diff options
Diffstat (limited to 'ChangeLog-99b')
-rw-r--r-- | ChangeLog-99b | 1453 |
1 files changed, 11 insertions, 1442 deletions
diff --git a/ChangeLog-99b b/ChangeLog-99b index 8af6d177972..84f142b8627 100644 --- a/ChangeLog-99b +++ b/ChangeLog-99b @@ -1,1440 +1,7 @@ -Mon Jul 26 12:32:41 1999 Douglas C. Schmidt <schmidt@cs.wustl.edu> - - * ACE version 4.6.46 released. - -Fri Jul 23 18:20:11 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/config-irix6.x-sgic++-nothreads.h: - Enabled namespaces with MIPSPro 7.2 or higher. - -Fri Jul 23 16:40:15 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/ACE.cpp: - Fixed typos in error return values for ACE::execname() - -Fri Jul 23 14:51:48 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * examples/Service_Configurator/IPC-tests/Handle_L_CODgram, - examples/Service_Configurator/IPC-tests/Handle_L_Dgram, - examples/Service_Configurator/IPC-tests/Handle_L_Pipe: - Cleaned this stuff up so that it builds correctly when - ACE_LACKS_UNIX_DOMAIN_SOCKETS is enabled. Thanks to - David for reporting this. - - * examples/Service_Configurator/IPC-tests/server/Handle_L_Pipe: - Made upper_case() a method in the Handle_L_Pipe class. - - * examples/Service_Configurator/IPC-tests/server/Handle_L_Pipe.cpp, - examples/Service_Configurator/IPC-tests/server/Handle_L_CODgram.cpp: - Removed the bizarre hacks for SunOS4. This stuff probably - hasn't been tested since 1992! - -Fri Jul 23 14:16:22 1999 Steve Huston <shuston@riverace.com> - - * ace/OS.i (ACE_OS::read): (only non-Win32), if call fails and - errno is EAGAIN, change it to EWOULDBLOCK. Improve portability for - platforms where EAGAIN and EWOULDBLOCK are different values and - both used, like HP-UX. - - * ace/Malloc_T.cpp (ACE_Malloc<> ctor): Added missing newline in error - message. - -Fri Jul 23 12:09:18 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/OS.i: Fixed some mismatched ifdef/endif comments. - - * ace/ACE.cpp (execname): Fixed some misuses of ACE_HAS_WIN32 - macros where they should have been ACE_WIN32. Thanks to Andy - for noticing this. - -Fri Jul 23 11:28:17 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * ace/Memory_Pool.cpp (init_acquire): shmat() returns -1 on - failure, not 0. Thanks to Kevin Liu - <kliu@fitch.math.uwaterloo.ca> for reporting this. - -Fri Jul 23 11:35:31 1999 David L. Levine <levine@cs.wustl.edu> - - * examples/Connection/non_blocking/tests_lsock_{acceptor, - connector.cpp}: disabled if ACE_LACKS_UNIX_DOMAIN_SOCKETS. - Thanks to Alain Magloire <alain@qnx.com> for reporting this. - - * examples/Connection/non_blocking/Makefile: removed unneeded - LIBS (=$(ACELIB)) and BUILD macro definitions. - -Fri Jul 23 10:07:05 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_qnx_neutrino.GNU: added - CROSS-COMPILE macro definition. Thanks to - Alain Magloire <alain@qnx.com> for reporting this. - -Fri Jul 23 08:58:56 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * ace/Service_Config.h: Clarified the fact that the - ACE_Service_Config uses the Monostate pattern, so there can only - be one of them per-process. Thanks to Ralf Kluthe - <kluthe@iti.mu-luebeck.de> for reporting this. - -Fri Jul 23 07:49:31 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_sunos5_sunc++.GNU: added - comment about options for producing 64-bit binaries. - Thanks to Jeff Franks <Jeffrey_Franks@i-o.com> for - raising this issue. - -Fri Jul 23 03:10:31 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Strategies_T.cpp (find_or_create_svc_handler_i): - * ace/Cached_Connect_Strategy_T.cpp (cached_connect): - - I had added the new_connection() method that lets go of the lock - for the duration of the connect(), but forgot to use it in the - above functions ;) It should be fine now. - - * tests/Cached_Conn_Test.cpp (cached_connect): - * tests/Cached_Accept_Conn_Test.cpp (cached_connect): - - Host of the remote address should be ACE_LOCALHOST instead of - ACE_DEFAULT_SERVER_HOST. - -Thu Jul 22 23:23:26 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Handle_Gobbler.h (class ACE_Handle_Gobbler): Inlined - declaration of the functions. - -Thu Jul 22 20:03:54 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * ace/Acceptor.cpp: For sanity's sake, initialize all the data - members to 0 in the ACE_Strategy_Acceptor constructor. Thanks - to Jody Hagins <jody@atdesk.com> for reporting this. - - * ace/Acceptor.cpp: The service_port_ data member in the - ACE_Strategy_Acceptor wasn't being initialized to 0. Thanks to - Jody Hagins <jody@atdesk.com> for reporting this. - - * ace/Acceptor.cpp: Modified the dump() and info() methods so that - don't do bad things if service_name_ and service_description_ - are NULL. Thanks to Jody Hagins <jody@atdesk.com> for this - suggestion. - - * ace/Message_Queue_T.h: Clarified what the errnos get set to when - timeouts occur or queues are deactivated. Thanks to - John Forest <forestj@res.raytheon.com> for suggesting this. - -Thu Jul 22 16:55:14 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/Stats.cpp: - Fixed message that had a %s but no argument! - -Thu Jul 22 15:12:18 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/Codeset_IBM1047.{h,cpp}: wrapped with #ifdef ACE_MVS - so that the code doesn't occupy any space on platforms - other than MVS. - -Thu Jul 22 14:22:41 1999 David L. Levine <levine@cs.wustl.edu> - - * examples/Shared_Malloc/Malloc.cpp: fixed explicit - template instantiations so that there are no duplications - with ACE_LACKS_SBRK and ACE_LACKS_SYSV_SHMEM. - -Thu Jul 22 12:17:26 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/config-lynxos.h: removed ACE_LACKS_FORK on PowerPC. - All of the ACE tests that fork now succeed. - -Thu Jul 22 11:54:58 1999 David L. Levine <levine@cs.wustl.edu> - - * tests/SV_Shared_Memory_Test.cpp (child): added a one - second sleep to give the parent a chance to create the - semaphore. This allows the test to succeed on LynxOS/PPC - (with fork ()). - - * tests/Process_Mutex_Test.cpp,Process_Strategy_Test.cpp (main): - added a one second sleep after spawning each server, to give it - a chance to start. Also, wrapped some string literals with - ASYS_TEXT. Again, success on LynxOS/PPC with fork (). - -Thu Jul 22 10:42:06 1999 Clarence M. Weaver <clarence_m_weaver@md.northgrum.com> - - * ACE-INSTALL.html: added steps for running the ACE tests - on VxWorks platforms that lack NFS. - -Thu Jul 22 10:25:31 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/Log_Record.cpp,SString.cpp: wrapped #include of - ace/streams.h with #ifndef ACE_LACKS_IOSTREAMS_TOTALLY. - [Bug 84] - -Thu Jul 22 09:55:10 1999 David L. Levine <levine@cs.wustl.edu> - - * docs/ACE-guidelines.html: updated commit message guideline - to show ChangeLogTag example. - -Thu Jul 22 02:15:46 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Handle_Gobbler (class ACE_Handle_Gobbler): New class that - gobbles up handles :) This class is useful when we need to - control the number of handles available for a process. Mostly - used for testing purposes. - - Note that even though the new files are in the ace directory, - they are not part of the ace library. The files only contain - inline functions and Handle_Gobbler.h is only included where - needed. - - * tests/Cached_Conn_Test.cpp: - * tests/Cached_Accept_Conn_Test.cpp: - - Added Handle_Gobbler to the test to reduce the iterations - required before handles run out and purging starts. - -Wed Jul 21 21:45:22 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Reactor.cpp: Applied ACE_NOTREACHED to several - run_*_event_loop methods. - - * docs/ACE-guidelines.html: Updated ASYS_TEXT usage. - - * ace/RB_Tree.cpp: - * tests/Cache_Map_Manager_Test.cpp: Applied ASYS_TEXT liberally. - -Wed Jul 21 21:19:36 1999 David L. Levine <levine@cs.wustl.edu> - - * ACE-INSTALL.html: added note about ACE support for - shared libraries on VxWorks. It's still under - development. - -Wed Jul 21 18:38:00 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Strategies_T.h (ACE_Cached_Connect_Strategy): Renamed - <connection_cache_> to <connection_map_>. - - * ace/Hash_Cache_Map_Manager_T.cpp (find): Finding and updating - the attributes is faster through the find() that uses an entry. - Therefore, for find (key, value) and find (key) we go to - find (key, entry) instead of going to the base class. - - * ace/Cached_Connect_Strategy_T.cpp: - - - ACE_Cached_Connect_Strategy_Ex::check_hint_i - - ACE_Cached_Connect_Strategy_Ex::find - - Update the caching attributes directly since we don't do a - find() on the cache map. - -Wed Jul 21 15:54:25 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/DLL.cpp: ACE_HAS_MOSTLY_UNICODE_APIS fixes. - - * ace/OS.i (ctime): - * ace/OS.cpp (ctime): One copy should only be available when - ACE_HAS_MOSTLY_UNICODE_APIS is defined. Thanks to Steve for - reporting this. - -Wed Jul 21 15:21:32 1999 David L. Levine <levine@cs.wustl.edu> - - * tests/Env_Value_Test.cpp (main): use the single-process - version if ACE_LACKS_FORK, in addition to if - ACE_HAS_NONSTATIC_OBJECT_MANAGER. - - * tests/Time_Service_Test.cpp (main): neuter if ACE_LACKS_FORK, - because the test uses ACE_Process::spawn (). - - Thanks to Riaz Syed <syed@Lynx.COM> for reporting these. - -Wed Jul 21 11:48:53 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_vxworks5.x_g++.GNU: - use full path to $(CC) in egcs version check, to avoid - warning if it is not on the user's PATH. - -Wed Jul 21 11:44:19 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/OS.h: - Simplified the definition of the ACE_THR_PRI_*_DEF macros from - (MIN+(MAX-MIN)/2) to ((MIN+MAX)/2), the new definition is not - only easier to read, makes less operations (and function calls) - but it is also safer when the the priorities are "backwards" and - PRI_MAX < PRI_MIN - -Wed Jul 21 11:40:31 1999 Steve Huston <shuston@riverace.com> - - * include/makeinclude/platform_aix4_cset++.GNU: - * include/makeinclude/platform_hpux_aCC.GNU: Replaced lines like - "debug=1" with ifeq (,$(debug)) debug=1 endif. - - * include/makeinclude/platform_aix4_g++.GNU: - Above change for variables, also added -Wl,-bbigtoc when linking - TAO_IDL, else it runs out of toc space. Don't do it for all links - because bigtoc introduces a run-time performance penalty. - -Wed Jul 21 10:25:07 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/OS.{h,cpp} (tss_base,tss_open): with TSS emulation, - no longer return indication of whether tss_base allocated - the ts_storage array. Instead, added an (optional) - argument that tss_open can use for that purpose. This - allows threads that haven't been spawned by ACE to use - ACE's TSS emulation. Thanks to David Hauck - <davidh@realtimeint.com> for reporting this and testing - the fix on NT. And, thanks to Terry Rosenbaum - <Terry.Rosenbaum@Radiology.MSU.edu> for originally reporting it. - -Wed Jul 21 09:34:09 1999 David L. Levine <levine@cs.wustl.edu> - - * examples/Shared_Malloc/test_position_independent_malloc.cpp (main): - removed declarations of unused argc and argv. - -Wed Jul 21 04:32:20 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Reactor.cpp: - * ace/Reactor.h: - * ace/Reactor.i: - * ace/Reactor_Impl.h: - * ace/Select_Reactor_T.cpp: - * ace/Select_Reactor_T.h: - * ace/Select_Reactor_T.i: - * ace/TP_Reactor.cpp: - * ace/WFMO_Reactor.cpp: - * ace/WFMO_Reactor.h: - * ace/WFMO_Reactor.i: The change was motivated by removing a race - condition in TP_Reactor during shutdown. There was a small time - gap between a thread running TP_Reactor event loop checking its - end_event_loop flag to the thread actually grabing the token in - TP_Reactor. Since the token return to its "normal" state after - broadcasting, this thread has no chance to notice that the - reactor has already been shutdowned and thus a deadlock occured. - In order to fix the problem, we must keep a state indicating - whether the reactor has been shutdowned or not in the reactor. - - An added benefit of adding this state is that every reactor now - has its own run_reactor_event_loop which mimic the behavior of - the static ACE_Reactor::run_event_loop but instead of working on - the singleton reactor, the added function work on an instance of - reactor object. In fact, all event_loop control functions have - been added. They are named *_reactor_event_loop to - differentiate their behavior. These newly added function use - the added state in the reactor to control the looping of event - handling. All run_reactor_event_loop methods take an extra - argument of a function pointer. The function will be called in - every iteration if it is set. - - Notice that the singleton event loop control functions are still - available. Although these function merely forward the call to - the singleton reactor's *_reactor_event_loop functions, you can - still use them to control the "main" event loop. In fact, it is - required that you use these functions to run the main event loop - because it facilitates Service_Configurator's "reconfigure" - feature. - - All reactors now also have two extra functions namely - "deactivated" which is used to query whether the reactor will - handle more incoming event or not, and "deactivate" which is - used to control this new feature. - - An side effect of the change is that, once a reactor get - deactivated, calls to handle_events will return -1 immediately. - You can use the "deactivated" method to check if there's - actually an error in the handle_events method or it simply - returned because the reactor has been deactivated. This is also - how the various "run_reactor_event_loop" handle the situation. - Notice that this should not cause any backward compatibility - problem becuase if a user is using his own end_event_loop flag, - he will not (and should not) invoke any more handle_events after - the flag has been set. Also, if he is using the global - run_event_loop methods (for controling singleton reactor,) they - should still behave the same. - - To summarize the change: - - 1. The run_event_loop control logic has been "pushed" down into - each individual reactor. The methods supporting the logic are - defined in ACE_Reactor and are: - - run_reactor_event_loop () - run_alertable_reactor_event_loop () - end_reactor_event_loop () - reactor_event_loop_done () - reset_reactor_event_loop () - - 2. The origianl run_event_loop methods in ACE_Reactor are still - working but they now use the run_reactor_event_loop methods in - the implementation of the reactor. - - 3. Reactors now keep a state to indicate handle_events should be - called or not. They can be controlled thru: - - deactivated () - deactivate () - - methods. - -Tue Jul 20 21:31:50 1999 Jim Rogers jrogers@viasoft.com - - * ace/config-mvs.h: added ACE_HAS_EXCEPTIONS. - -Tue Jul 20 21:23:19 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/OS.