diff options
Diffstat (limited to 'ACE/ChangeLogs/ChangeLog-06b')
| -rw-r--r-- | ACE/ChangeLogs/ChangeLog-06b | 8927 |
1 files changed, 0 insertions, 8927 deletions
diff --git a/ACE/ChangeLogs/ChangeLog-06b b/ACE/ChangeLogs/ChangeLog-06b deleted file mode 100644 index 2212ce3f6f0..00000000000 --- a/ACE/ChangeLogs/ChangeLog-06b +++ /dev/null @@ -1,8927 +0,0 @@ -Fri Nov 17 07:06:10 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/msvc_static_order.lst: - Added another RTScheduling example library project - -Fri Nov 17 06:47:10 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - Updated this file because tao_idl doesn't generate a S_T.inl - file anymore - -Thu Nov 16 20:38:10 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/msvc_static_order.lst: - Added new RTScheduling example library projects to fix compile - errors in the msvc static builds - -Wed Nov 15 23:23:10 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * protocols/Makefile.am: - - Add lost support for --disable-ace-tests and --disable-ace-examples - -Wed Nov 15 21:55:01 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * ace/Makefile.am: - - Add OS.inl to nobase_include_HEADERS. - -Wed Nov 15 19:08:39 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * ace/config-lite.h: - - Changed to not include <Cstd/stdcomp.h> on Solaris systems - when using the stlport4 C++ library. - -Wed Nov 15 15:33:10 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * apps/JAWS/clients/Caching/caching.mpc: - - Add exename definition so that an executable, rather than - a shared library, is built. - -Wed Nov 15 10:35:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * NEWS: - - Added entry for Iliyan Jeliazkov. - -Wed Nov 15 01:56:28 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * protocols/ace/HTBP/Makefile.am: - * protocols/ace/RMCast/Makefile.am: - * protocols/ace/TMCast/Makefile.am: - - Change includedir to omit the "protocols" subdirectory. - - The header files for these libraries are intended to be - installed in ${prefix}/include/ace/{HTBP,RMCast,TMCast} - not ${prefix}/include/protocols/ace/{HTBP,RMCast,TMCast}. - -Wed Nov 15 01:11:02 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * ace/Makefile.am: - - Add Numeric_Limits.h to nobase_include_HEADERS. - -Tue Nov 14 18:42:18 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.inl: - * ace/Service_Gestalt.cpp: - - After a gestalt gets close()-ed it is still a valid instance - however, because it is disassociated from its service repository - the situation becomes another special case the gestalt - implementation must account for. This change is in addition to - the bug fix introduced with the change from "Thu Nov 9 20:07:42 - UTC 2006". - -Tue Nov 14 14:30:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Select_Reactor_T.cpp: - Removed some not used macros that did cause problems when running - FOCUS - -Tue Nov 14 11:23:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Timer_Queue_T.cpp: - Added hook for FOCUS, const improvement - -Tue Nov 14 07:56:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: - Updated rtems information - -Mon Nov 13 00:17:19 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.cpp: - - Calling close() on a gestalt must not obliterate any static - service registrations. Those may still be needed if the gestalts - gets re-opened again, before its destructor runs. This change - fixes a problem (introduced with "Thu Nov 9 20:07:42 UTC 2006") - which broke this invariant. - -Sat Nov 11 18:37:12 UTC 2006 Chris Cleeland <cleeland_c@ociweb.com> - - * ace/Service_Gestalt.cpp (find_processed_static_svc): - - Added a check at the beginning of this function that drops out - if processed_static_svcs_ or the arg is a zero pointer. This - gets ride of core dumps in scoreboard builds running tests like - some of the TAO EC tests (basically anything which has a static - service defined and expects to find it), but there is still the - more fundamental problem as to why the static services aren't - there in the first place. - -Fri Nov 10 22:30:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * include/makeinclude/platform_vxworks6.3.GNU: - - Fixes for VxWorks 6.3 PPC builds. - -Fri Nov 10 21:26:41 UTC 2006 Chris Cleeland <cleeland_c@ociweb.com> - - * ace/OS_NS_Thread.h (THR_HAS_WTHREADS): - - Added THR_EXPLICIT_SCHED definition to ACE_HAS_WTHREADS clause - so that application-level code doesn't have to have #ifdefs to - figure out whether it's okay to use THR_EXPLICIT_SCHED. - - This change is part of a change to TAO from - Thu Nov 9 15:42:32 UTC 2006 Chris Cleeland <cleeland_c@ociweb.com> - -Fri Nov 10 15:02:26 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.cpp (close): - - Removing a forgotten debug statetement. - -Fri Nov 10 13:42:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * bin/vx_repo_link.pl: - - New perl 'program' to handle link phase for VxWorks 5.5.x - builds using -frepo option. - - * include/makeinclude/platform_vxworks5.5.x.GNU: - * include/makeinclude/rules.lib.GNU: - - Changes to enable building ACE/TAO apps for VxWorks 5.5.x SIMNT - targets using -frepo option. - -Fri Nov 10 12:53:32 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/make_release: - * docs/bczar/bczar.html: - - Minor update to bczar documentation to correct install paths, - minor bugfix to release script to correct PRF generation. - -Fri Nov 10 12:22:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/INET_Addr.cpp: - Fixed argument not used warnings in rtems no network build - -Fri Nov 10 10:06:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Reactor_Impl.h: - * ace/Dev_Poll_Reactor.{h,cpp}: - * ace/Select_Reactor_T.{h,cpp}: - * ace/WFMO_Reactor.{h,inl}: - Changed initialized to return a bool, that is the type we store - as member. - -Fri Nov 10 10:03:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_unistd.cpp: - Const changes and use prefix operators instead of postfix - - * ace/Process.cpp: - Use prefix increment instead of postfix - - * ace/Reactor.cpp (timer_queue): - No need to use an intermediate variable - -Fri Nov 10 05:47:15 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * docs/bczar/bczar.html - - Updates to release script documentation. - -Thu Nov 9 20:07:42 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.cpp: - * ace/Service_Gestalt.h: - * ace/Service_Gestalt.cpp: - - A follow up to my change from "Wed Nov 8 22:12:46". Adding the - close_svcs() call in ACE_Service_Config::close() ensured that - the repository and its contents gets properly finalized, but it - did not account for the potential subsequent re-opening of the - Service Config. This causes failures in open/close/open - scenarios, like Framework_Component_Test. - -Thu Nov 9 19:59:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_aix4_cset++.GNU: - * include/makeinclude/platform_integrity_ghs.GNU: - * include/makeinclude/platform_irix6.x_sgic++.GNU: - * include/makeinclude/platform_linux_pgi.GNU: - * include/makeinclude/platform_macosx_panther.GNU: - * include/makeinclude/platform_macosx_tiger.GNU: - * include/makeinclude/platform_netbsd.GNU: - * include/makeinclude/platform_openbsd.GNU: - * include/makeinclude/platform_osf1_4.x_rcc.GNU: - * include/makeinclude/platform_vxworks5.x_ghs.GNU: - Simplified these files using the ?= operator - -Thu Nov 09 16:10:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/Service_Config.cpp: - - Reverted my earlier change of Nov 6 and additionally removed - the ACE_Reactor::instance() call from the ACE_Service_Config - CTOR since it seems to be unnecessary here. - -Thu Nov 9 15:22:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/wrapper_macros.GNU: - Default threads to 1 if not set, thanks to Chad for helping out - with finding out why I had a problem in the rtems build - -Thu Nov 9 12:46:54 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * bin/MakeProjectCreator/config/acedefaults.mpb: - * bin/MakeProjectCreator/config/taodefaults.mpb: - - Add automake specific section to subtract out $(ACE_ROOT)/lib - from libpaths, as libraries are not copied to, or linked from, - there in automake builds. This removes unnecessary -L's from - being added to *_LDFLAGS, sometimes eliminating the need for - that automake variable altogether. - -Thu Nov 9 10:48:52 UTC 2006 Olli Savia <ops@iki.fi> - - * examples/IPC_SAP/DEV_SAP/reader/reader.cpp: - * examples/IPC_SAP/DEV_SAP/writer/writer.cpp: - Changed used baudrate to 19200. Extended reader example to handle - timeout. - -Thu Nov 9 10:50:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Timer_Heap_T.{h,cpp}: - Changed the preallocated constructor from int to bool. - -Thu Nov 9 09:25:48 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/TTY_IO.cpp: - Removed LynxOS 3.x specific workaround. - -Wed Nov 8 22:22:27 UTC 2006 Steve Huston <shuston@riverace.com> - - * bin/make_release: Re-applied this: - Sat Oct 28 20:10:42 UTC 2006 Steve Huston <shuston@riverace.com> - -Wed Nov 8 22:12:46 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.cpp: - * ace/Service_Gestalt.cpp: - - This resolves a problem with shutting down the Service - Configurator via the static ACE_Service_Config::close () method, - where the registered services' fini() methods were not being - called. For full details, see bugzilla# 2701. Thanks to Aaron - Scamehorn <Aaron dot Scamehorn at cogcap dot com> for discovering - it an providing a test case. - -Wed Nov 8 19:21:17 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/make_release: - - Updates to the release script to work with Subversion. - -Mon Nov 6 16:07:43 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/SOCK_IO.h (ACE_SOCK_IO): Clarified the behavior of - non-blocking I/O. Thanks to Robert Neumann <robert dot neumann - at gmail dot com> for motivating this change. - -Tue Nov 08 15:40:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/INET_Addr.inl: - - Enhanced operator <() to support IPv6 addresses as well. - Thanks to Andre Kostur (akostur at incognito dot com) for - submitting a patch for this. - -Wed Nov 8 14:24:02 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * configure.ac: - - Change ACE_LACKS_NET_IF_H feature test to #include <sys/types.h> - (if present). - -Wed Nov 8 13:45:07 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * examples/C++NPv2/C++NPv2.mpc: - - The AC_CLD project requires threads in order to build and run - correctly. - -Tue Nov 07 15:49:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/OS_NS_netdb.inl: - - Changes to accomodate VxWorks 6.x versions of gethostbyname and - gethostbyaddr. These are threadsafe as the x_r versions are but - require deallocating the returned hostent structure. - -Tue Nov 7 15:44:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Sock_Connect.cpp: - With VxWorks 6.3 in rtp mode with IPv6 enabled declare the - in6addr_ constants explicitly, these are exported by the - VxWorks 6.3 rtp header files but not implemented - -Tue Nov 07 13:31:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * include/makeinclude/platform_vxworks6.3.GNU: - - Fixed build settings for PPC32 builds with softfloat. - -Tue Nov 7 10:38:25 UTC 2006 Simon McQueen <sm@prismtech.com> - - * tests/Naming_Test.cpp: - - Moved an ::open out of an ACE_ASSERT. It wasn't getting performed on - some builds. - -Tue Nov 7 09:56:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * performance-tests/Misc/basic_func.{h,cpp}: - Renamed B to BB to resolve compile problems on rtems where we - have also a typedef B - -Tue Nov 7 07:52:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/ACE.h: - Also rtems defines log2, use the same workaround as with cygwin - - * include/makeinclude/platform_rtems.x_g++.GNU: - Removed -ansi - -Tue Nov 7 01:47:46 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Sock_Connect.cpp (count_interfaces): - Another scoreboard fix. - -Mon Nov 6 20:52:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * apps/Gateway/Gateway/Config_Files.{h,cpp}: - * apps/Gateway/Gateway/File_Parser.{h,cpp}: - * apps/Gateway/Gateway/Gateway.cpp: - Renamed class FP to FPRT. This resolves the compile errors on - rtems where a struct FP does exist - -Mon Nov 6 16:52:14 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Sock_Connect.cpp: - My fix for Solaris x86 broke HPUX. The trouble is that the - method ACE::count_interfaces() is only used by the "generic" - implementation of get_ip_interfaces(). Probably a better - solution is to remove count_interfaces() and move its - functionality into the get_ip_interfaces() implementation, but - that is a bigger change than I care to make at the moment. I - think it should be possible to do so in a way that also reduces - the footprint and improves performance by eliminating the - effectively duplicated code in count_interfaces(). - -Mon Nov 06 14:14:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/config-vxworks6.3.h: - - Fixed incorrect definition of IOV_MAX. - -Mon Nov 6 13:08:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-sunos5.5.h: - Added ACE_HAS_GETRUSAGE, thanks to Michael Klein - <michael dot klein at fazi dot de> for reporting that SunOS has - this method, fixes bugzilla 2694. - -Mon Nov 6 12:59:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Service_Configurator/Misc/main.cpp: - * examples/Service_Configurator/Misc/Service_Configurator_Misc.mpc: - * examples/Service_Configurator/Misc/svc.conf1 - Fixed shared library name, make working for embedded platforms. - Thanks to Abdul Sowayan <abdullah dot sowayan at lmco dot com> - for reporting this, fixes bugzilla bug 2698 - -Mon Nov 06 11:17:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/Service_Config.cpp: - - Implemented a fix for runtime problems in case - ACE_HAS_NONSTATIC_OBJECT_MANAGER is defined (seen on VxWorks 6.3 RTP). - -Sun Nov 5 14:14:51 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Sock_Connect.cpp: - Fix for Solaris x86. - -Sat Nov 4 12:46:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-sunos4*.h: - Removed these files, these are ancient and not supported anymore - -Fri Nov 3 18:12:36 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Unload_libACE.cpp (main): Reverted the ACE_OS::exit() from - this since it's not supposed to be an ACE test. Thanks to Simon - Massey for reporting this. - -Fri Nov 03 15:52:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * bin/PerlACE/ProcessVX_Unix.pm: - - Bugfix in cmdline composition. - -Fri Nov 3 15:15:43 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Sock_Connect.cpp: - Scoreboard cleanup. - -Fri Nov 03 10:54:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/OS_NS_sys_time.inl: - - Fixed conversion warning on OpenVMS. - -Fri Nov 3 00:39:04 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Sock_Connect.cpp: - One more optimization, this one to combine the methods for - evaluating the IPv6 or IPv4 enabled status. - -Fri Nov 3 00:32:27 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Select_Reactor_Base.inl: - * ace/Select_Reactor_Base.cpp: - Additional optimization for finding and unbinding event - handlers. The solution moves the test for handles in range into - the top-level find and unbind methods to allow the fix for bug - 2653 to continue to work in unbind while avoiding a double check - for handles in range during a find. - -Thu Nov 2 20:33:55 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/SOCK_Test_IPv6.cpp (spawn): - * tests/Pipe_Test.cpp (run_main): - * tests/SPIPE_Test.cpp (client): - * tests/TkReactor_Test.cpp (run_main): - * tests/MT_SOCK_Test.cpp (spawn): - * tests/Process_Strategy_Test.cpp (run_main): - * tests/Unload_libACE.cpp (main): - * tests/SOCK_Test.cpp (spawn): - * tests/Priority_Reactor_Test.cpp (run_main): - * tests/SOCK_Dgram_Test.cpp (spawn): - * tests/SOCK_SEQPACK_SCTP_Test.cpp (spawn_test): - * examples/Map_Manager/test_hash_map_manager.cpp (ACE_TMAIN): - * apps/mkcsregdb/mkcsregdb.cpp (fail): Replaced exit() with - ACE_OS::exit(). - -Wed Nov 01 20:04:00 US Central Time 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * Bronson Royce Schmidt checked in at 7 pounds, 15 ounces and - 21.25 inches long. - -Thu Nov 2 16:10:10 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Sock_Connect.cpp: - Cleaned up debugging code left in my previous premature commit. - -Thu Nov 2 15:42:29 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Sock_Connect.cpp: - * ace/Sock_Connect.h: - Added support for IPv6 detection on HPUX, AIX, Solaris, WinCE, - and VxWorks 5.5.x. - -Wed Nov 01 09:05:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/OS_NS_stdlib.inl: - - Fixed conversion warnings in atop() on OpenVMS. - -Wed Nov 01 08:35:10 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/OS_NS_sys_stat.h: - - Fixed Borland compile failure over def of ACE_Stat. - -Tue Oct 31 23:08:32 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Dev_Poll_Reactor.h: Note that the Linux epoll code implementation - is supported, but the /dev/poll implementation is still - experimental. - -Tue Oct 31 10:31:10 UTC 2006 Ossama Othman <ossama_othman at symantec.com> - - * ace/OS_NS_sys_stat.h (ACE_WSTAT_FUNC_NAME): - - Corrected wstat() function name for Borland. - -Mon Oct 30 19:57:20 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * bin/ciao_tests.lst: - - Added run_test.pl in CIAO/tests/IDL3/Events/Any to the list of - CIAO tests to be run automatically. - -Mon Oct 30 17:34:35 UTC 2006 Ossama Othman <ossama_othman at symantec.com> - - * ace/config-win32-common.h: - - Use 64-bit file offsets by default in the WIN64 case, similar to - what 64-bit UNIX systems do. - -Mon Oct 30 17:23:15 UTC 2006 Steve Huston <shuston@riverace.com> - - * include/makeinclude/wrapper_macros.GNU: Added 2 new build options, - dmalloc and mtrace. When specified at build time (e.g. make mtrace=1) - the PLATFORM_DMALLOC_CPPFLAGS and/or PLATFORM_MTRACE_CPPFLAGS values - are added to CPPFLAGS. For dmalloc, the PLATFORM_DMALLOC_LDFLAGS and - PLATFORM_DMALLOC_LIBS are added to LDFLAGS and LIBS, respectively. - Also added the ability to specify additional purify and quantify - command-line options by setting PLATFORM_PURIFY_OPTIONS and - PLATFORM_QUANTIFY_OPTIONS, respectively. Thank you to Howard - Finer for supplying these additions. - -Mon Oct 30 16:56:28 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/MakeProjectCreator/templates/gnu.mpd: - - The implementation for the 'targetoutdir' template variable was - incorrect. Libraries and executables were not symbolically linked - to the output location correctly. Also, 'targetoutdir' was not - taken into account for the 'libpaths' setting. - -Mon Oct 30 15:30:47 UTC 2006 Simon McQueen <sm@prismtech.com> - - * bin/tao_orb_tests.lst: - - Scheduled regression test for bug #2669. - -Mon Oct 30 14:31:42 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Select_Reactor_Base.inl: - Reverting the functional part of my previous commit until I - learn the correct way to manage the upper bound test portably. - - * bin/tao_orb_tests.lst: - Make the Bug 2654 test be ignored for minimum CORBA builds. - -Mon Oct 30 12:18:03 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Select_Reactor_Base.inl (unbind): - Added an additional validity check and a comment to ward off - future optimizations. - -Mon Oct 30 09:24:42 UTC 2006 Ossama Othman <ossama_othman at symantec.com> - - * ace/OS_NS_sys_stat.h (ACE_stat): - * ace/OS_NS_sys_stat.inl (stat, wstat): - - Only use Windows 64-bit stat structures and functions when - _FILE_OFFSET_BITS=64 is defined. Addresses 64-bit conversion - warnings. - -Sun Oct 30 05:31:01 UTC 2006 Ossama Othman <ossama_othman at symantec.com> - - * ace/Auto_Ptr.h (ACE_auto_ptr_reset): - - Adjusted function template arguments to handle case where the - auto_ptr element type and pointer type passed to this function - are not the same. Addresses argument deduction failures. - - * ace/FILE.h (ACE_FILE): - - 64-bit file offset fixes. s/off_t/ACE_OFF_T. - - * ace/Truncate.h (Truncator): - - Added explicit specializations for int and long for Borland. - Relying on ACE_{U}INT32 specializations is not sufficient. - - * tests/CDR_File_Test.cpp (run_test): - - A 64-bit file offset is not always enabled. Corrected offset - type. s/ACE_LOFF_T/ACE_OFF_T. - -Sun Oct 29 17:10:39 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - Added Bug 2654 regression test to suite. - -Sat Oct 28 20:50:56 UTC 2006 Steve Huston <shuston@riverace.com> - - * examples/Threads/thread_specific.cpp: Need to reinterpret_cast a - pointer back to an intptr_t. - -Sat Oct 28 20:10:42 UTC 2006 Steve Huston <shuston@riverace.com> - - * bin/make_release: Added Smartphone configurations to the WinCE - project files generated at release time. - -Sat Oct 28 13:50:01 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Select_Reactor_Base.cpp: - - Fix to address Win32-specific problems in bug 2653 fix. - -Sat Oct 28 01:48:27 UTC 2006 Steve Huston <shuston@riverace.com> - - * include/makeinclude/wrapper_macros.GNU: Add -L$(INSLIB) to LDFLAGS - if it was not there already. Allows link to work when built libraries - are moved to locations other than the default $(ACE_ROOT)/lib - location. - -Fri Oct 28 03:23:18 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * apps/JAWS2/JAWS/FILE.cpp: - * apps/JAWS2/JAWS/FILE.h: - * apps/JAWS2/JAWS/IO.cpp: - * apps/JAWS3/jaws3/FILE.cpp: - * apps/JAWS3/jaws3/FILE.h: - * examples/Threads/thread_manager.cpp: - * examples/Threads/cancel.cpp: - * examples/Threads/thread_specific.cpp: - * examples/C++NPv1/Process_Per_Connection_Logging_Server.cpp: - * performance-tests/Synch-Benchmarks/Perf_Test/Benchmark_Performance.cpp - Addressed 64-bit conversion warnings. - -Fri Oct 28 02:49:11 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * examples/APG/Shared_Memory/Mem_Map.cpp: - * examples/Mem_Map/file-reverse/file-reverse.cpp: - * examples/Web_Crawler/Mem_Map_Stream.cpp: - - Addressed 64-bit conversion warnings. - -Fri Oct 27 21:41:25 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * THANKS: - * ace/Select_Reactor_Base.inl: - * ace/Select_Reactor_Base.cpp: - * ace/TP_Reactor.inl: - * ace/TP_Reactor.cpp: - * tests/run_test.lst: - - The fix for bug 2653. Thanks to Scott Mitchell <rsm at progress - dot com> and Chris Reed <cr at progress dot com> for the - considerable effort they put in helping to isolate this problem. - -Fri Oct 27 21:28:36 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * tests/Bug_2653_Regression_Test.cpp: - * tests/tests.mpc: - - Added a new regression test. Without the accompaning fix, the - test coredump. If you comment out the call to test_for_crash, - the test will spin madly for 5 seconds, then abort. With the - pending fix to the TP reactor, the test has no errors. - -Fri Oct 27 19:48:23 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Filecache.cpp (ACE_Filecache_Object): - - Cast -1 to size_t. Ugly (and works), but currently necessary. - Addresses "passing negative value" warnings. - -Fri Oct 27 19:28:16 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/OS_NS_stdio.inl (flock_adjust_params): - * ace/OS_NS_sys_stat.inl (filesize): - - Fixed comparison of signed/unsigned warnings. - -Fri Oct 27 09:25:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Configuration_Import_Export.cpp: - Explicitly initialise pointer to 0 to silence warning in - gcc cygwi build - -Fri Oct 27 01:58:09 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - From Russell Mora <russell_mora at symantec dot com> - * ace/ACE.cpp: - * ace/ACE.h: - * ace/FILE.cpp: - * ace/File_Lock.h: - * ace/File_Lock.inl: - * ace/Filecache.cpp: - * ace/Filecache.h: - * ace/Local_Memory_Pool.cpp: - * ace/MEM_Acceptor.h: - * ace/MEM_Acceptor.inl: - * ace/MEM_IO.cpp: - * ace/MEM_IO.inl: - * ace/MEM_IO.h: - * ace/MMAP_Memory_Pool.cpp: - * ace/MMAP_Memory_Pool.h: - * ace/Mem_Map.cpp: - * ace/Mem_Map.h: - * ace/Mem_Map.inl: - * ace/OS_NS_macros.h: - * ace/OS_NS_stdio.cpp: - * ace/OS_NS_stdio.inl: - * ace/OS_NS_stdio.h: - * ace/OS_NS_sys_mman.h: - * ace/OS_NS_sys_mman.inl: - * ace/OS_NS_sys_stat.h: - * ace/OS_NS_sys_stat.inl: - * ace/OS_NS_unistd.cpp: - * ace/OS_NS_unistd.inl: - * ace/OS_NS_unistd.h: - * ace/Pagefile_Memory_Pool.inl: - * ace/Shared_Memory_MM.cpp: - * ace/Shared_Memory_MM.h: - * ace/Shared_Memory_MM.inl: - * ace/Shared_Memory_Pool.cpp: - * ace/Shared_Memory_Pool.h: - * ace/os_include/sys/os_types.h: - * ace/os_include/sys/os_stat.h: - * ace/os_include/os_unistd.h: - * examples/Web_Crawler/Iterators.cpp: - * examples/Web_Crawler/Mem_Map_Stream.cpp: - * examples/Web_Crawler/Mem_Map_Stream.h: - * examples/IPC_SAP/FILE_SAP/client.cpp: - - Added support for 64-bit file offsets on 32-bit platforms that - support the _FILE_OFFSET_BITS preprocessor symbol, as well as - Windows. Enable by defining the _FILE_OFFSET_BITS=64 - preprocessor symbol. ACE provides the necessary support on - Windows when it is defined with the assistance of a new - ACE_OFF_T typedef. - -Thu Oct 26 24:44:55 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Truncate.h (Truncator): - - Removed ACE_Export macros in the Borland case. They shouldn't - be necessary. - -Thu Oct 26 14:05:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_other_tests.lst: - Added orbsvcs/tests/Event/UDP test which tests the rtec - udp federation - -Thu Oct 26 11:19:45 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/config-lynxos.h: - Added include "ace/config-posix.h" and removed defines - that are already defined by config-posix.h. - -Thu Oct 26 10:29:49 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/config-posix.h: - Removed duplicate check of _POSIX_MESSAGE_PASSING. - -Thu Oct 26 07:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Truncate.h: - - Added partial specialization for when types are the same, fixes - compile problems in the ACE lib with Borland - -Thu Oct 26 07:01:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-vxworks5.x.h: - Added ACE_LACKS_INTPTR_T, VxWorks 5.5.x seems not to deliver this - type - -Thu Oct 26 00:31:56 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Auto_Ptr.h: - - Added missing "element_type" typedef. - - (ACE_auto_ptr_reset): - - Introduced this new function template to simplify the - ACE_AUTO_PTR_RESET macro implementation, to support - ACE's other auto_ptr-like class templates - (e.g. ACE_Auto_Array_Ptr<>), and to make auto_ptr - reset() operations convenient and easy for those needing the - ACE_AUTO_PTR_RESET functionality to. Users now need only - perform a call like: - - ACE_auto_ptr_reset (my_auto_ptr_instance, new foo); - - Note that ACE_AUTO_PTR_RESET support for ACE's other - auto_ptr-like classes isn't strictly necessary since they all - support the reset() method. - - Addresses build failures on platforms that define - ACE_LACKS_AUTO_PTR_RESET and in code that expects to use - ACE_AUTO_PTR_RESET on ACE_Auto_{Basic_}_Array<> instances. - -Wed Oct 25 22:42:03 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * tests/Integer_Truncate_Test.cpp: - - Remove "const" qualifiers from types used in - ACE_Utils::Truncate<>() function template parameters that are - deduced using argument dependent lookup. Borland appears to get - confused, and can't find the non-const template parameter - Truncate<>() specializations. - -Wed Oct 25 20:17:13 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Hash_Map_Manager_T.inl (ACE_Hash_Map_Iterator_Base_Ex): - - Removed commented ACE_TRACE macro entirely. Fuzz parses - commented out code, and still incorrectly reports an ACE_TRACE - mismatch. - - * ace/Truncate.h: - - Yet another attempt at placating Borland C++. Removed the - partial specialization for truncating values of the same type. - Typedefed the underlying Truncator structure template to further - simplify the function call. - -Wed Oct 25 18:31:47 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * examples/C++NPv1/Reactive_Logging_Server.h: - * examples/IPC_SAP/SOCK_SAP/CPP-inclient.cpp: - - Include "ace/Basic_Types.h" to pull in intptr_t typedef. - -Wed Oct 25 15:05:42 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_NS_stdlib.inl (atop): s/intptr/intptr_t/ in ACE_WIN64 case. - -Wed Oct 25 09:47:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/SOCK_Dgram.cpp: - Fixed compile error in non win32 part - -Wed Oct 25 09:24:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/ace_for_tao.mpc: - Added Rtems_init.c, for rtems without network support we - also use the ace_for_tao base project - -Wed Oct 25 08:38:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-rtems.h: - * ace/OS_NS_netdb.inl: - Added ACE_LACKS_GETHOSTBYADDR_R - -Wed Oct 25 08:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-rtems.h: - * ace/OS_NS_sys_socket.inl: - Added ACE_LACKS_GETSOCKOPT, ACE_LACKS_RECV, ACE_LACKS_SEND, - and ACE_LACKS_SENDTO - -Wed Oct 25 08:04:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Priority_Buffer_Test.cpp: - Fixed compile problem with gcc 2.96 - -Wed Oct 25 07:58:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-rtems.h: - * ace/OS_NS_sys_socket.inl: - Added ACE_LACKS_ACCEPT, ACE_LACKS_CONNECT, ACE_LACKS_GETPEERNAME, - ACE_LACKS_GETSOCKNAME - -Wed Oct 25 06:30:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/TTY_IO.cpp: - Fixed compile error on Vxworks - -Tue Oct 24 23:53:08 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-aix-4.x.h: Changed ACE_HAS_PTHREAD_CONTINUE to - ACE_HAS_PTHREAD_CONTINUE_NP and ACE_HAS_PTHREAD_SUSPEND to - ACE_HAS_PTHREAD_SUSPEND_NP. The non-_NP varieties return ENOTSUP; - the _NP varieties are the only documented APIs through AIX 5.3. - Thanks to Guy Peleg for raising this issue. - - * tests/Thread_Manager_Test.cpp: Altered the suspend/resume tests to - only allow ENOTSUP as a legit result if the config.h file does not - include ACE_HAS_... settings that indicate it should work. - -Tue Oct 24 23:59:10 UTC 2006 Ossama othman <ossama_othman at symantec dot com> - - * examples/Service_Configurator/IPC-tests/server/Handle_Timeout.inl - (init): - * examples/Shared_Malloc/test_position_independent_malloc.cpp - (parse_args): - - Use intptr_t instead of int when casting to void*. Addresses - "sizes do not match" warnings in 64-bit Intel compiler builds. - -Tue Oct 24 23:49:25 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Hash_Map_Manager_T.inl (ACE_Hash_Map_Iterator_Base_Ex): - - Commented out use of ACE_TRACE macro. The regular expression in - the bin/fuzz.pl script used to detect ACE_TRACE mismatch - incorrectly identifies a mismatch the ACE_TRACE question. Once - the fuzz script is fixed, the ACE_TRACE can be uncommented. - -Tue Oct 24 22:20:38 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Truncate.h: - - Rename truncate to trunc. Yet another attempt at fixing Borland - compiler brain damage. - -Tue Oct 24 22:01:41 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/OS_NS_stdlib.inl (atop): - - Use intptr_t type to fix 64-bit conversion to void* warnings. - - * ace/Select_Reactor_Base.h (max_handlep1): - * ace/Select_Reactor_Base.inl (max_handlep1): - - Changed return type to be ACE_HANDLE on UNIX platforms instead - of size_t. Addresses "loss of significant bits" warnings in - 64-bit Intel compiler builds. - - * ace/Select_Reactor_Base.cpp (dump): - - Corrected format specifiers for Windows configuration. - - * ace/Select_Reactor_T.cpp (work_pending): - (wait_for_multiple_events): - - Removed unnecessary casts and corrected type of "width" variable - that is passed to select(). - - * tests/Map_Manager_Test.cpp (test_cache_map_manager): - - Fixed "might be used uninitialized" warning. - -Tue Oct 24 17:46:40 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Truncate.h (Truncator): - - Export the Borland-specific specializations from the DLL. This - should address the "undefined symbol" link-time problem. - - * bin/MakeProjectCreator/config/acexml.mpb: - - Re-disabled ACEXML when ace_for_tao is enabled. ACEXML needs - the ACE_Configuration and memory map classes that are not found - in the ace_for_tao subset. - -Tue Oct 24 14:44:09 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/TTY_IO.cpp: - - Changed the default constructor to always initialise member - variables regardless of ACE_HAS_TERMIOS or ACE_HAS_TERMIO setting. - -Tue Oct 24 14:16:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/ACE.cpp: - Use some bools instead of int - - * ace/INET_Addr.cpp (set_port_number): - Added check for ACE_LACKS_HTONS - - * ace/INET_Addr.inl (get_port_number): - Added check for ACE_LACKS_NTOHS - -Tue Oct 24 13:54:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/netinet/os_in.h: - * ace/config-rtems.h: - Added ACE_LACKS_IP_MREQ - - * ace/INET_Addr.cpp (get_port_number_from_name): - Added ACE_LACKS_HTONS check - -Tue Oct 24 13:27:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Reactor.h: - Added an include of Time_Value.h, this is included implicitly - but this implicit include is not there when using FOCUS with - tp-reactor specialization - -Tue Oct 24 13:09:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Sock_Connect.cpp: - * ace/config-rtems.h: - Added ACE_LACKS_GET_BCAST_ADDR - -Tue Oct 24 13:00:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/os_socket.h: - * ace/config-rtems.h: - Added ACE_LACKS_SOCKADDR - -Tue Oct 24 12:54:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/os_netdb.h: - * ace/config-rtems.h: - Added ACE_LACKS_HOSTENT - -Tue Oct 24 12:09:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_arpa_inet.inl: - * ace/config-rtems.h: - Added ACE_LACKS_INET_ADDR and ACE_LACKS_INET_NTOA - -Tue Oct 24 12:00:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_socket.h: - Define MSG_OOB when it is not defined - -Tue Oct 24 11:17:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_socket.inl: - * ace/config-rtems.h: - Added ACE_LACKS_SOCKET, ACE_LACKS_SHUTDOWN and ACE_LACKS_SETSOCKOPT - -Tue Oct 24 11:08:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_select.inl: - * ace/config-rtems.h: - Added ACE_LACKS_SELECT - -Tue Oct 24 10:58:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/netinet/os_in.h: - * ace/config-rtems.h: - Added ACE_LACKS_IN_ADDR and ACE_LACKS_SOCKADDR_IN - -Tue Oct 24 06:31:05 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Numeric_Limits.h (ACE_Numeric_Limits): - - Added missing "char" specialization for the - ACE_LACKS_NUMERIC_LIMITS case. The "signed char" and - "unsigned char" specializations are not always suitable. - -Tue Oct 24 04:04:26 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Hash_Map_Manager_T.h (lock_): - - Placed "mutable" keyword to beginning of declaration. It is a - storage class specifier, not a CV qualifier, and belongs at the - beginning. Addresses g++ warnings. - - * ace/Sock_Connect.cpp (get_fqdn): - - Addressed "might be used uninitialized warnings. - -Mon Oct 23 02:28:42 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Array_Base.cpp: - * ace/Array_Base.h: - * ace/Array_Base.inl: - - Added STL-style traits, iterators and swap() method. - - (operator=): - - Improved exception-safety. - - * ace/Array_Map.h: - * ace/config-lite.h: - - Moved reverse_iterator declarations to a new - ACE_DECLARE_STL_REVERSE_ITERATORS so that other ACE containers - may use them in a way that minimizes platform-specific - declaration duplication. - - * ace/CDR_Stream.cpp (read_boolean_array): - - Addressed a Coverity CHECKED_RETURN error. There's no need to - check the return value of read_boolean() in this case. - Explicitly cast the return value to "void". - - Addressed Coverity DEADCODE error. No need to check for zero - "cont" pointer since the ACE_NEW_RETURN() call that initializes - it will return immediately. "cont" cannot be zero if the latter - call succeeds. - - * ace/CDR_Stream.inl (operator>>): - - Explicitly check return value of input stream read operations. - Addressed a Coverity CHECKED_RETURN errors. - - * ace/Configuration.cpp (new_section): - - Fixed Coverity USE_AFTER_FREE error. Memory was being accessed - after it was freed. - - * ace/Configuration_Import_Export.cpp: - - Addressed Coverity RESOURCE_LEAK errors. - - * ace/Default_Constants.h (ACE_DEFAULT_CLOSE_ALL_HANDLES): - - This symbol represents a boolean value. s/1/true. - - * ace/Dev_Poll_Reactor.cpp (bind): - * ace/WFMO_Reactor.cpp (bind_i): - - Disallow binding of a nil event handler pointer to the - repository. It is deferenced later on this method. - - * ace/Hash_Map_Manager_T.h: - * ace/Hash_Map_Manager_T.inl: - - Added STL-style traits and iterator accessors. - - (find): - - New method that returns an iterator as an "out" parameter. An - iterator return value could not be used due a conflict with an - existing find() method. - - (unbind): - - New method that unbinds the value pointed to by the given - iterator. - - (ACE_Hash_Map_Iterator_Base_Ex, ACE_Hash_Map_Iterator_Ex): - - New constructors that position the iterator at the given - ACE_Hash_Map_Entry. - - * ace/Lib_Find.cpp (ldfind): - - Place parentheses around ternary operator "?:" conditional to - address an operator precedence problem. In this case, an - expression such as "a + b ? 0 : 1" was used, which is equivalent - to "(a + b) ? 0 : 1" instead of the desired "a + (b ? 