() via a NULL - pointer! Thanks to Pedro Brandao <pbrandao@inescn.pt> for - reporting this. - -Sun Oct 22 18:01:34 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS_String.cpp: - - strtok_r_emulation wasn't protected by #if !defined - (ACE_HAS_REENTRANT_FUNCTIONS). - -Sun Oct 22 17:30:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/ACE.i: - - wcslen was used directly here, changed it to - ACE_OS_String::strlen. - - * ace/ace_wchar.inl: - - Moved around some of the inline functions to - remove warnings. - - * ace/config-vxworks5.x.h: - - Added all the ACE_LACKE_WCS* that VxWorks needs. - -Sun Oct 22 17:22:32 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/OS_Test.cpp: - - Borland complained about memchr (NULL, ...), since - it couldn't figure out whether to use the void * or - const void* verison of ACE_OS_String::memchr. - Added an explicit cast to help out. - -Sun Oct 22 17:09:26 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/MT_SOCK_Test.cpp: - - Missed an ACE_ALPHABET in a previous checkin. - - And based on this part of the description in the file: - - "Note that most of the connections will fail since - we're overrunning the size of the listen queue for - the acceptor-mode socket." - - I changed the connection_failed ACE_ERROR_RETURN to an - ACE_DEBUG and return. Now it will not show up as an error - via run_test.pl. - -Sun Oct 22 16:32:24 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/OS_String.cpp: - * ace/OS_String.h: - * ace/OS_String.inl: - * ace/config-linux-common.h: - - Renamed ACE_HAS_SAFE_WCSTOK to ACE_LACKS_WCSTOK to be - consistent with the rest of the ACE_LACKS_WCS* macros. - - Fixed a signature for wcsncmp_emulation, as it was taking - in const wchar_t as the first arg instead of const - wchar_t *. - - Added wcscspn_emulation and an implementation for - wcsncmp_emulation. - -Sun Oct 22 15:35:38 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Makefile.bor: - - Changed ACE_Sock_Connect.obj to Sock_Connect.obj, since - it appears that it has been renamed. - -Sun Oct 22 15:07:42 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/ace_dll.dsp: - - While playing with doxygen, I found some old .h files listed - in the project but no longer existed. Removed these. - - * ace/OS_String.cpp: - * ace/OS_String.h: - * ace/OS_String.inl: - - Added quite a few emulation macros for wcs* functions that - often do not exist on platforms. Since these wcs* functions - just play with wchar_t's, they are not difficult to put in. - Some, such as to_upper and strtod/l/ul are more difficult, - so they are not present on platforms without support for - them. - - Also updated the docs to doxygen style. Someday I'll get - around to writing a script to do this... - - * ace/config-win32-borland.h: - * ace/config-win32-msvc.h: - - Added a couple of macros: - - ACE_HAS_STRCASECMP_EQUIVALENT - - ACE_HAS_STRNCASECMP_EQUIVALENT - - ACE_ITOA_EQUIVALENT - - These are used to specify different names for functions - that differ in names across different compilers. - - * tests/OS_Test.cpp: - - Added tests to test OS_String methods so I can make sure - the emulations work. - - * tests/Buffer_Stream_Test.cpp: - * tests/Conn_Test.cpp: - * tests/MM_Shared_Memory_Test.cpp: - * tests/Mem_Map_Test.cpp: - * tests/Priority_Buffer_Test.cpp: - * tests/Priority_Reactor_Test.cpp: - * tests/Reactor_Performance_Test.cpp: - * tests/SOCK_Test.cpp: - * tests/SPIPE_Test.cpp: - * tests/test_config.h: - - In order to get the test_config.h to work when a project - has more than one .cpp file that includes it (and some - compilers give warnings about a static char[]) I moved - ACE_ALPHABET into the tests that need it. Also I changed - it to a static const char[] in places that allowed it. - - So it could be possible to move it back to test_config.h - now that it can be static