* (next_key,total_keys): with TSS_EMULATION, - added guard (using preallocated ACE_TSS_KEY_LOCK) - to access of next_key_. Uninlined both of these - functions because they now have the guards. Thanks - to Jake Hamby <jehamby@anobject.com> for reporting this. - -Tue Jul 20 21:11:02 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_vxworks5.x_g++.GNU: - fixed test of g++ for whether it's egcs. Thanks to - Erik Johannes <ejohannes@oresis.com> for reporting this. - (The symptom was that TAO's IORManipulation.cpp wouldn't - compile due to an internal compiler error. Adding - -fno-exceptions works around that.) - -Tue Jul 20 18:30:22 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Cached_Connect_Strategy_T.h: All functions in - ACE_Cached_Connect_Strategy_Ex and ACE_Cached_Connect_Strategy - that contain specific information about the map used by the - class should not be virtual. - - * ace/Future_Set.cpp (next_readable): Removed extra code that was - left behind from a debugging stage. Thanks to author John - Tucker <jtucker@infoglide.com> for helping with this. - -Tue Jul 20 12:02:47 1999 David L. Levine <levine@cs.wustl.edu> - - * examples/Threads/task_five.cpp (main): removed - ACE_NOTREACHED wrapper around the final return statement. - It's reachable now. - -Tue Jul 20 11:46:09 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * examples/Shared_Malloc/test_position_independent_malloc.cpp: - Completely #ifdef'd this example out until we get it working. - Thanks to David Levine for reporting this. - -Tue Jul 20 11:15:23 1999 David L. Levine <levine@cs.wustl.edu> - - * ACE-INSTALL.html: updated memory requirements discussions. - -Tue Jul 20 09:43:36 1999 Ossama Othman <othman@cs.wustl.edu> - - * ace/Future_Set.cpp (next_readable): - - Fixed unused variable warnings. - -Tue Jul 20 09:33:02 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/OS.{h,i} (sigwait): only use the DIGITAL_UNIX sigwait - massaging with cxx, not with any other compiler. Thanks to - Doug Anderson <dla@home.com> for suggesting this, because - egcs issued warnings with it. - -Tue Jul 20 07:06:28 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_osf1_4.0.GNU: - removed -ptr ptrepository. Let the compiler use its - default template repository. - - * include/makeinclude/rules.local.GNU (realclean): added - cxx_repository, for DU cxx. - -Tue Jul 20 07:01:57 1999 Doug Anderson <doug@clark.net> - - * ace/OS.i (sigwait,thr_sigsetmask): added support for - Cray UNICOS 9. - -Mon Jul 19 22:46:54 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Strategies_T.cpp - (ACE_Cached_Connect_Strategy::new_connection):: We use a reverse - lock to let go of the lock for the duration of the actual - connect. This will allow other threads to hack on the - connection cache while this thread creates the new connection. - Thanks to James Whitledge <jwhitledge@spyglass.com> for - suggesting this. - - * tests/Cached_Accept_Conn_Test.cpp - * tests/Cached_Conn_Test.cpp - * tests/Conn_Test.cpp - - Added reverse lock template instantiations. - - * ace/Synch_T: Since the reverse lock constructor that created the - regular lock was removed, there was no need for the - <delete_lock_> variable, and <lock_> could be changed into a - reference (instead of pointer). - -Mon Jul 19 23:04:15 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * netsvcs/lib/Logging_Strategy.cpp (init): The ACE_LOG_MSG->open() - call at the end of the ACE_Logging_Strategy::init() method - *must* specify the flags and logger key or an application will - not connect to the client logger daemon automatically when the - ACE_Log_Msg::LOGGER flag is set in a svc.conf file. Thanks to - Jerry De Master <jdemaste@ipdinc.com> for reporting this. - -Mon Jul 19 22:11:13 1999 David L. Levine <levine@cs.wustl.edu> - - * ACE-INSTALL.html: cleaned up instructions for building - on NT for VxWorks target. Thanks to Stephane Chatre - <schatre@oresis.com> for reporting that the instructions - were out of date. - -Mon Jul 19 21:38:16 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * tests/run_tests.vxworks: - Re-enable the Cached_Conn_Test and Cached_Accept_Conn_Test, it - works on Linux, NT, Solaris, IRIX and HP-UX so we should be - pretty close. Thanks to Irfan for fixing the problems with this - test. - -Mon Jul 19 21:26:57 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/OS.{h,i}: reverted sigwait () changes on Digital - Unix. They cause examples/Threads/process_manager.cpp - to not compile on DU 5.0 with cxx 6.2-021. See - Mon Jul 19 21:13:16 1999 Doug Anderson <doug@clark.net>. - -Mon Jul 19 21:13:16 1999 Doug Anderson <doug@clark.net> - - * ace/ACE.cpp,ace/Basic_Types.h,ace/CDR_Stream.{h,cpp}, - ace/SOCK_Dgram_Bcast.cpp,ace/config-cray.h, - tests/Basic_Types_Test.cpp,tests/Handle_Set_Test.cpp, - include/makeinclude/platform_cray.GNU - - I finally managed to get the latest and greatest to work on - the Cray. Still no CDR support, but that will be post 5.0. - - * ace/OS.{h,i}: added support for egcs on Digital Unix (4.0d & e). - Fixe problem with legacy workarounds in OS.* for sigwait() and - cxx/GNU that don't appear to be necessary on 4.0d/e/f with - neither cxx nor GNU. - - * include/makeinclude/platform_tru64_g++.GNU: added. Works - with ace/config-tru64.h. - -Mon Jul 19 19:46:50 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/OS.h: Added ACE_*_cast_*_ptr and ACE_*_cast_*_ref for const, - static, and reinterpret casts. - - * ace/Strategies_T.cpp (find): - * ace/Cached_Connect_Strategy_T.cpp (find): - - The comparison should be between ACE_PEER_CONNECTOR_ADDR and not - REFCOUNTED_HASH_RECYCLABLE_ADDRESS. - - * ace/Svc_Handler: Changed <state> to <recycle_state>. <state> - was too generic and in some cases a <state> method was already - present in classes deriving from Svc_Handler. - - Also, added a <recycle_state> accessor. - - The following files were effected: - - - ace/Cached_Connect_Strategy_T.cpp - - ace/Cached_Connect_Strategy_T.h - - ace/Caching_Utility_T.cpp - - ace/Strategies.h - - ace/Strategies.i - - ace/Strategies_T.cpp - - ace/Strategies_T.h - - ace/Strategies_T.i - - tests/Cached_Accept_Conn_Test.cpp - - tests/Cached_Conn_Test.cpp - -Mon Jul 19 20:08:15 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/Strategies_T.cpp: - Fixed lookup of connections in the Cached_Connect_Strategy, - thanks to Irfan for providing this fix. - -Mon Jul 19 12:26:20 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_vxworks5.x_g++.GNU, - ace/config-g++-common.h: - don't #define ACE_HAS_GNUG_PRE_2_8 with Tornado II's - g++. It's egcs, though it's version identifier says - it's 2.7. Thanks to Thomas Lockhart - <Thomas.G.Lockhart@jpl.nasa.gov> for