0 : 1)". - - Changed local variable name from "got_suffix" to "has_suffix" to - clarify how the variable is being used. - - * ace/OS_NS_Thread.inl (rw_unlock): - - Fixed potential deadlock in case where reference count is one. - Addresses a Coverity DEADLOCK error. - - * ace/OS_NS_dirent.cpp (scandir_emulation): - - Fixed potential dereferencing of zero pointer reported by the - Coverity FORWARD_NULL checker. - - * ace/OS_NS_fcntl.cpp (open): - - Return if SetFilePointer() call fails. Addresses Coverity - CHECKED_RETURN error. - - Added support for large files SetFilePointer() calls. Thanks to - Russell Mora for this fix. - - * ace/OS_NS_stdio.cpp (fopen): - - Fixed potential crash if null mode string is passed to this - method. - - * ace/OS_NS_stdio.inl (fdopen): - * ace/OS_NS_unistd.inl (isatty): - - No need to check for ACE_LACKS_INTPTR_T. Just use intptr_t in - all cases. ACE now supplies this typedef if the platform - doesn't. - - * ace/OS_TLI.h (t_bind): - * ace/OS_TLI.inl (t_bind): - - Work around HP aCC 3.63 namespace scoping problem using typedef - hack in use for other structures. - - * ace/Select_Reactor_Base.cpp: - * ace/Select_Reactor_Base.h: - * ace/Select_Reactor_Base.inl: - - Changed underlying event handler container from a simple array - to an ACE_Hash_Map_Manager<> on Windows, and to an - ACE_Array_Base<> on UNIX-like platforms. - - Use of an ACE_Hash_Map_Manager<> on Windows allows for faster - event handler searches. - - Use of an ACE_Array_Base<> on UNIX-like platforms allows us to - leverage advantages using a container (improved - exception-safety, container interface, etc), while still - retaining the direct file descriptor to event handler mapping at - essentially no cost to footprint relative to the previous - implementation. - - (bind): - - Disallow binding of a nil event handler pointer to the - repository. It is deferenced later on this method. Addresses a - potential crash. - - * ace/Service_Config.cpp (current): - - Simplified these methods. - - Removed global lock that was held when accessing TSS. Only - initialization of global implementation needs to synchronized - with that lock. - - Removed redundant accesses to TSS. - - * ace/Service_Config.h: - - Removed stray semi-colon. - - (impl_): - - This method no longer needs to return the pointer by reference. - - * ace/Service_Repository.cpp: - - Changed loop index type so that it matches repository size type - (an unsigned integer), and updated loop decrement to take into - account the index type is now unsigned. - - * ace/Sock_Connect.h (ipv4_enabled): - * ace/Sock_Connect.cpp (ipv4_enabled): - - New run-time check for existence IPv4 stack. Some hosts only - contain IPv6 stacks. - - From Sriram Chandramouli <sriram_chandramouli at symantec dot com> - (get_fqdn): - - New method that retrieves fully qualified domain. getaddrinfo() - doesn't always provide the desired results. - - * ace/Svc_Conf.y (ace_get_module): - * ace/Svc_Conf_y.cpp (ace_get_module): - - Simplified explicit casts. - - Prevent dereferencing of potentially null pointer. - - * ace/System_Time.cpp: - * ace/System_Time.h: - - No need to include "ace/Malloc_T.h", "ace/MMAP_Memory_Pool.h" - and "ace/Null_Mutex.h". Forward declarations are sufficient. - - Use time_t type instead of ACE_UINT32 when returning times, such - as those returned from ACE_OS::time(). Addresses lack of - precision issues. - - * ace/Thread_Manager.cpp (wait): - - Addressed a Coverity CHECKED_RETURN error. There's no need to - check the return value of ACE_Thread::join() in this case. - Explicitly cast the return value to "void". - - * ace/Time_Value.h (ACE_ONE_SECOND_IN_USECONDS): - - Made this constant type-safe. It is now of type suseconds_t, - instead of a preprocessor symbol definition. - - * ace/UUID.cpp (lock): - - Addressed USE_AFTER_FREE Coverity error. Do not return a - pointer to the previous lock. It may have been destroyed. - - * ace/UUID.h (lock): - - Changed return type from "ACE_SYNCH_MUTEX*" to "void". The - interface/implementation does not provide a means to safely - return the previous lock instance. - - * ace/ace_wchar.inl (convert): - - Corrected calculation of wide string length. - - * bin/MakeProjectCreator/config/acexml.mpb: - - No longer any need to explicitly disable ACEXML in the - ace_for_tao configuration. - - * include/makeinclude/platform_irix6.x_sgic++.GNU: - - SGI 7.3.x compiler doesn't have the 3506 error code. Do not - attempt to supress it. - -Mon Oct 23 22:24:01 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Unbounded_Queue.cpp (set): - - Explicitly initialize dummy value. Addresses "use of - uninitialized variable" g++ 4.1 warning when the variable is a - POD type. - - * ace/ace_for_tao.mpc: - - Removed unnecessary semaphore related files. - - Added missing ace/Truncate.h header. - -Mon Oct 23 21:44:32 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Truncate.h (Truncator): - - Borland doesn't appear to be able to grok the use of the - function call operator in the ACE_Utils::Truncate<> function - template. Replaced the Truncator<>::operator() function call - operator with a static Truncator<>::truncate() method instead. - -Mon Oct 23 13:29:15 UTC 2006 Olli Savia <ops@iki.fi> - - * ACE-INSTALL.html: - * ace/Mem_Map.h: - * ace/Mem_Map.inl: - * ace/Mem_Map.cpp: - * ace/OS_NS_Thread.inl: - * ace/OS_NS_sys_mman.inl: - * ace/Reverse_Lock_T.h: - * ace/Task_Ex_T.h: - * ace/Task_T.h: - * ace/config-lynxos.h: - * include/makeinclude/platform_lynxos.GNU: - - Removed LynxOS 3.x support. - -Mon Oct 23 12:59:35 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/generate_compile_stats.sh: - - Corrected a bug with my previous change. If config.h or - platform_macros.GNU were not present, the html would be incorrect. - I also added default.features to the list of configuration files - that are displayed. - -Mon Oct 23 06:34:24 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/os_include/os_pthread.h: Changed - - # if !defined (ACE_LACKS_RWLOCK_T) - - to - - # if !defined (ACE_LACKS_RWLOCK_T) && !defined (ACE_HAS_PTHREADS_UNIX98_EXT) - - to stop a compile error on Solaris after SunOS 5.8. Thanks to - Boris Kaminer <boris dot kaminer at gmail dot com> for reporting - this fix. - -Sun Oct 23 00:29:55 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Truncate.h: - - In the Borland-specific workaround, explicitly cast return value - of ACE_Numeric_Limits::max() of the smaller unsigned type to the - unsigned larger type to which it is being compared. Addresses - non-intuitive "comparison of signed and unsigned types" warning - issued by Borland compiler. Borland promotes unsigned char to - int, instead of unsigned int, when comparing to unsigned int. - This implicit type promotion behavior complies with the C++ - standard, but is a bit awkward nonetheless (IMHO :)). Thanks to - Johnny for pointing out Borland's explanation of the - non-intuitive warning. - -Sun Oct 22 08:11:25 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Token_Request_Reply.inl (token_name): - - Simplified use of ACE_Utils::Truncate<> function template. Only - the "TO" type must be explicitly specified. The "FROM" type is - deduced automatically by the compiler. Note that explicitly - specifying both template parameters is perfectly valid, too. - - * ace/Truncate.h (Truncator): - - Added char/int specializations for Borland to allow the - Integer_Truncate_Test to compile. All of the Borland-specific - Truncator specializations are meant to be a temporary workaround - until Borland C++ is capable of supporting the generalized - implementation. - -Sat Oct 21 14:49:21 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/TSS_T.{h cpp}: Added "volatile" keyword to once_ member - since it can be changed by other threads. Removed the lock - acquiring from the ts_object() methods - after once_ is checked, - there are no further needs for locks. When needed, the key_ is - set up by ts_init() which does acquire a lock. This avoids a lock - acquire/release cycle on all calls when initialization of the - ACE_TSS object is not required, speeding up performance. Thanks to - Guy Peleg <guy dot peleg at amdocs dot com> for reporting this and - to Russell Mora <russell_mora at symantec dot com> for assisting in - analyzing the situation and recommending the particular fixes. - - * ace/Logging_Strategy.h: Minor doxygen improvements. - - * ace/Sock_Connect.cpp: When checking for IPv6 interfaces on Windows, - there's two conditionally compiled blocks of code. One block was - built if ACE_HAS_IPV6 SIO_ADDRESS_LIST_QUERY are both defined; the - other was built if ACE_HAS_IPV6 is defined. Added the check for - SIO_ADDRESS_LIST_QUERY to the second case as well since it makes - no sense to run it if the first isn't built. Thanks to Andre - Kostur for this fix. Also see: - Mon Oct 16 14:24:25 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Task_T.h (reply): Clarified what happens on a reply(). Thanks - to Doug Schmidt for explaining the rationale for this. - - * ace/Token_Request_Reply.inl (token_name): VC7.1 had trouble grokking - the ACE_Utils::Truncate arguments in this case, so add the types - to a more complete specification of what template is desired. - -Sat Oct 21 15:19:46 UTC 2006 Steve Huston <shuston@riverace.com> - - * configure.ac: Removed the check for LINUX_VERSION that disabled - the check for epoll on Linux. Just let the check run - the macros - used checked the kernel/C version used to build the kernel instead - of the running kernel itself (I think...). In any event, the - use of the macros to disable the check prevented the check from - running on systems (such as RHEL 4) that do have epoll. Thank you - to Brian Campbell <becampbel at hotmail dot com> for researching - this area. - -Sat Oct 21 10:48:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * bin/PerlACE/RunTest.pm: - * bin/PerlACE/ProcessVX.pm: - * bin/PerlACE/ProcessVX_Unix.pm: - - Changes to accomodate running autotests for VxWorks RTPs on - Linux hosts. - -Sat Oct 21 04:34:33 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/OS_NS_stdlib.cpp: - - Always include "ace/Numeric_Limits.h" in the ACE_LACKS_MKSTEMP - case. There is no need to conditionally include it when "max" - is undefined. Addresses VxWorks build errors. - - * ace/Truncate.h: - - Only use workaround for Borland compiler versions less than or - equal to 0x582. Thanks to Johnny for the suggestion. - - * tests/Integer_Truncate_Test.cpp (sizeof_from_lt_sizeof_to): - - Explicitly cast types in inequality comparison to silence - comparison of signed and unsigned warnings. It's not clear why - the warning is occuring since both types are unsigned. - - Removed remaining set of redundant parentheses that were - missed in a previous commit. - -Fri Oct 20 12:55:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Service_Manger.{h,cpp}: - Changed debug_ member to a real bool - -Fri Oct 20 12:40:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_socket.inl (bind): - * ace/config-rtems.h: - Added ACE_LACKS_BIND - - * ace/Basic_Types.h: - Fixed compile problems with BCB6 - -Fri Oct 20 12:05:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Log_Record.inl: - Added check for ACE_LACKS_NTOHL and ACE_LACKS_HTONL - - * ace/config-rtems.h: - Added ACE_LACKS_NTOHL and ACE_LACKS_HTONL when no networking - is available - -Fri Oct 20 09:56:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/INET_Addr.h: - Small change to improve doxygen documentation - - * ace/OS_NS_sys_socket.inl (listen): - Added ACE_LACKS_LISTEN - - * ace/config-rtems.h: - Added ACE_LACKS_LISTEN - -Fri Oct 20 09:36:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/ciao_nodeapplicationmanager.mpb: - * bin/MakeProjectCreator/config/ciao_nodemanager_stub.mpb: - * bin/MakeProjectCreator/config/ciao_targetmanager_stub.mpb: - New base projects - - * bin/MakeProjectCreator/config/ciao_static_dnc_app.mpb: - Use the new base projects above - - Thanks to Abdul Sowayan <abdullah dot sowayan at lmco dot com> - for delivering these files - -Fri Oct 20 09:00:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_socket.h: - Defined AF_UNSPEC and PF_UNSPEC when these are not defined by the - OS header files. - -Fri Oct 20 08:55:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_uio.h: - When INADDR_ANY is not defined, then define it, we need it in several - of the ACE header files - -Fri Oct 20 08:44:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_uio.h: - We define iovec when ACE_WIN32 was defined, changed this to a new - define, ACE_LACKS_IOVEC to make it easier to use with other platforms - - * ace/config-win32-common.h: - Added ACE_LACKS_IOVEC - - * ace/config-rtems.h: - When no networking has been configured, also define ACE_LACKS_IOVEC - -Thu Oct 19 22:28:59 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * contrib/minizip/iowin32.c - * contrib/minizip/iowin32.h - Removed these files. - -Thu Oct 19 23:20:26 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * examples/IPC_SAP/SOCK_SAP/CPP-inserver.cpp: - * examples/IPC_SAP/SOCK_SAP/CPP-inclient.cpp: - * examples/C++NPv1/Reactive_Logging_Server.h: - - Fixed integer/void* conversion related warnings by using - the "intptr_t" type instead of "int" or "long". - -Thu Oct 19 22:58:02 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Truncate.h: - - Borland compilers currently cannot handle the template - meta-programming in this header. Work around the limitation by - explicitly specializing the ACE_Utils::Truncator<> structure - template for a select few 32- and 64-bit integers. More will be - added as needed. - -Thu Oct 19 17:25:16 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * tests/Integer_Truncate_Test.cpp: - - Removed redundant parentheses. - -Thu Oct 19 15:25:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/generate_rel_manpages: - Updated doxygen version - - * bin/MakeProjectCreator/templates/gnu.mpb: - Removed kylix support - -Thu Oct 19 13:19:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_main.h: - Added main for rtems - -Thu Oct 19 11:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Rtems_init.c: - Fixed empty translation unit warnings - -Thu Oct 19 08:27:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * include/makeinclude/platform_openvms.GNU: - Small change to create hardlinks for produced executables. - -Wed Oct 18 16:19:38 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * contrib/minizip/minizip.suo - * contrib/minizip/miniziplibrarycode.suo - Removed these files. - -Wed Oct 18 13:52:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_rtems.x_g++.GNU: - Major cleanup of old stuff and always define ACE_HAS_RTEMS - - * etc/rtems_init.c: - * ace/rtems_init.c: - Another move, for rtems we require some functions to always - be available, this defines our network settings and much more. - Because we always build static, we moved this to the ACE lib, - on rtems we define ACE_HAS_RTEMS and then compile the contents - of this file. This way the functions in this file are available - in any executable we link - - * ace/OS.h: - * ace/os_include/os_signal.h: - * ace/os_include/sys/os_types.h: - * ace/OS_NS_stdio.h: - * ace/OS_NS_time.inl: - Replaced __rtems__ with ACE_HAS_RTEMS - - * ace/ace.mpc: - Added rtems_init.c - -Wed Oct 18 13:01:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeincludes/rules.bin.GNU: - * include/makeincludes/rules.lib.GNU: - * include/makeincludes/wrapper_macros.GNU: - * include/makeincludes/platform_linux_borland.GNU: - Removed kylix support - -Wed Oct 18 10:05:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * apps/Gateway/Gateway/File_Parser.cpp: - Removed egcs workaround - -Wed Oct 18 08:07:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/rtems_init.c: - * etc/rtems_init.c: - Moved from tests to etc - -Tue Oct 18 06:27:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Reactor/Misc/test_signals_1.cpp: - * performance-tests/Misc/context_switch_time.cpp: - Renamed the global count variables to another name. These triggered - a bug in the VxWorks 5.5.2 compiler which we reported to windriver. - By renaming these we will get an error free compilation. - -Tue Oct 18 06:02:28 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Basic_Types.h (intptr_t, uintptr_t): - - Provided suitable typedefs for these standard types when - ACE_LACKS_INTPTR_T is defined. - - * ace/Event_Handler.cpp (~ACE_Event_Handler_var): - - Only use an ACE_Errno_Guard if the underlying ACE_Event_Handler - pointer is non-zero. Saves a TSS access in the zero - ACE_Event_Handler pointer case. - - (operator=): - - Made strongly exception safe. - - * ace/FILE_IO.cpp (send, recv): - * ace/Message_Queue.cpp (enqueue, dequeue): - * ace/Message_Queue_T.cpp: - * ace/OS_NS_unistd.inl (hostname): - - Explicitly supply the "int" "TO" type to the - ACE_Utils::Truncate<> function template. It must now be - explicitly specified since ACE_Utils::Truncate<> now supports - truncation to all integer types, not just "int". - - * ace/Global_Macros.h (ACE_STATIC_CONSTANT): - - New macro for in-class static integral constant initialization - that works around compilers that don't like such initialization. - In-class integral constants in this case are meant to be - compile-time constants so that they may be used as template - arguments, for example. BOOST provides a similar macro. - - * ace/OS_NS_stdlib.inl (atop): - * ace/OS_NS_stdio.cpp (fopen): - - Removed ACE_LACKS_INTPTR_T check. It is no longer needed. - - * ace/Truncate.h: - - Generalized the ACE_Utils::Truncate<> function template - implementation for all integer types. - - * tests/run_test.lst: - * tests/tests.mpc: - - Added new "Integer_Truncate_Test" test that verifies the - ACE_Utils::Truncate<> function template works for all type - size/sign scenarios. - -Tue Oct 17 21:28:11 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.cpp (parse_args_i): - - Checking for no arguments was missed inadvertently in the - original change-set from "Tue Oct 17 16:27:27 UTC 2006 ..." - -Tue Oct 17 21:09:36 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * contrib/minizip/ChangeLogUnzip - * contrib/minizip/crypt.h - * contrib/minizip/ioapi.c - * contrib/minizip/ioapi.h - * contrib/minizip/iowin32.c - * contrib/minizip/iowin32.h - * contrib/minizip/miniunz.c - * contrib/minizip/minizip.c - * contrib/minizip/minizip.mpc - * contrib/minizip/minizip_export.h - * contrib/minizip/mztools.c - * contrib/minizip/mztools.h - * contrib/minizip/unzip.c - * contrib/minizip/unzip.h - * contrib/minizip/zip.c - * contrib/minizip/zip.h - Added the minizip project. Thanks to Krishnakumar Balasubramanian - <kitty at dre dot vanderbilt dot edu> for providing this. - -Tue Oct 17 17:17:36 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/OS_NS_unistd.inl (truncate): SetFilePointerEx() returns a - BOOL not a DWORD (unlike SetFilePointer()..). Thanks to Boris - Kaminer <boris dot kaminer at gmail dot com> for reporting this. - -Tue Oct 17 16:27:27 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/DLL_Manager.cpp: - * ace/Parse_Node.cpp: - - Updated logging messages. - - * ace/Service_Gestalt.cpp: - - Updated logging messages to exclude the "e-word" causing false - positives in the Service_Config_Test. Updated comments. - -Tue Oct 17 12:35:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * apps/soreduce/Library.cpp: - taoversion is not a valid base project. Thanks to Manuel Traut - <manut dot de at googlemail dot com> for reporting this. - -Tue Oct 17 12:09:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: - Updated supported platforms by Remedy IT. New are VxWorks 5.5.2, - OpenVMS 8.3, RTEMS and LynxOS. - -Tue Oct 17 12:05:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-rtems.h: - Cleanup - - * include/makeinclude/plaform_rtems.x_g++.GNU: - Improved for rtems 4.6.6 - -Tue Oct 17 07:03:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-rtems.h: - * ace/OS_NS_unistd.inl: - Added support for ACE_LACKS_FDOPEN - -Mon Oct 16 21:29:57 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * THANKS: Updated Chris Reed's entry in the list. - -Mon Oct 16 20:04:59 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - This change fixes bug#2612. See the description for the gory - details. In a nutshell, the problem is caused by the fact that - static services, loaded from a DLL, as part of a dynamic service - initialization can not be safely finalized when the dynamic - service's DLL gets unloaded. This was historically handled by - simply not unloading DLLs, and by finalizing all services at - process exit. With the introduction of the local configurations - feature, this changed. - - The finalization order was supposed to be guaranteed by the - ordering of the services in the repository. It really was a - kludge, because when dynamically loading and removing services - the order can change. Then whenever the process tries to clean - up, the dynamic service's DLL can be unloaded (because it gets - finalized first), and the static services' destructor code would - become inaccessible. The situation results in a SEGV at exit(). - - * ace/DLL.h: - * ace/DLL.cpp: - - Made assignment operator return a non-const reference to make it - well-formed. See C++ Standard, section "[special]". - - * ace/DLL_Manager.cpp: - * ace/Dynamic_Service_Base.cpp (instance): - * ace/Parse_Node.cpp: - - Fixed formatting and logging - - * ace/Service_Config.cpp (parse_args_i): - - Fixing a bug related to parsing the command line options that - take arguments. Thanks to Aaron Scamehorn <Aaron dot Scamehorn - at cogcap dot com> for reporting it. - - * ace/Service_Gestalt.cpp: - - Added ACE_Service_Dynamic_Guard (formerly - ACE_Service_Type_Forward_Declaration_Guard) which helps to - resolve an issue with hybrid services, i.e. dynamic services, - accompanied by static services in the same DLL. Only automatic - instances of SDG are supposed to exist. Those are created during - (dynamic) service initialization and serve to: - - (a) Ensure the service we are loading is ordered last in the - repository, following any other services it may cause to - register, as part of its own registration. This is a common case - when loading dynamic services from DLLs - there are often static - initializers, which register static services. - - (b) The SDG instance destructor detects if the dynamic service - initialized successfully and "fixes-up" all the newly registered - static services, so that each holds a reference to the DLL, from - which they have originated. - - Updated comments and formatting. - - * ace/Service_Object.h (dll): - * ace/Service_Object.inl (dll): - - Added void dll (const ACE_DLL&) to make it possible to - "relocate" services registered through static initializers found - in DLL's code segment(s). - - * ace/Service_Object.cpp (fini): - - Modified ACE_Service_Type::fini() to ensure the DLL associated - with the service is properly closed. This in conjunction with - the gestalt changes above, makes it possible to safely and - completely unload a service. - - * ace/Service_Repository.h: - * ace/Service_Repository.cpp (remove_i,relocate_i,relocate): - - Added relocate(), which allows association of a (static) service - objects with a DLL. Per changes to ACE_Service_Gestalt, - above. Added a counterpart private relocate_i(), which does not - obtain locks. - - * tests/Service_Config_Test.cpp: - - Added a test that combines more than one argument-taking command - line option to guard against regression for the bug fixed in - ACE_Service_Config::parse_args_i (see above). - -Mon Oct 16 19:20:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-rtems.h: - Updated with more changes needed for rtems - - * ace/os_include/os_signal.h: - Added specific defines for rtems - - * ace/OS_NS_time.cpp (difftime): - Removed not used variable and get gmtime/gmtime_r from the - global scope - - * ace/OS_NS_string.{h,cpp,inl}: - Added ACE_LACKS_STRTOK_R to make the rtems port easier - -Mon Oct 16 14:24:25 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_NS_netdb.{cpp inl} (gethostbyname[_r], gethostbyaddr[_r]): - Check the specified string (name or addr) for unspecified or of - 0 length and return 0 if so - bypasses a system call and a possible - memory leak. Thanks to Howard Finer <hfiner at sonusnet dot com> - for these fixes. - - * ace/Sock_Connect.cpp (get_ip_interfaces): Don't report an error if - we can't find IPv6 addresses. It may be on a system with ACE_HAS_IPV6 - support compiled but IPv6 not configured/enabled. Thanks to Andre - Kostur <akostur at incognito dot com> for reporting this. - - * ace/WFMO_Reactor.cpp (ok_to_wait): In the ACE_HAS_WINCE block, - change fwaitAll arg from TRUE to FALSE - it must be FALSE on - WinCE (thru WinMobile 5) per MSDN. Thanks to Drew Reynaud - <drewreynaud at hotmail dot com> for this fix. - - * THANKS: Added Drew Reynaud to the Hall of Fame. - -Mon Oct 16 14:21:48 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/generate_compile_stats.sh: - - I accidentally deleted a letter in my previous commit. - -Mon Oct 16 14:02:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-rtems.h: - Added first set of missing defines - - * ace/os_include/sys/os_select.h: - * ace/os_include/sys/os_uio.h: - Removed rtems specific part, not needed anymore - - * ace/os_include/sys/os_types.h: - Added rtems check - - * include/makeinclude/plaform_rtems.x_g++.GNU: - Updated - -Mon Oct 16 13:47:25 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/MakeProjectCreator/modules/GNUACEWorkspaceCreator.pm: - - Use the create_command_line_string() method instead of printing - the command line directly. This opens the door for disabling this - functionality. - -Mon Oct 16 12:18:14 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * ace/OS_NS_Thread.inl: - * ace/os_include/os_semaphore.h: - - The feature that I put in for Gottwald Port on Mon Mar 6 18:25:41 - UTC 2006 causes a performance overhead for semaphores where - ACE_HAS_POSIX_SEM is defined (SunOS, HP-UX, MacOS X, etc.). For - those that are not interested in timed semaphores, this feature - provides nothing and is detemental to performance. Now, this - feature is disabled by default and must be manually enabled by - defining the ACE_HAS_POSIX_SEM_TIMEOUT_EMULATION macro. - -Mon Oct 16 12:06:10 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/generate_compile_stats.sh: - - Added some changes to make this script usable for software other - than ACE+TAO+CIAO. - - - Enhanced the usage message. - - Fixed minor HTML errors. - - Added options to set the base root directory and software title. - -Mon Oct 16 11:53:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/IOStream/client/IOStream_Client.mpc: - * examples/IOStream/server/IOStream_Server.mpc: - Added exename so that we create an exe, not a dll - - * examples/IOStream/client/.cvsignore: - * examples/IOStream/server/.cvsignore: - Removed duplicate lines - - * examples/IOStream/server/iostream_server.cpp: - Added needed include for building with VxWorks 5.5.2 - -Mon Oct 16 08:33:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/templates/ciao_static_dnc_app.mpb: - New base project for static dnc. - Thanks to Abdul Sowayan <abdullah dot sowayan at lmco dot com> - for creating this file - -Sun Oct 15 23:22:52 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - Added Bug_2683_Regression to the test suite. - -Fri Oct 13 18:26:19 UTC 2006 Andy Stafford <ajs3@cec.wustl.edu> - - * include/makeinclude/wrapper_macros.GNU: - - Added 'lksctp' as recognized option to sctp documentation. - -Fri Oct 13 18:13:00 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/cle.pl (getDefaultDomain): - - Fixed a bug where extra text would be part of the domain name when - the /etc/hosts file contains an alias after the fully qualified - host name. - -Fri Oct 13 12:32:35 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/PerlACE/Process_Unix.pm: - - Added support for remote testing on UNIX only. Currently, ssh or - rsh is used to run an executable on the remote machine. The - remote test capability is only enabled if the RemoteInformation - method is called on the PerlACE::Process object. For more - details, see the RemoteInformation method in this module. - -Fri Oct 13 10:25:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * apps/soreduce/Library.cpp: - aceversion is not a valid base project. Thanks to Manuel Traut - <manut dot de at googlemail dot com> for reporting this. - -Thu Oct 12 16:44:34 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Log_Record.cpp (operator<<): Zapped the nonconst_record and - const_cast since this is no longer needed as per the following - change: - - Sun Aug 12 00:34:42 2001 Steve Huston <shuston@riverace.com> - * ace/Log_Record.h: - * ace/Log_Record.i (msg_data()): Changed ACE_TCHAR *msg_data() - to: const ACE_TCHAR *msg_data (void) const. Makes it safer, - and conforms to all current usage inside ACE. Also makes - C++ Network Programming examples simpler :-) - - Thanks to Abdul Sowayan <abdullah dot sowayan at lmco dot com> - for reporting this. - -Thu Oct 12 15:45:24 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/MakeProjectCreator/templates/gnu.mpd: - - If a custom definition has an 'output_option' set, the order of - listing the input file changes such that it is listed before the - output_option setting in the build rule. - - * bin/auto_run_tests.pl: - - Added a new option, -l, which can be used to specify an arbitrary - list of tests. - -Thu Oct 12 12:20:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Svc_Conf.h: - Fixed incorrect placed ACE_BEGIN_VERSIONED_NAMESPACE_DECL macro. - Thanks to Alexander Libman <alexander dot libman at earthlink dot net> - for reporting this - -Wed Oct 11 15:24:04 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Log_Record.{h inl cpp}: Removed the set_msg_data_ptr() - method - it opens the possibility of putting a buffer in place - that can't be freed properly. All of the memory allocating and - deleting should be done internally to this class. The removed - method was only used internally. - Added a private msg_data_size_ member that keeps the allocated size - of msg_data_ in ACE_TCHARs (not bytes) to tell if the buffer needs - to be resized. Comparing to the string length is incorrect - it - causes buffer reallocating whenever a short logged message is - followed by a longer one, even if both are much shorter than the - allocated length of msg_data_. When a ctor allocates a msg_data_ - area, set the first spot to '\0' to a call to msg_data_len() won't - get random answers or a crash. - (msg_data (const ACE_TCHAR *)): Changed this method to return int, - not void. It also will now reallocate the msg_data_ buffer if needed - to hold the specified string. operator<< now uses msg_data() to set - the string instead of the now removed set_msg_data_ptr() method, - keeping all msg_data_ allocation and release in the same class. - (operator<<): Fix memory leak when setting a new msg_data_ - buffer. Thanks to Arturo Montes for letting us know about the leak. - - * docs/ACE-guidelines.html: Clarify the format of ChangeLog entries. - Remove the recommendation to always test on egcs; things should be - tested on at least Windows and Linux, but egcs is outdated. - -Tue Oct 10 18:46:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: - Document how to get a current copy from subversion - -Tue Oct 10 14:16:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/CDR_Stream.cpp: - Fixed incorrect comment - -Tue Oct 10 12:54:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * ace/config-win32-msvc.h: - Changed the MFC build run-time dll decorators to match - the current build system settings. - -Tue Oct 10 06:08:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/fuzz.pl: - Changed check for empty inline files to also check for empty - source files - - * bin/MakeProjectCreator/config/avoids_corba_e_compact.mpb: - * bin/MakeProjectCreator/config/avoids_corba_e_micro.mpb: - New avoids projects - - * etc/ace.doxygen: - Expand some more macros, should improve documentatation - - * html/README: - Updated download location to ISIS - -Tue Oct 10 06:03:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/ACE-bug-process.html: - Updated to mention svn - - * docs/Download.html: - Commented out the source packages that are not complete at this moment - -Mon Oct 09 09:07:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/generate_rel_manpages: - Fixed an error in this script and updated the location to which the - generated packages should be copied to - -Mon Oct 09 09:03:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/Download.html: - Updated with the new download locations - - * docs/bczar/bczar.html: - Updated documentation generation instructions - -Thu Oct 05 00:35:37 2006 William R. Otte wotte@dre.vanderbilt.edu - - * ACE version 5.5.3 released. - -Tue Oct 3 16:45:35 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * NEWS - Updated the NEWS to add the changes for ACE_Hash_MultiMap_Manager - class and executionmanager_stub and plan_generator base projects. - -Tue Oct 3 15:02:14 UTC 2006 Steve Huston <shuston@riverace.com> - - * NEWS: Described the ACE_Refcounted_Auto_Ptr effects of - Thu Jul 6 20:45:15 UTC 2006 Steve Huston <shuston@riverace.com> - -Mon Oct 2 12:21:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Log_Msg.cpp: - Const/bool improvements - -Mon Oct 2 11:26:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Log_Msg.{h,cpp}: - Changed is_set_ conditional values member to bool - -Sat Sep 30 14:00:33 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Cached_Allocator_Test.cpp: - * tests/Codecs_Test.cpp: - * tests/Enum_Interfaces_Test.cpp: - * tests/Get_Opt_Test.cpp (run_test): - * tests/Naming_Test.cpp: - * tests/Reactor_Notification_Queue_Test.cpp: - * tests/Reference_Counted_Event_Handler_Test.cpp: - * tests/Timer_Queue_Reference_Counting_Test.cpp: - Correct logging format for wide-char build. - -Fri Sep 29 18:54:29 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_NS_signal.inl (sigaction): In the !ACE_HAS_SIGACTION_CONSTP2 - case, use ACE_SIGACTION instead of "struct sigaction" to pick up - the funky HP hack. Also see: - Mon Jan 23 22:08:56 UTC 2006 Steve Huston <shuston@riverace.com> - -Fri Sep 29 10:00:04 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Web_Crawler/Web_Crawler.mpc: - Added avoids+=uses_wchar for borland and bmake, the Borland - compiler gives a warning which can't be easily suppressed and - by not building this project we do get a few new green builds - -Thu Sep 28 19:37:13 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * bin/diff-builds.pl: - Fixed the finding of the nearest earlier build for a given date. - -Thu Sep 28 16:53:14 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * bin\MakeProjectCreator\config\ciao_executionmanager_stub.mpb - * bin\MakeProjectCreator\config\ciao_plan_generator.mpb - Added the base projects for executionmanager_stub and plan_generator. - -Thu Sep 28 14:46:43 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * bin/diff-builds.pl: - - Updated to fix temporary files creation. Now all those are put - in the temporary files directory. Fixed the ability to compare - just two dates. - -Thu Sep 28 13:35:04 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-hpux-11.00.h: - Removed ACE_HAS_EXPLICIT_STATIC_TEMPLATE_MEMBER_INSTANTIATION that - was set for gcc, with this define set the ACE library doesn't - build. - - * ace/CDR_Stream.cpp: - Initialise pointer with 0 - - * ace/TP_Reactor.{h,inl}: - Changed owner from int to bool. - -Wed Sep 27 21:18:04 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Synch_Options.h: Corrected the table of possible Use_Reactor, - Timeout combinations - it was reversed. - - * ace/WFMO_Reactor.