const char[], but I'm not 100% - sure that will keep all compilers from complaining. - - * tests/run_test.lst: - - Added OS_Test to the list of tests run. - -Sun Oct 22 17:11:16 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * ace/Malloc_T.h: Added a default value of sizeof (T) to the - malloc() method of class ACE_Cached_Allocator. Thanks to - Shourya Sarcar <Shourya.Sarcar@geind.ge.com> for reporting this. - -Sun Oct 22 10:08:31 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile.bor: - Update file names for the ACE_ components. - -Sun Oct 22 08:57:33 2000 David L. Levine <levine@cs.wustl.edu> - - * tests/test_config.h: implemented these changes to remove - g++ warnings introduced by: - - Thu Oct 19 13:17:41 2000 Darrell Brunsch <brunsch@uci.edu> - - 1) Moved ace_file_stream typedef to end of file. - 2) Moved inline definition of ACE_Test_Output::output_file () - to before first use. - 3) Removed static qualifier from ACE_ALPHABET. - -Sat Oct 21 12:52:58 2000 Carlos O'Ryan <coryan@uci.edu> - - * Makefile*: - Update dependencies. - -Fri Oct 20 20:46:37 2000 Steve Huston <shuston@riverace.com> - - * ace/OS.i (chdir): Added !defined (AIX) to the !defined (ACE_WIN32) - && defined (<Visual Age C++>) to get it to build clean on AIX. - - * ace/ace-dll.icc: Added Base_Thread_Adapter, Thread_Adapter, - Flag_Manip, Handle_Ops, Init_ACE, Lib_Find, Sock_Connect, - OS_Log_Msg_Attributes, Malloc_Allocator, PI_Malloc, - Thread_Exit, Log_Msg_IPC, OS_Thread_Adapter, Log_Msg_Backend, - Thread_Control (h and cpp). - - * ace/Handle_Ops.cpp: - * ace/PI_Malloc.cpp: Corrected ACE_RCSID args. - -Fri Oct 20 16:38:55 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - Update MSVC projects - -Fri Oct 20 16:23:40 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/ACE.h: - Modified to include the renamed files. - -Fri Oct 20 14:13:43 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Sock_Connect.i: - * ace/Sock_Connect.h: - * ace/Sock_Connect.cpp: - - * ace/Lib_Find.i: - * ace/Lib_Find.h: - * ace/Lib_Find.cpp: - - * ace/Init_ACE.i: - * ace/Init_ACE.h: - * ace/Init_ACE.cpp: - - * ace/Handle_Ops.i: - * ace/Handle_Ops.h: - * ace/Handle_Ops.cpp: - - * ace/Flag_Manip.i: - * ace/Flag_Manip.h: - * ace/Flag_Manip.cpp: - Renamed the recently added ACE_* files as above to follow - the naming convention. - - * ace/IPC_SAP.h: - * ace/IO_SAP.h: - * ace/FILE_Connector.cpp: - * ace/FILE_Addr.h: - Changed the include files to reflect the above change. - - * ace/Makefile: - Added the new files and updated dependencies. - -Fri Oct 20 11:40:31 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Makefile: - Updated dependencies. - -Fri Oct 20 02:52:17 2000 Marina Spivak <marina@cs.wustl.edu> - - * bin/auto_run_tests.lst: - - Added RTCORBA Server_Declared test. - -Thu Oct 19 18:02:54 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/OS_Log_Msg_Attributes.h : - Modified u_long to 'unsigned long' as the compiler was - complaining. - -Thu Oct 19 13:52:49 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/Malloc_Allocator.cpp: - * ace/Naming_Context.cpp: - * ace/PI_Malloc.cpp: - - Needed to change some ACE_TEXT's to ACE_LIB_TEXT. - - * ace/OS.h: - * ace/OS_Log_Msg_Attributes.h: - - Moved the definition of ACE_OSTREAM_TYPE from OS.h to - OS_Log_Msg_Attributes.h, since some include paths in - ACE_LEGACY_MODE needed it there. - -Thu Oct 19 13:17:41 2000 Darrell Brunsch <brunsch@uci.edu> - - * tests/run_test.pl: - - Changed the method used to detect errors in a log file. Since - the log contains the log message type (such as LM_ERROR), - I changed the detection to look for this. - - * tests/ACE_Init_Test.cpp: - - Added ACE_START_TEST and ACE_END_TEST so the run_test.pl - script won't think ACE_Init_Test failed. - - * tests/test_config.h: - - With the way ACE_ALPHABET and some functions were