reporting this. [Bug 80] - -Mon Jul 19 09:50:31 1999 Steve Huston <shuston@riverace.com> - - * ace/Future.cpp: Added ACE_TYPENAME to template-qualified types. - * ace/Future.cpp: Added #include "ace/Containers.h" to be sure the - types necessary for this class is available at template link - time. - - * apps/gperf/src/Options.h: If on PharLap ETS, clear out its definition - of EOS to avoid compile errors. Thanks to David Hauck - <davidh@realtimeint.com> for this fix. - -Mon Jul 19 08:12:30 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_tru64_cxx.GNU: added support - for DU 4.0E and D. - -Mon Jul 19 08:08:25 1999 Steve Huston <shuston@riverace.com> - - * ace/config-hpux-(10 11).x-hpc++.h: In the aC++ section, if __HP_aCC - is not set, set it. Some parts of TAO rely on it, but it's not set - until the most recent versions of the compiler. Thanks to Brian - Wallis <Brian.Wallis@sr.com.au> for helping figure this out. - -Mon Jul 19 07:53:59 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/SOCK_Dgram_Mcast.cpp (open): added unnecessary const_cast - of mcast_addr to ACE_Addr &, so that Sun C++ 4.2 will compile - without error. Thanks to Andy for this fix. - -Sun Jul 18 21:36:04 1999 David L. Levine <levine@cs.wustl.edu> - - * OS.cpp (sched_params): wrapped DIGITAL_UNIX preprocessor - symbol in parenthesis. - -Sun Jul 18 14:30:54 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * tests/run_tests.sh: Reenabled the Future_Set_Test. Hopefully, - this will work on Solaris and the other platforms now! - - * ace/Future*: Added some fixes for Future_Set that should prevent - it from hanging indefinitely on certain platforms. Also, - removed the Future_Node.h and Future_Node.cpp files since they - are no longer used. Thanks to John Tucker - <jtucker@infoglide.com> for contributing these fixes. - - * ace/SOCK_Dgram_Mcast: Finished implementing the QoS-enabled APIs for - socket datagram multicast. This required refactoring quite a - bit of code and adding some new methods that take the - ACE_QoS_Params. - - * ace/SOCK_Dgram.cpp: Finished integrating the QoS support - for SOCK dgrams. - - * ace/SOCK_Dgram.cpp: Factored out common code in the - ACE_SOCK_Dgram constructors. - - * ace/SOCK_Connector: Finished implementing the QoS-enabled - connect() methods. To make this cleaner, I've created a new - "shared_open()" method that factors out the common code. - - * ace/SOCK_Acceptor: Finished implementing the QoS-enabled open() - method. To make this cleaner, I've created a new - "shared_open()" method that factors out the common code. - -Sat Jul 17 13:36:03 1999 Ossama Othman <othman@cs.wustl.edu> - - * ace/High_Res_Timer.cpp (get_cpuinfo): - - Removed extra `FILE *cpuinfo' declaration. - -Sat Jul 17 12:20:25 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * ace/High_Res_Timer: Added integrated support for Alpha and - Pentium that will eliminate the 5sec startup delay for Linux/PII - clients using the ACE high resolution timer. Thanks to Thomas - Lockhart <Thomas.G.Lockhart@jpl.nasa.gov> for these fixes. - - * netsvcs/servers/main.cpp (main): Fixed a typo caused by recent - integration of Jerry De Master's fixes. - -Fri Jul 16 19:41:05 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/ACE.cpp (out_of_handles): Added a function to check if a - process is out of handles (file descriptors). This is required - to avoid repeating this code in many places. Also, note that - all platform specific quirks can be captured in this function. - - * ace/Cached_Connect_Strategy_T.cpp (cached_connect): - * tests/Cached_Conn_Test.cpp (out_of_sockets_handler): - * tests/Cached_Accept_Conn_Test.cpp (out_of_sockets_handler): - - Changed the above to use the new ACE::out_of_handles() function. - - * ace/OS.h (ENFILE): Added ENFILE if missing on platform. In that - case, ENFILE will equal to EMFILE. - -Fri Jul 16 13:55:34 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * include/makeinclude/{build_dll.bor,build_exe.bor,compiler.bor}, - ace/{Makefile.bor,ace.rc}: - Updated these files to add version information to the DLLs so - when you right click them in windows explorer you can see what - version the file is (e.g., "4.6.45"). Thanks to Christopher - Kohlhoff <chris@kohlhoff.com> for contributing this. - - * netsvcs/servers/main.cpp (main): Disable debugging messages - if the -d option isn't given to ACE_Service_Config. Thanks to - Jerry De Master <jdemaste@ipdinc.com> for reporting this. - - * ace/config-win32-borland.h: Updated this file to enable some - necessary macros. Thanks to Christopher Kohlhoff - <chris@kohlhoff.com> for reporting this. - - * tests/DLL_Test.cpp, - tests/run_test.bat: Updated these to handle the appropriate - directory where the Borland tests are built. Thanks to - Christopher Kohlhoff <chris@kohlhoff.com> for reporting this. - - * netsvcs/lib/Logging_Strategy.cpp (init): When the application - program is started, none of the log output is sent to the - Client_Logging_Service. Further investigation shows that the - ACE_Log_Msg::open() method is responsible for connecting to the - client logger if the ACE_Log_Msg::LOGGER flag is set. However, - the connections never attempted because the ACE_Log_Msg::open() - method is called before the Logging_Strategy flags have been - parsed. This problem was fixed by rearranging the point at - which the ACE_LOG_MSG->open() method was called. Thanks to - Jerry De Master <jdemaste@ipdinc.com> for contributing this. - - * netsvcs/lib/Client_Logging_Handler.cpp (send), - netsvcs/lib/Log_Message_Receiver.cpp (log_output): - All log message output in ACE is done through one of the - ACE_Log_Record::print() methods. Each of the overloaded methods - has a <flags> argument as the second argument. However, all - uses of the ACE_Log_Record::print() method in - Log_Message_Receiver.cpp and Client_Logging_Handler.cpp pass - zero for the flags argument. To fix this, simply pass - ACE_Log_Msg::instance()->flags() instead of 0. Thanks to Jerry - De Master <jdemaste@ipdinc.com> for contributing this. - - * ace/Log_Record.cpp (format_msg): Added code for the new - ACE_Log_Record::format_msg() method that adds the date/time - stamp to the ACE_Log_Msg::VERBOSE_LITE output. Thanks to Jerry - De Master <jdemaste@ipdinc.com> for contributing this. - - * netsvcs/lib/Logging_Strategy.cpp (init): Make sure the - ACE_Log_Msg::VERBOSE_LITE flag is OR'd into the clr_flags() - call. Thanks to Jerry De Master <jdemaste@ipdinc.com> for - reporting this. - - * netsvcs/lib/Logging_Strategy.cpp (tokenize): Make sure that - the ACE_Log_Msg::VERBOSE_LITE flag is correctly parsed. Thanks - to Jerry De Master <jdemaste@ipdinc.com> for reporting this. - - * ace/Memory_Pool.h: Added a way to set the segment_size in - ACE_Shared_Memory_Pool_Options and ACE_Shared_Memory_Pool. - Thanks to Serge Kolgan <skolgan@cisco.com> for this fix. [This - feature was actually added a couple of beta releases ago, but - somehow the ChangeLog entry got lost.] - -Fri Jul 16 16:41:22 1999 Ossama Othman <othman@cs.wustl.edu> - - * ace/Stats.cpp (print_summary): - - Fixed format specifier that was expecting an integer argument. - It needed an additional