{h cpp}: Change the delete_signal_handler_, - delete_timer_queue_, delete_handler_rep_, delete_notify_handler_, - and open_for_business_ members from int to bool to more - accurately reflect their nature as yes/no indicators. - Also, initialize notify_handler_ to 0 in constructors to avoid - a UMR warning from Purify. - -Wed Sep 27 06:28:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/msvc_static_compile.pl: - Added protocols directory - -Tue Sep 26 09:49:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/dynamicany.mpb: - Derive from valuetype, needed to get AbstractBase::_nil() in a no - inline build. - -Mon Sep 25 16:29:42 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/ACE.cpp (send_n_i): Conditionally removed the check for - errno == ENOBUFS because Winsock send, or WSASend will try to - enqueue all buffer at once (to the kernel) and if not it will - return ENOBUFS, in which case the socket is always ready for - writing but actually winsock can never perform such write and we - have endless loop. Thanks to Aleksandar Vukajlovic <vukajlo at - finsoft dot co dot yu> for reporting this and suggesting a fix. - -Mon Sep 25 11:39:35 UTC 2006 Boris Kolpackov <boris@codesynthesis.com> - - * ace/Bound_Ptr.inl: - * ace/Bound_Ptr.h: - - Applied workarounds that should fix compile errors on BCB6. - Thanks to Martin Corino <mcorino at remedy.nl> for providing - the patch. - - * protocols/ace/TMCast/Messaging.hpp: - * protocols/ace/TMCast/TransactionController.hpp: - - Updated to use ACE_Strong_Bound_Ptr. - -Fri Sep 22 12:42:37 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * docs/index.html: Added a link to the http://www.acejoy.com ACE - developers website in China. Thanks to Winston Zhang <webmaster - at acejoy dot com> for creating this website. - -Tue Sep 19 12:27:47 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Unload_libACE.cpp: Fixed the formatting of this file so it - conformed to ACE guidelines. - -Tue Sep 19 02:54:20 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/FILE_Connector.h: Changed the documentation to reflect the fact - that mkstemp() is used, not mktemp(). - -Thu Sep 21 16:51:34 UTC 2006 Ciju John <johnc@ociweb.com> - - * ace/TP_Reactor.h: - * ace/TP_Reactor.cpp: - The TP_Reactor signal handling implementation is - incomplete. In the current implementation under certain - circumstances (event loop breaks out if interrupted. This is - controlled via ACE_Select_Reactor_T::restart()), a properly - handled OS signal can break the reactor event loop in - an unpredictable manner. This is due to the fact that we cannot - predict which thread will receive the OS signal. Untill the - TP_Reactor signal handling is fixed, I am commenting it out. - -Thu Sep 21 15:13:19 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - Added new bug 2188 regression test. - -Mon Sep 18 10:56:29 UTC 2006 Steve Huston <shuston@riverace.com> - - * include/makeinclude/platform_sunos5_sunc++.GNU: Add buildbits=64 - branch for Sun Studio 11 (Sun C++ 5.8). Thanks to Aleksandar - Vukajlovic <vukajlo at finsoft dot co dot yu> for this. - - * THANKS: Added Aleksandar Vukajlovic to the Hall of Fame. - -Thu Sep 14 23:01:16 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Bound_Ptr.inl: Tweak syntax to try and work around apparant - issues with BCB6. - -Thu Sep 14 21:48:39 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_NS_Thread.inl (thr_getspecific): Removed the ACE_Errno_Guard - around ::TlsGetValue() in the ACE_HAS_WTHREADS case. The comments - near the code said (paraphrasing) "...it was to protect against - ACE_Log_Msg::instance() overwriting the error value before it had a - chance to be logged; although ACE_ERROR et al already store the - error value before calling ACE_Log_Msg::instance(), there may be a - chance that other uses of ACE_Log_Msg don't protect this way." - I have a report that having the errno guard in place is taking - over 10% CPU during a customer system's run-time. This is way too - much to have in a hot path, especially for a mis-placed guard. - It seems a little random to be picking on ACE_OS::thr_getspecific() - when there are many more OS calls in the ACE_Log_Msg::instance() - call path. - If there are really cases outside of ACE_ERROR... et al, then the - guard should be worked into ACE_Log_Msg::instance() method. However, - rather than try that now, let's see if there are any real issues, - since OS-level calls should be setting errno when errors are - noticed, propagating from GetLastError() via - ACE_OS::set_errno_to_last_error () as needed. - Thanks to Kelly Hickel <kfh at mqsoftware dot com> for raising - this point. - -Wed Sep 13 18:25:37 UTC 2006 Boris Kolpackov <boris@codesynthesis.com> - - * ace/Bound_Ptr.h: - * protocols/ace/RMCast/Reassemble.cpp: - - Re-applied my changes (Wed Sep 6) on build czar's (Steve - Huston) request. - -Tue Sep 12 18:07:57 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * docs/svn/svn-prefs.reg - * docs/svn/config - - Added pdf,eps,ps as binary file types. - -Tue Sep 12 13:56:22 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * apps/JAWS3/small/SS_Service_Handler.h: Added #include - "ace/svc_export.h". Thanks to Andy Wang <ypwangandy at gmail - dot com> for reporting this. - -Tue Sep 12 13:56:09 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * apps/JAWS3/small/SS_State_DONE.h: - * apps/JAWS3/small/SS_State_READ.h: Added #include - "ace/Synch_Traits.h". Thanks to Andy Wang <ypwangandy at gmail - dot com> for reporting this. - -Thu Sep 7 19:43:02 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/FILE_Connector.h (ACE_FILE_Connector): Updated the - documentation to remove mention of the return values. Thanks to - Hieu Ngyuen <hieu dot t dot nguyen at vanderbilt dot edu> for - reporting this. - -Mon Sep 11 13:03:17 UTC 2006 Boris Kolpackov <boris@codesynthesis.com> - - * ace/Bound_Ptr.h: - * protocols/ace/RMCast/Reassemble.cpp: - - Reverted my previous changes (Wed Sep 6) since BCB2006 - is unable to handle the required C++ constructs. - -Sun Sep 10 09:23:04 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * include/makeinclude/platform_g++_common.GNU - - Committing a fix provided by Ossama Othman (Ossama_Othman at symantec dot com) - to fix a problem with symbol visibility when building static libraries. - -Fri Sep 8 21:11:20 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/CDR_Stream.cpp (ACE_OutputCDR::replace()): Fixed compile error - from apparant cut/paste error in ACE_ENABLE_SWAP_ON_WRITE path. - Thanks to JR Andreassen for reporting this. - - * ace/Bound_Ptr.h: Changed new constructor/assignment to use - dynamic_cast in order to be able to transport pointers through - that can go both up and down a class hierarchy. - (See Wed Sep 6 07:40:03 UTC 2006 Boris Kolpackov - <boris@codesynthesis.com>). - - * protocols/ace/RMCast/Protocol.h (clone(), clone_()): Replaced use of - naked pointers returned from clone_() with ACE_Strong_Bound_Ptr and - used new ACE_Strong_Bound_Ptr ctor to facilitate carrying a more- - refined type through a Profile_ptr and back. Closes up a window - for exception unsafeness noticed by Boris Kolpackov. Also see - Tue Aug 22 22:52:41 UTC 2006 Steve Huston <shuston@riverace.com> - -Fri Sep 8 18:25:17 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Synch_Options.{h cpp}: Improved the documentation for this - class, particularly the table that explains the behavior for the - valid combinations of reactor-usage and timeout. Thank you to - Andriy Gapon <avg at icyb dot net dot ua> for motivating this. - - Also changed the return type of operator[] from int to bool; its - meaning was always 1/0 but was overlooked in the sweep for this - type of thing last year. - - * NEWS: Added the ACE_Synch_Options::operator[] return type change. - - * THANKS: Added Andriy Gapon to the Hall of Fame. - -Wed Sep 6 07:40:03 UTC 2006 Boris Kolpackov <boris@codesynthesis.com> - - * ace/Bound_Ptr.h: - - Added new constructor and assignment operator templates that - support the naked pointer-like implicit conversion. - - * protocols/ace/RMCast/Reassemble.cpp: - - Use the new constructor to get rid of reinterpret_cast. - -Tue Sep 5 21:28:02 UTC 2006 Nishanth Shankaran <nshankar@dre.vanderbilt.edu> - - * bin/MakeProjectCreator/config/ciao_race_component.mpb: - * bin/MakeProjectCreator/config/ciao_race_default.mpb: - * bin/MakeProjectCreator/config/ciao_race_monitor.mpb: Added new - RACE specific base projects. - -Fri Sep 1 19:58:06 UTC 2006 Ciju John <johnc@ociweb.com> - - * bin/tao_orb_tests.lst: - Turn on nightly regression test for bug 1476. - -Fri Sep 1 13:49:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * bin/PerlACE/Run_Test.pm: - Increased process creation timeout for tests on openVMS. - -Thu Aug 31 01:45:06 UTC 2006 Sumant Tambe <sutambe@nospam> - - Added .xml files to it. - * docs/svn/config: - * docs/svn/svn-prefs.reg: - -Wed Aug 30 17:45:30 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - This change fixes bug#2648 - - * ace/Service_Object.cpp: - - Fixed formatting. - - * ace/Service_Repository.cpp: - - Updated logging. - - * ace/Service_Types.h: - * ace/Service_Types.cpp: - - Added ACE_Service_Object_Type::initialized_ member variable to - hold the result of calling init() on the ACE_Service_Object. The - value is later consulted in ACE_Service_Object_Type::fini() to - determine if ACE_Service_Object::fini() should be - called. C++NPv2 (page 120) says that fini() must be called if - and only if, init() succeeded, i.e. returned 0. - -Wed Aug 30 17:18:52 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * bin/MakeProjectCreator/config/ciao_config_handlers.mpb - Modified this file to fixed the "reference to `ACE_Singleton<CIAO - ::Config_Handlers::XML_Helper, ACE_Null_Mutex>::instance()'"warnings - with FC5_Static build. - -Wed Aug 30 13:34:50 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/make_release - - Updates to the make_release script to make it subversion compatible. - major updates to be aware of: - . Assumes the working copy it uses is flattened - . Must have DOC_ROOT set - . Script should be started from DOC_ROOT. - . Produces traditional (nested) releases - . Now produces a set of source-only kits in addition to - the traditional kits. - - IMPORTANT: The release makefiles have not been updated, and are - most likely broken. - - Producing a release is a two step process: - make_release -u # Update version numbers - make_release -k ace+tao+ciao # Create all kits. - -Wed Aug 30 12:19:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/CDR_Stream.{h,cpp}: - Use true/false consistently for the good_bit - -Wed Aug 30 05:57:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Service_Config_Test.cpp: - Fixed argument not used warnings - -Tue Aug 29 21:17:26 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Service_Config_DLL.{h cpp}: - * tests/Service_Config_Test.cpp: Added a test for ACE_Service_Config's - behavior when a dynamic service's init() hook returns -1 (failure). - This was precipitated by Bugzilla #2648. - -Tue Aug 29 18:07:39 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * examples/IPC_SAP/SSL_SAP/SSL_SAP.mpc: - * tests/SSL/tests.mpc: - - Disable MPC projects in these files that depend on features not - found in the "ACE for TAO" subset", i.e. ACE_Profile_Timer and - asynch I/O. - -Tue Aug 29 13:19:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added bug 2560 regression. It runs fine on windows but seem to - crash on linux exhibiting a bug that isn't fixed yet. - -Tue Aug 29 13:17:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/PerlACE/ProcessVX.pm: - Commented out the ProcessVX_Unix require, it resulted in - errors on test systems when they don't have Net::Telnet () - available. - -Tue Aug 29 11:16:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/PerlACE/Run_Test.pm: - * bin/PerlACE/ProcessVX_Unix.pm: - * tests/run_test.pl: - Changes for running tests on linux with vxworks 6.3 rtp model target - -Tue Aug 29 08:20:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/run_test.lst: - Don't run UUIDTest in ACE_FOR_TAO configuration - - * bin/ace-tests.lst: - Don't run wfmo reactor tests in ACE_FOR_TAO configuration - -Tue Aug 29 00:24:48 UTC 2006 Adam Mitz <mitza@ociweb.com> - - * docs/run_test.txt: - Fixed Perl bug in example (needed "" instead of '' to get $ variable - expansion) and added explanatory text on using $PerlACE::ACE_ROOT - and $PerlACE::TAO_ROOT. - -Tue Aug 29 00:26:56 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/SSL/ssl.mpc - * ace/SSL/ssl_for_tao.mpc - * bin/MakeProjectCreator/config/ssl_full.mpb: - * bin/MakeProjectCreator/config/ssl.mpb: - * bin/MakeProjectCreator/config/ssl_for_tao.mpb: - - Added support for new ACE_SSL subset "ACE_SSL_FOR_TAO" for - "ace_for_tao" configuration. The ACE_SSL_FOR_TAO subset merely - lacks asychronous I/O support. - -Tue Aug 29 00:04:29 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/ACE_export.h: - - SGI C++ 7.3 compiler does not support standard C++ "extern - inline" declarations. - - * ace/Log_Msg.cpp (log): - - Corrected scalar/vector operator delete() mismatch. The wide - string buffer was allocated as an array. Destroy it - accordingly. - - * ace/MMAP_Memory_Pool.cpp (ACE_MMAP_Memory_Pool): - - Prevent potential dereferencing of null pointer. - - * ace/Svc_Conf_Param.h: - - Corrected author information. - -Mon Aug 28 23:59:50 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - From Amnon Berger <AmnonAB at gmail dot com> - * ace/SSL/SSL_Context.h (ACE_SSL_Context): - * ace/SSL/SSL_Context.cpp (certificate): - - Added support for reading an X.509 certificate from memory - instead of a file. - -Mon Aug 28 19:33:50 UTC 2006 Sumant Tambe <sutambe@nospam> - - A quick perl script to create a change-log-entry date and time. - * bin/cltime.pl: - - Added *.cidl *.pcd *.cdd - * docs/svn/config: - * docs/svn/svn-prefs.reg: - -Mon Aug 28 14:14:39 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * bin/tao_other_tests.lst: - Added TAO/orbsvcs/tests/Security/Big_Request/run_test.pl to the - list of SSL-based tests. Thanks to Johnny Willemsen - <jwillemsen@remedy.nl> for pointig out that it was missing. - -Mon Aug 28 13:45:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Svc_Conf_Lexer.cpp: - Fixed variable not used warning when ACE_USES_WCHAR is not - defined - - * ace/Min_Max.h: - Removed ACE_LACKS_MIN_MAX_TEMPLATES and the code that was - there when this define was set, this define wasn't enabled - in any configuration file. - -Mon Aug 28 11:53:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taodefaults.mpb: - Added exceptions as base project, TAO requires native C++ exceptions. - -Mon Aug 28 11:45:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * protocols/tests/RMCast/RMCast.mpc: - Removed avoids usage, that is already in the base projects - -Mon Aug 28 11:13:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * etc/*.doxygen: - Set ignore prefix in several files to improve generated - documentation - -Mon Aug 28 11:08:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * etc/acexml.doxygen: - Ignore ACEXML_ as prefix - -Mon Aug 28 10:34:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/run_test.txt: - Updated to explain how the script works on VxWorks and extended - the exampl in such a way that if there is not an executable - it returns directly - -Mon Aug 28 10:07:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/run_test.lst: - Added UUIDTest, was just not listed in this file - - * bin/PerlACE/ProcessVX_Unix.pm: - New initial file for running tests for VxWorks from an unix host - - * bin/PerlACE/ProcessVX.pm: - On unix use the new file above - -Mon Aug 28 09:05:20 UTC 2006 Boris Kolpackov <boris@codesynthesis.com> - - * bin/ace_tests.lst: - - Copied test exclusion list from the Multicast_Test entry. - - * protocols/tests/RMCast/Protocol.h: - * protocols/tests/RMCast/run_test.pl: - - Adjusted the timeout value and message size. - -Sun Aug 27 11:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added twoway DII collocation test - -Sun Aug 27 00:03:07 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Filecache.cpp (ACE_Filecache_Object): Added a call to - - this->lock_.release (); - - to release the lock and so the same file can be fetched again. - Thanks to Liu Qian <qianliu dot gmail at com> for reporting - this. - -Sat Aug 26 17:40:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/ACE-porting.html: - ACE_VXWORKS is the define for VxWorks, exceptions are required - for TAO. - -Sat Aug 26 17:36:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/ACE-guidelines.html: - Updated some parts to mention svn instead of cvs - -Sat Aug 26 03:11:01 UTC 2006 Krishnakumar B <kitty@spam.invalid.domain> - - * docs/svn/svn-prefs.reg: - * docs/svn/config: - - Updates to add even more file types. - -Sat Aug 26 02:32:18 UTC 2006 Krishnakumar B <kitty@spam.invalid.domain> - - * docs/svn/svn-prefs.reg: - * docs/svn/config: - - Updated these files to add more entries to the list of globally - ignored extensions, and also to the list of file types which - should have automatic properties set. Also synced up the config and - svn-prefs.reg file. - -Fri Aug 25 22:01:38 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Sock_Connect.cpp (get_ip_interfaces, count_interfaces): - Corrected to extract IPv6 information correctly from ioctl()- - returned structures. - - * tests/Enum_Interfaces_Test.cpp: On error condition ACE_HAS_IPV6 but - no ipv6 interfaces, do the log as LM_ERROR so the scoreboard picks - it up. - -Fri Aug 24 12:55:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-vxworks6.3.h: - * ace/Process.cpp: - * ace/Task.cpp: - Fixes for building on vxworks 6.3 with diab 5.4 compiler - -Fri Aug 24 11:17:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/CDR_Size.cpp: - Added include of OS_Memory.h to get ACE_align_binary - -Thu Aug 24 21:40:24 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Process_Manual_Event_Test.cpp (acquire_release): Fixed a race - condition whereby while a parent and child were ping-ponging, it - was possible for one process to signal event1, wait on event2, then - signal event1 again before the other process could respond to - event1 being signaled and reset it. Thus, event1 was being reset - after the other process signaled it, ensuring a never-ending - wait at the end of the iterations. I changed the code to quickly - reset the event before printing the debug message, thereby closing - the window of opportunity a lot. But probably not completely - removing it... we'll have to see. - -Thu Aug 24 16:40:54 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-aix-4.x.h: Only need to set - ACE_HAS_EXPLICIT_STATIC_TEMPLATE_MEMBER_INSTANTIATION for g++ - prior to g++ 4.1 (have only heard report of this for g++ 4.1; - it may need adjustment for 4.0). Thanks to Nzer Zaidenberg - <scipioenterprises at yahoo dot com> for reporting the original - problem and testing this fix. - - * THANKS: Added Nzer Zaidenberg to the Hall of Fame. - -Thu Aug 24 16:33:22 UTC 2006 James H. Hill <hillj@isis.vanderbilt.edu> - - * docs/svn/config: - * docs/svn/svn-prefs.reg: - - Added auto-props for *.tex and *.cls files. - -Thu Aug 24 13:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-openvms.h - Removed ACE_HAS_SOCK_BUF_SIZE_MAX, not needed anymore with the - latest OpenVMS versions - -Thu Aug 24 08:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/DLL.h - * ace/Service_Repository.h - * ace/Service_Object.cpp - * ace/Parse_Node.cpp - * ace/Service_Object.inl - * ace/Parse_Node.h - * ace/Service_Repository.cpp - * ace/Service_Object.h - * ace/Service_Gestalt.h - * ace/Service_Gestalt.cpp - * ace/DLL_Manager.cpp - * ace/DLL.cpp - Reverted change from below on request of Iliyan because it broke - a set of tests that spawn a thread in init. - Tue Aug 22 21:43:34 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - -Wed Aug 23 08:32:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Sock_Connect.{h,cpp}: - Reverted the change below, VxWorks/OpenVMS/AIX all have problems - after this change - - Tue Aug 22 15:19:46 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - * ace/Sock_Connect.h: - * ace/Sock_Connect.cpp: - -Tue Aug 22 22:52:41 UTC 2006 Steve Huston <shuston@riverace.com> - - * protocols/ace/RMCast/Protocol.h: Replaced use of - ACE_Refcounted_Auto_Ptr with ACE_Strong_Bound_Ptr. Code formerly used - ACE_Refcounted_Auto_Ptr::release() with the assumption it had the - same semantics as auto_ptr::release() - stealing the pointer away. - This used to work by accident. The fixed ACE_Refcounted_Auto_Ptr's - release() method decrements the reference count which may cause the - object to be deleted, which is what happened all the time in the - RMCast protocol's clone() methods. Because all the clone() methods - relied on being able to take a more-derived pointer and use it as - a Profile*, changing the refcounted pointer type necessitated - changing the protected clone_() methods to return a naked pointer - rather than a refcounted one. This is acceptable because the naked - pointer is not exposed to class users; the public methods insert - the naked pointer into a ACE_Strong_Bound_Ptr before giving access - to it externally. - - * protocols/ace/RMCast/Acknowledge.cpp: - * protocols/ace/RMCast/Reassemble.cpp: With properly functioning - refcounted pointers (see above) these needed a few changes to - make proper use of the pointers while maintaining correct reference - counts. - -Tue Aug 22 21:43:34 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - This change fixes bug#2612. - - * ace/DLL.h: - * ace/DLL.cpp: - - Made assignment return 'T&' instead of 'const T&'. Adjusted code - formatting and logging messages. - - * ace/DLL_Manager.cpp: - - Fixed formatting. - - * ace/Parse_Node.h: - * ace/Parse_Node.cpp: - - Added accessor for the underlying ACE_DLL and adjusted logging. - - * ace/Service_Gestalt.h: - * ace/Service_Gestalt.cpp: - - Introduced ACE_Service_Type_DLL_Guard class. As dynamic service - objects are loaded, static service objects that come with the - same DLL code, may also be registered. The static services - however, have no information about the context in which their - registration takes place and their relation with the DLL is - lost. A situation can easily arise where the dynamic service - object is finalized and its DLL - unmapped, but some of the - DLL's static services may still be around. Their finalization - would not be possible, if the DLL counting that code is gone. - - The guard, when instantiated (on the stack) swaps out the - current Gestalt and its service repository, and replaces them - with a "sandbox" Gestalt, which has lifetime sufficient for just - the current service initialization. Note that this mechanism is - effective even in the case where a service initialization causes - re-entry into the Gestalt to load and initialize another service - as part of the initialization. - - If no action is taken (in case of an error, for instance) the - guard destructor will simply finalize all newly initialized - service(s) from the sandbox. On the other hand, if - initialization is successful the service object descriptors are - "relocated" to the original Gestalt. - - During the relocation, all static services are "fixed-up" to - hold a reference to the DLL they belong to. Effectively, a - static service, initialized via loading of a particular DLL, - becomes a dynamic service. Thus the order of service - finalization does not matter anymore because the DLL will be - help in memory until the last service object it provides is - finalized. - - * ace/Service_Object.h: - * ace/Service_Object.inl: - * ace/Service_Object.cpp: - - Added a relocate method to modify the ACE_DLL each SO has. - - * ace/Service_Repository.h: - * ace/Service_Repository.cpp: - - Befrended ACE_Service_Type_DLL_Guard. Adjusted logging. - - * tests/Service_Config_Test.cpp: - - Simplified test case. - -Tue Aug 22 18:37:48 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/MakeProjectCreator/config/ipv6.mpb: - - Building with IPv6 for WinCE requires an additional library for - rtti that is not required for the windows builds. - - * bin/ciao_tests.lst: - * bin/tao_orb_tests.lst: - * bin/tao_other_tests.lst: - Added ACE_FOR_TAO exclusions as necessary - -Tue Aug 22 15:19:46 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Sock_Connect.h: - * ace/Sock_Connect.cpp: - - Refactored interface discovery code. The methods used to find - IPv6 interface addresses is highly platform specific. So much so - that the single method was too confusing to maintain. With this - patch, the code has been refactored to put each - platform-specific variation in its own method, then only compile - the method needed by any given platform. I'm not sure this will - yield much of a footprint reduction, as all the existing code is - still present, it is just easier to keep track of what is going on. - - Also, when IPv6 is enabled, it might be on a platform that is - using IPv6 exclusively, ie, with no IPv4 enabled. At least one - third-party user of ACE requested a specific test for IPv4 - enablement status, thus I have added a new method for that test. - - * ace/config-linux.h: - Linux supplies the getipaddrs system call. - - * tests/Enum_Interfaces_Test.cpp: - Extended the test so that if ACE is built with ACE_HAS_IPV6 - defined, the test will expect at least one IPv6 endpoint or it - will fail. - -Tue Aug 22 11:41:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/UUID.{h,cpp}: - Implemented copy constructor - - * tests/UUIDTest.cpp: - Test copy constructor - -Tue Aug 22 09:36:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/UUID.{h,inl}: - Made the accessors const and reimplemented operator == and - operator !=. - - * tests/UUIDTest.cpp: - Test the operator != and ==. - -Mon Aug 21 20:46:43 UTC 2006 Steve Huston <shuston@riverace.com> - - * configure.ac: Modified the check for the net/if.h header file - to do a #include <sys/socket.h> if it exists. Allows the compile - of net/if.h to succeed, fixing a configure warning for a file that - exists but won't compile. Thanks to Don Meek <Don dot Meek at - windstream dot com> for reporting this problem. - - * THANKS: Added Don Meek to the Hall of Fame. - -Mon Aug 21 20:43:59 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.inl: - - Added the missing implementation of the static process_file () - method. Thanks to Lothar Werzinger <lothar@tradescape.biz> for - pointing that out. - -Mon Aug 21 20:06:38 Pacific Daylight Time 2006 Ossama Othman <ossama_othman at symantec dot com> - - From Russell Mora <russell_mora at symantec dot com>: - * ace/Thread_Exit.cpp (instance): - - Fixed race condition in generated code by stopping optimizer - from re-ordering static variable initializations by declaring - static "instance_" variable as "volatile". This problem was - observed on a quad-CPU Solaris 8 box. - -Mon Aug 21 18:31:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * etc/*.doxygen: - In all files except ace.doxygen added SHORT_NAMES=YES to let - doxygen generate shorter filenames, this should fix the problems - we had during the release with cpio not capable of handling - long filenames - -Mon Aug 21 17:46:55 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Numeric_Limits.h (ACE_Numeric_Limits): - - The Microsoft Platform SDK does not provide - std::numeric_limits<> specializations for 64 bit integers so we - need to explicitly provide ACE_Numeric_Limits<> specializations - to compensate for this deficiency. [Bug 2639] - - Unfortunately there is no way to tell if the platform SDK is - being used so we specialize for the ACE_WIN64 + MSVC++ 7.1 case, - which is the configuration that exhibits this problem. It also - happens to be a fairly isolated configuration since 64-bit - support in MSVC++ 7.1 was not very good to begin with. - -Mon Aug 21 15:35:26 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Timeprobe_T.cpp (print_times): More fixed for timestamp - inversion due to recording time from different threads. This - prints a warning message the first time it detects a timestamp - inversion. It also fixes various minor problems in the - timeprobe reports whenever the total number of timeprobes - exceeds 8192. Thanks to Patrick Rabau for contributing this. - - * ace/INET_Addr.h: Updated the documentation of the non-reentrant - versions of get_host_addr() and get_host_name() to clarify how - they can be used correctly (or better yet not used at all..). - Thanks to Ephy Levy <EphyLevy at gmail dot com> for motivating - this. - - * ace/Logging_Strategy.cpp: Added a destructor to delete - this->filename_, which is allocated in the constructor. Thanks - to Amnon AB <AmnonAB at gmail dot com> for reporting this. - - * ace/Acceptor.h: Made the suspend()/resume() methods of - ACE_Strategy_Acceptor public rather than protected. Thanks to - Amnon AB <AmnonAB at gmail dot com> for reporting this. - - * docs/Symbol_Versioning.html (HREF): Added a link to the Riverace - webpage that describes symbol versioning in ACE. Thanks to - Steve Huston for contributing this. - - * ace/Timeprobe_T.cpp (print_times): Revised the code so that if - there is a timestamp inversion due to recording time from - different threads, the time difference will be displayed as a - negative value instead of a gigantic positive value. Also - compute the time difference correctly when the timeprobe array - wraps around. Thanks to Patrick Rabau <pr2345 at gmail dot com> - for this fix. - - * ace/Containers_T.cpp, - * ace/Log_Msg.cpp, - * ace/CDR_Size.cpp, - * ace/CDR_Size.inl, - * ace/CDR_Stream.cpp: Fixed a bunch of problems with type punning. - Thanks to Thomas Girard <thomas.g.girard at free dot fr> for - reporting this. - - * ace/TSS_T.cpp (tryacquire): There was a missing '&' in - - ACE_Thread::getspecific (this->key_, temp); - - Thanks to Thomas Girard <thomas.g.girard at free dot fr> for - reporting this. - -Mon Aug 21 10:45:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/UUID.{h,cpp,inl}: - Extended UUID to make it possible to assign a string to a UUID - to set it a new value. - - * tests/UUIDTest.cpp: - Extended this test to test the new functionality above. - -Mon Aug 21 10:26:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/UUID.{h,cpp,inl}: - Changed node_release_ to a real bool and use ACE_ERROR instead - of ACE_DEBUG for a real error message - - * ace/OS_NS_unistd.cpp: - Const improvement - -Sat Aug 19 22:03:03 UTC 2006 Yan Dai <dai_y@ociweb.com> - - * bin/tao_orb_tests.lst - - Excluded the DII collocation oneway test from running in the minimum - CORBA build. - -Fri Aug 18 20:38:35 UTC 2006 Steve Huston <shuston@riverace.com> - - * examples/APG/Logging/Trace.h: Further changed use of __VA_ARGS__ - to remove the comma and the fixed FMT arg from the macros. This - may still trigger the same ISO C99 warning, but it doesn't fail. - -Fri Aug 18 17:22:08 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_NS_unistd.cpp (pwrite): Corrected the large-file-offset - logic to build correctly for Windows. Also moved the explicit - SetFilePointerEx() before WriteFile() to the non-ACE_HAS_WINNT4 - section - for ACE_HAS_WINNT4 the supplied OVERLAPPED structure - carries the file offset to write at. - -Fri Aug 18 17:08:35 UTC 2006 Yan Dai <dai_y@ociweb.com> - - * bin/tao_orb_tests.lst - - Added DII collocation oneway test. This is the regression test - for bugzilla #2545. - -Fri Aug 18 15:49:01 UTC 2006 Steve Huston <shuston@riverace.com> - - * examples/APG/Logging/Trace.h: Changed the varargs form used in the - macros from ##__VA_ARGS__ to ,__VA_ARGS__ in an effort to conform - to C99 and clear some compile warnings from gcc4. This effectively - prohibits leaving the varargs out completely from the macro - invocation but that use-case isn't exercised anyway. - - * bin/ace_tests.lst: Add protocols/tests/RMCast/run_test.pl to run - RMCast tests, except ACE_FOR_TAO builds. - -Fri Aug 18 12:03:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Don't run Bug_2510_Regression in a ST build, it uses threading - - * bin/MakeProjectCreator/templates/bor.mpd: - Don't print a message with the dummy realclean - -Thu Aug 17 14:20:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * apps/drwho/Binary_Search.cpp: - * apps/drwho/CM_Client.cpp: - * apps/drwho/CM_Server.cpp: - * apps/drwho/Hash_Table.cpp: - * apps/drwho/Options.cpp: - * apps/drwho/Options.h: - * apps/drwho/PMC_All.cpp: - * apps/drwho/PMC_Flo.cpp: - * apps/drwho/PMC_Ruser.cpp: - * apps/drwho/PMC_Usr.cpp: - * apps/drwho/PMS_All.cpp: - * apps/drwho/PMS_Flo.cpp: - * apps/drwho/PMS_Ruser.cpp: - * apps/drwho/PMS_Usr.cpp: - * apps/drwho/Protocol_Manger.cpp: - * apps/drwho/Protocol_Record.cpp: - * apps/drwho/Rwho_DB_Manager.cpp: - * apps/drwho/Search_Struct.cpp: - * apps/drwho/Single_Lookup.cpp: - * apps/Gateway/Gateway/Gateway.cpp: - * apps/Gateway/Gateway/Options.cpp: - * apps/Gateway/Gateway/Options.h: - - Renamed the enum ID "DEBUG" to "DEBUGGING", was causing problems - with MFC builds. - -Thu Aug 17 12:15:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/SString_Test.cpp: - Added a check whether npos is zero, if it is print an error, - with Intel C++ 9.0 with the Microsoft platform SDK this seems - to be zero - -Thu Aug 17 08:30:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/ciao_client_dnc.mpb: - * bin/MakeProjectCreator/config/ciao_deployment_stub.mpb: - * bin/MakeProjectCreator/config/ciao_deployment_svnt.mpb: - * bin/MakeProjectCreator/config/ciao_events_dnc.mpb: - * bin/MakeProjectCreator/config/ciao_rtevent_dnc.mpb: - Removed -Sc form the taoidlflags, it is deprecated already sometime - ago - -Wed Aug 16 22:19:47 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Acceptor.h: Doxygen improvements. - -Wed Aug 16 17:42:02 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Numeric_Limits.h: - - Explicitly #undef min and max before including <limits> in the - Mingw case to address global namespace pollution induced - conflicts. The previous attempt to address this problem where - the Windows-specific NOMINMAX preprocessor symbol was defined in - this header didn't work since the min and max macros were - already defined beforehand. In order for NOMINMAX to work as - expected, it would be have to be defined before any Windows - headers are included, or more preferrably on the - preprocessor/compiler command line. Unfortunately, that would - probably break some applications. - -Wed Aug 16 14:51:58 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * docs/svn/svn-prefs.reg - * docs/svn/config - - Updated to correct keywords values, and add use-commit-times - as a default option. - -Wed Aug 16 10:19:41 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Obstack_T.{cpp,inl}: - Moved grow_fast and freeze methods from inl to cpp file, - should fix the link problems with the RHAS IPv6 build - -Wed Aug 16 09:50:41 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Misc/Misc.mpc: - * examples/Misc/test_timestamp.cpp: - Simple small test for ACE::timestamp() - -Wed Aug 16 08:37:41 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/build_lib.bor: - * include/makeinclude/clean.bor: - * include/makeinclude/compiler.bor: - Removed support for the CBuilderX preview compiler, it has never - become more then preview and we don't test with it anymore - - * include/makeinclude/recurse.bor: - Removed from the repo, seems not to be used anymore - -Tue Aug 15 17:47:11 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Numeric_Limits.h: - - Windows defines min/max macros that interfere with the - numeric_limits::min/max() traits. Prevent those macros from - being defined by defining the Windows-specific NOMINMAX symbol - before any Windows headers are included. This is currently - only done for Mingw builds since the problem is not exhibited by - other Windows builds (e.g. MSVC++, Borland, etc). - -Tue Aug 15 16:17:56 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * ace/Hash_MultiMap_Manager_T.h - Made some minor changes to the documentation to use @a to refer to - argument names. - -Tue Aug 15 15:00:00 UTC 2006 Adam Mitz <mitza@ociweb.com> - - * ace/Event_Handler.cpp: - - Updated Event_Handler_var to use an ACE_Errno_Guard to - prevent overwriting the errno. - - * bin/tao_orb_tests.lst - - Added new test: TAO/tests/Oneway_Timeouts/run_test.pl - -Tue Aug 15 14:22:53 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * ace/Hash_MultiMap_Manager_T.h - Address the "unsupported xml/html tag" warnings in doxygen build. - -Tue Aug 15 08:36:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * bin/MakeProjectCreator/config/acedefaults.mpb - - Backout all of the recent MFC changes. - - * bin/MakeProjectCreator/config/ace_opt_mfc.mpb: - - DELETED. - -Mon Aug 14 23:34:11 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * ace/TP_Reactor.cpp - * ace/TP_Reactor.h - - Removed all register_handler member functions from these files. - This essentially enables the ability to register signal handlers in - the TP Reactor by implicitly delegating to the select reactor base - class. - - Other register_handler methods (not relating to signal handling) - were explicit delegations to the select reactor base class, present - only to silence virtual function hiding warnings created by - disabling the signal related methods. - -Mon Aug 14 16:14:56 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * test/Hash_Multi_Map_Manager_Test.cpp - Address the "no matching function for call" errors. - - * ace/Hash_MultiMap_Manager_T.h - Address the "unsupported xml/html tag" warnings in doxygen build. - -Mon Aug 14 16:00:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * bin/MakeProjectCreator/config/ace_mfc.mpb: - * bin/MakeProjectCreator/config/ace_opt_mfc.mpb: - - Backout the "Specific" guards. - -Mon Aug 14 15:20:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * bin/MakeProjectCreator/config/ace_mfc.mpb: - * bin/MakeProjectCreator/config/ace_opt_mfc.mpb: - - Specific the macro and don't inherit from mfc for opt. - -Mon Aug 14 15:43:12 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Makefile.am: Added Hash_Multi_Map_Manager_Test. - -Mon Aug 14 13:33:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/clean.bor: - Use .inl as file extension to clean instead of .i - - * bin/MakeProjectCreator/templates/bor.mpd: - Added empty realclean target - -Mon Aug 14 12:52:19 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * ACE-INSTALL.html: - - I added "-type gnuace" to the options where the default project - type was assumed. Eventually, MPC will not have a default project - type. - - * include/makeinclude/macros.GNU: - * include/makeinclude/rules.common.GNU: - * include/makeinclude/rules.local.GNU: - - Added a new target to remove only the binaries (executables and - shared libraries). - -Mon Aug 14 11:15:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/*.h: - Added /**/ to the include of config-all.h and ACE_export.h, this - way doxygen doesn't add these files to the include graphs which - make them much easier to read - -Sun Aug 13 23:30:02 UTC 2006 Yan Dai <dai_y@ociweb.com> - - * bin/tao_other_tests.lst: - - Excluded TAO POAManagerFactory and EndpointPolicy tests from - running in a minimum CORBA configuration. - -Sun Aug 13 22:21:55 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * test/Hash_Multi_Map_Manager_Test.cpp - Address the "Implicit conversion of string literal to char * is - deprecated" warnings. - -Sun Aug 13 15:25:17 UTC 2006 Boris Kolpackov <boris@codesynthesis.com> - - * bin/MakeProjectCreator/config/ciaocidldefaults.mpb: - - Removed '--' from the cidlc command line. - -Sun Aug 13 000:02:00 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * ace/Hash_MultiMap_Manager_T.inl - * ace/Hash_MultiMap_Manager_T.cpp - * ace/Hash_MultiMap_Manager_T.h - Change the name of these files into Hash_Multi_Map_Manager.