diff options
author | elliott_c <ocielliottc@users.noreply.github.com> | 2003-01-03 13:52:53 +0000 |
---|---|---|
committer | elliott_c <ocielliottc@users.noreply.github.com> | 2003-01-03 13:52:53 +0000 |
commit | acd07f55fad146acf1b84e5bc69c42a802748fc4 (patch) | |
tree | 0950ba6f579c4b62da7f4f8ccbd256eef6f86ffa | |
parent | f5922d948fd6b3dc4ef400ec129ad0d666d03c4a (diff) | |
download | ATCD-acd07f55fad146acf1b84e5bc69c42a802748fc4.tar.gz |
ChangeLogTag: Fri Jan 3 07:49:09 2003 Chad Elliott <elliott_c@ociweb.com>
-rw-r--r-- | ChangeLog | 387 | ||||
-rw-r--r-- | ChangeLogs/ChangeLog-03a | 387 | ||||
-rwxr-xr-x | bin/aix_shr | 6 |
3 files changed, 398 insertions, 382 deletions
diff --git a/ChangeLog b/ChangeLog index 758fcd2b0e3..c2629f2729d 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,21 +1,28 @@ +Fri Jan 3 07:49:09 2003 Chad Elliott <elliott_c@ociweb.com> + + * bin/aix_shr: + + Changed the name of the AIX template instantiation macro and + commented on why it's needed. + Thu Jan 2 17:44:29 2003 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/config-cygwin32-common.h: Set the custom export - macro's. These will be used when building dynamically with the - cygwin compiler. + * ace/config-cygwin32-common.h: Set the custom export + macro's. These will be used when building dynamically with the + cygwin compiler. - * bin/MakeProjectCreator/templates/gnu.mpd: Generate dllflags and - libflags into the GNU makefiles. These are important for the - Cygwin/MingW and Kylix compiler. + * bin/MakeProjectCreator/templates/gnu.mpd: Generate dllflags and + libflags into the GNU makefiles. These are important for the + Cygwin/MingW and Kylix compiler. Thu Jan 2 16:11:38 2003 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/config-freebsd.h: Fix use of macros for FreeBSD 4.6: + * ace/config-freebsd.h: Fix use of macros for FreeBSD 4.6: ACE_LACKS_RWLOCK_T, ACE_LACKS_READDIR_R, ACE_LACKS_SETSCHED, ACE_LACKS_PTHREAD_THR_SIGSETMASK. Do not define SCHED_RR, - SCHED_IO, SCHED_FIFO, SCHED_OTHER for FreeBSD >= 4.0, since they - are defined in sched.h. Thanks to Craig Rodrigues for providng - the patch. + SCHED_IO, SCHED_FIFO, SCHED_OTHER for FreeBSD >= 4.0, since they + are defined in sched.h. Thanks to Craig Rodrigues for providng + the patch. Thu Jan 2 13:33:01 2003 Chad Elliott <elliott_c@ociweb.com> @@ -24,36 +31,36 @@ Thu Jan 2 13:33:01 2003 Chad Elliott <elliott_c@ociweb.com> Added #include of ace/Log_Msg.h for ACE_ASSERT. * bin/aix_shr: - + Added -DTAO_INSTANTIATING_TEMPLATES to the link command line. Thu Jan 2 12:38:55 2003 Johnny Willemsen <jwillemsen@remedy.nl> - * ace/Makefile.ace: Set the dllflags in Makefile. This makes it - possible to do __declspec (dllexport) for the symbols in the - dll. This is needed for the Cygwin/MingW and Kylix compiler. + * ace/Makefile.ace: Set the dllflags in Makefile. This makes it + possible to do __declspec (dllexport) for the symbols in the + dll. This is needed for the Cygwin/MingW and Kylix compiler. - Thid change takes care of the static and dynamic library cases. + Thid change takes care of the static and dynamic library cases. Thu Jan 2 08:57:26 2003 Johnny Willemsen <jwillemsen@remedy.nl> - * include/makeinclude/platform_cygwin32.GNU: Add - --enable-auto-import to the linker flags. This will result in - the linker doing auto-import without giving warnings about - it. + * include/makeinclude/platform_cygwin32.GNU: Add + --enable-auto-import to the linker flags. This will result in + the linker doing auto-import without giving warnings about + it. Wed Jan 1 14:26:09 2003 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/topinfo_stats.sh: Forgot that we run this scripts in cron - environment. Set the path and other details that would help to - ease things. + * bin/topinfo_stats.sh: Forgot that we run this scripts in cron + environment. Set the path and other details that would help to + ease things. Tue Dec 31 21:21:26 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/generate_topinfo_charts.sh: - * bin/topinfo_stats.sh: Scripts to track the memory usage and plot - them. At present it works only for the simple test in - $TAO_ROOT/performance-tests/Memory/Single_Threaded. + * bin/generate_topinfo_charts.sh: + * bin/topinfo_stats.sh: Scripts to track the memory usage and plot + them. At present it works only for the simple test in + $TAO_ROOT/performance-tests/Memory/Single_Threaded. Tue Dec 31 17:57:19 2002 Steve Huston <shuston@riverace.com> @@ -142,42 +149,42 @@ Mon Dec 30 17:38:31 2002 Steve Huston <shuston@riverace.com> Mon Dec 30 15:25:26 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ACEXML/common/StrCharStream.cpp (open): - * ACEXML/common/FileCharStream.cpp (open): Changed to allow open - to succeed if automatic determination of encoding fails. This - allows the "Simple parsing" feature of ACEXML to work. Thanks - to Michael F"olsl <michael.foelsl@gmx.net> for reporting the - problem. + * ACEXML/common/StrCharStream.cpp (open): + * ACEXML/common/FileCharStream.cpp (open): Changed to allow open + to succeed if automatic determination of encoding fails. This + allows the "Simple parsing" feature of ACEXML to work. Thanks + to Michael F"olsl <michael.foelsl@gmx.net> for reporting the + problem. - * ACEXML/examples/SAXPrint/main.cpp: Added a new option to test - out ACEXML's simpling parsing feature. + * ACEXML/examples/SAXPrint/main.cpp: Added a new option to test + out ACEXML's simpling parsing feature. - * ACEXML/examples/SAXPrint/simple.svcconf.xml: Added this new - "simple" XML document to test out the "Simple" parsing feature - in ACEXML. + * ACEXML/examples/SAXPrint/simple.svcconf.xml: Added this new + "simple" XML document to test out the "Simple" parsing feature + in ACEXML. Mon Dec 30 12:21:33 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/MakeProjectCreator/config/taoversion.mpb: Fixed the version - information. This should automatically get updated during the - next beta/release. + * bin/MakeProjectCreator/config/taoversion.mpb: Fixed the version + information. This should automatically get updated during the + next beta/release. Sun Dec 29 18:27:38 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/auto_run_tests.lst: Prevented BiDirGIOP tests from running - in GIOP 10 builds. + * bin/auto_run_tests.lst: Prevented BiDirGIOP tests from running + in GIOP 10 builds. Sun Dec 29 16:43:03 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ACEXML/common/FileCharStream.cpp (close): Changed to reset the - file pointer, filename_ and encoding_ strings after they are - release. Thanks to Andrew Marlow <apm35@student.open.ac.uk> for - suggesting the change. + * ACEXML/common/FileCharStream.cpp (close): Changed to reset the + file pointer, filename_ and encoding_ strings after they are + release. Thanks to Andrew Marlow <apm35@student.open.ac.uk> for + suggesting the change. Sun Dec 29 00:47:58 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/auto_run_tests.lst: Security tests using SSLIOP will not be - run in static builds. + * bin/auto_run_tests.lst: Security tests using SSLIOP will not be + run in static builds. Sat Dec 28 23:05:51 2002 Steve Huston <shuston@riverace.com> @@ -186,42 +193,42 @@ Sat Dec 28 23:05:51 2002 Steve Huston <shuston@riverace.com> Sat Dec 28 13:16:24 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ace/MEM_IO.cpp (recv_buf): Reverted my previous hack in: + * ace/MEM_IO.cpp (recv_buf): Reverted my previous hack in: - Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> + Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> - and make the connection close handling more robust. + and make the connection close handling more robust. - * tests/MEM_Stream_Test.cpp: Reverted my previous change in: + * tests/MEM_Stream_Test.cpp: Reverted my previous change in: - Sat Dec 28 11:04:38 2002 Nanbor Wang <nanbor@cs.wustl.edu> + Sat Dec 28 11:04:38 2002 Nanbor Wang <nanbor@cs.wustl.edu> - It was not necessary. Added debug information for close - connection. + It was not necessary. Added debug information for close + connection. Sat Dec 28 11:04:38 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * tests/MEM_Stream_Test.cpp: Made sure this test ignore the EIO - error from the following change when a connection was closed. + * tests/MEM_Stream_Test.cpp: Made sure this test ignore the EIO + error from the following change when a connection was closed. - Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> + Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> Fri Dec 27 20:06:39 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ace/Malloc_T.h: - * ace/Malloc_T.i: - * ace/Malloc_T.cpp: Added a new ACE_Malloc_T::bad () method to - query whether a Malloc_T object was constructed successfully. - This approach, IMO, is sub-optimal compared to using an explicit - init method approach. The existing constructor signatures, - however, prevent us from doing so without breaking backward - compatibility. - - * ace/MEM_SAP.cpp: Now that there's a way to check whether the - underlying Malloc is functioning or not. Changed MEM_SAP to - check for proper initialization of underlying Malloc object and - clean up OS resources if the Malloc object is defected. This - change makes MEM_Stream transport mechanism more robust. + * ace/Malloc_T.h: + * ace/Malloc_T.i: + * ace/Malloc_T.cpp: Added a new ACE_Malloc_T::bad () method to + query whether a Malloc_T object was constructed successfully. + This approach, IMO, is sub-optimal compared to using an explicit + init method approach. The existing constructor signatures, + however, prevent us from doing so without breaking backward + compatibility. + + * ace/MEM_SAP.cpp: Now that there's a way to check whether the + underlying Malloc is functioning or not. Changed MEM_SAP to + check for proper initialization of underlying Malloc object and + clean up OS resources if the Malloc object is defected. This + change makes MEM_Stream transport mechanism more robust. Fri Dec 27 19:46:31 2002 Steve Huston <shuston@riverace.com> @@ -245,46 +252,46 @@ Fri Dec 27 17:43:29 2002 Steve Huston <shuston@riverace.com> Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ace/MEM_IO.cpp (recv_buf): Changed to return error (-1) and set - the errno to EIO when we receive 0 byte off the signaling socket - as this should never happen under normal circumstance. However, - on Win32 and Solaris, when one end of a connection closes the - socket while the other end is still reading the socket, the - active end will continuously receive several 0 bytes messages. - This confuses the hack of the MEM_IO mechanism and causes the - memory management routine to fall into a tight spinning - deadlock. + * ace/MEM_IO.cpp (recv_buf): Changed to return error (-1) and set + the errno to EIO when we receive 0 byte off the signaling socket + as this should never happen under normal circumstance. However, + on Win32 and Solaris, when one end of a connection closes the + socket while the other end is still reading the socket, the + active end will continuously receive several 0 bytes messages. + This confuses the hack of the MEM_IO mechanism and causes the + memory management routine to fall into a tight spinning + deadlock. Fri Dec 27 07:49:28 2002 Craig Rodrigues <crodrigu@bbn.com> - * ace/config-macosx.h: Remove temporary hacks for buggy + * ace/config-macosx.h: Remove temporary hacks for buggy gcc 3.1 compiler. Thanks to John Zorko for reminding me about - this. + this. - * examples/Service_Configurator/IPC-tests/server/Makefile: - * examples/Service_Configurator/Misc/Makefile: - * examples/Timer_Queue/Makefile: - Add ACELIBS to ACE_SHLIBS to fix MacOS X linking problem. + * examples/Service_Configurator/IPC-tests/server/Makefile: + * examples/Service_Configurator/Misc/Makefile: + * examples/Timer_Queue/Makefile: + Add ACELIBS to ACE_SHLIBS to fix MacOS X linking problem. Thu Dec 26 17:11:54 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/make_release: Fixed the following problems in the release - script which caused problems during the beta cutting process: + * bin/make_release: Fixed the following problems in the release + script which caused problems during the beta cutting process: - - cvs commits of TAO/Changelog TAO/VERSION etc. failed. The - failure was because we were also commiting taoversion.mpb but - from $TAO_ROOT as "cvs commit ChangeLog VERSION - other_files ../bin/Makeyada". Using ../ to indicate files in - higher directories is an unpardonable mistake for cvs - 1.10.x. It used to work 1.9.x. But with 1.10* versions the - protocol changed disallowing the ../ + - cvs commits of TAO/Changelog TAO/VERSION etc. failed. The + failure was because we were also commiting taoversion.mpb but + from $TAO_ROOT as "cvs commit ChangeLog VERSION + other_files ../bin/Makeyada". Using ../ to indicate files in + higher directories