percent specifier. - -Fri Jul 16 09:20:42 1999 Arturo Montes <mitosys@colomsat.net.co> - - * ace/OS.i: Patch the dl* family to keep ACE - compatible in SCO OpenServer 5.0.0, 5.0.2, 5.0.4 with 5.0.5 - - * ace/config-sco-5.0.0.h: Bring OpenServer up to date. - - * include/makeinclude/platform_sco5.0.0-fsu-pthread.GNU: - Use gcc-2.95. - -Fri Jul 16 11:50:24 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Log_Msg.cpp (log): - * tests/Basic_Types_Test.cpp (main): Reverted my previous - change. It was not a safe assumption. - -Fri Jul 16 02:28:10 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Log_Msg.cpp (log): Changed (*va_arg (argp,PTF))(), which - extracts a function pointer from va_list and makes a call using - the pointer pointer to ACE_reinterpret_cast (PTF, va_arg (argp, - long))(). This assumes the sizeof (long) is exactly equal to - the size of a function pointer. - - * tests/Basic_Types_Test.cpp (main): Narrow down the check to - ensure sizeof(long) == sizeof (a function pointer) (which is - true for all the platforms that ACE runs on.) - -Fri Jul 16 00:07:23 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * tests/Cached_Accept_Conn_Test.cpp (test_connection_management): - Changed the test to allow for both one listen endpoint or - multiple listen endpoints. - -Thu Jul 15 22:42:36 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Message_Queue.h: (ACE_MESSAGE_QUEUE_DEFAULT_TYPE): - * ace/Future_Set.{h,cpp} (ACE_Future_Set): Typedef'ed the - ACE_Message_Queue to ACE_MESSAGE_QUEUE_DEFAULT_TYPE and use the - typedef'ed name in ACE_Future_Set's contstructor to get around a - bug in vxworks' g++ compiler. - -Thu Jul 15 22:14:18 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Cached_Connect_Strategy_T.cpp - * ace/Cached_Connect_Strategy_T.h - * ace/Strategies_T.cpp - * ace/Strategies_T.h - - Changed find()'s signature to help g++. - - * tests/Cached_Accept_Conn_Test.cpp - * tests/Conn_Test.cpp - - Added template instantiations. - -Thu Jul 15 21:32:26 1999 David L. Levine <levine@cs.wustl.edu> - - * bin/create_ace_build: removed # -*- perl -*- on first - line so that t/csh users don't trip over it. Thanks - to Irfan for reporting this. - -Thu Jul 15 20:23:39 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * tests/Cached_Conn_Test.cpp (test_connection_management): Changed - the test to allow for both one listen endpoint or multiple - listen endpoints. - - * ace/Cached_Connect_Strategy_T.cpp: - * ace/Strategies_T.cpp - - - find: Since the comparison of - REFCOUNTED_HASH_RECYCLABLE_ADDRESS doesn't work very well when - trying to find an idle connection (because there are many idle - states), we now use the new backet iterator to find a idle - connection. - - - check_hint_i: Explicitly checks for idleness; previously it - was simple checking !closed. - - * ace/Strategies.h (ACE_Recyclable): Changed ACE_Recyclable::State - to ACE_Recyclable_State. - - * ace/Strategies_T.h (ACE_Refcounted_Hash_Recyclable): - - - Added operator!=(ACE_Refcounted_Hash_Recyclable<T>) - - Removed operator==(T) - - Added T &subject() - - Change operator==(ACE_Refcounted_Hash_Recyclable<T>) to check - for equality in state and subject. - - * ace/Caching_Utility_T.cpp (minimum): Entries marked - ACE_RECYCLABLE_PURGABLE_BUT_NOT_IDLE should also be considered - for purging. - - * ace/Cleanup_Strategies_T.cpp (cleanup): The hash map of the - connection cache may contain duplicates. Therefore, we are not - exactly sure which map entry will be removed when calling - unbind(). Therefore, we use the other unbind() which gives us - back the value of the entry that was removed. This way will - close the correct handler. - - * ace/Svc_Handler.cpp (state): Added a new method that allows the - user to explcitly set the state of the Svc_Handler. If the - Svc_Handler does not have a recycler, this method will have no - effect. - -Thu Jul 15 16:39:06 1999 Ossama Othman <othman@cs.wustl.edu> - - * Makefile: - * NEWS: - - The NEWS file is no longer a generated file so it was added to - the list of controlled files. - -Thu Jul 15 16:20:43 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/Service_Manager.cpp (init): Check to see the acceptor handle - is ACE_INVALID_HANDLE before the actual initialization to - prevent this service from being initialized twice. - (fini): Must call remove_handler with - ACE_Event_Handler::DONT_CALL since the service manager won't be - around when the reactor is being destroyed. Thanks to Gheorghe - Aprotosoaie <gaprotosoaie@paragonms.com> for reporting this. - -Thu Jul 15 16:26:28 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/OS.cpp (spa): use ACE_NEEDS_HUGE_THREAD_STACKSIZE - instead of hard-coded value. - - * ace/config-vxworks5.x.h: added default - ACE_NEEDS_HUGE_THREAD_STACKSIZE of 64000. - -Thu Jul 15 16:08:48 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/config-win32-common.h: Moved ACE_HAS_IP_MULTICAST down - *after* ACE_HAS_WINSOCK2 is defined. Thanks to Marina for - reporting this. - -Thu Jul 15 16:11:04 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/Basic_Types.h (ACE_U_LongLong): removed user-defined - conversion to ptr_arith_t. Replaced with new macro, - ACE_LONGLONG_TO_PTR. - -Thu Jul 15 15:31:49 1999 Ossama Othman <othman@cs.wustl.edu> - - * ace/ACE.cpp (timestamp): - * ace/High_Res_Timer.cpp (print_ave): - * ace/Local_Tokens.cpp (open): - * ace/Log_Msg.cpp (log): - * ace/Log_Record.cpp (format_msg): - * ace/Naming_Context.cpp (parse_args): - * ace/OS.cpp (unique_name): - - Fixed warnings and generated by gcc 2.95; mostly due to format - specifier mismatches and inefficient conversion operator usage. - -Thu Jul 15 14:21:32 1999 David L. Levine <levine@cs.wustl.edu> - - * docs/ACE-guidelines.html: reverted to the original - perl magic that goes at the top of every script. The - fix that was supposed to work on HP/UX doesn't. See - Thu Jul 08 22:15:50 1999 David L. Levine <levine@cs.wustl.edu> - -Thu Jul 15 12:19:44 1999 David L. Levine <levine@cs.wustl.edu> - - * bin/make_release: added -u option, to allow updating the - version stamps from any directory. Also, added suppression - of cvs checkout warning about empty ("New") directories. - - * Makefile: added -u to make_release invocations. - -Thu Jul 15 11:39:44 1999 David L. Levine <levine@cs.wustl.edu> - - * ACE version 4.6.45 released. - -Thu Jul 15 11:26:43 1999 David L. Levine <levine@cs.wustl.edu> - - * Makefile: made NEWS a dependency of release. - -Thu Jul 15 10:05:40 1999 David L. Levine <levine@cs.wustl.edu> - - * docs/ACE-guidelines.html: removed reference to - ACE_OS::scanf (). No such thing. - -Thu Jul 15 10:01:31 1999 Steve Huston <shuston@riverace.com> - - * ace/OS.h: TRANSMIT_FILE_BUFFERS are not defined on PharLap ETS - - it's a Microsoft-specific extension to Winsock 1.1. - - * ace/config-win32-common.h: ACE_HAS_IP_MULTICAST is only set for - Winsock 2. - - Thanks to David Hauck <davidh@realtimeint.com> for these fixes. - -Thu Jul 15 09:28:32 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_tru64_cxx.GNU: changed - 4.0f to 4.0F, because that's what sizer -v returns - on DU 4.0F. Thanks to Doug Anderson <dla@home.com> - for