defined, - they would cause duplicate symbols if test_config.h was - included in more than one object file. Changed them so - the symbols are static to the file. - - * tests/Reactors_Test.cpp: - - When the reactor was shutting down, it was returning a - information message with ACE_ERROR_RETURN. Since it really - wasn't an error condition, changed this to a ACE_DEBUG - and return 0. - - * tests/SOCK_Send_Recv_Test.cpp: - - The test was checking errno without checking a return value. - As a result, it was getting an out-of-date errno and - incorrectly interpreted that as an error. Fixed. - -Thu Oct 19 12:12:00 2000 Darrell Brunsch <brunsch@uci.edu> - - * ace/ACE_Lib_Find.h: - - Added missing static specifier for the wchar version of - strrepl. - -Thu Oct 19 12:03:41 2000 Darrell Brunsch <brunsch@uci.edu> - - * apps/JAWS2/HTTPU/HTTPU.DSP: - * apps/JAWS2/JAWS/jaws.dsp: - - These two projects were placing their DLL's in the wrong - directory. - -Thu Oct 19 14:23:52 2000 Steve Huston <shuston@riverace.com> - - * ace/ace-dll.icc: Removed extraneous ',' from Win32 section. - -Thu Oct 19 12:08:45 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/Cached_Connect_Strategy_T.cpp: Fixed compile errors with old - g++. - -Thu Oct 19 09:56:44 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Basic_Types.h: - Must #include stdio.h for platforms that lack 'long long' - -Thu Oct 19 09:50:30 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Log_Msg.cpp: - Fixed problems during TSS cleanup, the ipc_backend_ object was - used after being destroyed. Apparently this only happens on some - platforms, but fortunately David's build detected the bug. - -Thu Oct 19 09:07:47 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/RMCast/RMCast_Copy_On_Write.h: - Fixed problem in 'friend' declaration, most compilers let it go - through, but not the newer gcc snapshots. Thanks to Craig - Rodrigues <crodrigu@bbn.com> for providing this patch. - -Wed Oct 18 15:44:14 2000 Luther J Baker <luther@cs.wustl.edu> - - * OS.h: - * OS.i: - * OS.cpp: - - Modifications to sem_open debugging ACE_HAS_PACE over NT. - -Wed Oct 18 09:45:12 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * ace/Cached_Connect_Strategy_T.h : - * ace/Cached_Connect_Strategy_T.cpp: Added some fixes provided by - Edan Ayal <edana@bandwiz.com>. - -Tue Oct 17 23:35:33 2000 David L. Levine <levine@cs.wustl.edu> - - * Don't use ACE_TSS_EMULATION by default with LynxOS 3.1.0 - because it's not necessary, according to - Jessie Ragsdale <jessie.ragsdale@latuslw.com>, and breaks - split=1. Jessie was able to create upo to 118 natvie TSS keys, - so the 3-key limit of LynxOS 3.0.0 has been removed in 3.1.0. - [Bug 701] - -Tue Oct 17 00:21:41 2000 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - - * examples/Service_Configurator/IPC-tests/server/Handle_Timeout.h, - examples/Service_Configurator/IPC-tests/server/Handle_Timeout.i, - examples/Service_Configurator/IPC-tests/server/Handle_R_Dgram.h, - examples/Service_Configurator/IPC-tests/server/Handle_R_Dgram.i, - examples/Service_Configurator/IPC-tests/server/Handle_Broadcast.h, - examples/Service_Configurator/IPC-tests/server/Handle_Broadcast.i, - examples/Service_Configurator/IPC-tests/server/Handle_L_FIFO.h, - examples/Service_Configurator/IPC-tests/server/Handle_L_FIFO.i, - examples/Service_Configurator/IPC-tests/server/Handle_R_Stream.h, - examples/Service_Configurator/IPC-tests/server/Handle_R_Stream.i: - Changed *::get_handle(), *::handle_input(), and - *::handle_close() to use ACE_HANDLE and ACE_INVALID_HANDLE - instead of int and -1. Thanks to Mike Curtis - <mccurry@my-deja.com> for pointing this out. - -Mon Oct 16 14:05:13 2000 David L. Levine <levine@cs.wustl.edu> - - * ACE-INSTALL.html: removed spurious 3D characters from - VxWorks section. - -Mon Oct 16 11:17:08 