is an unpardonable mistake for cvs + 1.10.x. It used to work 1.9.x. But with 1.10* versions the + protocol changed disallowing the ../ - - we didnt set the file permissions on the generated files - properly + - we didnt set the file permissions on the generated files + properly - - the generated files from mpc were not added to the list of - files for which LF->CRLF conversions are supposed to be - masked during zip! + - the generated files from mpc were not added to the list of + files for which LF->CRLF conversions are supposed to be + masked during zip! Thu Dec 26 16:17:41 2002 Steve Huston <shuston@riverace.com> @@ -293,19 +300,19 @@ Thu Dec 26 16:17:41 2002 Steve Huston <shuston@riverace.com> Thu Dec 26 09:14:38 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * Makefile: Removed configure.in from the list of controlled files. + * Makefile: Removed configure.in from the list of controlled files. Thu Dec 26 07:36:10 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> * include/makeinclude/platform_freebsd_pthread.GNU: Removed for - sure now. This should have been removed in this "Thu Nov 14 - 22:47:45 UTC 2002 Craig Rodrigues <crodrigu@bbn.com>" - checkin. Again, thanks to Craig for catching this. + sure now. This should have been removed in this "Thu Nov 14 + 22:47:45 UTC 2002 Craig Rodrigues <crodrigu@bbn.com>" + checkin. Again, thanks to Craig for catching this. Thu Dec 26 07:31:42 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * apps/Gateway/Gateway/Makefile: Fixed a ttypo. Thanks to Craig - Rodrigues for catching this. + * apps/Gateway/Gateway/Makefile: Fixed a ttypo. Thanks to Craig + Rodrigues for catching this. Wed Dec 25 17:47:12 2002 Bala <bala@cs.wustl.edu> @@ -313,34 +320,34 @@ Wed Dec 25 17:47:12 2002 Bala <bala@cs.wustl.edu> Wed Dec 25 17:39:05 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/make_release: Fixed problems with generation of vcproj etc. + * bin/make_release: Fixed problems with generation of vcproj etc. Wed Dec 25 06:14:17 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> - * Merry Christmas to all. May we have peace and justice for all in 2003. + * Merry Christmas to all. May we have peace and justice for all in 2003. Tue Dec 24 17:29:11 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * docs/ACE-subsets.html: Documented the compiler options that - Steve Vranyes experimented and tested with to get a good - footprint reduction on Solaris. + * docs/ACE-subsets.html: Documented the compiler options that + Steve Vranyes experimented and tested with to get a good + footprint reduction on Solaris. Tue Dec 24 07:31:54 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Argv_Type_Converter.h: - * ace/Argv_Type_Converter.cpp: Reverted the following changes "Mon - Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com>" and - "Mon Dec 23 22:57:18 2002 Balachandran Natarajan - <bala@isis-server.isis.vanderbilt.edu>". The argument parsing in - TAO gets busted due to these changes and most of tests in our - daily builds fail. + * ace/Argv_Type_Converter.h: + * ace/Argv_Type_Converter.cpp: Reverted the following changes "Mon + Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com>" and + "Mon Dec 23 22:57:18 2002 Balachandran Natarajan + <bala@isis-server.isis.vanderbilt.edu>". The argument parsing in + TAO gets busted due to these changes and most of tests in our + daily builds fail. Mon Dec 23 22:57:18 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Argv_Type_Converter.h: Changed saved_argc_ to be of type int - as opposed to int &. This seems to be necessitated by the change - "Mon Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com>" - which gives compile errors on all platforms. + * ace/Argv_Type_Converter.h: Changed saved_argc_ to be of type int + as opposed to int &. This seems to be necessitated by the change + "Mon Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com>" + which gives compile errors on all platforms. Mon Dec 23 21:07:20 2002 Steve Huston <shuston@riverace.com> @@ -374,12 +381,12 @@ Mon Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com> Mon Dec 23 16:16:05 2002 Boris S Kolpackov <boris@isis-server.isis.vanderbilt.edu> - * bin/auto_run_tests.lst: Added - TAO/orbsvcs/tests/Security/Crash_Test to the list of tests. + * bin/auto_run_tests.lst: Added + TAO/orbsvcs/tests/Security/Crash_Test to the list of tests. Mon Dec 23 17:08:17 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * THANKS: Changed Boris's information in the Hall of Fame. + * THANKS: Changed Boris's information in the Hall of Fame. Mon Dec 23 12:13:51 2002 Steve Huston <shuston@riverace.com> @@ -435,46 +442,46 @@ Mon Dec 23 12:02:33 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> Sun Dec 22 11:41:56 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Connector.cpp: Fixed fuzz errors. + * ace/Connector.cpp: Fixed fuzz errors. Sun Dec 22 11:37:44 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * THANKS: Added Daniel White to the hall of fame! + * THANKS: Added Daniel White to the hall of fame! Sun Dec 22 11:31:53 2002 Craig Rodrigues <crodrigu@bbn.com> - * apps/Gateway/Gateway/Makefile: - * apps/Gateway/Peer/Makefile: Add $(ACELIB) to $(ACE_SHLIBS) so - that -lACE is added to the linkline when a shared library is - built. Solves linking problem MacOS X. Thanks to John Zorko and - Daniel white <ygor@comcast.net> for pointing this out. + * apps/Gateway/Gateway/Makefile: + * apps/Gateway/Peer/Makefile: Add $(ACELIB) to $(ACE_SHLIBS) so + that -lACE is added to the linkline when a shared library is + built. Solves linking problem MacOS X. Thanks to John Zorko and + Daniel white <ygor@comcast.net> for pointing this out. Sun Dec 22 10:25:35 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Connector.cpp: Fixed a warning with g++. + * ace/Connector.cpp: Fixed a warning with g++. Sun Dec 22 09:18:00 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Connector.cpp: - * ace/Connector.h: Added a lock to the ACE_Connector class to - make the following atomic + * ace/Connector.cpp: + * ace/Connector.h: Added a lock to the ACE_Connector class to + make the following atomic - . Registration of AST with the handler_map_ - . Registration of the ACE_Connector with the Reactor - . and the registration of timers with the Reactor's timer - queue. + . Registration of AST with the handler_map_ + . Registration of the ACE_Connector with the Reactor + . and the