*. - Add some comments to explain in brief the purpose of using - ACE_Hash_Multi_Map_Manager class and how to use it. - Remove the backward compatibility of ACE_Hash_Multi_Map_Manager class - and its associated iterator classes. - Remove "_Ex" on the names of ACE_Hash_Multi_Map_Manager class and its - associated iterator classes. - - * ace/Hash_MultiMap_Manager.h - Remove this file since it simply includes the - Hash_MultiMap_Manager_T.h file. - - * test/Hash_MultiMap_Manager_Test.cpp - Change the name of this file into Hash_Multi_Map_Manager_Test.cpp. - Address the "convert const char* to ACE_TCHAR*" errors. - - * ace/ace.mpc - Change the name of Hash_MultiMap_Manager project into - Hash_Multi_Map_Manager. - - * test/run_test.lst - * test/tests.mpc - Change the name of Hash_MultiMap_Manager_Test project into - Hash_Multi_Map_Manager_Test. - -Fri Aug 11 16:46:56 UTC 2006 Shanshan Jiang <shanshan.jiang@vanderbilt.edu> - - * ace/Hash_MultiMap_Manager.h - * ace/Hash_MultiMap_Manager_T.cpp - * ace/Hash_MultiMap_Manager_T.h - * ace/Hash_MultiMap_Manager_T.inl - Add these files which implement Hash_MultiMap_Manager. - - * test/Hash_MultiMap_Manager_Test.cpp - Add this test file for Hash_MultiMap_Manager. - - * ace/ace.mpc - Modify this file to include Hash_MultiMap_Manager. - - * test/run_test.lst - * test/tests.mpc - Modify these files to include Hash_MultiMap_Manager_Test. - -Fri Aug 11 16:37:34 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/CDR_File_Test.cpp: Corrected a signed/unsigned compare. - -Fri Aug 11 13:59:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/CDR_Size.cpp: - Add include of OS_Memory.h to get ACE_align_binary - - * ace/Sig_Handler.cpp: - Add include of Guard_T.h to get the guards - - * ace/config-vxworks6.3.h: - Add some code for the diab compiler but this is not complete - support, just some testing results. - -Fri Aug 11 07:17:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/svn/config: - * docs/svn/svn-prefs.reg: - Added settings for .idl and .pidl - -Thu Aug 10 13:14:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_other_tests.lst: - Disabled the performance tests on OpenVMS for the moment, they - result in too much error logs - -Thu Aug 10 10:33:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - Updated for -GX - -Thu Aug 10 09:38:13 UTC 2006 Simon McQueen <sm@prismtech.com> - - * bin/tao_other_tests.lst: - - Exclude Bug_2615_Regression from running on 'ACE for TAO' builds. - -Thu Aug 10 01:52:34 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-aix-4.x.h: Added ACE_HAS_TEMPLATE_TYPEDEFS for Visual - Age C++ 6 and higher. - -Wed Aug 9 14:15:18 UTC 2006 Ciju John <johnc@ociweb.com> - - * bin/MakeProjectCreator/config/csd_threadpool.mpb: - - Added a dependency upon threads.mpb as this project shouldn't be - built if thread support is not going to be built into TAO. - -Wed Aug 9 13:45:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * bin/MakeProjectCreator/config/acedefaults.mpb - - Typo, Remove the + character that came with the patch file. - -Wed Aug 9 12:17:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * bin/MakeProjectCreator/config/ace_opt_mfc.mpb - * bin/MakeProjectCreator/config/acedefaults.mpb - - Chad's new fix for MPC building MFC projects. - -Tue Aug 8 14:00:06 UTC 2006 Carlos O'Ryan <coryan@atdesk.com> - - * Merged in all the changes from the fix_bug_2540 branch, that is, - from revision 73857 to revision 73859. - - Thu Aug 3 21:49:27 UTC 2006 Carlos O'Ryan <coryan@atdesk.com> - - * ace/TP_Reactor.cpp: - * ace/Select_Reactor_T.cpp: - Avoid the problems described in bug 2540 by detecting the - infinite loop and recomputing the number of active handlers. - Thanks to Jody Hagins for correcting my original "fix." - -Tue Aug 8 14:05:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * bin/MakeProjectCreator/config/ace_mfc.mpb - * bin/MakeProjectCreator/config/acedefaults.mpb - - Reverted Chad's fix for MPC building MFC projects. - It seems that this causes non MFC builds to build as MFC - breaking the scoreboard. - -Tue Aug 8 09:30:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE.mwc: - Removed several directories which aren't existing anymore - -Tue Aug 8 09:28:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: - Removed some parts of the VxWorks building instructions because - it was really wrong what was said - -Tue Aug 8 09:15:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * ace/OS_Memory.h - * ace/config-win32-msvc.h - * bin/MakeProjectCreator/config/ace_mfc.mpb - * bin/MakeProjectCreator/config/acedefaults.mpb - - Chad's fix for MPC building MFC projects. Correction for - Thu Apr 27 21:04:08 UTC 2006 Steve Huston <shuston@riverace.com> - throw CMemoryException on out-of-memory condition. These - exceptions must be caught by pointer and thrown via calling - AfxThrowMemoryException (). - -Tue Aug 8 08:06:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Select_Reactor_Base.cpp: - * ace/Select_Reactor_T.cpp: - Const improvements, use bool when possible - -Tue Aug 8 07:46:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/generate_rel_manpages: - * bin/make_release: - Did a little bit of cleanup removing old things - -Tue Aug 8 07:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Select_Reactor_T.cpp (dump): - Removed usage of not needed operator++ on the handle set - -Mon Aug 7 08:02:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Handle_Set.{h,cpp}: - Removed operator++, it is deprecated and nothing more then a noop - -Fri Aug 4 07:57:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Asynch_IO.h: - * ace/Event_Handler.h: - * ace/Framework_Component_T.h: - * ace/Hash_Map_Manager_T.h: - * ace/IO_SAP.h: - * ace/Local_Tokens.h: - * ace/LSOCK_Connector.h: - * ace/Map_Manager.h: - * ace/Map_T.h: - * ace/WFM_Reactor.h: - Doxygen improvements - -Thu Aug 3 10:39:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Connector.h: - Improved documentation - -Wed Aug 2 22:54:24 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * THANKS: Added Eric Danielou to the hall of fame. - -Wed Aug 2 21:27:59 UTC 2006 Adam Mitz <mitza@ociweb.com> - - * bin/PerlACE/Run_Test.pm: - - Added two new packaged-scoped variables, $PerlACE::ACE_ROOT and - $PerlACE::TAO_ROOT. $PerlACE::TAO_ROOT is set to the TAO_ROOT - environment variable, if one exists. If not it is set to - $ACE_ROOT/TAO. $PerlACE::ACE_ROOT is the same as the ACE_ROOT - envrionment variable so it's just there for convenience and for - symmetry. - -Wed Aug 2 20:19:04 UTC 2006 Adam Mitz <mitza@ociweb.com> - - * bin/tao_other_tests.lst: - Added new test (TAO/orbsvcs/tests/ImplRepo/Bug_2604_Regression). - -Wed Aug 2 18:16:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Connector.cpp: - Const improvements - - * ace/Caching_Utility_T.h: - * ace/config-lite.h: - * ace/Event_Handler.h: - * ace/Framework_Component.h: - * ace/Lock.h: - * ace/Lock_Adapter_T.h: - * ace/Null_Mutex.h: - * ace/Object_Manager.h: - * ace/Object_Manager_Base.h: - * ace/Proactor.cpp: - * ace/Process_Manager.h: - * ace/Semaphore.h: - * ace/Thread_Exit.h: - Doxygen improvements - -Wed Aug 2 15:50:48 UTC 2006 Simon McQueen <sm@prismtech.com> - - * bin/tao_other_tests.lst: - - Schedule bugzilla #2615 regression test. - -Wed Aug 2 14:55:45 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * ACE-INSTALL.html: - - Updated the location of the MPC documentation. It was moved from - the MPC root directory to a docs directory. - - * bin/ChangeLogEditor/CVSFileLocator.pm: - * bin/ChangeLogEditor/ChangeLogEdit.pm: - * bin/ChangeLogEditor/ChangeLogEntry.pm: - * bin/ChangeLogEditor/FileLocator.pm: - * bin/ChangeLogEditor/SVNFileLocator.pm: - * bin/cle.pl: - - Added better support for subversion and it now reports errors from - cvs/svn. - - * bin/MakeProjectCreator/README: - - Updated to provide information on anonymously checking out MPC - using Subversion. - - * bin/make_release: - - I added "-type gnuace" to the options where the default project - type was assumed. Eventually, MPC will not have a default project - type. - -Wed Aug 2 13:47:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/ACE-development-process.html: - Document the fact that people should test on at least two - platforms and should be available for some days after they - commit a change. - -Wed Aug 2 11:16:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/Download.html: - Also put the download page from deuce.doc under svn control - so that we can update it for the svn conversion - -Wed Aug 2 11:07:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/bczar/bczar.html: - * docs/bczar/privileges.html: - Added documents that where on deuce.doc, the documentation - for the bczar how to create a release should really be handled - with care, so put them under svn control - -Tue Aug 1 20:23:50 UTC 2006 Adam Mitz <mitza@ociweb.com> - - * ace/config-macosx-tiger.h: - Added preliminary support for Mac OS X 10.4 on Intel CPU's. - -Tue Aug 1 20:00:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-vxworks6.3.h: - Enabled some features in ACE because vxworks 6.3 delivers more system - api's - -Mon Jul 31 19:23:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_stropts.inl (ioctl): - Check for ACE_HAS_IOCTL_INT_3_PARAM to make it easier to handle - the different VxWorks versions - - * config-vxworks5.x.h: - * config-vxworks6.2.h: - * config-vxworks6.3.h: - Added ACE_HAS_IOCTL_INT_3_PARAM - -Mon Jul 31 18:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * Containers_T.{h,cpp}: - * Hash_Map_Manager_T.{h,cpp}: - * Map_Manager.{h,inl}: - * Node.{h,cpp}: - Removed check for !ACE_HAS_BROKEN_NOOP_DTORS, this define is - not in any config file - - * config-unixware-7.1.0.udk.h: - Removed commented out define - - * Token_Collection.{h,cpp}: - * Token_Manager.{h,inl}: - Changed debug flag to be a bool - - * SOCK_Dgram_Bcast.h - * TP_Reactor.h - * Signal.h - * Thread.h - * Thread_Adapter.h - Documentation updates - -Mon Jul 31 14:57:08 UTC 2006 Steve Huston <shuston@riverace.com> - - * apps/Makefile.am: Removed hard-coded inclusion of gperf in the - SUBDIRS list; it's an optional piece; see also - Tue Jun 20 20:19:37 UTC 2006 Steve Huston <shuston@riverace.com> - Thanks to Robert Schwebel <robert at schwebel dot de> for this fix. - -Mon Jul 31 14:36:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * etc/*.doxygen: - Cleanup of these files - -Sat Jul 29 11:26:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-icc-common.h: - Added fix for Intel C++ 9.1 - -Fri Jul 28 15:32:27 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * ace/SSL/Makefile.am - * ace/QoS/Makefile.am - * performance-tests/Synch-Benchmarks/Base_Test/Makefile.am - * performance-tests/Synch-Benchmarks/Perf_Test/Makefile.am - * Kokyu/Makefile.am - * apps/JAWS/clients/Caching/Makefile.am - * ACEXML/apps/svcconf/Makefile.am - * ACEXML/common/Makefile.am - * ACEXML/parser/parser/Makefile.am - * ACEXML/examples/SAXPrint/Makefile.am - * websvcs/lib/Makefile.am - * examples/ASX/Event_Server/Event_Server/Makefile.am - * examples/ASX/UPIPE_Event_Server/Makefile.am - * examples/Service_Configurator/IPC-tests/server/Makefile.am - - Updated to comply with the .i->.inl changes from: - Thu Jul 27 20:30:26 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - -Fri Jul 28 12:17:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * Kokyu/DSRT_Dispatch_Item_T.inl: - * ACEXML/parser/parser/Parser.inl: - Fixed file tag - -Thu Jul 27 20:30:26 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * ace/SSL/SSL_SOCK_Connector.h - * ace/SSL/SSL_SOCK_Acceptor.inl - * ace/SSL/SSL_SOCK_Acceptor.cpp - * ace/SSL/SSL_SOCK_Connector.i - * ace/SSL/SSL_SOCK_Stream.h - * ace/SSL/SSL_SOCK_Stream.i - * ace/SSL/SSL_SOCK_Acceptor.h - * ace/SSL/SSL_SOCK_Acceptor.i - * ace/SSL/SSL_SOCK.cpp - * ace/SSL/SSL_SOCK.inl - * ace/SSL/SSL_SOCK.h - * ace/SSL/SSL_SOCK.i - * ace/SSL/SSL_SOCK_Connector.cpp - * ace/SSL/SSL_SOCK_Connector.inl - * ace/SSL/SSL_SOCK_Stream.cpp - * ace/SSL/SSL_SOCK_Stream.inl - * ace/QoS/SOCK_Dgram_Mcast_QoS.inl - * ace/QoS/SOCK_Dgram_Mcast_QoS.cpp - * ace/QoS/QoS_Session_Impl.cpp - * ace/QoS/QoS_Session_Impl.inl - * ace/QoS/SOCK_Dgram_Mcast_QoS.h - * ace/QoS/SOCK_Dgram_Mcast_QoS.i - * ace/QoS/QoS_Session_Impl.h - * ace/QoS/QoS_Session_Impl.i - * performance-tests/Synch-Benchmarks/Base_Test/Baseline_Test.cpp - * performance-tests/Synch-Benchmarks/Base_Test/Baseline_Test.inl - * performance-tests/Synch-Benchmarks/Base_Test/Baseline_Test.h - * performance-tests/Synch-Benchmarks/Base_Test/Baseline_Test.i - * performance-tests/Synch-Benchmarks/Perf_Test/Performance_Test_Options.inl - * performance-tests/Synch-Benchmarks/Perf_Test/Performance_Test_Options.h - * performance-tests/Synch-Benchmarks/Perf_Test/Performance_Test_Options.i - * Kokyu/Kokyu_dsrt.cpp - * Kokyu/Kokyu_dsrt.inl - * Kokyu/Kokyu.cpp - * Kokyu/Kokyu.inl - * Kokyu/Kokyu_defs.h - * Kokyu/Kokyu_defs.i - * Kokyu/Dispatcher_Impl.h - * Kokyu/Dispatcher_Impl.i - * Kokyu/Default_Dispatcher_Impl.h - * Kokyu/Default_Dispatcher_Impl.i - * Kokyu/Dispatch_Deferrer.h - * Kokyu/Dispatch_Deferrer.i - * Kokyu/DSRT_Dispatcher_Impl_T.h - * Kokyu/DSRT_Dispatcher_Impl_T.i - * Kokyu/DSRT_Dispatch_Item_T.h - * Kokyu/DSRT_Dispatch_Item_T.i - * Kokyu/Kokyu_dsrt.h - * Kokyu/Kokyu_defs.cpp - * Kokyu/Kokyu_defs.inl - * Kokyu/Kokyu_dsrt.i - * Kokyu/Kokyu.h - * Kokyu/Dispatcher_Impl.cpp - * Kokyu/Dispatcher_Impl.inl - * Kokyu/Kokyu.i - * Kokyu/Default_Dispatcher_Impl.cpp - * Kokyu/Default_Dispatcher_Impl.inl - * Kokyu/Dispatch_Deferrer.cpp - * Kokyu/Dispatch_Deferrer.inl - * Kokyu/DSRT_Dispatcher_Impl_T.cpp - * Kokyu/DSRT_Dispatcher_Impl_T.inl - * Kokyu/DSRT_Dispatch_Item_T.cpp - * Kokyu/DSRT_Dispatch_Item_T.inl - * apps/JAWS/clients/Caching/Local_Locator.inl - * apps/JAWS/clients/Caching/Local_Locator.cpp - * apps/JAWS/clients/Caching/URL_Properties.inl - * apps/JAWS/clients/Caching/URL_Properties.cpp - * apps/JAWS/clients/Caching/Local_Locator.h - * apps/JAWS/clients/Caching/Local_Locator.i - * apps/JAWS/clients/Caching/URL_Properties.h - * apps/JAWS/clients/Caching/URL_Properties.i - * apps/JAWS/clients/Caching/Locator_Request_Reply.cpp - * apps/JAWS/clients/Caching/Locator_Request_Reply.inl - * apps/JAWS/clients/Caching/Locator_Request_Reply.h - * apps/JAWS/clients/Caching/Locator_Request_Reply.i - * apps/JAWS2/HTTPU/http_response.h - * apps/JAWS2/HTTPU/http_response.i - * apps/JAWS2/HTTPU/http_base.inl - * apps/JAWS2/HTTPU/http_base.cpp - * apps/JAWS2/HTTPU/http_headers.h - * apps/JAWS2/HTTPU/http_headers.i - * apps/JAWS2/HTTPU/http_status.inl - * apps/JAWS2/HTTPU/http_status.cpp - * apps/JAWS2/HTTPU/parse_http_request.h - * apps/JAWS2/HTTPU/http_request.inl - * apps/JAWS2/HTTPU/parse_http_request.i - * apps/JAWS2/HTTPU/http_request.cpp - * apps/JAWS2/HTTPU/parse_http_response.h - * apps/JAWS2/HTTPU/http_response.cpp - * apps/JAWS2/HTTPU/parse_http_response.i - * apps/JAWS2/HTTPU/http_response.inl - * apps/JAWS2/HTTPU/http_headers.inl - * apps/JAWS2/HTTPU/http_headers.cpp - * apps/JAWS2/HTTPU/http_base.h - * apps/JAWS2/HTTPU/http_base.i - * apps/JAWS2/HTTPU/parse_http_request.inl - * apps/JAWS2/HTTPU/parse_http_request.cpp - * apps/JAWS2/HTTPU/http_status.h - * apps/JAWS2/HTTPU/parse_http_response.cpp - * apps/JAWS2/HTTPU/parse_http_response.inl - * apps/JAWS2/HTTPU/http_status.i - * apps/JAWS2/HTTPU/http_request.h - * apps/JAWS2/HTTPU/http_request.i - * ACEXML/apps/svcconf/Svcconf_Handler.cpp - * ACEXML/apps/svcconf/Svcconf_Handler.inl - * ACEXML/apps/svcconf/Svcconf_Handler.h - * ACEXML/apps/svcconf/Svcconf_Handler.i - * ACEXML/common/AttributesImpl.cpp - * ACEXML/common/LocatorImpl.cpp - * ACEXML/common/Env.cpp - * ACEXML/common/Exception.cpp - * ACEXML/common/SAXExceptions.h - * ACEXML/common/SAXExceptions.i - * ACEXML/common/AttributesImpl.h - * ACEXML/common/AttributesImpl.i - * ACEXML/common/XMLFilterImpl.cpp - * ACEXML/common/Env.h - * ACEXML/common/SAXExceptions.cpp - * ACEXML/common/Env.i - * ACEXML/common/Exception.h - * ACEXML/common/Exception.i - * ACEXML/common/AttributesImpl.inl - * ACEXML/common/LocatorImpl.inl - * ACEXML/common/Env.inl - * ACEXML/common/XMLFilterImpl.h - * ACEXML/common/XMLFilterImpl.i - * ACEXML/common/Exception.inl - * ACEXML/common/XMLFilterImpl.inl - * ACEXML/common/LocatorImpl.h - * ACEXML/common/LocatorImpl.i - * ACEXML/common/SAXExceptions.inl - * ACEXML/parser/debug_validator/Element_Tree.i - * ACEXML/parser/debug_validator/Element_Tree.cpp - * ACEXML/parser/debug_validator/Element_Tree.inl - * ACEXML/parser/debug_validator/Element_Tree.h - * ACEXML/parser/parser/Parser.h - * ACEXML/parser/parser/Parser.i - * ACEXML/parser/parser/Entity_Manager.cpp - * ACEXML/parser/parser/Entity_Manager.inl - * ACEXML/parser/parser/Entity_Manager.h - * ACEXML/parser/parser/Entity_Manager.i - * ACEXML/parser/parser/Parser.cpp - * ACEXML/parser/parser/Parser.inl - * ACEXML/examples/SAXPrint/SAXPrint_Handler.cpp - * ACEXML/examples/SAXPrint/SAXPrint_Handler.inl - * ACEXML/examples/SAXPrint/SAXPrint_Handler.h - * ACEXML/examples/SAXPrint/SAXPrint_Handler.i - * websvcs/lib/URL_Addr.h - * websvcs/lib/URL_Addr.i - * websvcs/lib/URL_Addr.cpp - * websvcs/lib/URL_Addr.inl - * examples/ASX/Event_Server/Event_Server/Options.h - * examples/ASX/Event_Server/Event_Server/Options.i - * examples/ASX/Event_Server/Event_Server/Options.inl - * examples/ASX/UPIPE_Event_Server/Options.h - * examples/ASX/UPIPE_Event_Server/Options.i - * examples/ASX/UPIPE_Event_Server/Options.inl - * examples/Service_Configurator/IPC-tests/server/Handle_Timeout.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_Timeout.inl - * examples/Service_Configurator/IPC-tests/server/Handle_L_SPIPE.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_L_SPIPE.inl - * examples/Service_Configurator/IPC-tests/server/Handle_L_CODgram.h - * examples/Service_Configurator/IPC-tests/server/Handle_L_CODgram.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_Broadcast.h - * examples/Service_Configurator/IPC-tests/server/Handle_Broadcast.i - * examples/Service_Configurator/IPC-tests/server/Handle_L_Pipe.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_L_Pipe.inl - * examples/Service_Configurator/IPC-tests/server/Handle_L_Dgram.h - * examples/Service_Configurator/IPC-tests/server/Handle_L_Dgram.i - * examples/Service_Configurator/IPC-tests/server/Handle_L_Stream.h - * examples/Service_Configurator/IPC-tests/server/Handle_L_Stream.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_R_Stream.h - * examples/Service_Configurator/IPC-tests/server/Handle_R_Stream.i - * examples/Service_Configurator/IPC-tests/server/Handle_L_CODgram.inl - * examples/Service_Configurator/IPC-tests/server/Handle_L_CODgram.cpp - * 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_L_FIFO.inl - * examples/Service_Configurator/IPC-tests/server/Handle_L_FIFO.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_L_SPIPE.h - * examples/Service_Configurator/IPC-tests/server/Handle_L_SPIPE.i - * examples/Service_Configurator/IPC-tests/server/Handle_Broadcast.inl - * examples/Service_Configurator/IPC-tests/server/Handle_Broadcast.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_L_Dgram.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_L_Dgram.inl - * examples/Service_Configurator/IPC-tests/server/Handle_L_Stream.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_L_Stream.inl - * examples/Service_Configurator/IPC-tests/server/Handle_R_Dgram.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_R_Dgram.inl - * examples/Service_Configurator/IPC-tests/server/Handle_R_Stream.cpp - * examples/Service_Configurator/IPC-tests/server/Handle_R_Stream.inl - * examples/Service_Configurator/IPC-tests/server/Handle_L_Pipe.h - * examples/Service_Configurator/IPC-tests/server/Handle_L_Pipe.i - - Moved the .i file to .inl (with history), changed the includes in the - .cpp and .h files to match, removed the .i file. - -Thu Jul 27 18:43:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: - Some more updates to refer to the subversion archive - -Thu Jul 27 13:57:31 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * docs/svn/config - - Fixed syntax errors. thanks to Adam Mitz for reporting them. - -Thu Jul 27 10:14:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - Committing some small changes I had pending to see how svn behaves now - with updates and status info on the build systems. - - * ace/Recursive_Thread_Mutex.{h,cpp}: - Changed removed_ to a bool - - * ace/Thread_Mutex.{h,cpp,inl}: - Changed removed_ to a bool and removed ACE_USES_OBSOLETE_GUARD_CLASSES - support. - -Thu Jul 27 07:37:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * etc/tao.doxygen: - * etc/tao_portableserver.doxygen: - Set ignore prefix - - * Kokyu/Dispatcher_Task.i - * Kokyu/Dispatcher_Task.inl - Renamed .i to .inl so that we can have a look how subversion stores - this in the history - - * Kokyu/Dispatcher_Task.cpp - * Kokyu/Dispatcher_Task.h - * Kokyu/Makefile.am - Updated for the change above. - -Wed Jul 26 07:25:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/sets-manager.py: - Fixed typo - -Wed Jul 26 07:25:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: - * html/Stats/index.shtml: - * docs/ACE-bug-process.html: - Updated documentation to point to the new svn repository - -Tue Jul 25 21:51:07 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/PythonACE/fuzz/max_project_len.py - * bin/PythonACE/fuzz/inline.py - - Fixed errors in these checks. - -Tue Jul 25 20:52:51 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/PythonACE/fuzz/cpp_inline.py - - Fixed this check, it was incorrectly matching ACE_INLINES inside - defines and comments. - -Tue Jul 25 20:16:26 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * docs/svn/svn-prefs.reg - * docs/svn/config - - Updated to add global-ignores to mimic .cvsignore functionality. - -Tue Jul 25 19:40:54 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/PythonACE/__init__.py - - This enables PythonACE to be loaded as a module. - - * bin/PythonACE/fuzz/check_includes.py - * bin/PythonACE/fuzz/no_conflict_markers.py - - Fixed these checks, they were incorrectly flagging errors. - - * bin/fuzz.py - - Command line fuzz client that uses the PythonACE fuzz module. - - * bin/sets-manager.py - - Corrected this file to point at DOC/Middleware instead of ACE/Middleware. - -Mon Jul 24 23:58:33 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * docs/svn/svn-prefs.reg - * docs/svn/config - - Updated these files to add svn:keywords to most text types. - -Mon Jul 24 23:53:53 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/create_ace_build.pl - - Fix provided by Steve Huston to prevent this script from building - links from .svn directories. - -Mon Jul 24 19:17:37 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/fuzz.pl - - Disabled the ID string check until the script can be fixed to ignore - .svn directories. - -Mon Jul 24 15:45:34 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * Subversion conversion completed at revision 73730 - -Sun Jul 23 15:44:43 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * Repository frozen for repository conversion - -Sun Jul 23 11:17:31 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/make_release: - Don't exclude the other mwc files for vc71/vc8, this will create then - some more workspace files which just contain a subset of the projects - -Sun Jul 23 09:25:12 UTC 2006 Olli Savia <ops@iki.fi> - - * NEWS: - Added a note about TTY_IO API change. - -Sat Jul 22 19:12:10 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * docs/svn/config - * docs/svn/svn-prefs.reg - - These are client side configuration files for Subversion. Currently, - the automate setting svn:eol-style and svn:executable properties on - newly added files. - -Sat Jul 22 18:28:33 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/sets-manager.py - - First cut of a branching script for subverion. Pass --help - to get documentation. - -Fri Jul 21 13:28:31 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - Updated because of -oS addition to tao_idl - -Fri Jul 21 13:17:31 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_unistd.cpp: - Const improvements and initialise pointer with 0 - -Thu Jul 20 21:44:31 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * tests/Service_Config_Test.cpp: - - In testLimits (), surrounded string literals with ACE_TEXT to - fix a build error on wchar builds. - -Thu Jul 20 16:42:04 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Basic_Types.h: - - Removed ACE_NTOHLL function. It was experimental, untested and - not ready for the ACE sources just yet. - - Removed "ace/os_include/netinet/os_in.h" include directive. Now - that ACE_NTOHLL has been removed it is no longer necessary. - - * ace/Numeric_Limits.h: - - Commented out "{unsigned} long long" specializations in the - ACE_LACKS_NUMERIC_LIMITS case. They're currently not used, and - we need - -Thu Jul 20 15:19:30 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Basic_Types.h: Added - - # include "ace/os_include/netinet/os_in.h" // Get ntohl() - - and added the inline keyword to ACE_NTOHLL() to fix the problems - introduced with - - Wed Jul 19 22:46:51 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - -Thu Jul 20 15:01:32 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Configuration.cpp: Made - ACE_Configuration_ExtId::operator==() and - ACE_Configuration_ExtId::operator!=() case insensitive. Thanks - to Domingos Monteiro <d dot monteiro at netia dot net> for this - contribution. - -Thu Jul 20 09:57:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/tao_no_iiop.mpb: - New feature to disable iiop in TAO - - * bin/MakeProjectCreator/config/global.features: - Set tao_no_iiop to 0 by default - -Thu Jul 20 09:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_Thread.h: - Doxygen improvements - -Thu Jul 20 08:55:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Mutex.{h,cpp,inl}: - Made the removed_ member a bool instead of an int - -Thu Jul 20 08:37:17 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/TTY_IO.cpp: - Fixed compile error on VxWorks. Thanks to Johnny for pointing - out the problem. - -Wed Jul 19 23:39:05 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/PythonACE/fuzz/__init__.py - * bin/PythonACE/fuzz/_generic_handler.py - * bin/PythonACE/fuzz/_mailer.py - * bin/PythonACE/fuzz/_singleton.py - * bin/PythonACE/fuzz/_warning_handler.py - * bin/PythonACE/fuzz/streams_include.py - - Changes necessary to support mailing (non-fatal) Fuzz errors to - the user that attempts a commit. streams_include is an example - of a fuzz check that sends warnings instead of stopping commits. - -Wed Jul 19 22:46:51 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Basic_Types.h (ACE_INT64, ACE_UINT64): - - Silence g++ "-pedantic" warnings regarding use of "long long" - type by preceding the typedef with the g++ __extension__ - keyword. - -Wed Jul 19 21:15:51 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * NEWS: - - Updated the file with a note announcing the interface change - from "Wed Jul 19 20:11:52 UTC 2006". - -Wed Jul 19 20:11:52 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/ACE.h: - * ace/ACE.cpp: - - Changing ACE::debug () to use boolean instead of char. - - * ace/Acceptor.cpp: - * ace/DLL_Manager.cpp: - * ace/Dynamic_Service_Dependency.cpp: - * ace/SOCK_Dgram_Bcast.cpp: - * ace/Service_Gestalt.cpp: - * ace/Service_Object.cpp: - * ace/Service_Repository.cpp: - * ace/Service_Types.cpp: - - Updating places where ACE::debug() return value is interpreted - as a integral type. - - * tests/Service_Config_Test.cpp: - - Fixing the test to correctly account for XML-based service - configurations. - -Wed Jul 19 15:53:09 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/PythonACE/fuzz/__init__.py - * bin/PythonACE/fuzz/_fuzz.py - * bin/PythonACE/fuzz/_generic_handler.py - * bin/PythonACE/fuzz/_path.py - * bin/PythonACE/fuzz/_types.py - * bin/PythonACE/fuzz/check_includes.py - * bin/PythonACE/fuzz/cpp_inline.py - * bin/PythonACE/fuzz/inline.py - * bin/PythonACE/fuzz/math_include.py - * bin/PythonACE/fuzz/max_filename.py - * bin/PythonACE/fuzz/max_project_len.py - * bin/PythonACE/fuzz/newline.py - * bin/PythonACE/fuzz/no_conflict_markers.py - * bin/PythonACE/fuzz/noncvs.py - * bin/PythonACE/fuzz/ptr_arith_t.py - * bin/PythonACE/fuzz/refcountservantbase.py - * bin/PythonACE/fuzz/verify_changelog.py - - This is a python module designed to eventually replace the - bin/fuzz.pl script. Each .py file in the fuzz directory that - does not have a _ at the beginning of its filename contains a - single check. New fuzz checks can be added to the system simply - by creating a new file in the fuzz directory. Care should be - taken to select a filename that does not conflict with an - existing python module. - - ptr_arith_t.py and inline.py are good examples that use the - generic handler to implement their checks, max_project_len is a - good example of a check that does not use the generic handler. - -Wed Jul 19 15:07:29 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/DLL_Manager.cpp: - - Fixing a build problem on HP and Windows, as a consequence of my - earlier change: "Tue Jul 18 19:17:54 UTC 2006". Moral: use - ACE_TEXT around literal expressions only. - -Wed Jul 19 13:55:39 UTC 2006 Olli Savia <ops@iki.fi> - - * examples/IPC_SAP/DEV_SAP/reader/reader.cpp: - * examples/IPC_SAP/DEV_SAP/writer/writer.cpp: - Updated to reflect changes below. - -Wed Jul 19 13:39:13 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/TTY_IO.h: - Changed type of rtsenb from 'int' to 'unsigned char'. - Updated documentation to note that 'parityenb' member is - now deprecated. - - * ace/TTY_IO.cpp: - Changed constructor of Serial_Params class to assign sensible - values to class members. - -Wed Jul 19 10:47:31 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Proactor_Timer_Test.cpp: Corrected compile warning re static. - -Wed Jul 19 08:00:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/ciao_config_handlers_base.mpb: - Use exceptions as base project - -Wed Jul 19 7:14:10 UTC 2006 Ossama Othman <ossama_othman@symantec.com> - - * ace/String_Base.h (size_type): - - Explicitly bring the ACE_String_Base_Const::size_type typedef - into the ACE_String_Base<> scope through a "using" declaration. - IBM Visual Age 6 and Sun Studio 10 have difficulty resolving the - name otherwise. - - * ace/SString.cpp: - - Replaced std::numeric_limits<>::max() call that was - inadvertently introduced with its ACE_Numeric_Limits<> - counterpart. The former is not available on VxWorks. - -Tue Jul 18 22:52:38 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Message_Queue_Test_Ex.cpp: Rearranged the pieces left out of - single-threaded builds; should correct single-threaded builds. - -Tue Jul 18 20:50:49 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * Happy Birthday to me!! - -Tue Jul 18 19:17:54 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - This change is in response to bug#2602, which was reported by - Lothar Werzinger <lothar at tradescape dot biz>. It fixes a - problem with trying to use XML as the service configuration file - format. The change also includes additional comments and - in-line documentation on the new ability to use local service - repositories. - - * ace/DLL_Manager.cpp: - - Updated the debugging output of DLL_Handle::open() to be able to - see _why_ a DLL open failed. For example, the errno message is - 'not found' in a case where one tries to open library A, which - in turn depends on library B, but B is missing. This change - allows one to see the underlying cause for the error - (via DLL_Handle::error(), if ACE_DEBUG=2, or greater is - present). - - * ace/Parse_Node.h: - * ace/Parse_Node.cpp: - - Moving the ACE_Service_Type_Factory here, from - Service_Config.{h,cpp} This class is only needed when - ACE_USES_CLASSIC_SVC_CONF is 1, i.e. when ACE is using the - non-XML configuration file format. Therefore, moving it to - Parse_Node.{h,cpp} causes it to compile only if needed. - - * ace/Service_Config.h: - - Removed the static get_xml_svc_conf() method, as it already has - the required implementation in the base class. - - * ace/Service_Gestalt.h: - * ace/Service_Gestalt.cpp: - - Moved the ACE_Service_Type_Factory class to Parse_Node.{h,cpp}. - Updated the documentation with a more clear description of the - class responsibilities. Corrected minor omissions. - -Tue Jul 18 18:12:17 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ACE-INSTALL.html: Updated the description of how to build - ACE+TAO using MPC. Thanks to Steve Totten for helping with - this. - -Tue Jul 18 19:04:42 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Numeric_Limits.h: - - Corrected short integer limit constants, i.e. "s/SHORT/SHRT/g". - - Try other known "long long" integer constants if - "{U}LLONG_{MIN,MAX}" are not available. Issue a preprocessor - error if all known cases have been exhausted. - -Tue Jul 18 16:55:31 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Makefile.am (libACE_la_SOURCES, nobase_include_HEADERS): - - Added new Sig_{Adapter,Handler}.* source and header files. - -Mon Jul 17 23:18:52 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/README: Documented the ACE_TIMEPROBE_ASSERTS_FIXED_SIZE macro. - -Mon Jul 17 23:16:55 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Timeprobe_T.cpp (timeprobe): Added a check to guard against - segfaults by wrapping the index around when it reaches the end - of the buffer. Thanks to Patrick Rabau <pr2345 at gmail dot - com> for this fix. - -Tue Jul 18 13:21:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added Bug_2595_Regression - -Tue Jul 18 12:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - Fixed the rules to decide whether a .inl file should be added - or not. Thanks to Chad Elliot for noticing this incorrect - rule. - -Tue Jul 18 09:59:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/PerlACE/ProcessVX_Win32.pm: - Increased delay factor from 2 to 3, downloading executables - seems to take longer now - -Tue Jul 18 09:54:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/PerlACE/Process_Unix.pm: - Removed chorus support - -Mon Jul 17 22:52:19 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Proactor.cpp (schedule_timer): Record the ACE_Proactor instance - that the timer is scheduled on in the ACE_Handler; makes this - behavior analogous to that in ACE_Reactor::schedule_timer(). - - * tests/Proactor_Timer_Test.cpp: Add a test for repeating timers and - canceling a repeating timer when it's likely that an expiration for - the timer is already queued when it's canceled. - -Mon Jul 17 16:06:57 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Unload_libACE.cpp: Alter the libACE suffix for HP-UX on - PA-RISC only. Also see: - Fri Jul 14 19:44:03 UTC 2006 john_c <johnc@ociweb.com> - -Mon Jul 17 22:11:42 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * examples/Threads/process_semaphore.cpp: - - Added #include of ace/OS_NS_stdlib.h to get ACE_OS::atoi(). - -Mon Jul 19 11:21:56 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/String_Base.cpp: - * ace/String_Base.inl: - - Explicitly qualify size_type parameters with "typename" to force - some compilers (Sun Studio 10) to recognize it as a type. - -Mon Jul 17 17:49:41 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * include/makeinclude/platform_hpux_aCC.GNU (SOFLAGS): - - Corrected shared library path for Itanium builds. - - * protocols/ace/HTBP/HTBP_ID_Requestor.cpp (get_HTID): - - Use new ACE_CString::size_type type to hold string - position/index value, instead of "int". Addresses "comparison - between signed and unsigned" warnings. - -Mon Jul 17 16:04:07 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * NEWS: - - Added information about the use of atomic operations on SPARC - Solaris. - -Mon Jul 17 15:09:21 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Refcounted_Auto_Ptr.