reporting this. - - * ace/config-tru64.h: only #define _LIBC_POLLUTION_H_ - with DU >= 5.0. Thanks to Doug Anderson <dla@home.com> - for reporting that it's not necessary on DU 4.0F. - -Thu Jul 15 09:15:11 1999 Steve Huston <shuston@riverace.com> - - * ace/config-aix-4.x.h: Corrected template settings for IBM C/C++ - 3.6 compiler. It does not get ACE_HAS_STD_TEMPLATE_SPECIALIZATION. - - * include/makeinclude/platform_aix4_cset++.GNU: Delete the - testAIXCompilerVersion.cpp file after getting the compiler version. - -Thu Jul 15 07:33:30 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/config-tru64.h,config-osf1.h,config-osf1-4.0.h, - include/makeinclude/platform_tru64_cxx.GNU,platform_osf1_cxx.GNU, - ACE-INSTALL.html: moved config-osf1.h to config-tru64.h, - and platform_osf1_cxx.GNU to platform_tru64_cxx.GNU. - Thanks to Doug Anderson <dla@home.com> for this suggestion. - -Wed Jul 14 22:23:10 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/config-cray.h: replaced ACE_HAS_TEMPLATE_INSTANTIATION_PRAGMA - with ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION. The standard - "template class" explicit instantiations work with version - 3.2.0.1 of the compiler. Thanks to Doug Anderson <dla@home.com> - for reporting this. - - * ace/config-ghs-common.h: use - ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION with Green Hills 1.8.9 - and later. Thanks to Carlos for suggesting that we give it - whirl. It worked :-) - -Wed Jul 14 22:03:08 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/config-osf1-4.0.h: enable pragma message disable declbutnotref - with Digital Unix 4.0f, as well as 5.0. Thanks to - Doug Anderson <dla@home.com> for reporting that - it's necessary. - - * ACE-INSTALL.html: updated config/platform info for - Digital UNIX 4.0f (and 5.0). - - [Bug 65] - -Wed Jul 14 19:44:21 1999 Steve Huston <shuston@riverace.com> - - * ace/Thread_Manager.cpp (spawn_i): Defer acquiring the thread - descriptor's lock until just before attempting to start the thread - to avoid places where the thread descriptor could be destroyed - while holding the lock (doing that is an undefined operation in - pthreads, and has particularly bad results on HP-UX 11). If the - thread creation does fail, release the lock before returning - the error indication (taking care to guard errno). Removed an - unreachable block of code which put the thread descriptor on the - wrong list anyway. - Removed the #if 1, #else, #endif around the auto_ptr for - new_thr_desc - if it's still working this close the ACE 5 release, - it wins. Besides, the use of the auto_ptr was ingrained in the - rest of the function, without the #if checks. - -Wed Jul 14 19:03:57 1999 David L. Levine <levine@cs.wustl.edu> - - * etc/purify.solaris2: added another UMR suppression for - __sigprocmask in the Solaris 2.6 threads library. Thanks - to Andy for reporting this. - -Wed Jul 14 16:30:48 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Hash_Map_Manager_T.h (class ACE_Hash_Map_Bucket_Iterator): - Added a new class that allows the iteration over the entries of - a of a particular bucket in a hash map. - - * tests/Hash_Map_Bucket_Iterator_Test.cpp: A test for the new - ACE_Hash_Map_Bucket_Iterator class. - -Wed Jul 14 16:37:33 1999 Ossama Othman <othman@cs.wustl.edu> - - * Makefile: - - Added Makefile.am and NEWS to the list of controlled files. The - file `NEWS' will get generated when a release is made. - - * TODO: - * acconfig.h: - * configure.in: - * configure: - - Minor ACE Configuration Project related updates and fixes. - -Wed Jul 14 13:38:12 1999 Steve Huston <shuston@riverace.com> - - * tests/Cached_Accept_Conn_Test.cpp: Moved defs of Client_Svc_Handler - and Server_Svc_Handler classes to new file, Cached_Accept_Conn_Test.h - to make AIX xlC happy. - - * tests/Cached_Conn_Test.cpp: Moved definition of Svc_Handler class - new file, Cached_Conn_Test.h to make AIX xlC happy. - - * tests/Cached_Accept_Conn_Test.cpp, Cached_Conn_Test.cpp: Don't - even try to build these on AIX C Set++ 3.1. It can't hack the - auto template instantiation - overwrites a template generator - file and tosses its cookies. ACE is still the best C++ compiler - test on the planet. - -Wed Jul 14 16:24:38 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/CDR_Stream.h: - * ace/CDR_Stream.cpp: - Added new method to efficiently steal the contents of a CDR - stream and store them in another. - -Wed Jul 14 12:05:13 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/rules.local.GNU (LN_S): on WIN32 - hosts, use cp -p instead of ln -s, because symbolic - links aren't supported. - -Wed Jul 14 11:22:31 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_vxworks5.x_g++.GNU, - ACE_INSTALL.html: with egcs (Tornado II) and without - exception handling, add -Wno-uninitialized to CCFLAGS. - The version of egcs that is provided with Tornado II - produces some spurious warnings about uninitialized - variables (including "this" :-). This was a problem - in older versions of egcs, but has since been fixed. - To avoid the warning, we always disable the warning - with VxWorks' egcs. - -Wed Jul 14 11:20:31 1999 Steve Huston <shuston@riverace.com> - - * ace/config-aix-4.x.h: Add ACE_LACKS_SETSCHED for pre-4.3 AIX. - -Wed Jul 14 11:15:24 1999 David L. Levine <levine@cs.wustl.edu> - - * examples/Shared_Malloc/Malloc.cpp: only explicitly - instantiate ACE_Read_Guard<ACE_SYNCH_MUTEX> and - ACE_Write_Guard<ACE_SYNCH_MUTEX> ifndef ACE_HAS_THREADS. - With threads, they're explicitly instantiated in - ace/Synch.cpp. - -Wed Jul 14 10:58:30 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * ace/CDR_Stream.h: - * ace/CDR_Stream.cpp: - Clarified the relationship between this implementation and the - original SunSoft version. - -Wed Jul 14 10:47:45 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_qnx_neutrino.GNU: - added ACE_HAS_GNUG_PRE_2_8 := 0, so that exception - handling will be disabled by default. - -Wed Jul 14 10:34:37 1999 David L. Levine <levine@cs.wustl.edu> - - * examples/IPC_SAP/SOCK_SAP/FD-unclient.cpp: fixed typo, - ACE_LACKS_UNIX_DOMAIN_SOCKETS instead of - ACE_HAS_UNIX_DOMAIN_SOCKETS. - -Wed Jul 14 10:22:29 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/Singleton.