registration of timers with the Reactor's timer + queue. - If they are not atomic, it throws open race conditions such as - the ones documented in BUG 1405. This is not complete fix for - BUG 1405, since it taxes applications using only one thread with a - lock. This is a short-term fix to prevent the race condition - from occuring. + If they are not atomic, it throws open race conditions such as + the ones documented in BUG 1405. This is not complete fix for + BUG 1405, since it taxes applications using only one thread with a + lock. This is a short-term fix to prevent the race condition + from occuring. - Other relevant changes include + Other relevant changes include - . the AST is now refcounted - . the AST cannot be deleted directly. The lifetime of the AST - can only be manipulated using the refcount on the AST. + . the AST is now refcounted + . the AST cannot be deleted directly. The lifetime of the AST + can only be manipulated using the refcount on the AST. Sat Dec 21 18:27:42 2002 Steve Huston <shuston@riverace.com> @@ -674,10 +681,10 @@ Thu Dec 19 18:47:29 2002 Steve Huston <shuston@riverace.com> Thu Dec 19 16:47:17 2002 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - * ace/SOCK_Dgram.h: - ace/SOCK_CODgram.h: Added metatype information for - ACE_SOCK_Dgram and ACE_SOCK_CODgram. Thanks to Sankaranarayanan - K V <sankar@migv.mot.com> for reporting this. + * ace/SOCK_Dgram.h: + ace/SOCK_CODgram.h: Added metatype information for + ACE_SOCK_Dgram and ACE_SOCK_CODgram. Thanks to Sankaranarayanan + K V <sankar@migv.mot.com> for reporting this. Thu Dec 19 10:02:56 2002 Rich Seibel <seibel_r@ociweb.com> @@ -903,39 +910,39 @@ Mon Dec 16 13:14:34 2002 Phil Mesnier <mesnier_p@ociweb.com> * ace/CDR_Stream.h: * ace/CDR_Stream.i: Preparing for codeset negotiation. The ACE input/output CDR - streams' support for wide character IO is enhanced by adding a - flag to allow or disallow wchar i/o regardless of whether a - translator is present. There are times when wchar is not - allowed, such as GIOP 1.0, or when a native codeset for wchar - was not specified. In CORBA, a native wchar codeset must be - specified, it may not be defaulted. Within the ACE CDR streams, - wchar i/o is allowed by default, to ensure backwards - compatibility. - - I fixed a few of the reading and writing methods to ensure that - the good_bit_ flag is cleared if an error occurs. This was not - happening in all cases, leading to errors elsewhere. - - I have extended the definition of the - ACE_[W]Char_Codeset_Translator classes to make them reference - counted objects. There are some codesets that support shifting - and may be stateful. By using reference counting, the CDR - streams don't have to worry about ownership of stateful - translators. + streams' support for wide character IO is enhanced by adding a + flag to allow or disallow wchar i/o regardless of whether a + translator is present. There are times when wchar is not + allowed, such as GIOP 1.0, or when a native codeset for wchar + was not specified. In CORBA, a native wchar codeset must be + specified, it may not be defaulted. Within the ACE CDR streams, + wchar i/o is allowed by default, to ensure backwards + compatibility. + + I fixed a few of the reading and writing methods to ensure that + the good_bit_ flag is cleared if an error occurs. This was not + happening in all cases, leading to errors elsewhere. + + I have extended the definition of the + ACE_[W]Char_Codeset_Translator classes to make them reference + counted objects. There are some codesets that support shifting + and may be stateful. By using reference counting, the CDR + streams don't have to worry about ownership of stateful + translators. Finally, the translators now have pure virtual methods that - return the native and translated codeset ID values. These values - refer to entries in the DCE Code and Character Set Registry, - which is the reference for CORBA codeset ids. Anyone - implementing a char or wchar codeset will have to supply a - numeric value for the native codeset (ncs) and the translated - codeset (tcs). NCS refers to the codeset of the text on the - application side of the stream, and TCS refers to the codset of - the text on the "wire" side of the stream. + return the native and translated codeset ID values. These values + refer to entries in the DCE Code and Character Set Registry, + which is the reference for CORBA codeset ids. Anyone + implementing a char or wchar codeset will have to supply a + numeric value for the native codeset (ncs) and the translated + codeset (tcs). NCS refers to the codeset of the text on the + application side of the stream, and TCS refers to the codset of + the text on the "wire" side of the stream. These changes are not everything that is required to support - codeset negotiation, but it is all the changes required to the - CDR streams. + codeset negotiation, but it is all the changes required to the + CDR streams. Sat Dec 14 08:56:23 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> @@ -1240,14 +1247,14 @@ Sun Dec 08 01:17:26 2002 Nanbor Wang <nanbor@cs.wustl.edu> Thu Dec 5 20:30:56 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> * ace/Event_Handler.cpp: Updated the implementation of - ACE_Event_Handler::read_adapter() so that the code is consistent - with the stated semantics. Thanks to Ron Muck <rlm@sdiusa.com> - for the fix. + ACE_Event_Handler::read_adapter() so that the code is consistent + with the stated semantics. Thanks to Ron Muck <rlm@sdiusa.com> + for the fix. * ace/OS_Dirent.inl (readdir_r): Enhanced the - ACE_OS_Directn::readdir_r() wrapper facade so that it returns 1 - (i.e., "done") when an error occurs. Thanks to Abhay Kulkarni - <Abhay.Kulkarni@veritas.com> for reporting this. + ACE_OS_Directn::readdir_r() wrapper facade so that it returns 1 + (i.e., "done") when an error occurs. Thanks to Abhay Kulkarni + <Abhay.Kulkarni@veritas.com> for reporting this. Sat Dec 7 19:54:12 2002 Steve Huston <shuston@riverace.com> @@ -1635,7 +1642,7 @@ Sat Nov 23 18:12:37 2002 Steve Huston <shuston@riverace.com> on AIX with g++. Without this, the compiler and linker don't get things done right and static data members in templates don't get instantiated correctly. Thanks to Max F. Bilyk <mbilyk@europe.com> - for reporting this problem. + for reporting this problem. * ace/SSL/Makefile: Set LIBS = $(ACELIB). Else it won't link on AIX, g++. diff --git