{h inl} (null): Changed the return type - of the null() method from int to bool in order to more accurately - reflect the true/false nature of the returned value. - - * NEWS: Noted the above change. - -Mon Jul 17 15:05:05 UTC 2006 Steve Huston <shuston@riverace.com> - - * examples/System_V_IPC/SV_Message_Queues/MQ_Server.cpp: Added - include "ace/Log_Msg.h" in the "non-supported platform" case. - Fixes compile error on Windows. - -Mon Jul 17 13:03:09 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * ace/ace_for_tao.mpc: - - Sig_Handler.cpp and Sig_Adapter.cpp are needed to link gperf. - -Mon Jul 17 12:49:34 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * netsvcs/servers/main.cpp: - - Added a missing #include for ace/Sig_Adapter.h. - -Mon Jul 17 11:57:26 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * ace/qt_reactor.mpb: - - Removed this file. It was no longer referenced by ace and it's - functionality was replaced by ace_qtreactor.mpc. - -Mon Jul 17 10:41:14 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Message_Queue_Test_Ex.cpp: Removed the VxWorks-specific - section from MQ_Ex_N_Tester::receiver() - it was apparantly - pulled in from another test example and didn't need to be here. - Should resolve compile errors on VxWorks. - -Mon Jul 18 09:32:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-vxworks6.3.h: - * include/makeinclude/platform_vxworks6.3.GNU: - New files for VxWorks 6.3 - - * ace/os_include/sys/os_types.h: - * ace/os_include/os_stdio.h: - * ace/OS_NS_unistd.inl: - * ace/OS_NS_arpa_inet.cpp: - Updated for VxWorks 6.3 - -Sun Jul 17 06:49:21 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * apps/soreduce/Library.cpp (set_path): - * apps/soreduce/Obj_Module.cpp (add_source): - * examples/Misc/test_sstring.cpp: - * examples/Web_Crawler/Command_Processor.cpp: - * examples/Web_Crawler/URL_Visitor.cpp: - - Use new ACE_CString::size_type type to hold string - position/index value, instead of "int". Addresses "comparison - between signed and unsigned" warnings. - - * examples/APG/Signals/SigAction.cpp: - * examples/System_V_IPC/SV_Message_Queues/MQ_Server.cpp: - * examples/System_V_IPC/SV_Message_Queues/TMQ_Server.cpp: - - Include "ace/OS_NS_stdlib.h" to pull in ACE_OS::exit() - prototype. - - * examples/APG/Signals/SigGuard.cpp: - * examples/APG/Signals/SigHandler.cpp: - * examples/APG/Signals/SigHandlers.cpp: - * examples/APG/ThreadManagement/Signals.cpp: - * examples/APG/ThreadManagement/Signals2.cpp: - * examples/ASX/CCM_App/SC_Server.cpp: - * examples/ASX/Event_Server/Event_Server/event_server.cpp: - * examples/ASX/UPIPE_Event_Server/event_server.cpp: - * examples/Connection/blocking/SPIPE-connector.h: - * examples/Connection/misc/Connection_Handler.cpp: - * examples/Connection/non_blocking/CPP-acceptor.h: - * examples/Connection/non_blocking/CPP-connector.h: - * examples/Service_Configurator/IPC-tests/server/server_test.cpp: - * netsvcs/lib/TS_Clerk_Handler.cpp: - - Include "ace/Sig_Handler.h" or "ace/Sig_Adapter.h" to pull in - ACE_Sig_Handler or ACE_Sig_Adapter class declaration, - respectively. Necessitated by "ace/Signal.h" refactoring. - -Sun Jul 16 23:14:53 UTC 2006 Ciju John <johnc@ociweb.com> - - * ace/String_Base_Const.cpp: - - Took out an extra sem-colon. It seems to make the - FC4_Versioned_Namespace build grumpy. - -Sat Jul 15 16:14:05 UTC 2006 Ciju John <johnc@ociweb.com> - - * NEWS: - - Added entry on HPUX library extension changes. - -Fri Jul 14 22:10:02 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Sig_Handler.cpp: - - Removed explicit template instantiations. They snuck back in - when Signal.* was refactored. Thanks to Johnny for pointing - this out. - -Fri Jul 14 21:32:02 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Numeric_Limits.h: - - "ace/ACE_Export.h" -> "ace/ACE_export.h". Darn Windows case - insensitivity. - - Removed unnecessary "ACE_Export" from ACE_Numeric_Limits primary - template. - - * ace/Registry.cpp (make_name): - - Use ACE_TString::size_type, not ssize_t, to hold string position - values. Addresses signed/unsigned comparison warnings. - - * ace/Shared_Memory_Pool.h: - - Include new "ace/Sig_Handler.h" header to pull in - ACE_Sig_Handler class declaration. - - * ace/String_Base.h: - - Added tips to documentation related to potentially user visible - changes. - -Fri Jul 14 21:10:11 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/Atomic_Op.h (ACE_Atomic_Op): - * ace/Atomic_Op.inl (ACE_Atomic_Op): - - Reverted change that made the constructor parameters for the - "long" type specialization be passed in by reference to const - instead of by value. Explicit template instantiation is no - longer performed by ACE so there is no need for the - specialization constructor parameters to match the primary - template constructor parameters. - - * ace/CDR_Base.cpp: - * ace/CDR_Stream.h: - * ace/CDR_Stream.inl: - - Const-correctness improvements. - - * ace/CDR_Base.h: - - Minor const related cosmetic changes. - - * ace/CDR_Base.inl (first_size, next_size): - - Quickly multiply by two using a bit shift. This is guaranteed - to work since the variable is an unsigned integer. - - * ace/CORBA_macros.h: - - Removed unnecessary instantiation of emulated exception - environment variable from native exception case. Reduces - footprint in code that uses the ACE emulated exception macros in - a native C++ exception configuration, and provides a slight - performance increase due to resulting tighter binaries. - - * ace/DLL_Manager.cpp (get_dll_names): - - ACE_TString::rfind() now return an ACE_TString::size_type type. - Use ACE_TString::size_type, not ssize_t, to hold the return - value of ACE_TString::rfind(). - - * ace/Dev_Poll_Reactor.cpp: - - Include new "ace/Sig_Handler.h" header to pull in - ACE_Sig_Handler class declaration. - - Added support for HP-UX devpoll headers. - - * ace/Event_Handler.cpp: - - No need to include "ace/Message_Block.h". - - * ace/MMAP_Memory_Pool.h: - * ace/Select_Reactor_T.cpp: - * ace/TP_Reactor.cpp: - * ace/WFMO_Reactor.inl: - - Include new "ace/Sig_Handler.h" header to pull in - ACE_Sig_Handler class declaration. - - * ace/Malloc_Base.h (ACE_Allocator): - - Defined a new "size_type" typedef (currently a "size_t") to - simplify potential allocator size type changes in the future. - It's introduction was motivated by the ACE string class size - type fixes described below. Analogous to the STL - std::allocator<>::size_type typedef. - - * ace/Numeric_Limits.h: - - Traits containing basic integer limits. Useful for - template-based code on platforms that lack - std::numeric_limits<>. These traits are not meant to be a - replacement for std::numeric_limits<>. Rather they are a crutch - until all ACE-supported platforms support - std::numeric_limits<>. [Bug 2451] - - * ace/OS_NS_dirent.inl: - - "__USE_POSIX" -> "_POSIX_SOURCE". The former is a C library - implementation-specific feature test macro/symbol. - - * ace/OS_NS_stdio.cpp: - * ace/OS_NS_stdio.h: - * ace/OS_NS_stdio.inl (get_win32_versioninfo): - - Fixed/added missing wide character support. - - * ace/OS_NS_stdlib.cpp (mkstemp_emulation): - - Clean up preprocessor conditional blocks by taking advantage of - new ACE_Numeric_Limits<> trait templates. - - * ace/OS_NS_stdlib.inl (atop): - - Address VC++ 64-bit warnings by taking advantage of standard - intptr_t type, if available. - - * ace/OS_NS_sys_socket.inl (closesocket): - - Do not shutdown the write end here. Doing so will break - applications that duplicate a handle on fork(), for example, and - expect to continue writing in the fork()ed process. - - * ace/OS_NS_sys_utsname.cpp (uname): - - Fixed garbage in wide string output. - - * ace/Object_Manager.cpp: - - No need to include "ace/Containers.h". - - Include newly refactored "ace/Sig_Adapter.h" instead of old - "ace/Signal.h". Removes dependency on the ACE_Reactor. Allows - service configurator-disabled subset containing - ACE_Object_Manager support to be built, e,g, MPC ace_svcconf=0 - configuration. - - * ace/Object_Manager_Base.cpp (init): - * ace/Process.cpp (pass_handle): - - Fixed Windows version info handling in wide string - configurations. - - * ace/Process_Manager.h (wait): - - Clarified that the timeout value is a relative time, not - absolute, in the documentation. - - * ace/SOCK_Acceptor.cpp (shared_open): - - Removed assertion that assumes all protocol families are - internet related. This method is invoked by some non-internet - protocol subclasses, such as ACE_LSOCK_Acceptor. - - * ace/SOCK_Stream.cpp (close): - - Reinstituted explicit close_writer() method call for the - ACE_WIN32 case since ACE_OS::close() no longer causes the writer - to be shutdown on Windows (nor should it!). - - * ace/Service_Config.cpp: - - Replaced old "ace/Signal.h" include directive with - "ace/Sig_Adapter.h". ACE_Sig_Handler support is unnecessary. - - Removed redundant "ace/Service_Config.h" include directive. - - * ace/Sig_Adapter.cpp: - * ace/Sig_Adapter.h: - * ace/Sig_Handler.cpp: - * ace/Sig_Handler.h: - * ace/Sig_Handler.inl: - * ace/Signal.cpp: - * ace/Signal.h: - * ace/Signal.inl: - - Refactored ACE_Sig_Adapter and ACE_Sig_Handler classes from - "ace/Signal.*" into their own dedicated source files. Allows - ACE_Sig_{Guard,Set} to be used without introducing a dependency - on the ACE_Reactor framework. - - * ace/String_Base_Const.cpp (npos): - * ace/String_Base_Const.h (npos): - - Introduced a new "size_type" typedef. This is consistent with - the standard C++ string class typdef/trait. - - The "no position" constant is now an unsigned integer of the - same "size" type used by the underlying ACE string allocator. - It is defined to be the maximum value the unsigned type can - handle (e.g. ACE_Numeric_Limits<size_type>::max()). This works - since a string length will always be less than or equal to - "max() - 1" since the final character in the string is the null - terminator. [Bug 2598] - - * ace/String_Base.cpp: - * ace/String_Base.h: - * ace/String_Base.inl: - * ace/SString.cpp: - * ace/SString.h: - * ace/SString.inl: - - All size parameter and return value types are now consistent - with the size type of the underlying allocator - (i.e. ACE_Allocator::size_type). Addresses inability of ACE - string classes to provide valid search index for string lengths - greater than the maximum value of ssize_t, a signed type. In - particular, the find(), rfind() and strstr() methods now return - an unsigned integer (size_t) instead of a signed one - (ssize_t). Affected classes include: - - ACE_CString - ACE_WString - ACE_TString - ACE_NS_WString - - Unless you have been explicitly using -1 instead of npos when - comparing the return value of find(), rfind() and strstr(), - and/or assigning the return value to ssize_t you should not see - any difference. A new size_type typedef has been added to the ACE - string class to aid developers. - - The ACE_String_Base<>::strstr() documentation as well as a - default rfind() argument erroneously referred to -1 instead of - npos. Those instances have been corrected. - - To summarize, a "no position" condition is denoted using the npos - constant, not -1. It can be referred directly by scoping it with - the appropriate string class (e.g. ACE_CString::npos, - ACE_WString::npos, etc). [Bug 2598] - - * ace/Task_T.h: - - Include "ace/IO_Cntl_Msg.h" to pull in ACE_IO_Cntl_Msg class - declaration. - - * ace/Time_Value.cpp (max_time): - - Cleaned up preprocessor conditional blocks by taking advantage - of ACE_Numeric_Limits trait templates. - - * ace/Timer_Queue_Adapters.cpp: - - Include "ace/Signal.h" to pull in ACE_Sig_Guard class - declaration. - - * ace/Timer_Queue_Adapters.h: - - Forward declare ACE_Sig_Set class. No need for the full blown - declaration. - - * ace/Token.cpp: - - Minor cosmetic change. - - * ace/ace_wchar.h (ACE_GetVersionEx): - - New wide string support for Windows "GetVersionEx{A,W}()" - functions. - - * ace/ace_wchar.inl: - - Improved const-correctness. - - * ace/config-aix-4.x.h: - * ace/config-macosx-panther.h: - * ace/config-macosx-tiger.h: - * ace/config-macosx.h: - - Fixed single-threaded configuration so that all multi-thread - support is disabled. - - * ace/ace.mpc: - - Added missing DLL_Manager.cpp source file, and newly refactored - Sig_{Adapter,Handler}.cpp source files. - - * ace/ace_for_tao.mpc: - - Explicitly require the ace_svcconf MPC feature. - - Added missing ACE_LACKS_ACE_TOKEN macro. The ace_for_tao subset - does not support ACE's token collections. - - * ace/svcconf.mpb: - - Do not define ACE_HAS_ACE_SVCCONF. There is no such macro in - the ACE sources. - - Explicitly remove ACE_LACKS_ACE_SVCCONF from the macro list. - - Moved "DLL_Manager.cpp" to the core ACE library source file - list. "DLL_Manager.cpp" is not an optional component, and must - always exist in ACE regardless of whether or not ACE's service - configurator support is enabled. - - * ace/token.mpb: - - Do not define ACE_HAS_ACE_TOKEN. There is no such macro in - the ACE sources. - - Explicitly remove ACE_LACKS_ACE_TOKEN from the macro list. - - * bin/MakeProjectCreator/config/ace_for_tao.mpb: - - Added missing ACE_LACKS_ACE_TOKEN macro. The ace_for_tao subset - does not support ACE's token collections. - - * tests/SString_Test.cpp: - - Corrected checks for "no position" return values to compare - against "npos" constant, not "-1". - -Fri Jul 14 20:57:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/INET_Addr.cpp: - Initialise some pointers explicitly with 0 - -Fri Jul 14 20:41:23 UTC 2006 john_c <johnc@ociweb.com> - - * bin/tao_orb_tests.lst: - - Add new test TAO/tests/Optimized_Connection to nightly builds. - -Fri Jul 14 20:39:41 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-vxworks6.2.h: Reverted this change: - Fri Jul 14 10:34:43 UTC 2006 Steve Huston <shuston@riverace.com> - It appears to not work, and I can't test anything else here. - -Fri Jul 14 18:53:32 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/MakeProjectCreator/config/global.features - * bin/MakeProjectCreator/config/stl.mpb - - Added an stl feature/base project for projects that use portions - of the STL that may not be universally supported. - -Fri Jul 14 19:44:03 UTC 2006 john_c <johnc@ociweb.com> - - * ace/config-hpux-11.00.h: - * include/makeinclude/platform_hpux_aCC.GNU: - - Changing the shared library extension for hpux ia64 to ".so". On - HP-UX 11i Version 1.5 the naming scheme is lib*.sl for PA and - lib*.so on IPF. - -Fri Jul 14 19:29:59 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_NS_stdio.cpp (DllMain): Add extern "C" to the DllMain - signature - enables it to be located and called at startup. - Thanks to Hubert Talbot <Hubert dot Talbot at criq dot qc dot ca> - for this fix. - -Fri Jul 14 19:13:06 UTC 2006 john_c <john@ociweb.com> - - * ace/Timer_Queue_Adapters.cpp: - - Include file was missing. Fixes compile error when - ACE_HAS_DEFERRED_TIMER_COMMANDS is defined. - -Fri Jul 14 19:05:15 UTC 2006 Ciju John <john_c@ociweb.com> - - * bin/tao_other_tests.lst: - - Adding the TAO/orbsvcs/tests/Event/Mcast/RTEC_MCast_Federated - test to nightly builds. - -Fri Jul 14 16:25:00 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - Added exclusions for Bug_2593_Regression test. - -Fri Jul 14 15:50:34 UTC 2006 Adam Mitz <mitza@ociweb.com> - - * bin/PerlACE/Process_Unix.pm: - * bin/PerlACE/Process_Win32.pm: - - Added an optional parameter to Process->Kill(). If the parameter is - true then the return value from the process is ignored. - -Fri Jul 14 15:45:04 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Message_Queue_Test_Ex.{h cpp}: Corrected for !ACE_HAS_THREADS. - -Fri Jul 14 15:39:53 UTC 2006 Adam Mitz <mitza@ociweb.com> - - * bin/tao_other_tests.lst: - Added new test (TAO/orbsvcs/tests/CosEvent/Timeout/run_test.pl). - -Fri Jul 14 10:34:43 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-vxworks6.2.h: Added ACE_LACKS_USECONDS_T to fix a - compile error. It's in the non-__RTP__ section since that's where - ACE_LACKS_SUSECONDS_T was, but I'm not sure this is correct. - -Fri Jul 14 07:26:46 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/Thread_Mutex.h: - * ace/ace_wchar.h: - * tests/SOCK_Dgram_Bcast_Test.cpp: - Fixed typos in comments. - - * m4/config_h.m4: - Removed references to ACE_HAS_TERM_IOCTLS. It's no longer - used in ACE/TAO/CIAO code. - -Fri Jul 14 02:11:21 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - Added new test for Bug 2593 to nightly suite. - -Thu Jul 13 20:40:10 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/config-integritySCA.h: - Removed references to ACE_HAS_ATTR_SETSCHED. It's no longer - used in ACE/TAO/CIAO code. - - * ace/config-rtems.h: - Removed references to ACE_HAS_SIGNAL_SAFE_OS_CALLS. It's no longer - used in ACE/TAO/CIAO code. - - * ace/config-sco-5.0.0-mit-pthread.h: - Removed references to ACE_HAS_T_OPMGMT. It's no longer - used in ACE/TAO/CIAO code. - -Thu Jul 13 10:35:39 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/Base_Thread_Adapter.cpp: - * ace/Cleanup.cpp: - * ace/OS.cpp: - * ace/OS_Errno.cpp: - * ace/OS_Log_Msg_Attributes.cpp: - * ace/OS_NS_Thread.cpp: - * ace/OS_NS_arpa_inet.cpp: - * ace/OS_NS_ctype.cpp: - * ace/OS_NS_dirent.cpp: - * ace/OS_NS_dlfcn.cpp: - * ace/OS_NS_errno.cpp: - * ace/OS_NS_fcntl.cpp: - * ace/OS_NS_math.cpp: - * ace/OS_NS_netdb.cpp: - * ace/OS_NS_poll.cpp: - * ace/OS_NS_pwd.cpp: - * ace/OS_NS_regex.cpp: - * ace/OS_NS_signal.cpp: - * ace/OS_NS_stdio.cpp: - * ace/OS_NS_stdlib.cpp: - * ace/OS_NS_string.cpp: - * ace/OS_NS_strings.cpp: - * ace/OS_NS_stropts.cpp: - * ace/OS_NS_sys_mman.cpp: - * ace/OS_NS_sys_msg.cpp: - * ace/OS_NS_sys_resource.cpp: - * ace/OS_NS_sys_select.cpp: - * ace/OS_NS_sys_sendfile.cpp: - * ace/OS_NS_sys_shm.cpp: - * ace/OS_NS_sys_socket.cpp: - * ace/OS_NS_sys_stat.cpp: - * ace/OS_NS_sys_time.cpp: - * ace/OS_NS_sys_uio.cpp: - * ace/OS_NS_sys_wait.cpp: - * ace/OS_NS_time.cpp: - * ace/OS_NS_unistd.cpp: - * ace/OS_NS_wchar.cpp: - * ace/OS_TLI.cpp: - * ace/Thread_Adapter.cpp: - * ace/Thread_Control.cpp: - * apps/JAWS2/HTTPU/parse_http_request.cpp: - * apps/JAWS2/HTTPU/parse_http_response.cpp: - - Fixed typo in comment. - -Thu Jul 13 09:51:33 UTC 2006 Olli Savia <ops@iki.fi> - - * configure.ac: - Removed references to ACE_HAS_EXCEPTION_H. It's no longer - used in ACE/TAO/CIAO code. - -Thu Jul 13 09:48:10 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/os_include/os_string.h: - * ace/os_include/os_stropts.h: - Fixed typo in comment. - -Thu Jul 13 09:32:54 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/config-vxworks6.2.h: - * configure.ac: - * m4/config_h.m4: - Removed references to ACE_HAS_USING_KEYWORD. It's no longer - used in ACE/TAO/CIAO code. - -Thu Jul 6 20:45:15 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Refcounted_Auto_Ptr.{h inl}: Repaired the ACE_Refcounted_Auto_Ptr - reset() and release() methods per Bugzilla #1925. They will both now - detach from the underlying ACE_Refcounted_Auto_Ptr_Rep object; - reset() will create a new one for the new pointer specified as its - argument. Generally modified the class to not make changes to the - underlying ACE_Refcounted_Auto_Ptr_Rep (such as with reset()). This - safely allows the get() method to avoid using the lock for reading - the pointer value since there's no way to change the underlying - rep's pointer. - Also cleaned up the Doxygen stuff for these classes. - - This work resulted in removal of the following methods in - ACE_Refcounted_Auto_Ptr_Rep (not ACE_Refcounted_Auto_Ptr): - - null () - - assign () - - release () - - reset () - - This significantly improves ACE_Refcounted_Auto_Ptr. This is not - a recommendation to [not] use it over ACE_Strong_Bound_Ptr. Just - responding to user reports. It may still be possibly/adviseable to - work further on this class, though I have no specific ideas or - reports at this time. - - * tests/Refcounted_Auto_Ptr_Test.{h cpp}: Added tests for release() - and reset() per code supplied in Bugzilla #1925. Removed code in - ACE_ASSERT since it doesn't run in Release builds; replaced with - explicit checks and ACE_ERROR. - -Wed Jul 12 20:32:49 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Message_Queue_Test_Ex.{h cpp}: Fixed compile errors on Linux. - -Wed Jul 12 19:15:58 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Message_Queue_T.{h cpp}: Added a new class, - ACE_Message_Queue_Ex_N<class ACE_MESSAGE_TYPE, ACE_SYNCH_DECL>, that - is similar to ACE_Message_Queue_Ex in that the object queued is a - template parameter. However, ACE_Message_Queue_Ex_N allows the - enqueueing and dequeueing of multiple chained objects at once. This - wasn't added to ACE_Message_Queue_Ex because the chained object - functionality requires the ACE_MESSAGE_TYPE class to have a - ACE_MESSAGE_TYPE *next (void) const method, analogous to - ACE_Message_Block::next(), to follow the chain and this would - probably break existing applications using ACE_Message_Queue_Ex. - - * tests/Message_Queue_Test_Ex.{h cpp}: Added test capability for - ACE_Message_Queue_Ex_N. - - * NEWS: Added description of the new class. - - Thanks to Guy Peleg <guy dot peleg at amdocs dot com> for - contributing this new class and its tests. - -Wed Jul 12 16:37:40 UTC 2006 Steve Huston <shuston@riverace.com> - - * etc/ace.doxygen: - * etc/ace_qos.doxygen: - * etc/ace_ssl.doxygen: Use the IGNORE_PREFIX directive to have Doxygen - ignore the ACE_, ACE_QoS_, ACE_SSL_ prefixes when alphabetizing - the class index. Thanks to Stefan Naewe <naewe dot s at atlas dot de> - for this idea. - - * THANKS: Added Stefan Naewe to the Hall of Fame. - -Wed Jul 12 14:30:16 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-lynxos.h: - * ace/config-netbsd.h: Changed ACE_SCANDIR_CMP_USES_VOIDPTR to - ACE_SCANDIR_CMP_USES_CONST_VOIDPTR - makes it match the decl in - ace/os_include/os_dirent.h and should resolve a compile error. - -Wed Jul 12 13:55:09 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Message_Queue.inl: Move the #include out of the namespace; see - Tue Jul 11 13:38:10 UTC 2006 Steve Huston <shuston@riverace.com> - -Wed Jul 12 10:15:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * include/makeinclude/platform_openvms.GNU: - Added fixes for SSL support. - - * ace/config-openvms.h: - Added ACE_HAS_PTHREAD_SCHEDPARAM. - - * ace/Global_Macros.h: - * ace/Lib_Find.cpp: - * ace/Service_Config.cpp: - * ace/Service_Config.h: - * ace/Service_Object.cpp: - * ace/Service_Object.h: - Moved OpenVMS specific dynamic service support class to a - better place. - -Wed Jul 12 07:13:01 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/FILE_Connector.cpp: - It is safe to use ACE_OS::mkstemp() on all platforms. - -Wed Jul 12 02:21:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/msvc_static_order.lst: - Build ObjRefTemplate before IORInterceptor - -Tue Jul 11 23:09:28 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * include/makeinclude/platform_hpux_aCC.GNU: - - The C compiler doesn't support the -mt option. So, add -mt to - CC_THR_DEFS and then later add CC_THR_DEFS to CCFLAGS. CCFLAGS - isn't used when compiling C files. - -Tue Jul 11 22:30:21 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-win32-common.h: - * ace/config-WinCE.h: Break the tight tie between UNICODE and - ACE_USES_WCHAR I introduced in - Thu Oct 6 17:56:19 2005 Steve Huston <shuston@riverace.com> - Now ACE_USES_WCHAR is always on for CE; non-CE Windows builds must - explicitly set ACE_USES_WCHAR independent of UNICODE. Thanks to - Russell Morra <russell_mora at symantec dot com> for explaining - the situation and submitting this fix. - -Tue Jul 11 22:20:14 UTC 2006 Steve Huston <shuston@riverace.com> - - * bin/generate_doxygen.pl: Add the ability to put the generated - HTML output in a specified location using the new -html_output - command line option. It alters the path specified for the - HTML_OUTPUT and GENERATE_TAGFILE Doxygen options. - -Tue Jul 11 22:09:56 UTC 2006 Steve Huston <shuston@riverace.com> - - * include/makeinclude/platform_linux.GNU: Added support for the - "buildbits=64" and "buildbits=32" make settings to target - 32- and 64-bit build targets, respectively. Thanks to Boris - Kaminer <boris dor kaminer at gmail dot com> for this addition. - -Tue Jul 11 22:22:04 UTC 2006 Steve Huston <shuston@riverace.com> - - * configure.ac: Change the minimum automake version to 1.9.6; this is - required to generate Makefiles that work properly on AIX. - - * m4/compiler.m4: For AIX, replace ACE_AIX_MAJOR_VERS and - ACE_AIX_MINOR_VERS compile macros with ACE_AIX_VERS as the - hand-edited platform_aix_ibm.GNU has and the code expects. - Also removed the AIX 3 and 4 code since AIX 3 and 4 are no longer - supported. Also, turned on -qtemplateregistry for xlC-type - compilers (Visual Age C++ 6 and XL C++ 7). - -Tue Jul 11 21:37:51 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_NS_stdio.cpp (snprintf, sprintf): In the wchar versions, - adjust the preprocessor checks to pick up the way that Solaris - indicates the presence of XPG5 capability. Allows these functions - to work, which allows ACE_Log_Msg to work also. - - * ace/OS_NS_unistd.{h inl} (ualarm): Changed u_long arguments and - return value to useconds_t to gain portability. - - * ace/OS_NS_dirent.{h inl cpp}: Made the selector and comparator - function pointer arguments to ACE_OS::scandir() and - scandir_emulation() be extern "C" to enforce their use with a - C RTL function. This resolves warnings from Solaris and makes usage - correct across all platforms. The function types are now in - typedefs ACE_SCANDIR_SELECTOR and ACE_SCANDIR_COMPARATOR, - respectively. The typedef was used previously only for Windows/MSVC - where the calling convention change had bit someone - now it's - uniform for all. - Also, folded in the ACE_SCANDIR_CMP_USES_VOIDPTR def here to - set up the proper typing with extern "C" for when it's needed to - cast inside the ACE_OS::scandir() implementation. The interface - offered by ACE_OS::scandir() is always defined in terms of - ACE_SCANDIR_SELECTOR and ACE_SCANDIR_COMPARATOR. - - * ace/Dirent_Selector.{h cpp}: Use the extern "C"-containing typedefs - for selector/comparator functions set up for ACE_OS::scandir() in - ACE_Dirent_Selector as well since they get passed through to ACE_OS. - - * apps/soreduce/Library.cpp: - * tests/Dirent_Test.cpp: Mark selector/comparator functions that'll be - passed to ACE_OS::scandir() as extern "C" to match new param types. - - * ace/config-aix-5.x.h: Added ACE_HAS_SCANDIR since AIX has scandir(), - as well as the needed ACE_SCANDIR_SEL_LACKS_CONST and - ACE_SCANDIR_CMP_USES_VOIDPTR to adjust to it's differences from - the common way scandir() is usually defined. - - * ace/config-hpux-11.00.h: Added ACE_HAS_SCANDIR. - - * ace/config-linux-common.h: Clarify that, despite the man page, - ACE_SCANDIR_CMP_USES_CONST_VOIDPTR is needed, at least on RHEL 3 - and 4. - - * ace/README: Added descriptions for ACE_HAS_SCANDIR and - ACE_SCANDIR_CMP_USES_VOIDPTR as well as the new settings - ACE_SCANDIR_SEL_LACKS_CONST and ACE_SCANDIR_CMP_USES_CONST_VOIDPTR. - - * NEWS: Summarize the above user-visible selector/comparator changes. - -Tue Jul 11 17:39:00 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/Basic_Types.h: - * ace/OS.h: - Removed unnecessary includes. - -Tue Jul 11 17:25:06 UTC 2006 Olli Savia <ops@iki.fi> - - * include/makeinclude/platform_rtems.x_g++.GNU: - * include/makeinclude/wrapper_macros.GNU: - - Removed references to obsolete ACE_HAS_GNUG_PRE_2_8 make - variable. Removed egcs checks. - -Tue Jul 11 16:55:57 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/TTY_IO.cpp: - Enable raw mode correctly on POSIX systems. Accept "none" as - valid parity value. Thanks to hongxing <hongxing777 at gmail dot com> - for the patch. - - * ace/TTY_IO.h: - Updated documentation to reflect changes above. - - * ace/os_include/sys/os_resource.h: - Fixed typo in comment. - -Tue Jul 11 16:38:29 UTC 2006 Olli Savia <ops@iki.fi> - - * ace/config-linux-common.h: - * ace/os_include/os_signal.h: - * ace/os_include/os_stropts.h: - - Changed ACE_HAS_ISASTREAM_PROTO to ACE_HAS_ISASTREAM_PROTOTYPE and - ACE_HAS_PTHREAD_SIGMASK_PROTO to ACE_HAS_PTHREAD_SIGMASK_PROTOTYPE - to better follow ACE naming conventions. - -Tue Jul 11 13:38:10 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Message_Queue.inl: Include "ace/Guard_T.h" in the ACE_WIN32 - section to avoid compile errors with Release builds. - - * ace/TSS_T.h: Improved documentation. - - * tests/ACE_Init_TestDlg.cpp: Added comment block at the top to explain - the file's purpose. - - * tests/Task_Test.cpp: Added a smoke test for the new - ACE_Task::last_thread() method. - -Thu Jul 6 12:48:58 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/OS_NS_sys_socket.inl (sendto): Initialize "bytes_sent" to 0 - to avoid random values. Thanks to Winston Zhang <winston_ace at - kubao-inc dot com> for reporting this. - -Wed Jul 5 20:32:20 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/WFMO_Reactor.cpp (open): Added some more code to prevent - memory leaks when the ACE_WFMO_Reactor is reopened. Thanks to - Sonicfly Zhou <zhoucn at txsec dot com> for this fix. - -Tue Jul 11 12:19:23 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * etc/tao.doxygen: - Added .idl to FILE_PATTERNS - -Tue Jul 11 12:01:12 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * ace/Atomic_Op.cpp: - * ace/Atomic_Op_Sparc.h: - * ace/Atomic_Op_Sparc.c: - * include/makeinclude/platform_sunos5_sunc++.GNU: - - Added a SPARC assembler implementation of functions needed for use - within the ACE_Atomic_Op specializations. The implementation will - only be used if atomic_ops_sparc is set to 1 in - platform_macros.GNU. It also requires that v8plus or higher be - used with SunCC. - - * ace/Makefile.am: - * ace/ace.mpc: - - Added the new Atomic_Op_Sparc.c. - - * ace/CDR_Base.cpp: - - In grow(), call the clone_nocopy() method on the ACE_Data_Block - with the new size. Previously, the clone_nocopy() would be called - and then directly after that, size() would be called on the cloned - data block which would re-allocate to the new size and copy the - previous contents. - - In the same method, we avoid constructing an ACE_Message_Block to - avoid reference count manipulating the data block's reference - count. - - Also, in consolidate(), avoid attempting to copy data if the - destination and source are the same. [RT 8372] - - * ace/CDR_Stream.h: - * ace/CDR_Stream.cpp: - - Added a constructor that takes an ACE_Data_Block* which can be - used to avoid an allocation while constructing the - ACE_Message_Block held by the ACE_OutputCDR. [RT 8372] - - * ace/Message_Block.h: - * ace/Message_Block.cpp: - * ace/Message_Block_T.h: - * ace/Message_Block_T.cpp: - - Added a new parameter to the ACE_Data_Block::clone_nocopy() method - that allows the user copy the data block, but increase the size - all in one step. [RT 8372] - - * ace/OS_NS_string.inl: - - Added an unrolled implementation of memset, only active if - ACE_HAS_SLOW_MEMSET is defined, that supports up to 16 bytes - without calling into the memset() C function. - - * ace/config-sunos5.5.h: - - SunCC 5.x supports new(nothrow_t). So, I have enable the use of - it when exceptions are enabled. [RT 8372] - - * ace/OS_NS_Thread.h: - * ace/os_include/os_pthread.h: - - Added macro definitions required to build ACE and TAO on Solaris - without using the pthreads library. - - * include/makeinclude/platform_aix_ibm.GNU: - * include/makeinclude/platform_hpux_aCC.GNU: - * include/makeinclude/platform_irix6.x_sgic++.GNU: - - Added 64-bit options to the CFLAGS macros to deal with the newly - added C file in ACE. [RT 8372] - -Tue Jul 11 01:47:23 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Log_Msg_Test.cpp: - initialise pointer with 0 - -Tue Jul 11 01:37:23 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_unistd.h: - Layout change - -Tue Jul 11 01:09:23 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - If -oA has been used as idlflags, don't add the A files - -Tue Jul 11 00:50:23 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/generate_compile_stats.sh: - Updated to also handle redhat version information - -Sun Jul 09 05:53:23 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE version 5.5.2 released. - -Fri Jul 7 08:07:50 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: Small improvements - -Thu Jul 6 19:57:50 UTC 2006 Steve Huston <shuston@riverace.com> - - * ACE-INSTALL.html: Improvements and corrections motivated by - Abdul Sowayan <abdullah dot sowayan at lmco dot com>. - -Wed Jul 5 19:15:37 UTC 2006 Steve Huston <shuston@riverace.com> - - * ACE-INSTALL.html: Cleaned up further; clarified role of DOC and - support companies and use of PROBLEM-REPORT-FORM. - - * NEWS: Noted removal of Chorus and pSOS. - -Tue Jul 4 04:10:11 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Memcpy_Test.cpp: Added a short summary of this test. - Thanks to Arvind Krishna for contributing it. - -Tue Jul 4 01:05:47 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Memcpy_Test.cpp (testit): ACE-ified this test. - -Tue Jul 4 01:03:52 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/CE_fostream.cpp: - * tests/Service_Config_Test.cpp: Improved the description of these - tests. - -Mon Jul 3 14:42:45 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * NEWS: Updated the NEWS file with information about using multiple - service repositories. - -Mon Jul 3 03:25:43 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * examples/APG/Svc_Config/HA_Configurable_Server_Dynamic.cpp - (ACE_TMAIN): Changed the reactor event loop to explicitly pass - in ACE_Reactor::check_reconfiguration(). Thanks to Michael Reed - <mnr102 at yahoo dot com> for reporting this. - -Mon Jul 3 03:15:14 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * examples/Service_Configurator/IPC-tests/server/server_test.cpp - (ACE_TMAIN): Changed the reactor event loop to explicitly pass - in ACE_Reactor::check_reconfiguration(). - -Fri Jun 30 13:40:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/msvc_static_order.lst: - Added CSD_Framework and CSD_ThreadPool - -Fri Jun 30 09:33:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/generate_rel_manpages: - Updated doxygen version we are using for the release - -Fri Jun 30 07:29:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoexe.mpb: - Added corba/e and core_minimum_corba as base projects so that - the executables that can be build with CORBA/e and minimum - corba do have set the correct flags during compilation - -Thu Jun 29 17:30:52 UTC 2006 Steve Huston <shuston@riverace.com> - - * NEWS: Added a note about the logging changes from - Tue May 16 22:23:55 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - -Thu Jun 29 10:05:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Signal_Test.cpp: - Layout changes - -Thu Jun 29 09:13:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Containers_T.cpp: - Initialise pointer with 0 - -Thu Jun 29 08:52:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_g++_common.GNU: - * include/makeinclude/platform_osf1_4.x_cxx.GNU: - * include/makeinclude/platform_sunos5_ghs.GNU: - * include/makeinclude/platform_sunos5_sunc++.GNU: - * include/makeinclude/wrapper_macros.GNU: - Removed support for explicit templates, we don't support that - anymore - -Thu Jun 29 08:34:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/ACE-guidelines.html: - Document that the STL generic algorithms are safe to be used and - made some other small updates - -Wed Jun 28 10:03:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/ace_for_tao.mpc: - Added missing file because of the sendfile addition to TAO - -Wed Jun 28 09:20:08 UTC 2006 Boris Kolpackov <boris@kolpackov.net> - - * protocols/ace/RMCast/Link.cpp: - * protocols/ace/RMCast/Link.h: - * protocols/ace/RMCast/Socket.cpp: - Added explicit calls to close functions for sockets and pipes. - Thanks to finalpatch <fengli@gmail.com> for reporting this. - -Wed Jun 28 07:10:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * docs/ACE-guidelines.html: - Document that template template arguments are not possible to - be used because HP aCC 3.70 doesn't support it. - -Tue Jun 27 13:06:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Unbounded_Queue.cpp: - Explicitly initialise pointers with 0 - -Tue Jun 27 08:59:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/ace_for_tao.mpc: - Added MMAP_Memory_Pool.* because TAO uses these for sendfile - support - -Tue Jun 27 06:56:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_unistd.inl: - * ace/config-WinCE.h: - WinCE doesn't seem to have SetFilePointerEx, so introduced - ACE_LACKS_SETFILEPOINTEREX which we check in OS_NS_unistd.inl, - if this is set we fallback to SetFilePointer - -Tue Jun 27 06:47:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/CDR_File_Test.cpp: - Fixed 64bit conversion warning - -Tue Jun 27 06:34:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taolib.mpb: - Added tao_versioning_idl_defaults as base - -Mon Jun 26 13:27:00 UTC 2006 Simon Massey <sma@prismtech.com> - - * NEWS: - Added PICurrent refactoring note. Moved from NEWS to TAO/NEWS - -Mon Jun 26 08:27:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/DLL_Manager.h: - Reexport ACE_DLL_Handle, the ACE_DLL_Manager returns a pointer - to this type in its interfaces. Thanks to Patrick Bennett - <Patrick dot Bennett at inin dot com> for reporting this - -Mon Jun 26 08:17:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * include/makeinclude/platform_openvms.GNU: - Some compileswitch changes. - - * ace/config-openvms.h: - Added multicast filtering setting. - -Sun Jun 25 11:54:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * include/makeinclude/platform_openvms.GNU: - Some compileswitch changes. - -Sun Jun 25 08:36:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/msvc_mpc_auto_compile.pl: - Build gperf and TAO_IDL before tao - -Sat Jun 24 14:02:15 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Lib_Find.cpp (open_temp_file): In the ACE_WIN32 section, - 'perm' is not used, so use ACE_UNUSED_ARG on it. - - * tests/Hash_Map_Manager_Test.cpp: Use const ACE_TCHAR with literal - string, not non-const. Resolves compile warning on some compilers. - -Fri Jun 23 21:03:37 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * ace/Lib_Find.cpp (open_temp_file): - - Added FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE - for the 'perms' argument of the Windows version, which - eliminates a open_temp_file() failure in the IDL compiler - on a file that was previously opened/closed with fopen/fclose. - Thanks to Krishnakumar Balasubramanian <kitty@dre.vanderbilt.edu> - for figuring out the fix. - -Fri Jun 23 16:01:31 UTC 2006 Steve Huston <shuston@riverace.com> - - * examples/Reactor/Proactor/test_proactor.cpp (open): When calling - ACE_OS::open(), don't supply a very UNIXy-looking 0644 value for - the 'perms' argument. Let it default to something reasonable for - the platform. Using 0644 on Windows passes something Windows doesn't - like, and the test fails. Thanks very much to William Cote - <William dot Cote at hp dot com> for fixing this. - - * THANKS: Added William Cote to the Hall of Fame. - -Fri Jun 23 13:20:21 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - Reverted the change below, it triggers pidl recompilation for all pidl files - Fri Jun 23 07:36:21 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - Added .pidl to the inputext. This should be the way how things should - run fine again in TAO but it seems somewhere ther is something missing. - -Fri Jun 23 07:36:21 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - Added .pidl to the inputext. This should be the way how things should - run fine again in TAO but it seems somewhere ther is something missing. - -Thu Jun 22 21:05:20 UTC 2006 Krishnakumar B <kitty@dre.vanderbilt.edu> - - * ace/OS_NS_fcntl.h (ACE_DEFAULT_OPEN_PERMS): Set the value of - ACE_DEFAULT_OPEN_PERMS from ACE_DEFAULT_FILE_PERMS instead of - 0. This ensures that users of ACE_OS::open() who don't override - the default argument get a sensible default. - -Thu Jun 22 15:08:01 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/SSL/SSL_Context.