*: added ACE_Unmanaged_Singleton and - ACE_Unmanaged_TSS_Singleton template class. They're - identical to ACE_Singleton and ACE_TSS_Singleton, - respectively, except that they don't register for - cleanup with the ACE_Object_Manager. Therefore, - the application is responsible for explicitly destroying - the instance after it is no longer needed (if it wants to - avoid memory leaks, at least). The close () static member - function must be used to explicitly destroy the Singleton. - Thanks to Valery Arkhangorodsky <valerya@servicesoft.com> - for suggesting this and providing the code. - -Wed Jul 14 10:01:47 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/OS.h (ACE_{,U_}ONE_SECOND_IN*): made the trailing - L or U upper case, for readability. Thanks to Chris - for noticing this. - -Wed Jul 14 09:40:37 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/wrapper_macros.GNU: when setting - ACE_HAS_GNUG_PRE_2_8, if not g++: don't override - the value to null if it was non-null. Also, clarified - use of ACE_HAS_GNUG_PRE_2_8: 1 for g++ < 2.8.0, - 0 for g++ >= 2.8.0, undefined for non-g++. - -Wed Jul 14 09:29:57 1999 David L. Levine <levine@cs.wustl.edu> - - * tests/Makefile (realclean): remove log/compilations.log*. - -Tue Jul 13 22:25:18 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * include/makeinclude/wrapper_macros.GNU: - Disable the "unitialized" warnings with EGCS+exceptions, it - gives too many bogus warnings to take care of them. - -Tue Jul 13 22:16:17 1999 Carlos O'Ryan <coryan@cs.wustl.edu> - - * bin/ACEutils.pm: - Wait for non empty files, sometimes the process creates the - files but does not have enough time to fill its contents. - -Tue Jul 13 21:32:44 1999 Ossama Othman <othman@cs.wustl.edu> - - * acconfig.h: - * ltconfig: - * ltmain.sh: - * ace/Makefile.am: - * tests/Makefile.am: - - ACE Configuration Project updates. Synchronized with current - version of ACE in CVS repository. - -Tue Jul 13 15:22:51 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_vxworks5.x_g++.GNU: - set ACE_HAS_GNUG_PRE_2_8 to 0 with egcs (Tornado II). - -Tue Jul 13 15:15:43 1999 Steve Huston <shuston@riverace.com> - - * ace/ace-dll.icc, ace-lib.icc: Corrected source file set; now builds - (still with some warnings) on AIX, Visual Age C++ 4. - -Tue Jul 13 15:11:28 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * bin/run_all_win32.pl: Fixed program output redirection problem. - Unfortunately, the fix requires one to have either 'sh' or '4nt' - to redirect output from another perl script correctly. - -Tue Jul 13 15:04:35 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_qnx_neutrino.GNU: - removed full paths from commands, because they're - prefixed with i386-nto-. - - * ace/config-qnx-neutrino.h: updated to build with egcs. - -Tue Jul 13 15:03:01 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/Pipe.cpp (open): moved declaration of local variable - "one" inside the ACE_LACKS_TCP_NODELAY #ifndef block. - -Tue Jul 13 14:57:56 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/SV_Semaphore_Simple.{h,cpp} (init,open,name_2_key): - added static cast of ACE_INVALID_SEM_KEY to key_t. key_t - is unsigned on QNX Neutrino, but we want to keep - ACE_INVALID_SEM_KEY as -1. - -Tue Jul 13 11:57:53 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/config-win32.h: Changed CApplication to CWinApp. Thanks to - John Morey <jmorey@tbi.com> for pointing this out. - -Tue Jul 13 11:13:52 1999 Steve Huston <shuston@riverace.com> - - * ACE_INSTALL.html: Added some info for building on AIX with - Visual Age C++. - - * ace/config-aix-4.3.x.h: New file - these version-specific configs - are particularly useful with Visual Age C++, which has its own - funky make system that doesn't use the platform_macros.GNU files, - and so doesn't pick up the ACE_AIX_MAJOR/MINOR_VERS settings which - are used in config-aix-4.x.h. - - * ace/config-aix-4.x.h: Corrected template settings for Visual Age C++ - and added ACE_HAS_TYPENAME for VAC++ also. - -Tue Jul 13 10:40:31 1999 Irfan Pyarali <irfan@cs.wustl.edu> - - * ace/Cached_Connect_Strategy_T.cpp (purge_connections): There is - no need to set the purge percentage again. - -Tue Jul 13 10:39:57 1999 Matthew J Braun <mjb2@cs.wustl.edu> - - * ace/Containers_T.h,i: Altered private member list_ of - DLList_Iterator to be a pointer rather than a reference. This - allows the reset method to work because it was overwriting the - contents of list_ rather than iterating over a different list - -Tue Jul 13 09:14:31 1999 Steve Huston <shuston@riverace.com> - - * ace/Cached_Connect_Strategy_T.cpp (dtor): Added ACE_TYPENAME - to 'iter' declaration since its type depends on template parameters. - -Tue Jul 13 09:11:24 1999 David L. Levine <levine@cs.wustl.edu> - - * include/makeinclude/platform_vxworks5.x_ghs.GNU: - moved LIBS setting for Green Hills 1.8.9 from - x86-only to all targets. Thanks to Dale Wood - <dale_wood@Mitel.com> for confirming that this - is necessary for PowerPC targets. - -Tue Jul 13 08:18:56 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/config-osf1.h,config-osf1-4.0.h, - include/makeinclude/platform_osf1_cxx.GNU: - added another significant figure to the version identifer, - e.g., 0x40F and 0x500. [Bug 65] - -Tue Jul 13 06:45:53 1999 Steve Huston <shuston@riverace.com> - - * examples/Shared_Malloc/test_position_independent_malloc.cpp: Moved - definition of Dummy_Data to test_position_independent_malloc.h, a - new file, to satisfy AIX IBM C++ template instantiation. - - * include/makeinclude/platform_aix4_cset++.GNU: Turn on/off warnings - based on compiler version, not OS version. Thanks to Craig - Rodrigues <rodrigc@mediaone.net> for sending the code in! - - * THANKS: Added Craig Rodrigues to the hall of fame. - - * ace/config-aix-4.1.x.h, config-aix-4.2.x.h: Added setting for - ACE_AIX_MAJOR_VERS and ACE_AIX_MINOR_VERS if they weren't set - already. Eases backwards compatibility for those who don't use - the ACE build macros (horrors! :-). Thanks to Jeff Donner - <jdonner@schedsys.com> for this suggestion. - -Mon Jul 12 18:24:34 1999 Kirthika Parameswaran <kirthika@cs.wustl.edu> - - * tests/Cached_Accept_Conn_Test.cpp: - This tests explicit purging from the server's point of view. - When the acceptor in either its accept_svc_handler () or open () - runs out of descriptors, purging of unused svc_handlers from the - connecton cache is requested. - - * tests/Cached_Conn_Test.cpp: - Defined ACE_HAS_BROKEN_EXTENDED_TEMPLATES for the - Caching_Strategy_Adapter. If this condition is true, then only - Least Recently Used Caching Strategy is available. This was - necessary for compilers which dont accept long template lists, - for example GNUC version < 8.0. Total credit for this hack - to get through the old GNUC compiler goes to Irfan - <irfan@cs.wustl.edu> of the DOC group. - - * tests/Makefile: - Added Cached_Accept_Conn_Test.cpp. - - * tests/Makefile: - * tests/run_tests.sh: - * tests/run_pharlap_tests.bat: - * tests/run_tests.bat: - * tests/run_tests.psosim: - * tests/run_tests.vxworks: - Added Cached_Conn_Test, Map_Manager_Test and - Cache_Map_Manager_Test to the above. - -Mon Jul 12 17:31:28 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * bin/run_all_win32.pl: Fixed the missing execution output problem - and simplified how the script gets invoked. See comments for - more details. - - * ace/config-win32.h: - * ace/Object_Manager.h: Updated some unclear comments regarding - the use and instantiation of non-static object manager. Thanks - to John Morey and David for pointing this out. - - * ace/ace_dll.dsp: Merged additional MFC configuration. Thanks to - John Morey <jmorey@tbi.com> for submitting the patch. - -Mon Jul 12 15:27:37 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/config-osf1.h,config-osf1-4.0.h, - include/makeinclude/platform_osf1_cxx.GNU, - ACE-INSTALL.html: added support for DU 4.0f. It needs the - same #defines that we use on DU 5.0. [Bug 65] - -Mon Jul 12 14:41:14 1999 David L. Levine <levine@cs.wustl.edu> - - * ace/config-win32.h,Object_Manager.h: cleaned up comments - about the non-static ACE_Object_Manager. Thanks to - John Morey <jmorey@tbi.com> for reporting this. - -Mon Jul 12 13:54:34 1999 Kirthika Parameswaran <kirthika@cs.wustl.edu> - - * ace/Cache_Map_Manager_T.