a/ChangeLogs/ChangeLog-03a b/ChangeLogs/ChangeLog-03a index 758fcd2b0e3..c2629f2729d 100644 --- a/ChangeLogs/ChangeLog-03a +++ b/ChangeLogs/ChangeLog-03a @@ -1,21 +1,28 @@ +Fri Jan 3 07:49:09 2003 Chad Elliott <elliott_c@ociweb.com> + + * bin/aix_shr: + + Changed the name of the AIX template instantiation macro and + commented on why it's needed. + Thu Jan 2 17:44:29 2003 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/config-cygwin32-common.h: Set the custom export - macro's. These will be used when building dynamically with the - cygwin compiler. + * ace/config-cygwin32-common.h: Set the custom export + macro's. These will be used when building dynamically with the + cygwin compiler. - * bin/MakeProjectCreator/templates/gnu.mpd: Generate dllflags and - libflags into the GNU makefiles. These are important for the - Cygwin/MingW and Kylix compiler. + * bin/MakeProjectCreator/templates/gnu.mpd: Generate dllflags and + libflags into the GNU makefiles. These are important for the + Cygwin/MingW and Kylix compiler. Thu Jan 2 16:11:38 2003 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/config-freebsd.h: Fix use of macros for FreeBSD 4.6: + * ace/config-freebsd.h: Fix use of macros for FreeBSD 4.6: ACE_LACKS_RWLOCK_T, ACE_LACKS_READDIR_R, ACE_LACKS_SETSCHED, ACE_LACKS_PTHREAD_THR_SIGSETMASK. Do not define SCHED_RR, - SCHED_IO, SCHED_FIFO, SCHED_OTHER for FreeBSD >= 4.0, since they - are defined in sched.h. Thanks to Craig Rodrigues for providng - the patch. + SCHED_IO, SCHED_FIFO, SCHED_OTHER for FreeBSD >= 4.0, since they + are defined in sched.h. Thanks to Craig Rodrigues for providng + the patch. Thu Jan 2 13:33:01 2003 Chad Elliott <elliott_c@ociweb.com> @@ -24,36 +31,36 @@ Thu Jan 2 13:33:01 2003 Chad Elliott <elliott_c@ociweb.com> Added #include of ace/Log_Msg.h for ACE_ASSERT. * bin/aix_shr: - + Added -DTAO_INSTANTIATING_TEMPLATES to the link command line. Thu Jan 2 12:38:55 2003 Johnny Willemsen <jwillemsen@remedy.nl> - * ace/Makefile.ace: Set the dllflags in Makefile. This makes it - possible to do __declspec (dllexport) for the symbols in the - dll. This is needed for the Cygwin/MingW and Kylix compiler. + * ace/Makefile.ace: Set the dllflags in Makefile. This makes it + possible to do __declspec (dllexport) for the symbols in the + dll. This is needed for the Cygwin/MingW and Kylix compiler. - Thid change takes care of the static and dynamic library cases. + Thid change takes care of the static and dynamic library cases. Thu Jan 2 08:57:26 2003 Johnny Willemsen <jwillemsen@remedy.nl> - * include/makeinclude/platform_cygwin32.GNU: Add - --enable-auto-import to the linker flags. This will result in - the linker doing auto-import without giving warnings about - it. + * include/makeinclude/platform_cygwin32.GNU: Add + --enable-auto-import to the linker flags. This will result in + the linker doing auto-import without giving warnings about + it. Wed Jan 1 14:26:09 2003 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/topinfo_stats.sh: Forgot that we run this scripts in cron - environment. Set the path and other details that would help to - ease things. + * bin/topinfo_stats.sh: Forgot that we run this scripts in cron + environment. Set the path and other details that would help to + ease things. Tue Dec 31 21:21:26 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/generate_topinfo_charts.sh: - * bin/topinfo_stats.sh: Scripts to track the memory usage and plot - them. At present it works only for the simple test in - $TAO_ROOT/performance-tests/Memory/Single_Threaded. + * bin/generate_topinfo_charts.sh: + * bin/topinfo_stats.sh: Scripts to track the memory usage and plot + them. At present it works only for the simple test in + $TAO_ROOT/performance-tests/Memory/Single_Threaded. Tue Dec 31 17:57:19 2002 Steve Huston <shuston@riverace.com> @@ -142,42 +149,42 @@ Mon Dec 30 17:38:31 2002 Steve Huston <shuston@riverace.com> Mon Dec 30 15:25:26 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ACEXML/common/StrCharStream.cpp (open): - * ACEXML/common/FileCharStream.cpp (open): Changed to allow open - to succeed if automatic determination of encoding fails. This - allows the "Simple parsing" feature of ACEXML to work. Thanks - to Michael F"olsl <michael.foelsl@gmx.net> for reporting the - problem. + * ACEXML/common/StrCharStream.cpp (open): + * ACEXML/common/FileCharStream.cpp (open): Changed to allow open + to succeed if automatic determination of encoding fails. This + allows the "Simple parsing" feature of ACEXML to work. Thanks + to Michael F"olsl <michael.foelsl@gmx.net> for reporting the + problem. - * ACEXML/examples/SAXPrint/main.cpp: Added a new option to test - out ACEXML's simpling parsing feature. + * ACEXML/examples/SAXPrint/main.cpp: Added a new option to test + out ACEXML's simpling parsing feature. - * ACEXML/examples/SAXPrint/simple.svcconf.xml: Added this new - "simple" XML document to test out the "Simple" parsing feature - in ACEXML. + * ACEXML/examples/SAXPrint/simple.svcconf.xml: Added this new + "simple" XML document to test out the "Simple" parsing feature + in ACEXML. Mon Dec 30 12:21:33 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/MakeProjectCreator/config/taoversion.mpb: Fixed the version - information. This should automatically get updated during the - next beta/release. + * bin/MakeProjectCreator/config/taoversion.mpb: Fixed the version + information. This should automatically get updated during the + next beta/release. Sun Dec 29 18:27:38 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/auto_run_tests.lst: Prevented BiDirGIOP tests from running - in GIOP 10 builds. + * bin/auto_run_tests.lst: Prevented BiDirGIOP tests from running + in GIOP 10 builds. Sun Dec 29 16:43:03 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ACEXML/common/FileCharStream.cpp (close): Changed to reset the - file pointer, filename_ and encoding_ strings after they are - release. Thanks to Andrew Marlow <apm35@student.open.ac.uk> for - suggesting the change. + * ACEXML/common/FileCharStream.cpp (close): Changed to reset the + file pointer, filename_ and encoding_ strings after they are + release. Thanks to Andrew Marlow <apm35@student.open.ac.uk> for + suggesting the change. Sun Dec 29 00:47:58 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/auto_run_tests.lst: Security tests using SSLIOP will not be - run in static builds. + * bin/auto_run_tests.lst: Security