{h cpp} (load_trusted_ca): Added a third - argument, bool use_env_defaults, which defaults to true. If false, - the ca_file and ca_dir arguments do not get ACE defaults applied - to them if passed as 0. This allows a user to specify a 0 file or - directory to the underlying SSL_CTX_load_verify_locations() function. - If a ca_dir is supplied, attempt to add its CA list to the - X509 names in addition to that of ca_file. - - (seed_file): If the file loads ok, check RAND_status() to see if - there's "enough" entropy. - - * ace/SSL/SSL_Context.inl: - * ace/SSL/SSL_Asynch_Stream.cpp: - * ace/SSL/SSL_SOCK_Stream.cpp: When setting up SSL_CTX, set the verify - mode as well. When creating a new socket object, let it inherit the - verify mode from the SSL context rather than making a call back to - the context to retrieve the verify mode and setting it with a - separate SSL call. - -Thu Jun 22 14:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/msvc_static_order.lst: - Build the portableserver earlier in the list of libs - -Thu Jun 22 05:29:48 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.cpp: - - The initialize_i() method would not remove a service, which - failed to initialize, in case the ACE_NOLOGGING is defined. For - some reason the two blocks of code were switched around. Thanks - to Martin Corino <mcorino at remedy dot nl> for pointing out the - problem. - -Wed Jun 21 21:29:04 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-lynxos.h: Added ACE_LACKS_STRPTIME. This should fix a - compile error on LynxOS. - -Wed Jun 21 15:45:32 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Hash_Map_Manager_Test.cpp: Fix incorrect concatenating of - text literals. - -Wed Jun 21 07:10:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Misc/test_trace.cpp: - Fixed fuzz error - -Tue Jun 20 23:20:17 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Framework_Component_DLL.cpp: Added a description of this test. - -Tue Jun 20 22:34:43 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Service_Config_DLL.cpp: Added a comment block at the top - to describe this file's purpose. - -Tue Jun 20 22:09:54 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_NS_time.cpp (strptime): Added a const_cast<char*> wherever - a return of something based on the const char *buf was done; related - to: - Tue Jun 20 20:19:37 UTC 2006 Steve Huston <shuston@riverace.com> - - * NEWS: Noted that the signature of ACE_OS::strptime() changed, also - related to the entry above. - -Tue Jun 20 20:19:37 UTC 2006 Steve Huston <shuston@riverace.com> - - * configure.ac: - * ace/README: - * ace/OS_NS_time.h: - * ace/OS_NS_time.inl: - * ace/OS_NS_time.cpp: - * m4/config_h.m4: The ACE_OS::strptime() wrapper was previously not - compiled in unless ACE_HAS_STRPTIME was set. If this was set and - ACE_LACKS_NATIVE_STRPTIME was also set, an emulation was supplied. - This has been changed so that: - - ACE_OS::strptime() is always supplied. - - The emulation is used if ACE_LACKS_STRPTIME is set. - - The first arg to ACE_OS::strptime() is now "const char *" - instead of "char *" as the native strptime() is. - - ACE_OS::strptime() is inlineable. - ACE_HAS_STRPTIME is no longer used, and ACE_LACKS_NATIVE_STRPTIME - is changed to ACE_LACKS_STRPTIME. If ACE_LACKS_STRPTIME is set - and the user decides not to include the emulation, then - ACE_REFUSE_STRPTIME_EMULATION should be set, which will cause - ACE_NOTSUP_RETURN to be used instead of the emulation. - - * ace/config-borland-common.h: - * ace/config-vxworks5.x.h: - * ace/config-vxworks6.2.h: - * ace/config-win32-dmc.h: - * ace/config-win32-ghs.h: - * ace/config-win32-mingw.h: - * ace/config-win32-msvc-7.h: - * ace/config-win32-msvc-8.h: - * ace/config-win32-visualage.h: Removed ACE_HAS_STRPTIME; changed - ACE_LACKS_NATIVE_STRPTIME to ACE_LACKS_STRPTIME. - - * ace/config-cygwin32.h: - * ace/config-doxygen.h: - * ace/config-irix6.x-common.h: - * ace/config-linux-common.h: - * ace/config-netbsd.h: - * ace/config-openvms.h: - * ace/config-sunos5.5.h: - * ace/config-tandem-nsk-mips-v3.h: Removed ACE_HAS_STRPTIME. - - * ace/config-rtems.h: Changed ACE_LACKS_NATIVE_STRPTIME to - ACE_LACKS_STRPTIME. - NOTE!! RTEMS users should add #define ACE_REFUSE_STRPTIME_EMULATION - to the local config.h file to avoid bringing in the emulation, - maintaining behavior in previous ACE versions. - - For additional background on why this arrangement was put in place - originally, please see: - Tue Aug 10 14:46:12 1999 David L. Levine <levine@cs.wustl.edu> - It makes sense, but adds a unique requirement for porting to new - platforms that one must know to explicitly enable an existing wrapper, - even if its natively available. This has tripped up a number of porting - efforts (including mine) and now it's consistent with other wrappers. - - * m4/ace.m4: Add --enable-wince to set up the BUILD_WINCE setting - that's inserted into the Makefile.am files for the wince MPC setting. - - * ace/ace.mpc: Replace config-win32-msvc-6.h with config-win32-msvc-8.h - - * bin/bin/mpc: Removed the verbatim for automake that adds - TESTS=auto_run_tests.pl - it was an early attempt at getting the - test ability to work that never was right. - - * */Makefile.am: Regenerated with MPC. This effectively changes the - version scheme based on the changes in: - Thu May 18 16:37:46 UTC 2006 Steve Huston <shuston@riverace.com> - - * apps/Makefile.am: Hand-added code to avoid building gperf if - --disable-gperf was specified to configure. Thanks to Robert Schwebel - <robert at schwebel dot de> for this fix. - - * apps/gperf/tests/Makefile.am: This one doesn't get close to being - regenerated with MPC. The "stamp" items require hand-tweaking as - given in previous entries for this file. - - * ace/Makefile.am: Regenerated with MPC and hand-tweaked to add needed - "nobase_include_HEADERS =" line. - - * configure.ac: Added the following new AC_CONFIG_FILES entries: - - * ACEXML/tests/util/Makefile.am: - * apps/JAWS2/Makefile.am: - * apps/JAWS2/JAWS/Makefile.am: - * examples/ASX/Event_Server/Makefile.am: - * examples/ASX/Event_Server/Event_Server/Makefile.am: - * examples/ASX/Event_Server/Transceiver/Makefile.am: - * examples/ASX/UPIPE_Event_Server/Makefile.am: - * examples/Map_Manager/Makefile.am: - * examples/Reactor/TP_Reactor/Makefile.am: - * protocols/tests/Makefile.am: - * protocols/tests/RMCast/Makefile.am: New files, generated with MPC. - - * THANKS: Added Robert Schwebel to the Hall of Fame. - -Tue Jun 20 19:25:20 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Task.{h inl cpp}: Moved the public data members (thr_count_, - thr_mgr_, flags_, grp_id_, lock_) from public to a new protected - section. There was a comment that said these should be private, but - no reason given why they weren't. - Removed the thr_count_dec() method since it is no longer used - anywhere in ACE. Moved its functionality into cleanup() where it - also sets the new last_thread_id_ member (see below). - Added a new member last_thread_id_ which holds the thread ID - (ACE_thread_t) of the last thread to exit svc() processing. Before - any threads exit, the value is 0; it's set to 0 in the constructor - and in activate(). - The new accessor, ACE_thread_t last_thread (void) const, retrieves - the value of last_thread_id_. It's recommended that a thread in - close() check the last_thread_id_ value, rather than the thr_count(), - when deciding if it's safe to destroy the object. Since thr_count - can be seen as 0 in multiple threads concurrently, it's not safe - to rely on for destroying a thread. This fixes Bugzilla #2339. - Thanks to Howard Finer <hfiner@sonusnet.com> for this idea. - - * NEWS: Added a note about the new ACE_Task::last_thread() method. - -Tue Jun 20 18:55:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * tests/Based_Pointer_Test.cpp: - * tests/Based_Pointer_Test_Lib.cpp: - * tests/CDR_File_Test.cpp: - * tests/DLL_Test.cpp: - * tests/Malloc_Test.cpp: - * tests/Process_Manager_Test.cpp: - Several fixes to make OpenVMS happy. - -Tue Jun 20 18:04:45 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * examples/Log_Msg/test_ostream.cpp (ACE_TMAIN): Couldn't used - std::getline() for some odd reason.. Fixed by using - ifstream::getline() instead. Thanks to Johnny for reporting - this. - -Tue Jun 20 14:31:05 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * examples/Misc/test_trace.cpp (recursive): Simplified the - ACE_TRACE method name to make fuzz happy. - -Tue Jun 20 13:50:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Hash_Map_Manager_Test.cpp: - Fixed unicode compile problems - -Tue Jun 20 12:51:19 UTC 2006 James H. Hill <hillj@isis.vanderbilt.edu> - - * bin/ChangeLogEditor/FileLocatorFactory.pm: - - Now supports the standard hack for subversion on Windows when - SVN_ASP_DOT_NET_HACK environment variable is set. Instead - of placing repo info in a .svn file, it is placed in _svn. - -Tue Jun 20 12:31:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * ace/os_include/os_time.h: - * ace/os_include/sys/os_types.h: - Fixed type definitions for OpenVMS. - - * ace/config-openvms.h: - Updated build configuration for OpenVMS. - - * ace/Lib_Find{.h .cpp}: - * ace/DLL_Manager.cpp: - * ace/Service_Config{.h .cpp}: - * ace/Global_Macros.h: - * ace/Strategies_T.cpp: - * ace/Logging_Strategy.cpp: - Added support for dynamic loading scheme of service configurator - framework under OpenVMS. - - * ace/OS_NS_stdio.h: - Additional include for OpenVMS. - -Tue Jun 20 08:23:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taolib.mpb: - Add corba subset projects as base - - * bin/MakeProjectCreator/config/core_minimum_corba.mpb: - New file for usage within the core libs - - * bin/MakeProjectCreator/config/core_minimum_corba.mpb: - Make minimum corba a feature - -Tue Jun 20 07:42:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Log_Msg/test_callback.cpp: - Fixed typo - -Tue Jun 20 07:04:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/global.features: - CORBA subset configurations are disabled by default - -Mon Jun 19 18:03:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * include/makeinclude/platform_openvms.GNU: - Changed template instantiation model. - - * include/makeinclude/rules.local.GNU: - Change to enable adding custom REALCLEAN_FILES in platform file. - -Mon Jun 19 15:11:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * bin/auto_run_tests.pl: - Changed to enable spawning Perl testscripts on OpenVMS. - -Mon Jun 19 14:26:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: - Replaced dsp/dsw with sln/vcproj. Thanks to Abdullah Sowayan - <abdullah dot sowayan at lmco dot com> for reporting this - -Mon Jun 19 13:14:54 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Hash_Map_Manager_Test.cpp (test_two_allocators): Fixed a typo. - Thanks to Abdul for reporting this. - - * tests/Hash_Map_Manager_Test.cpp (test_two_allocators): Added a - test for the two allocator version of ACE_Hash_Map_Manager. - Thanks to Jarek for contributing this. - -Mon Jun 19 02:55:48 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * examples/Log_Msg/test_ostream.cpp (ACE_TMAIN): Enhanced the test to - show the ofstream output better. - -Mon Jun 19 01:31:46 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * examples/Log_Msg/test_callback.cpp (ACE_TMAIN): Zapped the - extraneous '\n' from the messages. - -Mon Jun 19 01:06:49 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * examples/Misc/test_trace.cpp: Enhanced to the the ACE_TRACE macro. - -Sun Jun 18 16:58:20 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * netsvcs/clients/Logger/indirect_logging.cpp (ACE_TMAIN): Cleaned - up the test to work if none of the daemons are running. - -Sun Jun 18 16:44:14 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Signal.{inl,cpp}: Moved the destructors out of the *.inl - file into the *.cpp file since inline destructors aren't really - a good idea. - -Sun Jun 18 16:35:13 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * netsvcs/lib/Client_Logging_Handler.cpp (send): Printout the type - of log message as a diagnostic to make it easier to see the - filtering that's going on! - -Fri Jun 16 17:48:18 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/CDR_Size.cpp (adjust): Added a fix for the case where - ACE_LACKS_CDR_ALIGNMENT is undefined. Thanks to Aleksandar - Vukajlovic <vukajlo at finsoft dot co dot yu> for contributing - this. - -Fri Jun 16 13:57:53 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Ping_Socket.cpp: Replaced the use of char, short, int, - etc. for struct ip and icmp with the equivalent - ACE_UNIT[8,12,32] types to ensure a fixed size. Thanks to Paul - Daugherty <Paul at nextalk dot com> and Robert S. Iakobashvili - <coroberti at gmail dot com> for reporting this and contributing - a fix. - -Mon Jun 19 11:37:17 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_unistd.{h,cpp,inl}: - Changed the (f)truncate method to use ACE_LOFF_T and use - SetFilePointerEx on windows to fix 64bit warnings - -Mon Jun 19 10:08:17 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/diff-builds.pl: - Changed this script not to die when we can't find results for - a certain date, then just take the closest results. - -Mon Jun 19 10:00:17 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Service_Gestalt.cpp: - * ace/Acceptor.cpp: - Const improvement - -Mon Jun 19 09:56:17 UTC 2006 Ossama Othman <ossama_othman at symantec dot com> - - * ace/OS_NS_sys_sendfile.cpp (sendfile_emulation): - - Verify the ACE_OS::mmap() call succeeded prior to continuing. - - Use ACE_OS::send() instead of ACE_OS::write() on Windows. The - latter doesn't provide the same semantics on Windows as on - Unix. Addresses an ACE_OS::sendfile() failure on Windows. - -Tue Jun 13 02:10:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/run_tests.lst: - Don't run netlink test in ACE_FOR_TAO configuration - -Mon Jun 12 17:00:02 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * include/makeinclude/wrapper_macros.GNU: - - Set winnt to 0 if not explicitly set by a platform_*.GNU - makefile. - -Mon Jun 12 16:48:19 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/SV_Shared_Memory_Test.cpp (run_main): Moved a couple of - calls outside of ACE_ASSERT since they don't belong in there. - -Mon Jun 12 14:34:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_stat.h: - Fixed mingw definition of ACE_stat - -Mon Jun 12 10:24:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/vc6.features: - Removed this file, wer don't support vc6 anymore - -Mon Jun 12 01:30:03 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Thread_Manager.cpp (wait): Changed the flag parameters to be - bool rathet than int. Thanks to Johnny for this suggestion. - -Sat Jun 10 22:32:03 UTC 2006 Steve Huston <shuston@riverace.com> - - * ACE-INSTALL.html: Corrected a bunch of platform build things and - hopefully clarified the sections a bit. - - * bin/make_release: Removed the search for, and chmod for, dsp/dsw - files - they're for MSVC6 which is no longer generated. - -Sat Jun 10 10:07:32 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Thread_Manager.cpp: - Fix for compiler errors. - -Sat Jun 10 04:31:54 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * bin/MakeProjectCreator/config/automake.features: - - Define MPC feature variable "wxWindows" to 1. - - * bin/MakeProjectCreator/config/wxwindows.mpb: - - Add automake specific definitions for compile_flags, - linkflags, and pure_libs. - -Fri Jun 9 19:25:19 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Thread_Manager.h: Added a new parameter to wait() that - determines whether relative or absolute time is used. Thanks to - Sonicfly Zhou <zhoucn at txsec dot com> for this idea. - -Fri Jun 9 14:06:52 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Hash_Map_With_Allocator_T.{h,inl,cpp}: Enhanced this class - so that it compiles with the new "two allocator" - ACE_Hash_Map_Manager_Ex class. This class needs to be improved - so that it actually leverages the two allocator features, but - that's something for the future. - -Fri Jun 9 13:15:38 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Hash_Map_Manager_T.h: Clarified the meaning of the - constructor and open() parameters. - -Fri Jun 9 13:07:40 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Hash_Map_Manager_T.cpp (open): If the user doesn't supply an - entry_alloc pointer then use the table_alloc pointer. - -Fri Jun 9 13:00:22 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Hash_Map_Manager_T.cpp (open): Zapped the use of ACE_ASSERT - and replaced it with and if statement. We shouldn't be using - ACE_ASSERT in ACE library code.. - -Fri Jun 9 12:53:48 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Hash_Map_Manager_T.{h,inl,cpp}: Added support for a separate - allocator for creating the entries in the chain of - ACE_Hash_Map_Entry objects in the buckets. There are now two - allocators, one for the bucket table and the other for the - entries in the buckets. Thanks to Jaroslaw Nozderko <jaroslaw - dot nozderko at polkomtel dot com dot pl> for this fix. - -Fri Jun 9 14:04:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/ace_for_tao.mpc: - Updated because ARGV and Arg_Shifter are template files now - -Fri Jun 9 13:58:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/FILE_IO.cpp: - Const improvements - - * ace/MMAP_Memory_Pool.cpp: - Added static cast from ACE_LOFF_T to size_t and const improvements - -Fri Jun 9 13:58:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/FILE_IO.cpp: - Added static cast from ACE_LOFF_T to size_t - -Fri Jun 9 13:42:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_stat.inl: - Use special stat call for vc71 - -Fri Jun 9 13:33:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Connection_Recycling_Strategy.h: - Doxygen improvement - -Fri Jun 9 13:12:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_unistd.{h,cpp}: - Use ACE_LOFF_T in the pwrite interface - -Fri Jun 9 12:47:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_stat.h: - With vc71 use _stati64 for ACE_stat, not __stat64. The first has - time as 32bit, the second time as 64 bit (__time64_t) which causes - problems because the time interfaces of ACE use time_t which is - 32bit with vc71 but 64bit with vc8. - -Fri Jun 9 10:29:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Configuration_Import_Export.cpp: - Initialise pointer with 0 - -Fri Jun 9 10:13:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Arg_Shifter.{h,cpp}: - Also refactored ACE_Arg_Shifter to ACE_Arg_Shifter<CHAR_TYPE> and - made ACE_Arg_Shifter a typedef. The refactoring fo ACE_ARGV resulted - in a drop of the TAO Hello client executable of about 2.5Kb which is - nice for such a small change. - - * ace/ace.mpc: - * ace/Makefile.am: - Updated because Arg_Shifter.cpp is not a template file - -Fri Jun 9 09:54:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-linux-common.h: - Added ACE_HAS_NETLINK, all linux versions do support netlink sockets. - This closes bugzilla bug 2317. Thanks to Robert Iakobashvili - <coroberti at gmail dot com> and Razi Ben-Yehuda - <raziebe at gmail dot com> for creating the netlink wrappers - -Fri Jun 9 08:16:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/SOCK_Netlink_Test.cpp: - New test for netlink sockets. Thanks to Robert Iakobashvili - <coroberti at gmail dot com> for creating this test. - - * tests/tests.mpc: - * tests/run_test.lst: - Added new netlink test - -Fri Jun 9 07:55:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/ARGV.{h,cpp,inl}: - - Refactored ACE_ARGV to ACE_ARGV_T which is a template, where - ACE_ARGV is now a typedef of ACE_ARGV_T<ACE_TCHAR>. Taken the - idea from the wchar work of David Giovannini <giovannd at ociweb - dot com> - - * ace/ace.mpc: - * ace/Makefile.am: - - ARGV.cpp added to the template files, added the new netlink files - - * ace/OS_NS_sys_stat.{h,inl}: - * ace/MMAP_Memory_Pool{h,cpp}: - * ace/Mem_Map.{h,cpp,inl}: - * ace/Filecache.{h,cpp}: - * ace/FILE_IO.cpp: - * ace/FILE.h: - - Use ACE_LOFF_T for the filesize so that we can support files - bigger then 2Gb on windows and other platforms. Also use for - Borland the 64bit stat structure and api's. This fixes bugzilla - bug 1286, thanks to Ruslan Zasukhin <sunshine at public dot - kherson dot ua> for reporting this. Also fixes part of bugzilla - bug 2359, thanks to Olivier Brunet <o dot brunet at free dot fr> - for reporting this. - -Thu Jun 8 17:13:42 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/String_Base.h, - ace/Array_Map.h: Removed the @deprecated - comments to make Ossama happy ;-) This fixes bugid 2558. - -Thu Jun 8 12:44:46 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/OS_NS_Thread.inl (recursive_mutex_lock): Cleaned up the - structure of the code here so that we don't try to release the - nesting lock unless it was acquired successfully. Thanks to - J.T. for reporting this. - -Thu Jun 8 12:37:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/corba_e_compact.mpb: - * bin/MakeProjectCreator/config/corba_e_micro.mpb: - New base projects for CORBA/e micro and compact - -Thu Jun 8 12:28:49 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/DependencyGenerator/DependencyEditor.pm: - * bin/DependencyGenerator/DependencyGenerator.pm: - * bin/DependencyGenerator/GNUDependencyWriter.pm: - * bin/DependencyGenerator/GNUObjectGenerator.pm: - * bin/DependencyGenerator/Preprocessor.pm: - * bin/depgen.pl: - - Committing performance improvements which take off about 7% from - the dependency generation time for the whole ACE_wrappers tree. - -Thu Jun 8 11:57:58 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * ace/OS_NS_dirent.cpp: - * ace/OS_NS_dirent.h: - - Don't compile the {opendir,readir,closedir,scandir}_emulation() - functions if {opendir,readir,closedir,scandir} are provided by - the OS. - -Thu Jun 8 11:29:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Netlink_Addr.h: - * ace/SOCK_SEQPACK_Association.h: - Doxygen fixes - - * ace/OS_NS_Thread.inl: - Const improvement - -Thu Jun 8 11:37:59 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * ace/config-netbsd.h: - - Define ACE_HAS_SCANDIR and ACE_SCANDIR_CMP_VOIDPTR. - -Wed Jun 7 23:02:53 UTC 2006 Krishnakumar B <kitty@nospam.invalid.domain> - - * tests/Recursive_Mutex_Test.cpp: Use ACE_TEST_MUTEX instead of - ACE_Recursive_Thread_Mutex so that we use the right version of - mutex (ACE_Process_Mutex) on Windows. Thanks to Doug for the - suggestion. This should fix the Recursive_Mutex_Test failing on - Windows. - -Wed Jun 7 21:42:36 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/MakeProjectCreator/config/ciao_config_handlers.mpb - - Fixed static library ordering problem. - -Wed Jun 7 13:34:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-doxygen.h: - Added ACE_HAS_NETLINK so that we get documentation for the new - netlink classes - -Wed Jun 7 13:31:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/run_test.lst: - Removed chorus - -Wed Jun 7 13:16:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_socket.h: - When ACE_HAS_NETLINK is defined, include more files specific - for this protocol. - -Wed Jun 7 12:56:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Netlink_Addr.{h,cpp,inl}: - * ace/SOCK_Netlink.{h,cpp,inl}: - New files for netlink sockets. Thanks to Robert Iakobashvili - <coroberti at gmail dot com> and Razi Ben-Yehuda - <raziebe at gmail dot com> for creating these new files. - These files will be added to the ACE mpc later this week, with - the files in the repo it is much easier to test them on - multiple linux version. See bugzilla bug 2317 for more info. - -Wed Jun 7 12:28:18 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/MakeProjectCreator/templates/gnu.mpd: - - Changed the message displayed when a project isn't build when a - particular feature is disabled. Many users thought that this - message was an error, I believe, because of the word 'missing'. - -Wed Jun 7 11:46:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-suncc-common.h: - New config file for the SunCC compiler, idea is to refactor - all things related to the Sun C++ compiler to this file. This is - to make it easier to support Sun C++ on Linux - -Wed Jun 7 10:45:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_linux_suncc.GNU: - First version for using the Sun compiler on Linux. The compiler - itself is an alpha version, so no guarantees at this moment - -Wed Jun 7 10:07:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Strategies_T.cpp: - Initialise pointer with 0 and const improvement - -Wed Jun 7 09:19:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - When using -SS we do generate an empty S.h file, so add S - to the header_outputext so that we zap this file during a clean - -Wed Jun 7 09:02:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_aix_g++.GNU: - Set X11 and Xt defines, thanks to Peter Grotrian - <peter dot grotrian at pdv-fs dot de> for reporting this. This fixes - bugzilla 2556 - - * ace/config-sunos5.5.h: - Added fixes to build with threads=0, this fixes bugzilla bug 2555. Thanks - to Peter Grotrian <peter dot grotrian at pdv-fs dot de> for reporting - this. - - * ace/config-aix-4.x.h: - Added fixes to build with threads=0, this fixes bugzilla bug 2553. Thanks - to Peter Grotrian <peter dot grotrian at pdv-fs dot de> for reporting - this. - - * include/makeinclude/platform_chorus.GNU: - * include/makeinclude/platform_chorus4.x_g++.GNU: - * include/makeinclude/platform_chorus_ghs.GNU: - * include/makeinclude/platform_psos_diab.GNU: - * include/makeinclude/platform_psos_diab_ppc.GNU: - * include/makeinclude/platform_psos_tm.GNU: - * include/makeinclude/platform_psosim_g++.GNU: - Removed these files, psos and chorus are not supported anymore - -Wed Jun 7 08:27:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_Thread.cpp: - Added needed check for VxWorks with task api - - * ace/os_include/sys/os_stat.h: - Removed workarounds for old Borland compiler versions - -Wed Jun 7 03:25:00 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/MakeProjectCreator/config/ciao_config_handlers_base.mpb - * bin/MakeProjectCreator/config/ciao_config_handlers_base.mpc - - Corrected the file extension. - -Tue Jun 6 21:11:37 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/WFMO_Reactor.cpp (open): make sure we delete the - notify_handler_ before reassigning it to prevent resource leaks. - Thanks to Sonicfly Zhou <zhoucn at txsec dot com> for suggesting - this. - -Tue Jun 6 21:41:02 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Service_Gestalt.{h cpp}: - * ace/Dynamic_Service_Base.cpp: - * ace/Dynamic_Service_Dependency.cpp: - * ace/Service_Config.cpp: - * ace/Service_Object.cpp: - * ace/Svc_Conf_Lexer.cpp: Add missing #include directives. - -Tue Jun 6 18:09:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_Thread.cpp: - Reverted a change for VxWorks, broke other platforms, will make - a new change tomorrow. - -Tue Jun 6 17:14:51 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/MakeProjectCreator/modules/GNUACEProjectCreator.pm: - - Cleaned up the fill_value() method and removed an unused method. - - * bin/MakeProjectCreator/modules/GNUACEWorkspaceCreator.pm: - - Use DirectoryManager::mpc_basename() instead of - File::Basename::basename(). - -Tue Jun 6 15:43:32 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/MakeProjectCreator/templates/gnu.mpd - - Modified to explicitly state the name of a project being skipped - due to missing features or libraries. - -Tue Jun 6 03:57:08 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * ace/Hash_Map_Manager_T.cpp: - * ace/Hash_Map_Manager_T.h: - - Added key() and item() methods to ACE_Hash_Map_Entry<>, to be - like ACE_RB_Tree_Node<>. This makes it easier to switch code - to and from Hash_Maps and RB_Trees. Mark direct access to - the ext_id_ and int_id_ fields @deprecated. - - * ace/Hash_Map_Manager_T.cpp: - * ace/Hash_Map_Manager_T.h: - * ace/Hash_Map_Manager_T.inl: - * ace/RB_Tree.h: - * ace/RB_Tree.inl: - - Added operator->() for iterators to be more like STL API. - -Mon Jun 5 20:27:42 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * bin/MakeProjectCreator/config/ciao_config_handlers.mpb - * bin/MakeProjectCreator/config/ciao_config_handlers_base.mpc - - Restructured to clean up config handlers MPC files. - -Sun Jun 4 13:38:21 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Array_Map_Test.cpp: Fixed all uses of empty() to be - is_empty(). - -Sun Jun 4 13:35:25 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Array_Map.{h,inl}, - * ace/Intrusive_List.{h,inl}, - * ace/String_Base.{h,inl}: Changed empty() to is_empty(), which - makes more sense. Kept empty() around for the time being as a - @deprecated method. Thanks to Domingos Monteiro <d dot monteiro - at netia dot net> for this suggestion. - -Sun Jun 4 00:51:38 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/CString.{h,inl,cpp}: Added some performance improvements for ACE_CString: - - . Introduced new operators == and != to compare ACE_CString - and const char *. This results in at least a 10-fold performance - improvement, due to avoiding dynamic memory allocation. - - . Streamlined the existing operator == between two ACE_CString's. - (15% improvement in some cases, due to avoiding some of the - overhead of compare()) - - Also cleaned up some of the doxygen comments, e.g., the - ACE_CString comparison operators must have been returning an - integer at some point (?), but it's not the case anymore. - Thanks to Patrick Rabau <pr2345 at gmail dot com> for these - patches. - -Fri Jun 2 19:52:19 UTC 2006 Yan Dai <dai_y@ociweb.com> - - * bin/tao_orb_tests.lst: - - Removed DII_Collocation_Tests. - -Fri Jun 2 16:29:55 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Select_Reactor_T.cpp (open): Added an error message - explicitly noting that the notification pipe open failed, which - should make debugging easier when this occurs, e.g., due to a - disabled network interface. Thanks to David Highley <dhighley - at highley-recommended dot com> for this suggestion. - -Fri Jun 2 12:20:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidl3toidl2defaults.mpb: - Updated to make cross compilation possible - -Fri Jun 2 12:12:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/templates/gnu.mpd: - Added TAO_IDL3_TO_IDL2 and TAO_IDL3_TO_IDL2_DEP as first step to - run idl3_to_idl2 in a cross build environment - -Fri Jun 2 11:51:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_vxworks6.2.GNU: - Updated for VxWorks 6.2 - -Fri Jun 2 11:33:36 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Recursive_Mutex_Test.cpp (test_recursion_depth): The - get_nesting_level() only works on certain OS platforms, so don't - use it as part of the test. - -Fri Jun 2 10:13:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Configuration.cpp: - Fixed operator== - - * tests/Config_Test.{h,cpp}: - Added an explicit test for operator==. Thanks to Popeye Cai - <popeye dot cai at utstar dot com> for reporting this. This - fixes bugzilla bug 2550 - -Fri Jun 2 10:02:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_Thread.cpp: - Improved VxWorks 6.2 support - - * ace/Svc_Handler.cpp: - 64bit change and initialise pointer with 0 - - * ace/Process.cpp: - Added support for VxWorks 6.x in rtp mode - -Fri Jun 2 09:42:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_Thread.inl: - Added necessary code for VxWorks 6.2 - -Thu Jun 1 23:13:58 UTC 2006 Yan Dai <dai_y@ociweb.com> - - * bin/tao_orb_tests.lst: - - Added oneway and twoway DII_Collocation_Tests. - -Thu Jun 1 20:53:31 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * tests/Recursive_Mutex_Test.cpp (run_main): - - Added ACE_UNUSED_ARG macros to eliminate 'unused parameter' - warnings on single threaded builds. - -Thu Jun 1 16:54:56 UTC 2006 Yan Dai <dai_y@ociweb.com> - - * bin/tao_orb_tests.lst: - - Added new DII_Collocation_Tests. - -Thu Jun 1 13:59:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/FILE.cpp: - Const improvement - -Thu Jun 1 12:15:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Dirent_Test.cpp: - Updated VxWorks specific checks, removed chorus checks - -Thu Jun 1 11:03:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/Threads/thread_pool.cpp: - Fixed 64bit conversion warnings - -Thu Jun 1 09:39:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/diff-builds.pl: - Corrected usage of strftime - -Thu Jun 1 09:19:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Default_Constants.h: - Define the platform suffix for vxworks - - * ace/Sock_Connect.cpp: - Added include for VxWorks and removed some unneeded checks - for vxworks >= 0x600 - - * ace/ACE.h: - Doxygen improvement - - * ace/config-vxworks6.2.h: - Updated with more settings to better match vxworks 6.2 - - * ace/Process_Manager.cpp: - Added specific check for vxworks - - * ace/OS_NS_stdlib.inl: - Emulate putenv with setenv when putenv is not available but - setenv is. - -Thu Jun 1 02:26:03 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Service_Gestalt.h: - * ace/Service_Gestalt.cpp: - - Reworked the way static svc descriptors are kept so that if DLLs - include static initializers, they won't cause a problem if the - DLL is later unloaded and reloaded. There is currently no - dependency replationship between these "static" services and the - DLL that contains them, so there is still a slight risk that - pointers held after the DLL is unloaded might be referenced. But - the only thing that actually uses these pointers is the service - gestalt itself if it is processing an explicit static directive - and the desired service object had not yet been loaded in that - configuration context. - -Wed May 31 20:49:03 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * examples/APG/Logging/Trace.h: - - Added '##' to several macro definitions in this file, at the - suggestion of Steve Huston <shuston@riverace.com> to get - rid of warnings on GCC4 platforms. - -Wed May 31 18:18:37 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * bin/MakeProjectCreator/config/ciao_config_handlers.mpb: - - Shortened some line lengths, to make the file more readable. - -Wed May 31 16:36:43 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/tests.mpc: - * tests/run_test.lst: Don't build or run Unload_libACE on WinCE. The - code is purposely leaving out ACE stuff and the ACE stuff is needed - for WinCE, unless there's lots of conditional code added to duplicate - the workings on ACE_TMAIN on CE. - -Wed May 31 14:33:16 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * ace/OS_NS_Thread.inl (recursive_mutex_lock): - - Added additional ACE_UNUSED_ARG for the timeout parameter, to - eliminate warnings in single-threaded builds. - -Wed May 31 13:05:05 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Argv_Type_Converter.cpp (align_char_with_wchar): ACE-ified this code. - -Wed May 31 14:00:16 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * tests/Recursive_Mutex_Test.cpp (run_main): - - Moved '#if defined (ACE_HAS_THREADS)' up a few lines to include - a variable defined inside the same guard at the top of the file, - to eliminate an 'undefined' error on single-threaded builds. - -Wed May 31 13:37:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added Bug_2542_Regression, this should run without problems - -Wed May 31 11:44:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_time.inl: - Fixed unreachable code warning - -Wed May 31 09:23:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/run_test.lst: - Changes for ACE_FOR_TAO configuration - -Wed May 31 08:19:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Bound_Ptr_Test.cpp: - Initalise pointers with 0 - -Wed May 31 08:16:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Recursive_Mutex_Test.cpp: - Added missing include to get ACE_OS::sleep - -Tue May 30 21:56:52 UTC 2006 Steve Huston <shuston@riverace.com> - - * ACE-INSTALL.html: Corrected the AIX section; removed descriptions of - old compilers that are no longer supported and suggested the config - and platform_macros files to use. Thanks to Michelle Zheng - <Michelle dot Zheng at dowjones dot com> for motivating this. - - * THANKS: Added Michelle Zheng to the Hall of Fame. - -Tue May 30 19:17:40 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Dynamic_Service.h: - * ace/Dynamic_Service.inl: - * ace/Dynamic_Service.cpp: - * ace/Dynamic_Service_Base.h: - * ace/Dynamic_Service_Base.cpp: - * ace/Dynamic_Service_Dependency.cpp: - - In order to ensure proper separation of contexts, this patch - gives the dynamic service instance lookup the ability to - optionally exclude the global configuration context. - - * ace/Service_Gestalt.h: - * ace/Service_Gestalt.cpp: - - To allow explicitly configured static service objects, those for - which a directive such as 'static XYZ "foo bar"', to be - independently configured in different configuration contexts, - these changes retain any static service descriptor for service - XYZ, then looks for a previous instance of XYZ in the local - context only, otherwise it uses the previously retained static - service descriptor to create a new instance. - - -Tue May 30 18:24:54 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/DLL.h (ACE_DLL): Updated the documentation of open() to explain the - lookup scheme when a filename is given. Thanks to Phlip - <phlip2005 at gmail dot com> for motivating this. - -Tue May 30 13:06:42 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/ACE.inl: - * ace/Base_Thread_Adapter.cpp: - * ace/Default_Constants.h: - * ace/Global_Macros.h: - * ace/OS.h: - * ace/OS_NS_Thread.cpp: - * ace/OS_NS_Thread.h: - * ace/OS_NS_netdb.inl: - * ace/OS_NS_signal.inl: - * ace/OS_NS_sys_socket.inl: - * ace/OS_NS_time.inl: - * ace/OS_main.cpp: - * ace/OS_main.h: - * ace/Object_Manager_Base.h: - * ace/Process_Manager.cpp: - * ace/Process_Semaphore.cpp: - * ace/Process_Semaphore.h: - * ace/Process_Semaphore.inl: - * ace/Recursive_Thread_Mutex.cpp: - * ace/SV_Semaphore_Simple.h: - * ace/TLI_Acceptor.cpp: - * ace/Thread_Adapter.cpp: - * ace/Token.h: - * ace/config-macros.h: Removed all vestiges of PSOS. - -Tue May 30 12:22:53 UTC 2006 Chad Elliott <elliott_c@ociweb.com> - - * bin/mpc.pl: - * bin/mwc.pl: - - Added a load-time check for the minimum perl version (5.6). - -Tue May 30 12:18:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Log_Msg.cpp: - * ace/OS_NS_dirent.inl: - * ace/OS_NS_netdb.inl: - * ace/OS_NS_pwd.inl: - * ace/OS_NS_stdlib.inl: - * ace/OS_NS_Thread.cpp: - * ace/OS_NS_Thread.inl: - * ace/OS_NS_time.cpp: - * ace/OS_NS_time.inl: - * ace/Thread_Manager.cpp: - * ace/os_include/os_pthread.h: - * ace/os_include/os_stdlib.h: - Removed support for HPUX 10.20 - - * ace/os_include/net/os_if.h: - * ace/os_include/netinet/os_tcp.h: - Removed psos support - - * ace/os_include/sys/os_types.h: - Removed ACE_LACKS_FLOATING_POINT - -Tue May 30 11:57:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-hpux-10.x-g++.h - * ace/config-hpux-10.x-hpc++.h - * ace/config-hpux-10.x.h - Removed these files, not supported anymore - - * ace/OS_NS_time.inl: - Removed check for ACE_CTIME_R_RETURNS_INT, it was for hpux 10.20 - -Tue May 30 11:57:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-osf1-4.0.h: - Removed, contents merged to tru64 file - - * ace/config-tru64.h: - Added osf1-4.0 file contents, be aware that tru64 support hasn't - been tested for a long time. - -Tue May 30 11:45:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Recursive_Mutex_Test.cpp: - Fixed compile error - -Tue May 30 11:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/os_stdio.h: - Removed psos support - - * ace/IOStream.h: - Removed checks for ACE_LACKS_CHAR_STAR_RIGHT_SHIFTS, that was only - set for chorus which is removed - - * ace/Basic_Types.h: - * ace/Log_Msg.cpp: - * ace/Profile_Timer.