{h,i,cpp}: - Defined ACE_T1 and ACE_T2 which contain the template parameter - list and added ACE_HAS_BROKEN_EXTENDED_TEMPLATES macro where - needed. This takes care of the template list for compilers which - cant accept a long template list, for example: GCC 2.7.2. - Also changed the purge() to now call the <caching_utility> for - purging entries form the cache. - - * ace/Hash_Cache_Map_Manager_T.h: - Added ACE_HAS_BROKEN_EXTENDED_TEMPLATES to complement the - Cache_Map_Manager class changes. - - * ace/Cached_Connect_Strategy_T.cpp: - Changed dtor to take care of the - ACE_HAS_BROKEN_EXTENDED_TEMPLATES definition. - - * ace/Cached_Connect_Strategy_T.h: - Removed CONNECTION_CACHE_ITERATOR typedef. - - * ace/Caching_Strategies_T.{h,i,cpp}: - Changed template parameter list and removed dependency on the - CONTAINER. Removed clear_cache() method. - - * ace/Caching_Utility_T.{h,i,cpp}: - Added default cleanup strategies to every kind of - caching_utiltity. ACE_Recyclable_Hanlder_Caching_Utility no - longer derives form ACE_Pair_Caching_Utility. - - * ace/Cleanup_Strategies_T.{h,i,cpp}: - Changed ACE_Svc_Cleanup_Strategy to - ACE_Recyclable_Handler_Cleanup_Strategy. - - * tests/Cache_Map_Manager_Test.cpp: - * tests/Map_Manager_Test.cpp: - Made changes to template instantiations. - - * tests/Cached_Conn_Test.cpp: - Added out_of_sockets_handler() which will do the purging. - Added a check for ENOBUFS which arose on NT and ENOENT which - arose on Linux + Solaris. Defined - ACE_HAS_BROKEN_EXTENDED_TEMPLATES for GCC versions < GCC 2.8. - -Mon Jul 12 12:14:45 1999 David L. Levine <levine@cs.wustl.edu> - - * docs/ACE-development-process.html: added this page, as - a start towards a less chaotic development process. - -Mon Jul 12 10:05:31 1999 Steve Huston <shuston@riverace.com> - - * ace/Event_Handler_T.i: Added ACE_TYPENAME on template-dependent types - in function signatures. Quiets HP-UX aC++. - - * ace/Based_Pointer_Repository.h: - * ace/Based_Pointer_T.h: Added #include "ace/OS.h" to pick up def. of - ACE_Export when instantiating templates with AIX IBM C++. - - * include/makeinclude/platform_hpux_aCC.GNU: Removed definition for - CPP_LOCATION - use aCC - cpp has too many problems, and a fix for - the pragma warnings was checked into TAO_IDL. - -Mon Jul 12 09:26:29 1999 David L. Levine <levine@cs.wustl.edu> - - * testsr/Makefile: don't build libDLL_Test.so if not building - shared libraries. This simplifies the build on static-libs - only platforms, and can be used by run_tests.sh, below. - [Bug 62] - - * tests/run_tests.sh: don't run DLL_Test if libDLL_Test.so - was not built. - -Mon Jul 12 07:48:04 1999 David L. Levine <levine@cs.wustl.edu> - - * docs/ACE-guidelines.html: updated guideline under CVS - section to match the other one about always testing - on egcs. - -Sun Jul 11 23:59:52 1999 nanbor <nanbor@cs.wustl.edu> - - * ACE version 4.6.44 released. - -Sat Jul 10 18:05:41 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * ace/config-kcc-common.h: Added ACE_NEW_THROWS_EXCEPTIONS. - Thanks to David for pointing this out. - -Sat Jul 10 16:22:00 1999 Chris Gill <cdgill@tango.cs.wustl.edu> - - * ace/config-win32-common.h: wrapped ACE_HAS_TYPENAME_KEYWORD so it is - only set for MSVC++ 6.0 or higher. This is to work around a compiler - bug in MSVC++ 5.0 when trying to compile certain forms with the - typename keyword in a template method argument list. - -Sat Jul 10 13:16:50 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - - * ace/Malloc: Changed the type of pointer used by ACE_Name_Node - from void * to char * so that we can integrate this cleanly with - ACE_Based_Pointer<>, which will fail if we use void since we - can't return from a "void" function! - - * tests/Thread_Pool_Test.cpp (open): Fixed another somewhat bogus - warning from EGCS related to initialization of loop variables. - - * tests/New_Fail_Test.cpp (main): Make sure to initialize pointers - so that EGCS doesn't complain. - - * tests/MT_Reactor_Timer_Test.cpp (main): Fixed a signed/unsigned - mismatch. - - * tests: Added appropriate header comments for all the *.h files. - - * tests/Cache_Map_Manager_Test.cpp: Capitlized the hash_key - class to be Hash_Key, which is standard ACE programming style. - - * ace/Service_Config.cpp: Added yet another set of - changes to open_i() so that the -d option is handled - appropriately. Thanks to Jonathan Reis - <reis@minniemouse.cemax.com> for this fix. - - * ace/SOCK_Acceptor.h: Clarified the parameters for - ACE_SOCK_Acceptor's constructor and open() methods. Thanks to - Ron MacKenzie <ronm@SLAC.Stanford.EDU> for motivating this. - -Sat Jul 10 13:56:30 1999 Nanbor Wang <nanbor@cs.wustl.edu> - - * tests/MT_Reactor_Timer_Test.cpp (main): Fixed signed/unsigned - comparison. +Sat Jul 17 11:17:18 1999 Alexander Babu Arulanthu <alex@cs.wustl.edu> + + * tao/CDR_Stream.cpp + * tao/CDR_Stream.h : Integrated Carlos's memory leak fixes from the main branch. Sat Jul 10 08:01:36 1999 David L. Levine <levine@cs.wustl.edu> @@ -2958,7 +1525,8 @@ Sun Jun 27 18:03:24 1999 Douglas C. Schmidt <schmidt@danzon.cs.wustl.edu> Sun Jun 27 20:50:02 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - Added support for IBM's VisualAge C++ compiler. + Added support for IBM's VisualAge C++ compiler. Thanks to Rene + Matteau <matteau@ca.ibm.com> for contributing the following fixes: * ace: Added a new config-visualage.h file that automatically includes the proper configurations for IBM's VisualAge C++ @@ -3437,7 +2005,7 @@ Thu Jun 24 22:42:44 1999 David L. Levine <levine@cs.wustl.edu> preprocessor usage so that it's not in the middle of the ACE_reintepret_cast. -Thu Jun 24 22:18:52 1999 Doug Anderson <dla@home.com> +Thu Jun 24 22:18:52 1999 Doug Anderson <doug@clark.net> * ace/CDR_Stream.*,ace/SOCK_Dgram_Mcast.cpp,ace/config-cray.h: updates for Cray UNICOS port. @@ -6923,7 +5491,8 @@ Tue Apr 20 13:55:34 1999 Joe Hoffert <joeh@cs.wustl.edu> Tue Apr 20 13:30:51 1999 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> * ace/ace.icc: Added a configuration file for creating a shared - library with Visual Age C++. + library with Visual Age C++. Thanks to Rene Matteau + <matteau@ca.ibm.com> for this. Mon Apr 19 23:50:25 1999 Irfan Pyarali <irfan@cs.wustl.edu> @@ -9230,7 +7799,7 @@ Wed Mar 10 11:05:39 1999 David L. Levine <levine@cs.wustl.edu> it was missing. * include/makeinclude/platform_osf1_4.0.GNU: disable msg 1136 with - cxx 6.1-029. Thanks to Doug Anderson <dla@home.com> for + cxx 6.1-029. Thanks to Doug Anderson <doug@clark.net> for reporting this. Wed Mar 10 10:57:16 1999 Carlos O'Ryan <coryan@cs.wustl.edu> @@ -11215,7 +9784,7 @@ Fri Jan 29 14:49:37 1999 Carlos O'Ryan <coryan@cs.wustl.edu> * tests/Message_Queue_Test.cpp: * tests/SV_Shared_Memory_Test.cpp: * tests/run_tests.sh: - Thanks to Doug Anderson <dla@home.com> for this port of ACE to + Thanks to Doug Anderson <doug@clark.net> for this port of ACE to Cray machines. Fri Jan 29 13:51:40 1999 Nanbor Wang <nanbor@cs.wustl.edu> |