diff options
author | levine <levine@ae88bc3d-4319-0410-8dbf-d08b4c9d3795> | 1998-10-16 19:30:59 +0000 |
---|---|---|
committer | levine <levine@ae88bc3d-4319-0410-8dbf-d08b4c9d3795> | 1998-10-16 19:30:59 +0000 |
commit | 53c42868e6ae47dbed131b4bf2f1efb2a37bc5a9 (patch) | |
tree | e1ad3d28d1c44aa196e4dca4604f1aeb949642b0 | |
parent | 14cff64f3a41ef1cce7c8d07ccdc9528c7e73d1a (diff) | |
download | ATCD-53c42868e6ae47dbed131b4bf2f1efb2a37bc5a9.tar.gz |
*** empty log message ***
-rw-r--r-- | ChangeLog-98b | 95 |
1 files changed, 49 insertions, 46 deletions
diff --git a/ChangeLog-98b b/ChangeLog-98b index a3afefb6935..7a69d6268ae 100644 --- a/ChangeLog-98b +++ b/ChangeLog-98b @@ -8,75 +8,75 @@ Thu Oct 15 23:50:41 1998 Douglas C. Schmidt <schmidt@cs.wustl.edu> Thu Oct 15 22:05:02 1998 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - * ace/ACE.cpp (unique_name): Moved the arguments around in the - sprintf() in order to generate a better "unique name". Thanks - to Bob McWhirter for suggesting this. + * ace/ACE.cpp (unique_name): Moved the arguments around in the + sprintf() in order to generate a better "unique name". Thanks + to Bob McWhirter for suggesting this. Thu Oct 15 21:20:36 1998 Irfan Pyarali <irfan@cs.wustl.edu> - * ace/Hash_Map_Manager.cpp (operator-- and operator++): - * ace/Map_Manager.cpp (operator-- and operator++): + * ace/Hash_Map_Manager.cpp (operator-- and operator++): + * ace/Map_Manager.cpp (operator-- and operator++): - Fixed by value and by reference return values. + Fixed by value and by reference return values. Thu Oct 15 20:16:46 1998 James CE Johnson <jcej@chiroptera.tragus.org> - * docs/ACE-FMM.html: Updated with Bob McWhirter's latest changes. - Discussion of race conditions when spawning threads from within a constructor. + * docs/ACE-FMM.html: Updated with Bob McWhirter's latest changes. + Discussion of race conditions when spawning threads from within a constructor. Thu Oct 15 18:30:58 1998 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu> - * ace/OS.h: Added ACE_HTONS(x) macros to complement ACE_HTONL(x) - macros. Thanks to Balaji Srinivasan <balaji@cplane.com> - for reporting this. + * ace/OS.h: Added ACE_HTONS(x) macros to complement ACE_HTONL(x) + macros. Thanks to Balaji Srinivasan <balaji@cplane.com> + for reporting this. - * ace/[Hash_]Map_Manager.cpp: Postfix and prefix were inverted. - They are correct now. Thanks to Balaji Srinivasan - <balaji@cplane.com> for reporting this. + * ace/[Hash_]Map_Manager.cpp: Postfix and prefix were inverted. + They are correct now. Thanks to Balaji Srinivasan + <balaji@cplane.com> for reporting this. Thu Oct 15 18:02:51 1998 Darrell Brunsch <brunsch@cs.wustl.edu> - * ace/config-win32.h: For non-winnt systems, define - ACE_LACKS_RECVMSG and ACE_LACKS_SENDMSG. Thanks to Bill Hall - <hts@gte.net> for pointing this out. BTW, Bill was the 400th - individual to contribute an ACE/TAO bug report, thanks to - everyone for all their help over the years! + * ace/config-win32.h: For non-winnt systems, define + ACE_LACKS_RECVMSG and ACE_LACKS_SENDMSG. Thanks to Bill Hall + <hts@gte.net> for pointing this out. BTW, Bill was the 400th + individual to contribute an ACE/TAO bug report, thanks to + everyone for all their help over the years! Thu Oct 15 16:50:09 1998 Nanbor Wang <nanbor@cs.wustl.edu> - * ace/SOCK_Dgram_Mcast.i (set_option): Need to convert the - <optval> from char to int for Win32 because the shortest length - of <optval> must be at least sizeof (int) on Win32. Thanks - to Johan Lundin <johan@lundin.com> for reporting this. + * ace/SOCK_Dgram_Mcast.i (set_option): Need to convert the + <optval> from char to int for Win32 because the shortest length + of <optval> must be at least sizeof (int) on Win32. Thanks + to Johan Lundin <johan@lundin.com> for reporting this. Thu Oct 15 16:40:41 1998 Steve Huston <shuston@riverace.com> - * tests/Thread_Manager_Test.cpp: Allow suspend/resume to be tried on - pthreads, but don't allow any error other than ENOTSUP. This - catches a bug in Thread_Manager that lost the error values on an - error. + * tests/Thread_Manager_Test.cpp: Allow suspend/resume to be tried on + pthreads, but don't allow any error other than ENOTSUP. This + catches a bug in Thread_Manager that lost the error values on an + error. - * ace/Thread_Manager.cpp (ACE_THR_OP macro): Don't remove a thread - from the active threads list if an operation fails with ENOTSUP. - This was causing calls to suspend_grp/resume_grp to erase the list - of threads in a group on pthreads, then wait_grp didn't wait because - it thought there weren't any more threads. + * ace/Thread_Manager.cpp (ACE_THR_OP macro): Don't remove a thread + from the active threads list if an operation fails with ENOTSUP. + This was causing calls to suspend_grp/resume_grp to erase the list + of threads in a group on pthreads, then wait_grp didn't wait because + it thought there weren't any more threads. Thu Oct 15 12:43:46 1998 Nanbor Wang <nanbor@cs.wustl.edu> - * ace/Task.cpp (activate): Changed to activate new active objects - using the same grp_id if this task has been activated before and - no new grp_id is specified in activate method. Also, the task's - grp_id will only get updated if it has never been updated even - if <force_active> is set. This prevent the original <grp_id_> - get lost but the consequence if that if users select to - force_activate new tasks with a different grp_id, the new grp_id - will get lost and the only way to wait for these threads is to - either perform a wait for all threads, or wait for each - individual thread by waiting on the thread handles returned by - the activate method. Thanks to Bob McWhirter - <bob@netwrench.com> for motivating this change. - (ACE_Task_Base): Changed to initialize <grp_id_> to -1. + * ace/Task.cpp (activate): Changed to activate new active objects + using the same grp_id if this task has been activated before and + no new grp_id is specified in activate method. Also, the task's + grp_id will only get updated if it has never been updated even + if <force_active> is set. This prevent the original <grp_id_> + get lost but the consequence if that if users select to + force_activate new tasks with a different grp_id, the new grp_id + will get lost and the only way to wait for these threads is to + either perform a wait for all threads, or wait for each + individual thread by waiting on the thread handles returned by + the activate method. Thanks to Bob McWhirter + <bob@netwrench.com> for motivating this change. + (ACE_Task_Base): Changed to initialize <grp_id_> to -1. Thu Oct 15 10:07:13 1998 David L. Levine <levine@cs.wustl.edu> @@ -88,7 +88,10 @@ Thu Oct 15 09:54:07 1998 David L. Levine <levine@cs.wustl.edu> it doesn't halt on g++/Solaris or DU 4.0. * include/makeinclude/wrapper_macros.GNU: for g++, added - -DACE_LACKS_PRAGMA_ONCE to CCFLAGS. + -DACE_LACKS_PRAGMA_ONCE to CCFLAGS. Thanks to Kirk Ellet + <kellet@mdc.com> for pointing out that is has to be done + in CCFLAGS rather than as a #define in OS.h, because + ace/Synch_T.h uses the pragma before config.h is #included. Wed Oct 14 23:20:45 1998 Douglas C. Schmidt <schmidt@cs.wustl.edu> |