tests using SSLIOP will not be + run in static builds. Sat Dec 28 23:05:51 2002 Steve Huston <shuston@riverace.com> @@ -186,42 +193,42 @@ Sat Dec 28 23:05:51 2002 Steve Huston <shuston@riverace.com> Sat Dec 28 13:16:24 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ace/MEM_IO.cpp (recv_buf): Reverted my previous hack in: + * ace/MEM_IO.cpp (recv_buf): Reverted my previous hack in: - Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> + Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> - and make the connection close handling more robust. + and make the connection close handling more robust. - * tests/MEM_Stream_Test.cpp: Reverted my previous change in: + * tests/MEM_Stream_Test.cpp: Reverted my previous change in: - Sat Dec 28 11:04:38 2002 Nanbor Wang <nanbor@cs.wustl.edu> + Sat Dec 28 11:04:38 2002 Nanbor Wang <nanbor@cs.wustl.edu> - It was not necessary. Added debug information for close - connection. + It was not necessary. Added debug information for close + connection. Sat Dec 28 11:04:38 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * tests/MEM_Stream_Test.cpp: Made sure this test ignore the EIO - error from the following change when a connection was closed. + * tests/MEM_Stream_Test.cpp: Made sure this test ignore the EIO + error from the following change when a connection was closed. - Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> + Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> Fri Dec 27 20:06:39 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ace/Malloc_T.h: - * ace/Malloc_T.i: - * ace/Malloc_T.cpp: Added a new ACE_Malloc_T::bad () method to - query whether a Malloc_T object was constructed successfully. - This approach, IMO, is sub-optimal compared to using an explicit - init method approach. The existing constructor signatures, - however, prevent us from doing so without breaking backward - compatibility. - - * ace/MEM_SAP.cpp: Now that there's a way to check whether the - underlying Malloc is functioning or not. Changed MEM_SAP to - check for proper initialization of underlying Malloc object and - clean up OS resources if the Malloc object is defected. This - change makes MEM_Stream transport mechanism more robust. + * ace/Malloc_T.h: + * ace/Malloc_T.i: + * ace/Malloc_T.cpp: Added a new ACE_Malloc_T::bad () method to + query whether a Malloc_T object was constructed successfully. + This approach, IMO, is sub-optimal compared to using an explicit + init method approach. The existing constructor signatures, + however, prevent us from doing so without breaking backward + compatibility. + + * ace/MEM_SAP.cpp: Now that there's a way to check whether the + underlying Malloc is functioning or not. Changed MEM_SAP to + check for proper initialization of underlying Malloc object and + clean up OS resources if the Malloc object is defected. This + change makes MEM_Stream transport mechanism more robust. Fri Dec 27 19:46:31 2002 Steve Huston <shuston@riverace.com> @@ -245,46 +252,46 @@ Fri Dec 27 17:43:29 2002 Steve Huston <shuston@riverace.com> Fri Dec 27 08:38:29 2002 Nanbor Wang <nanbor@cs.wustl.edu> - * ace/MEM_IO.cpp (recv_buf): Changed to return error (-1) and set - the errno to EIO when we receive 0 byte off the signaling socket - as this should never happen under normal circumstance. However, - on Win32 and Solaris, when one end of a connection closes the - socket while the other end is still reading the socket, the - active end will continuously receive several 0 bytes messages. - This confuses the hack of the MEM_IO mechanism and causes the - memory management routine to fall into a tight spinning - deadlock. + * ace/MEM_IO.cpp (recv_buf): Changed to return error (-1) and set + the errno to EIO when we receive 0 byte off the signaling socket + as this should never happen under normal circumstance. However, + on Win32 and Solaris, when one end of a connection closes the + socket while the other end is still reading the socket, the + active end will continuously receive several 0 bytes messages. + This confuses the hack of the MEM_IO mechanism and causes the + memory management routine to fall into a tight spinning + deadlock. Fri Dec 27 07:49:28 2002 Craig Rodrigues <crodrigu@bbn.com> - * ace/config-macosx.h: Remove temporary hacks for buggy + * ace/config-macosx.h: Remove temporary hacks for buggy gcc 3.1 compiler. Thanks to John Zorko for reminding me about - this. + this. - * examples/Service_Configurator/IPC-tests/server/Makefile: - * examples/Service_Configurator/Misc/Makefile: - * examples/Timer_Queue/Makefile: - Add ACELIBS to ACE_SHLIBS to fix MacOS X linking problem. + * examples/Service_Configurator/IPC-tests/server/Makefile: + * examples/Service_Configurator/Misc/Makefile: + * examples/Timer_Queue/Makefile: + Add ACELIBS to ACE_SHLIBS to fix MacOS X linking problem. Thu Dec 26 17:11:54 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/make_release: Fixed the following problems in the release - script which caused problems during the beta cutting process: + * bin/make_release: Fixed the following problems in the release + script which caused problems during the beta cutting process: - - cvs commits of TAO/Changelog TAO/VERSION etc. failed. The - failure was because we were also commiting taoversion.mpb but - from $TAO_ROOT as "cvs commit ChangeLog VERSION - other_files ../bin/Makeyada". Using ../ to indicate files in - higher directories is an unpardonable mistake for cvs - 1.10.x. It used to work 1.9.x. But with 1.10* versions the - protocol changed disallowing the ../ + - cvs commits of TAO/Changelog TAO/VERSION etc. failed. The + failure was because we were also commiting taoversion.mpb but + from $TAO_ROOT as "cvs commit ChangeLog VERSION + other_files ../bin/Makeyada". Using ../ to indicate files in + higher directories is an unpardonable mistake for cvs + 1.10.x. It used to work 1.9.x. But with 1.10* versions the + protocol changed disallowing the ../ - - we didnt set the file permissions on the generated files - properly + - we didnt set the file permissions on the generated files + properly - - the generated files from mpc were not added to the list of - files for which LF->CRLF conversions are supposed to be - masked during zip! + - the generated files from mpc were not added to the list of + files for which LF->CRLF conversions are supposed to be + masked during zip! Thu Dec 26 16:17:41 2002 Steve Huston <shuston@riverace.com> @@ -293,19 +300,19 @@ Thu Dec 26 16:17:41 2002 Steve Huston <shuston@riverace.com> Thu Dec 26 09:14:38 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * Makefile: Removed configure.in from the list of controlled files. + * Makefile: Removed configure.in from the list of controlled files. Thu Dec 26 07:36:10 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> * include/makeinclude/platform_freebsd_pthread.GNU: Removed for - sure now. This should have been removed in this "Thu Nov 14 - 22:47:45 UTC 2002 Craig Rodrigues <crodrigu@bbn.com>" - checkin. Again, thanks to Craig for catching this. + sure now. This should have been removed in this "Thu Nov 14 + 22:47:45 UTC 2002 Craig Rodrigues <crodrigu@bbn.com>" + checkin. Again, thanks to Craig for catching this. Thu Dec 26 07:31:42 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * apps/Gateway/Gateway/Makefile: Fixed a ttypo. Thanks to Craig - Rodrigues for catching this. + * apps/Gateway/Gateway/Makefile: Fixed a ttypo. Thanks to Craig + Rodrigues for catching this. Wed Dec 25 17:47:12 2002 Bala <bala@cs.wustl.edu> @@ -313,34 +320,34 @@ Wed Dec 25 17:47:12 2002 Bala <bala@cs.wustl.edu> Wed Dec 25 17:39:05 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * bin/make_release: Fixed problems with generation of vcproj etc. + * bin/make_release: Fixed problems with generation of vcproj etc. Wed Dec 25 06:14:17 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> - * Merry Christmas to all. May we have peace and justice for all in 2003. + * Merry Christmas to all. May we have peace and justice for all in 2003. Tue Dec 24 17:29:11 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * docs/ACE-subsets.html: Documented the compiler options that - Steve Vranyes experimented and tested with to get a good - footprint reduction on Solaris. + * docs/ACE-subsets.html: Documented the compiler options that + Steve Vranyes experimented and tested with to get a good + footprint reduction on Solaris. Tue Dec 24 07:31:54 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Argv_Type_Converter.h: - * ace/Argv_Type_Converter.cpp: Reverted the following changes "Mon - Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com>" and - "Mon Dec 23 22:57:18 2002 Balachandran Natarajan - <bala@isis-server.isis.vanderbilt.edu>". The argument parsing in - TAO gets busted due to these changes and most of tests in our - daily builds fail. + * ace/Argv_Type_Converter.h: + * ace/Argv_Type_Converter.cpp: Reverted the following changes "Mon + Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com>" and + "Mon Dec 23 22:57:18 2002 Balachandran Natarajan + <bala@isis-server.isis.vanderbilt.edu>". The argument parsing in + TAO gets busted due to these changes and most of tests in our + daily builds fail. Mon Dec 23 22:57:18 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Argv_Type_Converter.h: Changed saved_argc_ to be of type int - as opposed to int &. This seems to be necessitated by the change - "Mon Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com>" - which gives compile errors on all platforms. + * ace/Argv_Type_Converter.h: Changed saved_argc_ to be of type int + as opposed to int &. This seems to be necessitated by the change + "Mon Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com>" + which gives compile errors on all platforms. Mon Dec 23 21:07:20 2002 Steve Huston <shuston@riverace.com> @@ -374,12 +381,12 @@ Mon Dec 23 17:26:42 2002 Steve Huston <shuston@riverace.com> Mon Dec 23 16:16:05 2002 Boris S Kolpackov <boris@isis-server.isis.vanderbilt.edu> - * bin/auto_run_tests.lst: Added - TAO/orbsvcs/tests/Security/Crash_Test to the list of tests. + * bin/auto_run_tests.lst: Added + TAO/orbsvcs/tests/Security/Crash_Test to the list of tests. Mon Dec 23 17:08:17 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * THANKS: Changed Boris's information in the Hall of Fame. + * THANKS: Changed Boris's information in the Hall of Fame. Mon Dec 23 12:13:51 2002 Steve Huston <shuston@riverace.com> @@ -435,46 +442,46 @@ Mon Dec 23 12:02:33 UTC 2002 Johnny Willemsen <jwillemsen@remedy.nl> Sun Dec 22 11:41:56 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Connector.cpp: Fixed fuzz errors. + * ace/Connector.cpp: Fixed fuzz errors. Sun Dec 22 11:37:44 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * THANKS: Added Daniel White to the hall of fame! + * THANKS: Added Daniel White to the hall of fame! Sun Dec 22 11:31:53 2002 Craig Rodrigues <crodrigu@bbn.com> - * apps/Gateway/Gateway/Makefile: - * apps/Gateway/Peer/Makefile: Add $(ACELIB) to $(ACE_SHLIBS) so - that -lACE is added to the linkline when a shared library is - built. Solves linking problem MacOS X. Thanks to John Zorko and - Daniel white <ygor@comcast.net> for pointing this out. + * apps/Gateway/Gateway/Makefile: + * apps/Gateway/Peer/Makefile: Add $(ACELIB) to $(ACE_SHLIBS) so + that -lACE is added to the linkline when a shared library is + built. Solves linking problem MacOS X. Thanks to John Zorko and + Daniel white <ygor@comcast.net> for pointing this out. Sun Dec 22 10:25:35 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Connector.cpp: Fixed a warning with g++. + * ace/Connector.cpp: Fixed a warning with g++. Sun Dec 22 09:18:00 2002 Balachandran Natarajan <bala@isis-server.isis.vanderbilt.edu> - * ace/Connector.cpp: - * ace/Connector.h: Added a lock to the ACE_Connector class to - make the following atomic + * ace/Connector.cpp: + * ace/Connector.h: Added a lock to the ACE_Connector class to + make the following atomic - . Registration of AST with the handler_map_ - . Registration of the ACE_Connector with the Reactor - . and the registration of timers with the Reactor's timer - queue. + . Registration of AST with the handler_map_ + . Registration of the ACE_Connector with the Reactor + . and the registration of timers with the Reactor's timer + queue. - If they are not atomic, it throws open race conditions such as - the ones documented in BUG 1405. This is not complete fix for - BUG 1405, since it taxes applications using only one thread with a - lock. This is a short-term fix to prevent the race condition - from occuring. + If they are not atomic, it throws open race conditions such as + the ones documented in BUG 1405. This is not complete fix for + BUG 1405, since it taxes applications using only one thread with a + lock. This is a short-term fix to prevent the race condition + from occuring. - Other