{h,cpp}: - Removed checks for ACE_LACKS_FLOATING_POINT. Was only set for chorus - which is now removed. - - * ace/OS_NS_Thread.cpp: - Removed chorus support - - * ace/config-chorus.h: - Removed. - - * ace/INET_Addr.cpp - * ace/Lib_Find.cpp - * ace/Local_Tokens.cpp - * ace/Log_Record.cpp - * ace/Parse_Node.cpp - Removed check for ACE_HAS_BROKEN_CONDITIONAL_STRING_CASTS, is only - set for psos which we don't support anymore - - * ace/os_include/os_signal.h: - Removed check for ACE_SIGINFO_IS_SIGINFO_T, only set for psos - - * ace/config-psos-diab-mips.h - * ace/config-psos-diab-ppc.h - * ace/config-psos-diab.h - * ace/config-psos-tm.h - * ace/config-psosim-g++.h - Removed from the repo, psos support has been removed - -Tue May 30 10:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/CDR_Stream.cpp: - Use true/false, when demarshaling a string, check the result of the - demarshaling of the length of the string, if that failed, just return - directly instead of trying to demarshal a random lenght string - - * ace/Task.h: - * ace/System_Time.h: - Doxygen improvement - - * ace/ACE.cpp: - * ace/SOCK_Dgram_Bcast.cpp: - * ace/Sock_Connect.cpp: - * ace/Process.cpp: - * ace/OS_NS_unistd.cpp: - * ace/Mem_Map.cpp: - Removed chorus support - - * ace/config-fsu-pthread.h: - Removed this file, this configuration isn't supported and used for - ages - - * ace/Condition_Thread_Mutex.cpp: - * ace/Recursive_Thread_Mutex.cpp: - Removed fsu pthread support - - * ace/Process_Manager.cpp: - Const improvement - - * ace/OS.inl: - * ace/OS_NS_sys_socket.inl: - * ace/OS_NS_stdlib.cpp: - * ace/OS_NS_stdlib.inl: - * ace/OS_NS_Thread.cpp: - Removed psos support - - * ace/OS_NS_unistd.inl: - * ace/OS_NS_stdio.{h,cpp,inl}: - * ace/Handle_Set.h: - * ace/OS_NS_Thread.{h,inl}: - Removed psos and chorus support - - * ace/Thread_Manager.{h,cpp,inl}: - Made terminated_ a real bool, removed chorus support, check for - ACE_VXWORKS instead of VXWORKS - - * ace/OS_NS_sys_time.inl: - Check for ACE_VXWORKS instead of VXWORKS - - * ace/MMAP_Memory_Pool.{h,cpp}: - Removed chorus support, use bool internally and added an option to - get each time an unique mmap filename - - * ace/config-linux-common.h: - Added ACE_HAS_VOIDPTR_MMAP and removed KCC support - - * ace/ace_wchar.inl - * ace/Default_Constants.h - * ace/INET_Addr.h - * ace/IPC_SAP.cpp - * ace/Message_Queue.cpp - * ace/Message_Queue.h - * ace/Message_Queue.inl - * ace/Message_Queue_T.cpp - * ace/Message_Queue_T.h - * ace/OS.h - * ace/OS_NS_macros.h - * ace/OS_NS_stdio.h - * ace/OS_NS_stdio.inl - * ace/OS_NS_stdlib.inl - * ace/OS_NS_Thread.cpp - * ace/OS_NS_Thread.h - * ace/OS_NS_Thread.inl - * ace/OS_NS_time.inl - * ace/Recursive_Thread_Mutex.cpp - * ace/Sched_Params.cpp - * ace/Select_Reactor_T.cpp - * ace/SOCK_Dgram_Bcast.cpp - * ace/Thread_Manager.h - * ace/Token.h - Check for ACE_VXWORKS instead of VXWORKS - -Mon May 29 22:08:50 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/OS_NS_Thread.inl (recursive_mutex_lock): *Really* make the compiler - happy now. Thanks Olli! - -Mon May 29 21:12:50 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/OS_NS_Thread.inl (recursive_mutex_lock): Added a const_cast - for ACE_Time_Value * to make the compiler happy when passing the - timeout parameter to ACE_OS::cond_timed_wait(). Thanks to Olli - Savia for helping here. - -Mon May 29 20:30:44 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * We recently added the 2100th contributor to ACE+TAO+CIAO!!! The - 2000th contributor was added in August of 2005, this means that - over 100 different people have contributed to the project in the - past 9 months! - -Mon May 29 19:50:54 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * examples/Service_Configurator/IPC-tests/server/server_test.cpp (ACE_TMAIN): - Changed - - ACE_Reactor::instance ()->run_reactor_event_loop () - - to - - ACE_Reactor::run_event_loop () - - so that reconfiguration happens properly after a SIGHUP or - SIGINT occurs. Thanks to Kevin Hu <Xiaotao dot Hu at ge dot - com> for motivating this. - -Mon May 29 18:10:50 UTC 2006 Steve Huston <shuston@riverace.com> - - * tests/Signal_Test.cpp: - * tests/Thread_Manager_Test.cpp: Add missing ACE_TEXT to a string. - -Mon May 29 18:44:05 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Recursive_Mutex_Test.cpp: Enhanced this test to exercise - the new timed feature of ACE_Recursive_Thread_Mutex. - -Mon May 29 18:16:31 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Recursive_Thread_Mutex.h: Clarified that get_nesting_level() - may return -1 on some platforms when ACE_HAS_RECURSIVE_MUTEXES - is enabled. - -Mon May 29 17:45:02 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Recursive_Thread_Mutex.inl: Moved some of the short accessor - methods from the Recursive_Thread_Mutex.cpp file to here. - - * ace/Recursive_Thread_Mutex.h, - ace/Recursive_Thread_Mutex.inl: Added support for time recursive - mutexes, thanks also to Abdullah. - - * ace/OS_NS_Thread.h, - * ace/OS_NS_Thread.inl: Added support for timed recursive mutexes. - Thanks to Abdullah Sowayan <abdullah dot sowayan at lmco dot - com> for contributing this. - -Mon May 29 17:30:52 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/OS_NS_Thread.h: Updated the documentation to clarify why - recursive mutexes can't be used for ACE_OS::mutex_lock(). - -Mon May 29 18:27:08 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * examples/DLL/test_dll.cpp: - - Changed instances of a C-style cast from void* to - pointer-to-function to a two-step reinterpret_cast - using ptrdiff_t as an intermediate. C++ forbids casting - directly between pointer-to-object and pointer-to-function, - and some of the newer compilers are catching it. - -Sat May 27 15:03:07 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * netsvcs/servers/main.cpp (ACE_TMAIN): Removed the following code: - - if (ACE::debug () == 0) - ACE_LOG_MSG->priority_mask (~LM_DEBUG, ACE_Log_Msg::PROCESS); - - since (1) it makes debugging hard and (2) it's simply wrong - since it's blowing away anything that's set by the - ACE_Logging_Strategy in the svc.conf file. - -Sat May 27 14:37:14 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Log_Record.cpp (print): Only print out the log records if the - corresponding type is enabled for logging. - -Sat May 27 14:35:21 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * netsvcs/servers/main.cpp: Enable the ignore_debug_flag in the - ACE_Service_Config::open() call so that we can override the - default ACE_Log_Priority settings in the svc.conf file. - -Sat May 27 11:36:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * performance-tests/SCTP/hist.cpp: - Fixed compile error - -Fri May 26 22:02:48 UTC 2006 Steve Huston <shuston@riverace.com> - - * examples/IPC_SAP/SPIPE_SAP/consumer_msg.cpp: - * examples/IPC_SAP/SPIPE_SAP/consumer_read.cpp: - * examples/IPC_SAP/SPIPE_SAP/server.cpp: Wide-char fixes; there's no - wide-char variant for ACE_OS::fdetach(). - - * examples/IPC_SAP/TLI_SAP/CPP-client.cpp: - * examples/IPC_SAP/TLI_SAP/db-client.cpp: - * examples/IPC_SAP/TLI_SAP/ftp-client.cpp: - * examples/Connection/non_blocking/test_tli_acceptor.cpp: - * examples/Connection/non_blocking/test_tli_connector.cpp: - Wide-char/narrow-char fixes. - -Fri May 26 21:47:56 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * netsvcs/lib/Client_Logging_Handler.cpp, - * netsvcs/lib/Server_Logging_Handler_T.cpp, - * ace/Log_Msg_IPC.cpp: Removed the special case code for Windows - since it seems like ACE (or Windows) now handles this properly. - -Fri May 26 21:20:35 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Log_Record.cpp (operator>>): Add one when allocating the buffer. - -Fri May 26 13:16:48 UTC 2006 Nanbor Wang <nanbor@exothermic.txcorp.com> - - * ace/DLL_Manager.cpp: - - Fixed incorrect usage of ACE::debug check. Thanks to Johnny for - reporting this. - -Fri May 26 15:04:55 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * ASNMP/tests/Counter_Test.cpp: - * ASNMP/tests/Gauge_Test.cpp: - * performance-tests/SCTP/hist.cpp: - - Added static_cast<>s to eliminate implicit conversion warnings. - -Fri May 26 12:42:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Naming_Test.cpp: - Replaced assert with if statement, resolved argument not used - warning in BCB6 release builds - -Fri May 26 10:26:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Bug_1890_Regression_Test.cpp: - * tests/Bug_2540_Regression_Test.cpp: - New tests made by Carlos O'Ryan for the given bug numbers. - - * tests/run_test.lst: - * tests/tests.mpc: - Added the new tests above - -Fri May 26 09:53:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/PerlACE/Run_Test.pm: - Increased wait_interval_for_process_creation from 5 to 10 seconds. - On the slower systems some tests failed because the startup time - is longer then 5 seconds, especially if several configuration files - are read and saved during startup - -Fri May 26 08:21:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/diff-builds.pl: - Fixed typo in comment - -Thu May 25 04:58:44 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * tests/Tokens_Test.cpp (run_main): Don't try to spawn the - netsvcs/server/main if it's not built. - -Thu May 25 03:53:47 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * m4/config_h.m4: - - Removed AH_TEMPLATE for ACE_LACKS_TCP_H. This was replaced - with ACE_LACKS_NETINET_TCP_H long ago. - -Wed May 24 22:19:43 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Repository.{h, inl, cpp}: - - Changed size arguments type to size_t in ctor, open, - current_size, total_size and the corresponding members. - ACE_Service_Repository_Iterator::next_ now starts from 0 instead - of -1 to avoid warnings about signed/unsigned comparisons. - - * ace/Service_Gestalt.cpp: - - Reverted the change "Wed May 24 14:00:43 UTC 2006 Iliyan - Jeliazkov <iliyan@ociweb.com>", since now the - Service_repository::instance() takes size_t argument. - - * performance-tests/Synch-Benchmarks/Base_Test/Synch_Benchmarks_Base_Test.mpc: - * performance-tests/Synch-Benchmarks/Perf_Test/Synch_Benchmarks_Perf_Test.mpc: - - Fixed and error whereby the DLL's generated by this two project - files did not match the name used in the svc configuration - files, causing the test to crash because of missing DLL. - -Wed May 24 13:08:08 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Message_Queue.h: Added #include "ace/Thread_Mutex.h" on platforms - that have ACE_Message_Queue_NT available; that class has a - ACE_Thread_Mutex member so the include is needed to see the class. - Thanks to Guy Peleg <guype at amdocs dot com> for this fix. - -Wed May 24 17:39:47 UTC 2006 Douglas C. Schmidt <schmidt@dre.vanderbilt.edu> - - * ace/Remote_Tokens.cpp (request_reply): Added a helpful error - message to make it easy to track down problems with TSS. - -Wed May 24 17:22:31 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * configure.ac: - - Change feature test to define ACE_LACKS_NATIVE_STRPTIME if - system doesn't support strptime(), as it seems ACE doesn't - follow its feature test macro naming conventions when it comes - to strptime(). - - Changed ACE_LACKS_STRPTIME_PROTOTYPE feature test to use - AC_CHECK_DECL instead of AC_EGREP_HEADER. - -Wed May 24 16:33:59 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * configure.ac: - * m4/config_h.m4: - - Changed ACE_LACKS_STRPTIME and ACE_LACKS_STRPTIME_PROTOTYPE - feature tests to use the same idiom as similar tests. This - Avoids requiring an AH_TEMPLATE for each feature test macro - in config_h.m4. - -Wed May 24 09:07:53 2006 Douglas C. Schmidt <schmidt@cse.wustl.edu> - - * ace/Activation_Queue.cpp, - * ace/Activation_Queue.inl: Fixed the queue() "setter" method - so that it doesn't leak memory. Thanks to David Chu <beyonddc - dot storage at gmail dot com> for providing this fix. - -Wed May 24 14:00:43 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.cpp: - - Added a cast from size_t to int in the ctor to match the - interface of ::instance() and to quiet the 64bit compilers, - complaining about loss of precision in the implicit conversion. - -Tue May 23 19:21:55 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * tests/Signal_Test.cpp: - - Reset SIGHUP handler to SIG_DFL to support cases where SIGHUP - is ignored (running under nohup(1), etc.). - -Tue May 23 10:55:59 2006 Douglas C. Schmidt <schmidt@cse.wustl.edu> - - * netsvcs/lib/Client_Logging_Handler.cpp (handle_input): - * netsvcs/lib/Server_Logging_Handler_T.cpp - (handle_logging_record): Cleaned up the code so that it should - have no memory leaks and will also build cleanly on all the OS - platforms. - - * ace/Log_Msg_IPC.cpp: Fix support for ACE_HAS_STREAM_PIPES. - -Tue May 23 13:05:03 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Log_Msg.cpp: - Fixing a memory leak found by valgrind. - - * ace/Timeprobe.cpp: - Added missing include - this wouldn't comple with both noinline - and ACE_COMPILE_TIMEPROPE set. - -Mon May 22 21:13:15 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * bin/diff-builds.pl: - - Updated to allow diff between the tests failures today and on a - specified date, for the clean builds (or, all the builds from - the build score). - -Mon May 22 19:06:41 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Log_Record.cpp (operator<<, operator>>): Have to use a different - CDR array method when ACE_USES_WCHAR. - -Mon May 22 09:04:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * netsvcs/lib/Client_Logging_Handler.cpp: - Fixed compile error on Solaris - -Mon May 22 04:53:29 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * tests/Signal_Test.cpp: - - The change: - Sat May 20 15:10:55 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - Was not enough to get this test working on all platforms. - - Revert to using ACE_Sig_Guard in worker_child(). Handle the - SIGINT race condition by raising it if and only if the child - was not shut down by the parent. - -Sun May 21 13:46:49 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * tests/Service_Config_Test.cpp: - - Fixing Windows test failures, because errno gets overriden - between Service Repository::insert() and testLimits(). Most - likely, it happens in ace_yyparse (), however the existing - mechanism for reporting parser errors does not allow for - reliably propagating the cause of the error. Removing the test - for the cause of the error and simply reporting the error, if it - had occured. - -Sat May 20 15:10:55 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * tests/Signal_Test.cpp: - - Changed synchronous tests to block signals at the top level. - This had been done with a ACE_Sig_Guard, but that lead to a race - condition between when the guard was release and the delivery of - the SIGINT the client sends to itself. The SIGINT would then be - handled by the default handler, which terminated the process - before the "Ending" message was written to the log. - - #if'd out the ACE_DEBUG, etc. invocations from handle_signal(). - These result in signal-unsafe functions being called when the - handler is called asynchronously. A braver person would have - just removed them entirely. - -Fri May 19 20:08:29 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-sunos5.6.h: - * ace/os_include/sys/os_mman.h: Moved the declaration of ::madvise() - from config-sunos5.6.h to os_mman.h, to be compiled only when - ACE_LACKS_MADVISE_PROTOTYPE is defined. Set - ACE_LACKS_MADVISE_PROTOTYPE in config-sunos5.6.h where it used to - declare madvise(). This allows autoconf-detected - ACE_LACKS_MADVISE_PROTOTYPE platforms (such as Solaris 9) to - compile correctly. - -Fri May 19 17:53:51 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.cpp: - - Fixed a defect in process_directive () which was replacing the - ENOSPC errno with EINVAL. Added #ifndef ACE_NLOGGING / #endif - around logging statements. - - * ace/Service_Repository.cpp: - - Minor reformating. - - * tests/Service_Config_DLL.cpp: - - Added two independent service objects, to be used in the - testLimits test. See testLimits(). - - * tests/Service_Config_Test.cpp: - - Reworked the testLimits() to avoid printing the word "error", - which confuses the test driver into thinking the test failed. - -Fri May 19 17:51:33 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Service_Gestalt.cpp: - - Elminated memory leak found by valgrind. - -Fri May 19 17:09:51 UTC 2006 Jeff Parsons <j.parsons@vanderbilt.edu> - - * tests/Service_Config_Test.cpp: - - Fixed a warning due to assignment of a string literal to a - non-const variable. - -Thu May 18 17:42:09 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * bin/diff-builds.pl: - - Contributing a little tool to find the difference in the failing - tests between two dates, on the same build. Or between two - builds, on a particular date. Thanks to Johnny Willemsen - <jwillemsen@remedy.nl> for mining and hosting the data used by - this script. - -Thu May 18 17:33:57 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * configure.ac: - * m4/config_h.m4: - - Changed feature test for socklen_t to #include <sys/types.h> - before <sys/socket.h>. Use cannonical AC_CHECK_TYPE and - AC_DEFINE with a third argument instead of a AH_TEMPLATE for - ACE_HAS_SOCKLEN_T. Fixes configuration on OpenBSD. Thanks to - Vladimir Panov <gbr at voidland dot org> for reporting this bug. - -Thu May 18 16:37:46 UTC 2006 Steve Huston <shuston@riverace.com> - - * configure.ac: Set a ACE_VERSION_NAME variable with the version - string (canonical major.minor.beta), which may include a fix kit id. - - * bin/MakeProjectCreator/config/acedefaults.mpb: Change am_version from - libtool's -version-info format (current:revision:age) to the ACE - release number format (major.minor.beta) and rename it am_release. - This is intended for use with the libtool's -release option and will - result in library names that correspond to the ACE kit number. - The -version-info option uses interface numbers meant to indicate a - series of public API numbers. However, with ACE, things change - pretty often. Rather than try to come up with a way of managing - interface numbers, just use the release numbering that users know - and are used to. - -Thu May 18 15:15:16 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * bin/tao_orb_tests.lst: - - Removed Limits from the list of ORB tests because it is now part - of ACE. - - * tests/Service_Config_Test.cpp: - - Added functionality to test gestalt size limits. - -Thu May 18 15:21:33 UTC 2006 Steve Huston <shuston@riverace.com> - - * NEWS: - * ace/Task.{h inl cpp}: Reverted the changes from: - Fri Apr 14 15:53:21 UTC 2006 Steve Huston <shuston@riverace.com> - Tasks that delete themselves from within close() would invalidate - the lock that ACE_Task held across the close() call and perform - undefined and illegal actions on the deleted lock upon return. - Bugzilla #2339 remains unresolved. - -Wed May 17 20:31:11 UTC 2006 Roopa Pundaleeka <roopa@txcorp.com> - - * ace/DLL_Manager.cpp: - * ace/DLL_Manager.h: - The dll handle is added to the to the - handle_vector_ only after the library is opened - successfully. This way, the vector does not have redundant - handles. - - * ace/Object_Manager.cpp: - ACE_DLL_Manager::close_singleton is called explicitly from the - ACE_Object_Manager to make sure ACE_DLL_Manager gets cleaned up, - even if there was an error while opening a lib file. - - Thanks to Timo Kuosmanen <ttkuosma at cc dot jyu dot fi> for - reporting this. - -Wed May 17 14:07:33 2006 Douglas C. Schmidt <schmidt@cse.wustl.edu> - - * netsvcs/lib/Server_Logging_Handler_T.cpp: Improved the code so - that it should compile and run properly on Windows. Thanks to - Phil Mesnier for reporting this. - -Wed May 17 19:07:43 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Dynamic_Service_Dependency.h: Doxygen correction; use @c to - set a word in "typewriter" font, but for known class names, there's - no need to do this. - -Wed May 17 17:56:49 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * examples/C++NPv1/Logging_Client.cpp - * examples/C++NPv1/Logging_Handler.cpp - * examples/C++NPv2/Logging_Handler.cpp - - Removed now-redundant CDR insertion/extration operators now - defined in ace/Log_Record.{h,cpp} - -Wed May 17 09:35:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_ipc.h: - * ace/os_include/sys/os_socket.h: - * ace/os_include/sys/os_stat.h: - * ace/os_include/sys/os_types.h: - Removed psos support - - * ace/os_include/sys/os_mman.h: - Removed madvice prototype and ACE_LACKS_MADVISE_PROTOTYPE, it - is not set in any config file - -Wed May 17 02:23:31 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Log_Msg_IPC.cpp: - - Fix typo exposed on systems with ACE_HAS_STREAM_PIPES defined. - I'll leave it to Will & Doug to do whatever the TO DO said to do. - -Tue May 16 22:23:55 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - * ace/Log_Msg_IPC.cpp - * ace/Log_Record.cpp - * ace/Log_Record.h - * ace/Log_Record.inl - * netsvcs/lib/Client_Logging_Handler.cpp - * netsvcs/lib/Server_Logging_Handler_T.cpp - - Fixed some problems that stemmed from changing ACE_Log_Msg - msg_data_ from a fixed-size array to a dynamically-sized array. - Thanks to Keith Muzzioli <Keith_W_Muzzioli at raytheon dot com>, - John Black <John_K_Black at Raytheon dot com>, and Mitsuo - Fukasawa <fukasawa_mitsuo at nifty dot com> for reporting this - and to Doug Schmidt for helping to fix it! - -Tue May 16 13:20:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_time.inl: - * ace/OS_NS_time.{h,cpp,inl}: - Removed chorus and psos support - - * ace/OS_Thread_Adapter.cpp: - Removed psos support - -Tue May 16 13:09:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_wait.{h,inl}: - Remove chorus support - - * ace/OS_NS_stropts.inl: - * ace/OS_NS_sys_stat.inl: - * ace/OS_NS_string.inl: - Removed psos support - -Tue May 16 13:02:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Log_Msg.cpp: - * ace/OS_NS_arpa_inet.{h,cpp,inl}: - * ace/OS_NS_dirent.inl: - Removed psos support - - * ace/INET_Addr.cpp: - * ace/config-macros.h: - Removed chorus support - - * ace/Unbounded_Queue.h: - Removed remark about chorus - - * ace/config-cray.h: - Removed explicit template instantiation define, we don't have that - anymore - - * ace/config-hpux-11.00.h: - * ace/config-sunos5.7.h: - * ace/config-rtems.h: - * ace/config-osf1-4.0.h: - Removed kcc support - -Mon May 15 11:30:14 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_other_tests.lst: - Added more tests that can't run in ACE_FOR_TAO configations due - to dependency on something that avoids ACE_FOR_TAO. - -Mon May 15 01:15:45 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - * bin/tao_other_tests.lst: - Added !ACE_FOR_TAO to those tests for which MPC skips generating - executables because it directly or indirectly avoids ACE_FOR_TAO. - Indirect avoidance occurs when a project derives from a base - project that avoids ACE_FOR_TAO. - -Thu May 11 12:37:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Message_Block.h: - Doxygen improvements - -Thu May 11 11:15:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Dev_Poll_Reactor.cpp: - Removed chorus support - -Thu May 11 07:19:12 UTC 2006 Kees van Marle <kvmarle@remedy.nl> - - * bin/tao_orb_tests.lst: - Added Bug_1254_Regression - -Wed May 10 14:10:02 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-win32-msvc-7.h: MSVC 7.1 operator new throws an exception - on an out-of-memory condition, but only when the standard C++ - library is used; the older CRT new() doesn't. - -Tue May 9 11:38:30 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/orbsvcslib.mpb: - Add -GT to the idlflags, all orbsvcs then do generate the TIE files, - if people then use them, they get them out of the box. - -Tue May 9 07:36:30 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/MakeProjectCreator/config/taoidldefaults.mpb: - Removed -Sc, added -GT, see bugzilla bug 2525 for more info - -Tue May 9 07:25:30 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Condition_T.{h,cpp,inl}: - * ace/Local_Name_Space_T.cpp: - * ace/Configuration.cpp: - * ace/Sched_Params.cpp: - * ace/Mutex.{h,cpp,inl}: - * ace/Select_Reactor_T.cpp: - * ace/Signal.cpp: - Removed Chorus support - - * ace/CDR_Stream.cpp: - Small layout change - - * ace/Event_Handler.cpp: - Removed psos support - - * ace/Message_Block.h: - Doxygen improvements - - * ace/Message_Block.cpp: - Const improvement - -Sun May 7 21:03:30 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.{h,cpp}: - - Weeding out the remaining issues with single threaded - builds. The implementation now relies on the ACE_TSS_* macros to - abstract from the differences among build styles and platform - support for TSS. - - * ace/Service_Gestalt.{h,cpp}: - - Eliminated commented out code. Minor reformatting in places. - - * ace/TSS_T.{h,inl,cpp}: - - Added some comments to help others avoid a few pitfalls. Made - ts_init() a non-const, which eliminates a few const_cast's. - -Sun May 7 07:02:51 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.h: - * ace/Service_Config.cpp: - - Refactored the TSS usage (again) to fix a nasty - order-of-initialization problem. The gist of which is that, if - the ptr (to a template class instance) is a static member, in - static builds, it will be initialized by the process prologue - code *after* another static initializer has had the chance to - use and assign it a value. The solution was to use a method - scope static instance, which C++ guarantees to be initialized by - the (first) exit from that method. - - * ace/Service_Gestalt.h: - * ace/Service_Gestalt.cpp: - - Refactored the construction to avoid using straight new in - constructor initializer lists. Replaced new with ACE_NEW_* where - appropriate. - -Sat May 6 18:08:13 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * configure.ac: - - Add feature tests for ACE_LACKS_ASCTIME, ACE_LACKS_ACSTIME_R, - ACE_LACKS_GETCWD, ACE_LACKS_GMTIME, ACE_LACKS_GMTIME_R, - ACE_LACKS_LOCALTIME, and ACE_LACKS_STRFTIME. - -Sat May 6 09:52:06 2006 Douglas C. Schmidt <schmidt@cse.wustl.edu> - - * examples/DLL/test_dll.cpp: Renamed the DLLs and removed the "./" - from the pathname. Thanks to Sam and Roopa for reporting this - and tracking it down, respectively. - -Sat May 6 10:23:13 UTC 2006 Steve Huston <shuston@riverace.com> - - * include/makeinclude/platform_sunos5_sunc++.GNU: Add -lrt -laio to - the LIBS. The Proactor framework needs this; I'm not sure why it - never showed up until static builds were run. - -Tue Apr 25 10:22:16 2006 Douglas C. Schmidt <schmidt@cse.wustl.edu> - - * ace/OS_NS_string.cpp (strerror): Do a sanity check to make sure - errnum >= 0 and < _sys_nerr and set errno to EINVAL if things - are not sane. Thanks to Aleksandar Vukajlovic <vukajlo at - finsoft dot co dot yu> for this fix. - -Fri May 5 19:37:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Default_Constants.h: - Added a compile time check that ACE_DEFAULT_CDR_BUFSIZE must be - defined to a value bigger then 0, it is 0 we get in a never ending - loop in the CDR code - -Fri May 5 17:01:32 UTC 2006 Steve Huston <shuston@riverace.com> - - * examples/Service_Configurator/IPC-tests/client/local_spipe_client_test.cpp: - * examples/Service_Configurator/IPC-tests/client/remote_thr_stream_client_test.cpp: - * examples/Service_Configurator/IPC-tests/server/Handle_L_SPIPE.{cpp i h}: - Wide-char corrections. - -Fri May 5 13:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/FIFO.cpp: - * ace/os_include/os_stddef.h: - Removed psos support - - * ace/config-WinCE.h: - * ace/os_include/os_stddef.h: - Moved setting of ACE_LACKS_STDDEF_H and ACE_LACKS_PTRDIFF_T to the - WinCE file - -Fri May 5 13:08:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_openvms.GNU: - Changed the way templates are handled - -Fri May 5 11:42:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Reactor_Token_T.h: - Removed workaround for vc6 - -Fri May 5 08:38:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/MMAP_Memory_Pool.{h,cpp}: - Made guess_on_fault_ a real bool - -Fri May 5 08:09:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/os_dirent.h: - * ace/os_include/os_errno.h: - * ace/os_include/os_netdb.h: - * ace/os_include/os_stdint.h: - * ace/os_include/os_time.h: - * ace/os_include/os_string.h: - * ace/os_include/os_stdlib.h: - Removed psos support - - * ace/os_include/os_dlfcn.h: - * ace/os_include/os_signal.h: - Removed psos and kcc support - - * ace/os_include/os_pthread.h: - Removed chorus support - - * ace/os_include/os_unistd.h: - Removed chorus and psos support - -Fri May 5 07:59:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_utsname.cpp: - * ace/OS.h - Removed psos and chorus support - - * ace/OS_NS_fcntl.cpp: - * ace/Lib_Find.cpp: - * ace/OS_NS_netdb.inl: - Removed psos support - - * ace/Mem_Map.cpp: - Removed chorus support - - * ace/IOStream.h: - Removed checks for KCC compiler - -Fri May 5 07:42:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Base_Thread_Adapter.h: - * ace/SPIPE_Connector.cpp: - * ace/SOCK_Dgram.cpp: - * ace/Select_Reactor_T.cpp: - * ace/Handle_Set.cpp: - Removed code for psos, support for this platform will be removed - - * ace/CDR_Base.h: - Removed check for KCC compiler - - * ace/Malloc_Allocator.h: - Doxygen improvement - -Thu May 4 22:11:06 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Message_Block.h: Made ACE_Data_Block::release_i() virtual. - Thanks to Gary Fernandez <gary dot fernandez at ironmountain dot - com> for this change. - -Thu May 4 21:07:10 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/SPIPE_Acceptor.cpp (create_new_instance, close): Have to use - ACE_TEXT_ALWAYS_CHAR to pass pipe name on wide-char w/ - ACE_HAS_STREAM_PIPES. Thanks to Gary Fernandez <gary dot - fernandez at ironmountain dot com> for this fix and to Yasser - Zabuair <yzabuair at hotmail dot com> for also reporting it. This - fixes bugid 2541. - - * THANKS: Added Gary and Yasser to the Hall of Fame. - -Thu May 4 08:50:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - Reverted the change below, broke the sendfile test and Steve - Huston changed things in such a way that sendfile is only - used on Solaris 9 and newer - - Tue May 2 09:34:54 Pacific Daylight Time 2006 <ossama_othman at symantec dot com> - * ace/OS_NS_sys_sendfile.inl (sendfile): - Removed "::" global scope qualifier from platform sendfile() - call. It caused problem on Solaris 8. Thanks to Simon Massey - and Johnny Willemsen for analyzing the problem and suggesting - this fix. - -Thu May 4 08:36:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/ace_for_tao.mpc: - Added OS_NS_sys_sendfile - -Thu May 4 00:43:32 UTC 2006 J.T. Conklin <jtc@acorntoolworks.com> - - * ace/Makefile.am: - - Add Dynamic_Service_Dependency.{cpp,h} to appropriate lists. - -Wed May 3 16:02:10 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.h: - * ace/Service_Config.cpp: - - Fixed a problem due to an attempt to access current_ before it - has been initialized (a static ACE_TSS instance). The current_ - is now just a pointer to an ACE_TSS instance, which gets - dynamicaly allocated upon invoking the current() method. The - deallocation of that instance is left up to the Object Manager - with which the dynamic ACE_TSS instance is registered. - - * ace/TSS_T.h: - - Added a ACE_TSS_SET () to complete the set of TSS access - operations, which should work regardless of threading, TSS - emulation and TSS availability. - -Wed May 3 11:52:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Mem_Map.inl: - Small const improvements - -Wed May 3 10:39:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Mem_Map.{h,cpp,inl}: - Made close_handle_ a bool - -Wed May 3 08:56:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-macros.h: - Removed support for KCC - -Wed May 3 08:14:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_sendfile.h: - Export sendfile_emulation to resolve link problems in the - windows builds - -Wed May 3 03:36:23 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Svc_Conf_y.cpp: - - Moved the ACE_BEGIN_VERSIONED_NAMESPACE_DECL macro to include - ace_yyparse. - -Tue May 2 09:34:54 Pacific Daylight Time 2006 <ossama_othman at symantec dot com> - - * ace/OS_NS_sys_sendfile.inl (sendfile): - - Removed "::" global scope qualifier from platform sendfile() - call. It caused problem on Solaris 8. Thanks to Simon Massey - and Johnny Willemsen for analyzing the problem and suggesting - this fix. - -Tue May 2 10:34:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-WinCE.h: - Added ACE_LACKS_GETCWD, ACE_LACKS_ASCTIME, ACE_LACKS_ASCTIME_R, - ACE_LACKS_GMTIME, ACE_LACKS_GMTIME_R, ACE_LACKS_LOCALTIME and - ACE_LACKS_STRFTIME. These macros are new and will be used soon - in the ACE_OS layer. - -Mon May 1 19:49:45 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/ARGV.{h i cpp}: Filled out more of the documentation in this - class, specifically mentioning that the class will enforce the - "whitespace separates tokens, except when quoted" rule of command - line parsing that's consistent with ACE_Get_Opt, etc. Also removed - documentation of non-existant restrictions such as after using - add(), can't use buf(). - Removed the State enum since the only value ever used was ITERATIVE, - and changed "int state_" to "bool iterative_" (and removed the - state() accessor). Changed substitute_env_args_ from int to bool. - - * ace/OS_NS_unistd.