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile.am: - Fixed misnamed file. - -Mon Oct 16 11:07:34 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Log_Msg.cpp: - Fixed memory leak, or rather, make sure that the default Log_Msg - IPC backend is destroyed when the Log_Msg class is closed. - - * ace/Object_Manager.cpp: - Fixed typo in #endif, the final comment was not really a - comment. - -Sun Oct 15 19:14:13 2000 Marina Spivak <marina@cs.wustl.edu> - - * bin/auto_run_tests.lst: - Excluded tests/RTCORBA/Client_Propagated from running in - Exceptions build since it's a linux build. - -Sat Oct 14 15:04:16 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.h: - * ace/Naming_Context.h: - Static services must be exported from the library, so they - require the right export directives. - Added a new ACE_STATIC_SVC_DECLARE_EXPORT() macro for that - purpose and use it for the Naming_Context. - -Sat Oct 14 09:55:43 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile.bor: - Fixed typo in file name. - -Sat Oct 14 03:32:54 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/ACE_Flag_Manip.h: - * ace/Service_Templates.h: - Added a missing header file. - -Fri Oct 13 22:40:27 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - Add new files to the project. - - * ace/Log_Msg.cpp: - Fixed typo in management of structured exception handling - callbacks. - -Fri Oct 13 21:49:54 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Makefile: - * ace/Makefile.am: - * ace/Makefile.bor: - Add the new Log_Msg files to the Makefiles. - - * ace/ACE_Lib_Find.cpp: - Add ACE_RCSID macro - - * ace/config-all.h: - * ace/OS_Log_Msg_Attributes.h: - * ace/OS_Log_Msg_Attributes.inl: - * ace/OS_Log_Msg_Attributes.cpp: - * ace/Base_Thread_Adapter.h: - * ace/Base_Thread_Adapter.cpp: - We pre-allocate the Log_Msg attributes in the Thread_Adapter, to - minimize memory allocations during thread creation. - - * ace/Log_Msg_Callback.h: - * ace/Log_Msg_Callback.cpp: - Move the Log_Msg callback interface to its own file. - - * ace/Log_Msg_Backend.h: - * ace/Log_Msg_Backend.cpp: - Define the interface for Log_Msg backend strategies. - - * ace/Log_Msg_IPC.h: - * ace/Log_Msg_IPC.cpp: - Implement a Log_Msg backend based on IPC. - The intention is to decouple Log_Msg from IPC components, thus - Log_Msg only uses the abstract class ACE_Log_Msg_Backend. - In the future we could implement other backends, such as logging - to syslog or the NT event log directly. - We need to use a dynamically loaded factory to complete the - decoupling, but that would take a little more effort. - - * ace/Log_Msg.h: - * ace/Log_Msg.cpp: - Use the attributes directly. - Use the new Log_Msg_Backend classes. - - * tests/Log_Msg_Test.cpp: - * examples/Log_Msg/test_callback.cpp: - #include Log_Msg_Callback.h, the file is not included unless - ACE_LEGACY_MODE is set. - - * tests/Malloc_Test.cpp: - Add missing #include - - * examples/Shared_Malloc/test_position_independent_malloc.h: - * examples/Shared_Malloc/test_position_independent_malloc.cpp: - Reordered #includes to make them more standard. - -Fri Oct 13 22:34:59 2000 Marina Spivak <marina@cs.wustl.edu> - - * bin/auto_run_tests.lst: - - Added new TAO/tests/RTCORBA/Client_Protocol test to the list. - Removed TAO/tests/Endpoint_Per_Priority because - TAO::Client_Priority_Policy is now disabled in builds by - default. Enabled TAO/tests/RTCORBA/Server_Protocol for linux - builds (it was disabled for some reason). - -Fri Oct 13 19:12:31 2000 Steve Huston <shuston@riverace.com> - - * ace/Malloc_T.h: Added #include "ace/Malloc.h" because the definition - of ACE_Control_Block is needed, and it's in Malloc.h. - -Fri Oct 13 18:47:51 2000 Steve Huston <shuston@riverace.com> - - * ace/Caching_Strategies_T.cpp: - * ace/Dynamic_Service.cpp: - * ace/Timer_Hash_T.cpp: - * ace/Timer_Wheel_T.cpp: - * netsvcs/lib/Log_Message_Receiver.cpp: Added #include "ace/Log_Msg.h" - else IBM C++ can't compile the template at instantiation time. - - * tests/Process_Strategy_Test.h: Added #include "ace/File_Lock.h" to - see ACE_File_Lock, so IBM C++ can compile the templates. - -Fri Oct 13 18:43:24 2000 Steve Huston <shuston@riverace.com> - - * tests/MM_Shared_Memory_Test.cpp: Changed SYNCHRONIZER ctor for - ACE_SV_Semaphore_Simple case to specify - ACE_SV_Semaphore_Simple::ACE_CREATE, not ACE_SV_Semaphore_Complex:: - ACE_CREATE. Else it got compile failures on Solaris, Sun CC, - in PACE build (why in just PACE build? Dunno...) - -Fri Oct 13 13:57:40 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Process_Manager.h : - * ace/Process_Manager.cpp : - - * ace/Service_Config.h : - * ace/Service_Config.cpp: - - Removed dependency of Service_Config on Process_Manager by - registering the Process_Manager singleton with the - Object_Manager when it is instantiated. This way we need not call - ACE_Process_Manager::close_singleton () in Service_Config - implementation. - -Mon Oct 9 10:34:18 2000 Ossama Othman <ossama@uci.edu> - - * configure.in: - - Corrected comments regarding the ACE_HAS_STREAM_PIPES run-time - test. - -Fri Oct 13 11:41:29 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Makefile.bor (OBJFILES): - * ace/Makefile.am (libACE_Utils_la_SOURCES): - Updated to include the new files. - - * examples/Shared_Malloc/test_position_independent_malloc.cpp: - Included a header file. - -Fri Oct 13 11:43:56 2000 David L. Levine <levine@cs.wustl.edu> - - * ACE-INSTALL.html: updated supported platforms list. - -Fri Oct 13 11:39:54 2000 David L. Levine <levine@cs.wustl.edu> - - * ACE-INSTALL.html: added Jaffar Shaikh's <Jaffar_Shaikh@Mitel.COM> - notes for building for VxWorks target on NT host. - -Fri Oct 13 06:26:18 2000 Balachandran Natarajan <bala@cs.wustl.edu> - - * examples/RMCast/Send_File: - Added makefiles for Send_File multicast example. - - * include/makeinclude/build_dll.bor: - * include/makeinclude/build_exe.bor: - * include/makeinclude/build_lib.bor: - Added support for compilation of C source files. - - * include/makeinclude/ace_flags.bor: - Added IORManip and FaultTolerance libraries for TAO. - - Thanks to Christopher Kohlhoff for providing these fixes. - -Thu Oct 12 23:33:21 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/Synch.h: - * ace/Synch_T.h: - * ace/Test_and_Set.cpp: - Fixed order of includes to compile under gcc. - -Thu Oct 12 22:58:13 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/ACE_Lib_Find.h: - * ace/ACE_Lib_Find.cpp: - * ace/ACE.cpp: - Move the wchar version of strrepl and strsplit_r to the - ACE_Lib_Find class, the other versions where there already. - - * ace/FILE_IO.h: - * ace/FILE_Connector.cpp: - Add missing #include in FILE_IO.h that is used in FILE_IO.i - - * ace/Test_and_Set.h: - * ace/Test_and_Set.cpp: - Fixed the guards for the .cpp file. Add the #include for - platforms that requires template source visible, and add the - #pragma implementation for (really) old g++ versions. - - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - Move Test_and_Set.cpp to the "Template Files" folder, and - disable compilation for it. - -Thu Oct 12 23:09:17 2000 Joe Hoffert <joeh@cs.wustl.edu> - - * ace/OS.i: - Added a !defined (__Lynx__) check for ACE_HAS_PACE in the - ACE_OS::mmap function. ACE does some special things for Lynx - for mmap since Lynx's mmap is deficient. Probably should - move this fix from ACE to PACE at some point. - -Thu Oct 12 19:38:00 2000 Ossama Othman <ossama@uci.edu> - - * ace/ace_dll.dsp: - * ace/ace_lib.dsp: - - Added new files introduced by Priyanka's footprint reduction - effort. - -Thu Oct 12 18:19:46 2000 Priyanka Gontla <pgontla@ece.uci.edu> - - * ace/Makefile: - Made the corresponding changes. - - * ace/System_Time.cpp: - * ace/UPIPE_Connector.cpp : - * ace/Timer_Queue_T.h : - * ace/Task.cpp : - * ace/POSIX_Asynch_IO.cpp: - * ace/OS.h: - * ace/Memory_Pool.cpp: - * ace/MEM_SAP.h: - * ace/MEM_Acceptor.cpp: - * ace/Logging_Strategy.cpp: - * ace/FIFO.h: - * ace/DLL.cpp: - * ace/IO_SAP.h: - * ace/Malloc_T.h : - Cosmetic changes. Ex: Change the included header files to the - new header files and similar changes. - - * ace/Naming_Context.h: - * ace/Naming_Context.cpp: - * ace/Object_Manager.cpp: - - Removed the dependency of Object_Manager on Naming_Context. - - * ace/Test_and_Set.h: - * ace/Test_and_Set.cpp: - * ace/Test_and_Set.i: - - Moved the class Test_and_Set from Synch_T to here. This helped - remove the dependency on Event_Handler. - - * ace/Service_Templates.h: - * ace/Service_Templates.cpp: - - Moved the template instantiations from Service_Config.cpp and - Service_Types.cpp since these templates were not needed by the - respective files. - - * ace/Malloc_Allocator.h : - * ace/Malloc_Allocator.cpp: - * ace/Malloc_Allocator.i : - - * ace/PI_Malloc.h : - * ace/PI_Malloc.cpp: - * ace/PI_Malloc.i : - - * ace/Malloc.h : - * ace/Malloc.cpp : - * ace/Malloc.i : - - Moved the ACE_PI_Control_Block, ACE_New_Allocator and - ACE_Static_Allocator_Base classes to new files: PI_Malloc and - Malloc_Allocator. - - This move helps in reduction in size of Malloc.i since the - template instantiations which were included only for - ACE_PI_Control_Block are now moved to the new file: PI_Malloc. - In addition, is the effect of clear and minimized dependencies. - - * ace/ACE.h : - * ace/ACE.cpp : - * ace/ACE.i : - - Moved several functions in ACE to new classes to minimize the - dependencies on and by ACE. Each of the following files - have the listed functions. - - * ace/Malloc_Instantiations.cpp: - Moved the template instantiations to this file. - - * ace/ACE_Sock_Connect.h : - * ace/ACE_Sock_Connect.cpp : - * ace/ACE_Sock_Connect.i : - - Has the socket connection establishment function calls : - bind_port, get_bcast_addr, get_ip_interfaces, count_interfaces - and get_handle. - - Helps remove the dependency on INET_Addr. - - * ace/ACE_Handle_Ops.h: - * ace/ACE_Handle_Ops.cpp: - * ace/ACE_Handle_Ops.i: - - Has the <handle_timed_open> operation on handles. - Helps remove the dependency of FILE_Connector, SPIPE_COnenctor, - UPIPE_Conenctpr and DEV_Connector on ACE. - - * ace/ACE_Lib_Find.h : - * ace/ACE_Lib_Find.cpp : - * ace/ACE_Lib_Find.i : - - Includes all the functions to search and open shared - libraries and realted functions : - ldfind, ldopen, ldname, get_temp_dir, open_temp_file, - strrepl, strsplit_r. - - Removes the dependency of DLL, Logging_Strategy and FILE_Addr on - ACE. Effects few other classes too but doesnt remove the - dependency on ACE for those classes. - - * ace/ACE_Flag_Manip.h : - * ace/ACE_Flag_Manip.cpp: - * ace/ACE_Flag_Manip.i : - - The Flag Manipulation functions: - set_flags (), clr_flags () and get_flags (). - - Removes the dependency of IO_SAP and IPC_SAP on ACE. - - * ace/ACE_Init.h : - * ace/ACE_Init.cpp: - * ace/ACE_Init.i : - - The init () and fini () functions which are used - for the initialization and shutting down of ACE. - - Removes dependency on Object_Manager. - -Thu Oct 12 15:41:46 2000 Carlos O'Ryan <coryan@uci.edu> - - * include/makeinclude/rules.bin.GNU: - List the LDFLAGS before the name of the program, this seems to - work better for some of our sponsors. - -Thu Oct 12 15:40:51 2000 Carlos O'Ryan <coryan@uci.edu> - - * ace/OS.cpp (ACE_OS_GUARD): - Use explicit names for the OS guards, otherwise the SGI compiler - (7.30 + several patches) crashes. Thanks to Philip Miller - <pwmiller@sarnoff.com> for reporting the problem and providing - the patch. - -Thu Oct 12 12:15:23 2000 Angelo <corsaro@cs.wustl.edu> - - * ACE version 5.1.10 released. |