relevant changes include + Other relevant changes include - . the AST is now refcounted - . the AST cannot be deleted directly. The lifetime of the AST - can only be manipulated using the refcount on the AST. + . the AST is now refcounted + . the AST cannot be deleted directly. The lifetime of the AST + can only be manipulated using the refcount on the AST. Sat Dec 21 18:27:42 2002 Steve Huston <shuston@riverace.com> @@ -674,10 +681,10 @@ Thu Dec 19 18:47:29 2002 Steve Huston <shuston@riverace.com> Thu Dec 19 16:47:17 2002 Douglas C. Schmidt <schmidt@ace.cs.wustl.edu> - * ace/SOCK_Dgram.h: - ace/SOCK_CODgram.h: Added metatype information for - ACE_SOCK_Dgram and ACE_SOCK_CODgram. Thanks to Sankaranarayanan - K V <sankar@migv.mot.com> for reporting this. + * ace/SOCK_Dgram.h: + ace/SOCK_CODgram.h: Added metatype information for + ACE_SOCK_Dgram and ACE_SOCK_CODgram. Thanks to Sankaranarayanan + K V <sankar@migv.mot.com> for reporting this. Thu Dec 19 10:02:56 2002 Rich Seibel <seibel_r@ociweb.com> @@ -903,39 +910,39 @@ Mon Dec 16 13:14:34 2002 Phil Mesnier <mesnier_p@ociweb.com> * ace/CDR_Stream.h: * ace/CDR_Stream.i: Preparing for codeset negotiation. The ACE input/output CDR - streams' support for wide character IO is enhanced by adding a - flag to allow or disallow wchar i/o regardless of whether a - translator is present. There are times when wchar is not - allowed, such as GIOP 1.0, or when a native codeset for wchar - was not specified. In CORBA, a native wchar codeset must be - specified, it may not be defaulted. Within the ACE CDR streams, - wchar i/o is allowed by default, to ensure backwards - compatibility. - - I fixed a few of the reading and writing methods to ensure that - the good_bit_ flag is cleared if an error occurs. This was not - happening in all cases, leading to errors elsewhere. - - I have extended the definition of the - ACE_[W]Char_Codeset_Translator classes to make them reference - counted objects. There are some codesets that support shifting - and may be stateful. By using reference counting, the CDR - streams don't have to worry about ownership of stateful - translators. + streams' support for wide character IO is enhanced by adding a + flag to allow or disallow wchar i/o regardless of whether a + translator is present. There are times when wchar is not + allowed, such as GIOP 1.0, or when a native codeset for wchar + was not specified. In CORBA, a native wchar codeset must be + specified, it may not be defaulted. Within the ACE CDR streams, + wchar i/o is allowed by default, to ensure backwards + compatibility. + + I fixed a few of the reading and writing methods to ensure that + the good_bit_ flag is cleared if an error occurs. This was not + happening in all cases, leading to errors elsewhere. + + I have extended the definition of the + ACE_[W]Char_Codeset_Translator classes to make them reference + counted objects. There are some codesets that support shifting + and may be stateful. By using reference counting, the CDR + streams don't have to worry about ownership of stateful + translators. Finally, the translators now have pure virtual methods that - return the native and translated codeset ID values. These values - refer to entries in the DCE Code and Character Set Registry, - which is the reference for CORBA codeset ids. Anyone - implementing a char or wchar codeset will have to supply a - numeric value for the native codeset (ncs) and the translated - codeset (tcs). NCS refers to the codeset of the text on the - application side of the stream, and TCS refers to the codset of - the text on the "wire" side of the stream. + return the native and translated codeset ID values. These values + refer to entries in the DCE Code and Character Set Registry, + which is the reference for CORBA codeset ids. Anyone + implementing a char or wchar codeset will have to supply a + numeric value for the native codeset (ncs) and the translated + codeset (tcs). NCS refers to the codeset of the text on the + application side of the stream, and TCS refers to the codset of + the text on the "wire" side of the stream. These changes are not everything that is required to support - codeset negotiation, but it is all the changes required to the - CDR streams. + codeset negotiation, but it is all the changes required to the + CDR streams. Sat Dec 14 08:56:23 2002 Douglas C. Schmidt <schmidt@tango.doc.wustl.edu> @@ -1240,14 +1247,14 @@ Sun Dec 08 01:17:26 2002 Nanbor Wang <nanbor@cs.wustl.edu> Thu Dec 5 20:30:56 2002 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu> * ace/Event_Handler.cpp: Updated the implementation of - ACE_Event_Handler::read_adapter() so that the code is consistent - with the stated semantics. Thanks to Ron Muck <rlm@sdiusa.com> - for the fix. + ACE_Event_Handler::read_adapter() so that the code is consistent + with the stated semantics. Thanks to Ron Muck <rlm@sdiusa.com> + for the fix. * ace/OS_Dirent.inl (readdir_r): Enhanced the - ACE_OS_Directn::readdir_r() wrapper facade so that it returns 1 - (i.e., "done") when an error occurs. Thanks to Abhay Kulkarni - <Abhay.Kulkarni@veritas.com> for reporting this. + ACE_OS_Directn::readdir_r() wrapper facade so that it returns 1 + (i.e., "done") when an error occurs. Thanks to Abhay Kulkarni + <Abhay.Kulkarni@veritas.com> for reporting this. Sat Dec 7 19:54:12 2002 Steve Huston <shuston@riverace.com> @@ -1635,7 +1642,7 @@ Sat Nov 23 18:12:37 2002 Steve Huston <shuston@riverace.com> on AIX with g++. Without this, the compiler and linker don't get things done right and static data members in templates don't get instantiated correctly. Thanks to Max F. Bilyk <mbilyk@europe.com> - for reporting this problem. + for reporting this problem. * ace/SSL/Makefile: Set LIBS = $(ACELIB). Else it won't link on AIX, g++. diff --git a/bin/aix_shr b/bin/aix_shr index f76770f74a3..16daded8e02 100755 --- a/bin/aix_shr +++ b/bin/aix_shr @@ -14,8 +14,10 @@ for possible in "$TMPDIR" "$TEMP" "$TMP"; do fi done -## Send the output to a temporary file and save the return code -$@ -DTAO_INSTANTIATING_TEMPLATES > "$usrtmp/$output" 2>&1 +## Send the output to a temporary file and save the return code. +## This macro is being passed during the to avoid #include errors +## while instantiating templates. +$@ -DACE_VACPP_INSTANTIATING_TEMPLATES > "$usrtmp/$output" 2>&1 status=$? ## Print the output to stdout, but filter duplicate symbol warnings from |