{h cpp} (argv_to_string, string_to_argv): Changed - the substitute_env_args parameter from "int" to "bool" to better - reflect what it is, a yes/no indicator. - -Mon May 1 06:53:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_sunos5_g++.GNU: - * include/makeinclude/platform_sunos5_sunc++.GNU: - Also for solaris 10 link the sendfile lib - -Sun Apr 30 04:21:37 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Svc_Conf_y.cpp: - What I meant to commit last time. - -Sun Apr 30 02:21:16 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Svc_Conf_y.cpp: - Eliminating more warnings from the scoreboard. I'm not sure the - best way to add this patch to ace/svcconf.mpb, since it affects - more than just gnuace targets. The easiest would be to store the - diff between this commit and the previous version and run patch, - but I'm not sure how to integrate that solution. - -Sat Apr 29 13:58:28 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.cpp: - More fuzz cleanup. - -Sat Apr 29 13:17:24 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/ACE.cpp: - Fuzz cleanup. - - * THANKS: - Added credits. - -Fri Apr 28 22:31:52 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.cpp: - - Removed an ';' after ACE_TRACE, which was confusing GCC4 with - "hidden visibility" enabled. - -Fri Apr 28 21:22:53 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.cpp: - * ace/Service_Gestalt.cpp: - - More re-formatting to weed out those pesky tabs. - -Fri Apr 28 20:57:19 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-sunos5.8.h: - * ace/config-sunos5.9.h: Moved ACE_HAS_SENDFILE from Solaris 8 to - Solaris 9. Solaris 8 supports sendfilev[64] but not sendfile, - at least not without some patches I don't know about. - -Fri Apr 28 20:13:41 UTC 2006 Steve Huston <shuston@riverace.com> - - * include/makeinclude/platform_sunos5_g++.GNU: - * include/makeinclude/platform_sunos5_sunc++.GNU: - Solaris 9 puts sendfile() is a separate library, so LIBS needs - -lsendfile on Solaris 9. - -Fri Apr 28 18:03:11 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/ACE.cpp: - * ace/DLL.cpp: - * ace/DLL_Manager.cpp: - * ace/Dynamic_Service_Base.cpp: - * ace/Dynamic_Service_Dependency.cpp: - * ace/Parse_Node.cpp: - * ace/Service_Config.inl: - * ace/Service_Config.cpp: - * ace/Service_Gestalt.cpp: - * ace/Service_Object.cpp: - * ace/Service_Repository.cpp: - * ace/Service_Types.cpp: - * ace/Shared_Object.cpp: - - Applied formatting style consistent with ACE convention. - -Fri Apr 28 15:48:25 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Service_Config.cpp (open_i): ACE_DEFAULT_SVC_CONF shouldn't - be enclosed in ACE_LIB_TEXT since it's defined in terms of - ACE_LIB_TEXT already. - -Fri Apr 28 15:49:21 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Makefile.am: - Added new Service_Gestalt files to the appropriate lists. - -Fri Apr 28 15:42:59 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/Service_Config.cpp: - Fixed a problem killing some wchar builds. - - * ace/Service_Gestalt.h: - Cleaned up the long-lines of the UNIMPLEMENTED functions. - -Fri Apr 28 15:26:53 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Timer_Queue_Adapters.h: Clarified the use and restrictions with - ACE_Async_Timer_Queue_Adapter. - -Fri Apr 28 15:22:35 UTC 2006 Steve Huston <shuston@riverace.com> - - * apps/JAWS2/JAWS/Concurrency.cpp: Removed unused variables. - -Fri Apr 28 14:54:04 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-macros.h: HP aC++ needs to see the statement in - ACE_NOTREACHED up to the version 6 series (Integrity). Also see - Tue Apr 11 13:12:36 UTC 2006 Simon McQueen <sm@prismtech.com> - -Fri Apr 28 14:31:48 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/Service_Gestalt.h: Removed extraneous ';' from - ACE_UNIMPLEMENTED_FUNC usage. - - * m4/ace.m4: Fixed incorrect variable referenced in case during - --enable-uses-wchar. Thanks to Paul Robinson <probinson at - gmail dot com> for reporting this. - -Fri Apr 28 14:03:50 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Dynamic_Service.h: - * ace/Parse_Node.h: - - Fixed the usage of ACE_UNIMPLEMENTED_FUNC to remove the trailing - semicolon. It was causing build errors on compilers that define - the macro as empty. - -Fri Apr 28 13:55:01 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.h: - - Fixed the usage of ACE_UNIMPLEMENTED_FUNC to remove the trailing - semicolon. It was causing build errors on compilers that define - the macro as empty. - -Fri Apr 28 13:29:08 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * bin/tao_orb_tests.lst: - - Updated the conditions for running the ORB-specific service - repository tests. - -Fri Apr 28 13:23:57 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Config.h: - - Updated the declaration of current_ to match the definition in - its use of ACE_TSS_TYPE macro, as they are not equivalent for - a single-threaded builds. - -Fri Apr 28 13:21:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Process_Mutex_Test.cpp: - When spawn fails log an error and exit instead of using an assert - which causes an abort of the exe - -Fri Apr 28 13:15:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Pipe_Test.cpp: - When spawn fails log an error and exit instead of using an assert - which causes an abort of the exe - -Fri Apr 28 12:49:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-openvms.h: - Added __SIGNED_INT_TIME_T so that we get a signed time_t that - matches the POSIX spec that ACE expects. - -Fri Apr 28 12:23:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-openvms.h: - Removed comments, macros are documented in generally, added - ACE_LACKS_PERFECT_MULTICAST_FILTERING - -Fri Apr 28 11:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/PerlACE/Process_VMS.pm: - New file to be able to run the tests automatically on OpenVMS - - * bin/PerlACE/Process.pm: - When running on OpenVMS use Process_VMS - - * ace/config-openvms.h: - Added ACE_HAS_SIGISMEMBER_BUG - - * tests/run_test.pl: - If ACE_ROOT is not set, add ../bin to lib - -Fri Apr 28 09:23:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Process_Manager_Test.cpp: - Also create a logfile for each child process started - -Fri Apr 28 08:06:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/run_test.lst: - Added Sendfile_Test - -Fri Apr 28 08:04:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Sendfile_Test.cpp: - Added missing include - -Thu Apr 27 21:04:08 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/OS_Memory.h: Allow configs to set ACE_bad_alloc to avoid - extending the if-elif chain here further. - - * ace/config-win32-msvc.h: Regardless of MSVC version, MFC changes - operator new to throw CMemoryException on out-of-memory condition. - Version-specific settings still control the setting of - ACE_HAS_NEW_NOTHROW. Thanks to Mohit Kapoor <Mohit dot Kapoor at - siemens dot com> for this fix. - - * ace/Select_Reactor_T.cpp (handle_error): HACK ALERT! Some Linux - kernels will return from select() with an "invalid" errno value, - ERESTARTNOHAND, when compiled with profiling enabled. This works - around the problem. This is a known kernel bug; see Red Hat Bugzilla - https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=161468 and - https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=123648 as well - as Riverace support issue #167 and Riverace's complaint to Red Hat, - support issue #17841808. - - * THANKS: Added Mohit Kapoor to the Hall of Fame. - -Thu Apr 27 20:14:52 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Service_Gestalt.cpp: - - Fixed a complaint from a ACE_NDEBUG builds about unrefrenced - variable (only used in ACE_ASSERT). - -Thu Apr 27 18:33:18 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Aio_Platform_Test.cpp: - When AIO is not available, just print an info message and exit - - * tests/run_test.lst: - Added Aio_Platform_Test, it was not in this file at all - -Thu Apr 27 18:21:17 UTC 2006 Steve Huston <shuston@riverace.com> - - * ace/config-aix-5.x.h: Added ACE_HAS_SIGSUSPEND. - -Thu Apr 27 14:57:10 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Svc_Conf.y: - * ace/Svc_Conf_Tokens.h: - * ace/Svc_Conf_y.cpp: - - Fixed fuzz complaint about #include inside a versioned namespace - in Svc_Conf_y.cpp. It was due to a code that Bison generates a - part of the parser, from Svc_Conf.y. Changed the versioned - namespace to encloses the non-generated code only, while the - rest of the parser seems to be designed so that multiple parses - can coexist without name clashes. - -Thu Apr 27 14:29:07 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Dynamic_Service_Dependency.h: - * ace/Dynamic_Service_Dependency.cpp: - - Updated to remove an empty inl file. - - * ace/Dynamic_Service_Dependency.inl: - - Removed this file. - -Thu Apr 27 14:00:18 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Sock_Connect.cpp: - Implemented get_ip_interfaces for OpenVMS - -Thu Apr 27 11:19:18 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Array_Map.h: - * ace/ATM_Addr.h: - * ace/Atomic_Op_T.h: - * ace/Base_Thread_Adapter.h: - * ace/Cached_Connect_Strategy_T.h: - * ace/Capabilities.h: - * ace/Default_Constants.h: - * ace/Hash_Map_Manager_T.h: - * ace/Log_Msg_Callback.h: - * ace/Log_Msg_IPC.h: - * ace/Map_Manager.h: - * ace/POSIX_Asynch_IO.h: - * ace/POSIX_Proactor.h: - * ace/RB_Tree.h: - * ace/Strategies_T.h: - * ace/Timer_Heap_T.h: - * ace/Timer_List_T.h: - * ace/Timer_Queue_Adapters.h: - * ace/Timer_Queue_T.h: - * ace/Timer_Wheel_T.h: - Fixed doxygen warnings - -Thu Apr 27 01:49:18 UTC 2006 Ossama Othman <ossama@dre.vanderbilt.edu> - - * ace/OS_NS_sys_sendfile.h (sendfile, sendfile_emulation): - * ace/OS_NS_sys_sendfile.cpp (sendfile_emulation): - - ACE_LOFF_T -> off_t. Expected ACE_LOFF_T behavior is still - under scrutiny. - - * ace/OS_NS_sys_sendfile.inl: - - Added missing <sys/sendfile.h> include directive. Addresses - infinite recursion at run-time when ACE_HAS_SENDFILE is defined. - - (sendfile): - - ACE_LOFF_T -> off_t, as above. - -Wed Apr 26 22:00:21 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - * ace/Parse_Node.h: - * ace/Service_Gestalt.h: - - Correcting a build problem, caused by the extraneous semicolons - after ACE_UNIMPLEMENTED_FUNC - -Wed Apr 26 20:21:49 UTC 2006 Iliyan Jeliazkov <iliyan@ociweb.com> - - The motivation for these changes was to enable support in ACE - for multiple instances of Service Configuration Repository, or - more appropriately - "Gestalt" (from the German word, meaning - something that's more than the sum of its parts). This is - feature is necessary to enable support for ORB-specific, or more - generally - service-specific sets of services, i.e. to be able - to have dynamically loaded services to use their own private set - of service objects. In the context of the implementation I will - use "gestalt" as a synonym for service configuration context or - service configuration repository. - - In order to accomplish this, the "instance"-related - functionality of the former ACE_Service_Config class has been - moved to another class - ACE_Service_Gestalt. The - ACE_Service_Config retains all static interfaces and those - members, pertaining to the concept of process-wide configuration - state. The service config, as known today has been retained as a - concept, but in its specialized semantics as holding a - process-wide configuration state. The primary concept expressing - an instance of configuration context is the gestalt. The - implementation actually uses a specialization of - ACE_Service_Gestalt in order to implement ACE_Service_Config as - a singleton, special case of configuration context. - - The static methods in ACE_Service_Config provide access to - process-wide gestalt (ubergestalt, anyone?) instance. - - For more details, here is the history of all included changes, - with their motivation and explanation. It is ordered - alphabetically, by the names of the changed files: - - * ace/ACE.cpp: - - Added to the condition evaluating ACE::debug() to include a - check if environmental variable ACE_DEBUG has been set. - - Updated the debug() function to call getenv() only once by using - a method-local static variable to hold the value from the - environment. - - * ace/DLL.h: - * ace/DLL.cpp: - - Implemented an operator=, needed in order to implement the - ACE_Service_Configuration_Guard class. - - * ace/DLL_Manager.cpp: - - Improved the error diagnostics so that a "file not found" would - not mask a "symbol not defined" error when loading a DLL. - - * ace/Dynamic_Service.h: - - "Privatized" copy ctor and operator= for ACE_Dynamic_Service - objects since it is not designed for such operations. - - * ace/Dynamic_Service.inl: - - Added overloaded method instance() for ACE_ANTI_TCHAR. - - * ace/Dynamic_Service.cpp: - - Changes to enable instantiation of a service, based on the - registration in a particular service repository. Adding a method - that takes ACE_Service_Gestalt* additional parameter. - - * ace/Dynamic_Service_Base.h: - * ace/Dynamic_Service_Base.cpp: - - Factored out a find_i() method to be used by the friend - ACE_Dynamic_Service_Dependency class. It implements the specific - lookup policy that extends the search to the global repository - if the named service object can not be found locally. - - Added instance method, allowing the caller to explicitly specify - the gestalt. - - * ace/Dynamic_Service_Dependency.h: - * ace/Dynamic_Service_Dependency.inl: - * ace/Dynamic_Service_Dependency.cpp: - - A newly introduced class, whose instances are designed to be - embedded inside instances of other classes, typically created by - DLL-based factories (service objects). This provides a way to - declare dependency on a specific service, thus avoiding order of - initialization issues with objects whose implementation code - resides in dynamically loaded services. - - It is disastrous to have dynamically loadable services create - and give away ownership of objects and then be unloaded before - all those instances have been deleted. Normally the code for - such objects classes resides within the TEXT segment of the DLL, - which implements the service. If a service gets removed, its DLL - may be unmapped from memory and then any attempt to invoke a - method on the said objects will cause SEGV. - - Such instances must contain a member of - ACE_Dynamic_Service_Dependency initialized with the service they - depend on. @code ACE_Dynamic_Service_Dependency's constructor - and destructor are "magical" - they work by maintaining the - underlying dynamic service's DLL reference count. - - The problem was not present before because the one Service - Repository outlived any other user objects. Now that an - un-loadable service can own its configuration, the sequence of - service finalization has been reversed. - - Updated the implementation of ACE_Dynamic_Service_Dependency - class to ease debugging. ACE_Dynamic_Service_Dependency gets - initialized whenever a dependent service gets loaded, to help us - keep our access to dependent instances after their DLL is gone - (by upping the ref count on the DLL). - - Updated usage of gestalt's instance() to current(). Removed - unnecessary comments. - - * ace/Global_Macros.h: - - Changed the ACE_STATIC_SVC_DEFINE to use the new - ACE_Service_Config::insert () method, which allows to streamline - the Service Configuration interface and promote hiding the - storage used for the static service descriptors. - - * ace/Parse_Node.h: - * ace/Parse_Node.cpp: - - Changes to aid in keeping track of the gestalt, while deep in - the guts of the svc.conf parser. - - Replaced the ACE_Service_Gestalt & in the method signatures with - ACE_Service_Gestalt * to unify the interfaces. - - Updated the static function node's logic to use only - ACE_Service_Gestalt public members and to no longer rely on - knowledge about the internal storage representation of service - repository. - - Removed a static_cast<> in ACE_Static_Function_Node::symbol() - that was causing problems on some platforms. It was not even - necessary as both sides were of the same type. - - * ace/Service_Config.{h,inl,cpp}: - - This separates the responsibilities between the service gestalt, - which represents an _instance_ of configuration information, and - the service config, which represents a special, process-wide, - global configuration repository. Since both these entities - represent the same concept - service configuration state, but - from different aspects, the ACE_Service_Config now inherits - (publicly) from ACE_Service_Gestalt. - - Lots of instance-related code moved to ACE_Service_Gestalt - Replaced the use of char with ACE_ANTI_TCHAR and the appropriate - ACE_TEXT_* macros. - - Moved the parsing of command-line options, pertaining to the - process as a whole, here - in the parse_args_i () (protected, - virtual). - - In order to resolve a problem caused by the inability of a - dependent static service to determine which gestalt they need to - register with, the concept of "current" was separated from the - "global" gestalt. The current gestalt is pointed to by a - pointer, stored in thread-specific storage (TSS) and manipulated - by the ACE_Service_Config_Guard class that implements the - "resource acquisition is initialization" idiom. - - Fixed an error in the separation of responsibilities between the - instance gestalt and the global (ubergestalt), during - initialization. The bug can cause infinite recursion, when - initializing an ORB, housed in a dynamic service. The singleton - gestalt (i.e. ACE_Service_Config) must be initialized through - open_i(), before open_i() can proceed initializing the specific - gestalt instance. The difficulty is in that some use cases - reverse the order of initialization by calling - ACE_Service_Config::open () first and then, at some point call - the instance gestalt open_i(). The solution is to use the - is_initialized_ member together with an explicit call to - ACE_Service_Config::global()->open_i(). To ease debugging of - the process of registering static service, I have changes the - ACE_Service_Config::static_svcs() to return - ACE_Service_Gestalt::instance (), instead. Thus all the - ACE_STATIC_SVC* macros are still working (no need to change - existing code), but now there is a convenient place to debug - - see the newly added ACE_Service_Gestalt::insert () method. The - header file no longer declares ACE_STATIC_SVCS, - ACE_STATIC_SVCS_ITERATOR, ACE_SVC_QUEUE and - ACE_SVC_QUEUE_ITERATOR. Those are now protected typedefs in - ACE_Service_Gestalt, where they are actually used. - - Added new insert () method to use instead of the sequence - static_svc ()->insert () which unnecessary exposes the internal - storage structure of Service Configurator. Made no_static_svcs_ - an instance member, instead of a class (static) member, thus - providing for ability to set it per instance, not globally. - - Added default values for open_i() and declared private copy ctor - and assignment operator for ACE_Service_Config_Guard, since it - is not intended to be copied. - - Similarly, moved the parts of the open_i() functionality that - did not belong in every instance of the gestalt. - ACE_Service_Config is again a process-wide Singleton,registered - with ACE_Object_Manager. - - Eliminated old #if 0-ed code in Service_Config.cpp. - - * ace/Service_Gestalt.h: - * ace/Service_Gestalt.inl: - * ace/Service_Gestalt.cpp: - - Encapsulates the idea of Service Configuration repository, or - configuration context. - - Changed open_i to be able to detect if it is working on the - special, process-wide instance (ubergestalt) or a "plain" - gestalt instance, since there are things that must only be done - once per process - like signal handler, logging key and - daemon-ization of the whole process. - - Added an ignore_static_svcs parameter to - ACE_Service_Gestalt::open_i in order to preserve the prior - behavior of the singleton ACE_Service_Gestalt instance. Even - though its no_static_svcs_ member was being initialized with 1, - some clients (TAO) needs to be able to override it. Since now a - call to open_i on an instance gestalt may precede in time the - call to open_i on the ubergestalt, we need a mechanism to allow - the no_static_svcs_ value to be overridden. - - Added the insert()'s implementation (described above). Added a - singleton() method to return the process-wide gestalt singleton - (or ubergestalt?:). Used in open_i() to account for the fact - that now a Service Gestalt may be called to open () prior to - open () on the singleton instance, which would typically contain - the bulk of the static service object descriptors. - - There was a problem with "hybrid" service objects, i.e. dynamic - SO which contains static SO (dependent services). The dependent - services are typically registered via static ctor or static - initializer at the point of loading of the dynamic SO's DLL. As - such, they end up registering before the dynamic SO. Upon - finalization of the service repository, the dynamic SO is - finalized first, thus unloading its DLL and unmapping the TEXT - segment, where the dependent static services code is. When - next, the static SO is to be finalized its memory is no longer - at valid address. The change ensures the dynamic service is - registered *before* its dependent static services and will - therefore be finalized last. The solution is based upon the - concept of forward declaring a dynamic service, named after the - configuration file being processed. Before processing a file, - we simply check for an existing service with the same name and - "pass" if we find one. - - Fixed the ACE_Service_Config_Guard's detection of the case where - the forward service declaration was not replaced by a proper - declaration as expected. It now properly removes the forward - decl, not the "good" instance. - - Cleared up the TSS management issue, causing executable to crash - on startup. The problem was due to an order-of-instantiation - issue, involving the TSS-based pointer to the current - gestalt. Used ACE_Object_Manager's shutting_down and starting_up - methods to determine if it is safe to use the static ACE_TSS - instance. - - Added const-ness for find(). Removed prototypes for initialize, - resume, suspend and remove that used plain char type and - conditionally compiled in only for WinCE and - ACE_USES_WCHAR. Service_Config still declares and defines these - (using ACE_ANTI_TCHAR). Updated to resolve problems when built - with ACE_USES_WCHAR. - - Declared private copy ctor and assignment operator for - ACE_Service_Gestalt and ACE_Service_Type_Factory, since their - instances are not intended to be copied. Added an - ACE_UNIMPLEMENTED_FUNC macro for the copy-ctor and assignment - operator of ACE_Service_Type_Factory. - - Reorganized to streamline the header file and remove unnecessary - declarations and to add a decl for find_static_svc_descriptor(). - - Changed the open() method's implementation to incorporate the - additional parameter for open_i(). - - Added the new find_static_svc_descriptor() member, which allows - ACE_Static_Function_Node to not need to know how the static - services are registered in the service repository. - - * ace/Service_Object.h: - * ace/Service_Object.inl: - * ace/Service_Object.cpp: - - Exposed the dll_ member through dll () const method because the - reference is needed by the new ACE_Dynamic_Service_Dependency - class. Fixed the dump () method. - - * ace/Service_Repository.h: - * ace/Service_Repository.cpp: - - Added const-ness for the find* () methods. Added debugging. The - remove method changes the order of services registrations in the - repository, which may break the Service Repository's invariant, - imposed by the need to correctly finalize dynamic services with - their own static services. - - For example, consider a case where a static service (S0) is - registered in the repository. Then a DLL-based service (D1) - registers its own static service (S1). So far the order is - "finalization-compatible": S0 < D1 < S1, i.e. services will be - finalized in reverse order and D1's DLL will still be available - when finalizing S1. Now let's remove (S0). Currently, after - this operation the repository end up in this order: S1 < D1, - which is trivially not finalization-compatible. - - Instead of replacing the pointer with the last ptr to a svc in - the array (repository) and then decrementing the current size of - the repo, the solution actually shifts left the rest of the - array. - - * ace/Service_Types.cpp: - * ace/Shared_Object.cpp: - - Removed some and updated existing debug statements. Eliminated - dead partial instantiation code. - - * ace/Svc_Conf.y: - * ace/Svc_Conf.h: - * ace/Svc_Conf_Tokens.h: - * ace/Svc_Conf_y.cpp: - - Changed to allow propagation of the knowledge about which - gestalt is currently being used for service registrations. - - Factored the ACE_Svc_Conf_Param class out in its own translation - unit. - - * ace/Svc_Conf_Param.h: - - Replaced the ACE_Service_Gestalt & in the method signatures with - ACE_Service_Gestalt * - - * ace/svcconf.mpb: - - Updated the svcconf.mpb rule to allow building with Bison 2.0 - and up. The change adds #ifdef/#endif around the generated - header file. A YYSTYPE_IS_DECLARED define is added in order to - make sure the token declarations from the Svc_Conf_Tokens.h are - used instead of those generated in the Svc_Conf_y.cpp file. - - Added the new Dynamic_Service_Dependency.* files. - - * bin/tao_orb_tests.lst: - - Adding the new tests to the automated test suite to run during - the nightly builds. - -Wed Apr 26 17:13:57 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - Added test for Parallel_Connect_Strategy. This test will take - about 9 minutes to run to completion. - -Wed Apr 26 14:22:31 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - Added tests for new POA features. - - * bin/MakeProjectCreator/config/endpointpolicy.mpb: - Applications wishing to use the TAO EndpointPolicy should add - this base project to their MPC files. This will include - libTAO_EndpointPolicy in their link lines. - -Wed Apr 26 14:10:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_hpux_kcc.GNU: - * include/makeinclude/platform_irix6.x_kcc.GNU: - * include/makeinclude/platform_linux_kcc.GNU: - * include/makeinclude/platform_osf1_4.x_kcc.GNU: - * include/makeinclude/platform_sunos5_kcc.GNU: - * include/makeinclude/platform_tru64_kcc.GNU: - Removed these files which are for the KAI C++ compiler. This - compiler doesn't exist anymore and full support for KAI C++ - has been removed now from ACE/TAO. See bugzilla bug 2007 - -Wed Apr 26 13:03:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added TAO/tests/POA/Bug_2511_Regression - -Wed Apr 26 12:52:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Process_Manager.cpp: - Fixed fuzz errors and use const when possible - -Wed Apr 26 12:02:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/CDR_Test.cpp: - Boolean is now size 1 or bigger, the marshaling code will take - care that a Boolean is marshaled as octet of size 1 - -Wed Apr 26 11:28:28 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * bin/tao_orb_tests.lst: - Added TAO/tests/OBV/truncatable. - -Wed Apr 26 10:05:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added TAO/tests/OBV - -Wed Apr 26 09:56:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added TAO/tests/Portable_Interceptors/Bug_2510_Regression - -Wed Apr 26 07:21:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_vxworks5.5.x.GNU: - Increased template-depth to 50, hopefully this is large enough to - also build CIAO - -Tue Apr 25 16:44:04 UTC 2006 Phil Mesnier <mesnier_p@ociweb.com> - - * ace/CDR_Stream.h: - * ace/CDR_Stream.cpp: - Added code to support rewinding the output CDR to some previous - location and updating a value there. This is required to support - marshaling offset values were the target of the offset is not - known until the final offset target is actually reached. - -Tue Apr 25 13:50:12 UTC 2006 Martin Corino <mcorino@remedy.nl> - - * tests/Signal_Test.cpp: - Fixed this test for (modern) Linux and Solaris at least. - The crux is blocking signals before calling sigwait() when doing - synchronous signal handling. - -Tue Apr 25 09:35:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Base_Thread_Adapter.h: - * ace/Cleanup.h: - Fixed some errors I made - -Tue Apr 25 09:23:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ACE-INSTALL.html: - Updated supported platforms and use vc8 for the MPC examples - -Tue Apr 25 08:49:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/CDR_Base.h: - Always use bool for Boolean, see bugzilla 2514 for more info - -Tue Apr 25 06:33:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Log_Msg.h: - Put back ACE_TSS_CLEANUP_NAME that got lost by accident with the - vc6 cleanup. Thanks to Olli Savia for reporting this. - -Mon Apr 24 17:39:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/run_test.lst: - Added Signal_Test. Thanks to Martin Corino for notificing that - we don't run this test on any platform. - -Mon Apr 24 10:17:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added TAO/tests/Bug_1676_Regression - -Mon Apr 24 09:57:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_vxworks5.5.x.GNU: - Increase template depth to resolve compile errors when building - TAO with the new sequence implementation - -Mon Apr 24 09:17:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_types.h: - Added ACE_LOFF_T for LynxOS. Thanks to Simon Massey for delivering - the patch - -Fri Apr 21 16:28:26 UTC 2006 William R. Otte <wotte@dre.vanderbilt.edu> - - Merge from ARMS Escher repository. - - * ace/OS_NS_Thread.cpp - - Updated sched_params() to not ignore the process id passed in. - Also enhanced the Windows implementation to get a handle to the - process specified before applying the priority class changes. - - * ace/Process_Manager.cpp - * ace/Process_Manager.h - - Added methods set_scheduler and set_scheduler_all - to set scheduling parameters for process(es) managed by this - class. - - * bin/MakeProjectCreator/config/ciao_config_handlers.mpb - - Fixed build ordering problem in CIAO. - -Fri Apr 21 07:41:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/Caching_Utility_T.cpp: - Removed invalid added const - -Thu Apr 20 19:27:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_types.h: - Added ACE_LOFF_T for the BSD variants, thanks to J.T. Conklin - for delivering the patch - -Thu Apr 20 18:16:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-macros.h: - Removed setting of ACE_ANY_OPS_USE_NAMESPACE again. This causes - problems when building CIAO. Thanks to Jeff Parsons for finding - this. - -Thu Apr 20 11:46:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/ACE_Init_Test.rc: - Removed another msvc specific part - -Thu Apr 20 11:06:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/fuzz.pl: - Added a check for deprecated macros, first one to check is - ACE_HAS_EXPLICIT_TEMPLATE_INSTANTIATION - - * tests/*.cpp: - * protocols/*.cpp: - * ACEXML/*.cpp: - * performance-tests/*.cpp - * netsvcs/*.cpp - Removed all explicit template instantiations - - * ace/os_include/sys/os_types.h: - Added ACE_LOFF_T typedef for VxWorks 6.2 and newer - -Thu Apr 20 07:23:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/os_include/sys/os_types.h: - Always define ACE_LOFF_T so that we can use it in the sendfile - wrapper - -Thu Apr 20 07:17:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/svcconf.mpb: - Removed Service_Templates.cpp - -Wed Apr 19 19:18:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/*: - Removed explicit template support - -Wed Apr 19 15:09:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Bug_2368_Regression_Test.cpp: - Fixed compile error and compile warning - -Wed Apr 19 14:35:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/OS_NS_sys_sendfile.h: - Updated include to get size_t and ACE_LOFF_T to fix compile - errors on Solaris - -Wed Apr 19 13:31:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * examples/APG/Logging/Use_Multiple_Sinks.cpp: - * examples/APG/Logging/Use_Ostream.cpp: - * examples/APG/Logging/LogManager.h: - Removed vc6 workarounds - -Wed Apr 19 13:13:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/Bug_2368_Regression_Test.cpp: - * tests/run_test.lst: - * tests/tests.mpc: - Added regression for bug 2368, this will fail because this bug - isn't addressed yet - -Wed Apr 19 13:08:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * include/makeinclude/platform_sunos5_g++.GNU: - Use -shared instead of -G when building shared libraries. Thanks to - Alan L Batongbacal <alanlb at vt dot edu> for reporting this. This - fixes bugzilla bug 1848. - -Wed Apr 19 12:11:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * tests/ACE_Init_Test.rc: - Removed msvc specific lines - - * tests/tests.mpc: - The ACE_Init_Test should use its resource file. Fixes bugzilla - 2495 - - * ace/Capabilities.{h,cpp}: - Prefix increment instead of postfix - - * ace/Hash_Map_Manager_T.{h,cpp}: - Removed workaround for suncc 4.2 - - * ace/Object_Manager.h: - * ace/Parse_Node.h: - * ace/POSIX_Proactor.h: - * ace/Thread_Mutex.h: - Doxygen improvements - - * ace/Parse_Node.cpp: - Use prefix increment, initialise pointers with 0 and don't print - an error message when we can't a dll, this gives false messages - when using TAO. This fixes bugzilla 2383. Thanks to Ming Xiong - for making a patch. - - * ace/Proactor.h: - Removed all virtual keywords, not needed for this class. Fixes - bugzilla 2178. - - * ace/Thread_Manager.h: - Removed virtual keywords, not needed and safes footprint. Fixes - bugzilla 2179. - - * bin/fuzz.pl: - Added more msvc checks to detect code constructors for vc6 or - vc7.0 - -Wed Apr 19 11:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - Merged ACE changes from sendfile branch. Also made a lot of const - improvements throughout ACE. - - Tue Apr 4 10:20:36 2006 Ossama Othman <ossama@dre.vanderbilt.edu> - - * ace/Mem_Map.cpp (open): - - Only copy the filename to the ACE_Mem_Map::filename_ member - if the call to open() succeeds. - - * ace/PI_Malloc.h (ACE_Malloc_Header): - - Added missing private and undefined copy constructor. The - corresponding assignment operator was already there. - - Wed Mar 29 14:51:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/config-sunos5.8.h: - Solaris 8 and newer also support sendfile. It seems HPUX also - supports it but with a different signature and semantics, so just - don't try it there yet. - - Wed Mar 29 13:06:12 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * ace/MMAP_Memory_Pool.{h,cpp,inl}: - Added new .inl file to implement the new mmap accessors, include - this new file in the header and implementation file - - Wed Mar 29 08:00:12 2006 Ossama Othman <ossama@dre.vanderbilt.edu> - - * ace/OS_NS_sys_sendfile.cpp - * ace/OS_NS_sys_sendfile.h - * ace/OS_NS_sys_sendfile.inl - New sendfile wrapper - - * ace/MMAP_Memory_Pool.h: - Added accessors for mmap - - * ace/ace.mpc - Added OS_NS_sys_sendfile - - * ace/config-linux-common.h - Added ACE_HAS_SENDFILE - -Wed Apr 19 07:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - Merged ACE changes from sequpdate3 branch - - * ace/config-osf1-3.2.h: - Removed, very old - - * ace/config-tru64.h: - Removed include of config-osf1-3.2.h - - * ace/OS_Dirent.cpp: - * ace/OS_Memory.{cpp,inl}: - * ace/OS_String.{cpp,inl}: - * ace/Sync.cpp: - Removed these files, where not build for a long time already - - * ace/ace.mpc: - Removed files above - - * ace/OS_Memory.h: - * ace/OS_String.h: - Removed include of inline files - - * ace/config-win32-msvc6.h: - * ace/config-kcc-common.h: - Removed these files, vc6 and kcc are deprecated - - * ace/config-irix6.x-kcc.h: - * ace/config-osf1-4.0.h: - * ace/config-sunos5.5.h: - Removed support for KCC - - * ace/Asynch_IO_Impl.h: - Removed workarounds for KCC - - * ace/config-win32-msvc.h: - Removed support for vc6 and vc7.0 - - * ace/Array_Base.cpp: - * ace/Array_Map.{h,cpp}: - * ace/Base_Thread_Adapter.h: - * ace/Cleanup.h: - * ace/Log_Msg.h: - * ace/Versioned_Namespace.h: - * ace/Get_Opt.cpp: - * ace/WIN32_Proactor.cpp: - * ace/Message_Queue.cpp: - * ace/CDR_Base.h: - Removed work arounds for vc6 - - * ace/README - * ace/config-win32-msvc6.h: - * ace/Global_Macros.h: - Deprecated ACE_HAS_BROKEN_NAMESPACES, we drop msvc6 and I am - removing the workarounds for this compiler one by one as part of the - new sequence implementation. Because we deprecate this also - ACE_NESTED_CLASS is deprecated - - * ace/config-win32-msvc6.h: - * ace/config-win32-msvc7.h: - * ace/config-win32-msvc8.h: - Removed ACE_HAS_BROKEN_IMPLICIT_CONST_CAST, not used in ACE - - * ace/config-win32-msvc6.h: - * ace/config-win32-visualage.h: - * ace/Registry.h - Removed ACE_HAS_BROKEN_NESTED_TEMPLATES, only used for msvc6 and - visual age, both are being dropped - -Wed Apr 19 07:48:12 UTC 2006 Johnny Willemsen <jwillemsen@remedy.nl> - - * bin/tao_orb_tests.lst: - Added Bug_2503_Regression - -Tue Apr 18 20:49:16 2006 Wallace Zhang <zhangw@ociweb.com> - - * ACE version 5.5.1 released. - -Local Variables: -mode: change-log -add-log-time-format: (lambda () (progn (setq tz (getenv "TZ")) (set-time-zone-rule "UTC") (setq time (format-time-string "%a %b %e %H:%M:%S %Z %Y" (current-time))) (set-time-zone-rule tz) time)) -indent-tabs-mode: nil -End: |