| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
| |
* tests/Bug_3319_Regression_Test.cpp:
Fix the test to be consistent with revised service config
behavior. The test had assumed that ACE_Service_Config::open()
would return 0 even if svc.conf were not present, but since it
was setting the ignore_default_svc_conf_file to 0, the restore
behavior of the service configurator is to fail with ENOENT in
that case.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* ace/TSS_T.inl:
* ace/TSS_T.h:
* ace/TSS_T.cpp:
Mostly fixed bug 3539. It doesn't fix a case when main thread
doesn't own ACE_TSS but still uses it. Fixing the later is not
straight forward and will require a lot more efforts.
Also rearranged the code so that it has less duplicates.
* tests/run_test.lst:
Marked the test as fixed.
|
|
|
|
|
|
|
| |
* tests/run_test.lst:
* tests/Bug_3539_Regression_Test.cpp:
* tests/tests.mpc:
Added a test for bug 3539 as non-fixed.
|
|
|
|
|
| |
* *.mpc:
Use dynamicflags += instead of dynamicflags =
|
|
|
|
|
| |
* tests/Log_Msg_Test.cpp:
Added nul terminator at the end of a string that was missing.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
* ace/OS_NS_dirent.cpp:
Small code improvements to closedir
* tests/FlReactor_Test.cpp:
Small changes
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* tests/Dev_Poll_Reactor_Test.cpp:
Fixed 2 memory leaks in this test. Previously both client and
server svc handlers were leaked. After fixing that it turned out
that client is not able to finish properly as its reactor is not
shutdown. The later was happening because server was reaching
desired exit condition faster and once it get closed client could
never get a chance to reach its own exit condition. So, in short,
there is no point in this test to run client side reactor if the
client's svc handler got closed.
|
|
|
|
|
|
|
|
|
| |
* examples/QOS/Diffserv/run_test.pl:
* examples/Reactor/TP_Reactor/run_test.pl:
Use TestTarget
* tests/OS_Test.cpp:
Prefix increment, const changes
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* tests/SSL/SSL_Asynch_Stream_Test.cpp:
* tests/SSL/Bug_2912_Regression_Test.cpp:
* tests/Proactor_UDP_Test.cpp:
* tests/MT_Reference_Counted_Event_Handler_Test.cpp:
* tests/Proactor_Test_IPV6.cpp:
* tests/Bug_2740_Regression_Test.cpp:
* tests/Dev_Poll_Reactor_Test.cpp:
* tests/Proactor_Test.cpp:
* tests/TP_Reactor_Test.cpp:
All these tests need to block SIGPIPE signal. POSIX defines that
newly created threads must inherit signal mask from the creating
thread. The later doesn't happen at least on Solaris 10. With this
change each new thread in the above tests will block SIGPIPE signal
individually.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* ace/Service_Types.h
* ace/Service_Types.inl
* ace/Service_Types.cpp
Added service_type_ attr to expose the type of service being
managed to the ACE_Service_Repository. This allows the ASR to
manage the lifecycle of the ACE_Module and ACE_Stream in order to
avoid a double delete of ACE_Module at shutdown.
ACE_Stream_Type::fini was modified to not call ACE_Module_Type::fini
as this will results in a double delete when ACE_Service_Repository::fini
is called. ACE_Stream_Type::remove(module_name) was modified to not call
ACE_Module_Type::fini as this will results in a double delete
when ACE_Service_Repository::fini is called also.
* ace/Service_Repository.cpp
Modified ASR::fini to iterate over the service_array_ twice.
ACE_Service_Type::fini is called on all ACE_Stream_Type and
ACE_Service_Object_Type instances first. Then, fini is called on
all ACE_Modules_Type instances. All calls to fini are done in the
order the services appear in the ASR::service_array_ except for
the grouping described here. The calls to ACE_Module_Type::fini
must be done last because ACE_Stream_Type::fini accesses the Modules
so they must not be deleted by a call to ACE_Module_Type::fini
before that.
* tests/Bug_3334_Regresssion_Test.cpp
Added call to ACE_Service_Repository::fini_svcs() to capture all
debug output before the application exits.
* tests/run_tests.lst
This fixes Bugzilla #3334
|
|
|
|
|
| |
* tests/Thread_Timer_Queue_Adapter_Test.cpp:
Fixed compilation problems with single threaded builds
|
|
|
|
|
| |
* tests/MT_NonBlocking_Connect_Test.cpp:
Fixed compilation problem for single threaded builds.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
* tests/NDDS_Timer_Test.cpp:
Resolved test issues.
|
| |
|
|
|
|
|
|
|
| |
* tests/NDDS_Timer_Test.cpp:
* tests/run_test.lst:
* tests/tests.mpc:
Added timer test which uses the ndds timers.
|
| |
|
|
|
|
|
| |
* tests/run_test.pl:
Added -l option which allows to run tests selectively.
|
|
|
|
|
|
| |
* tests/MT_NonBlocking_Connect_Test.cpp:
And another attempt to workaround WFMO reactor. The previous
changes could break the test on other platforms.
|
|
|
|
|
| |
* tests/MT_NonBlocking_Connect_Test.cpp:
Fixed a typo in previous commit.
|
|
|
|
|
|
| |
* tests/MT_NonBlocking_Connect_Test.cpp:
Fixed a test on Windows for WFMO reactor which doesn't support
work_pending() function.
|
|
|
|
|
| |
* tests/NonBlocking_Conn_Test.cpp:
Made logging a bit more verbose.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Committed the changes that I had to revert before x.7.7.
* ace/Strategies_T.cpp:
Changed the code so that close() is called in case of failure
instead of destroy().
* ace/Connector.cpp:
* ace/Connector.h:
Changed the fix for bug#3731. Now NBCH adds a reference in
constructor and removes it in destructor if the SVC_HANDLER that
it owns is reference counted. This is a cleaner solution than the
one used before.
* tests/NonBlocking_Conn_Test.h:
* tests/Process_Strategy_Test.cpp:
* tests/NonBlocking_Conn_Test.cpp:
Extended NonBlocking_Conn_Test and changed Process_Strategy_Test
because of the change in Strategies_T.cpp.
* tests/MT_NonBlocking_Connect_Test.cpp:
* tests/tests.mpc:
* tests/run_test.lst:
Added a new test. This can be a reproducer for the bug#3731 which
currently doesn't have its own test.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* ace/High_Res_Timer.cpp
* ace/High_Res_Timer.h
* ace/High_Res_Timer.inl
Use 64bit calculations to increase our precision. If you want to have the 32bit
calculations, add ACE_USE_WINDOWS_32BIT_HIGH_RES_TIMER_CALCULATIONS as define. If
the scoreboard doesn't show any platforms requiring 32bit, we will remove that code
before the next micro release goes out
Thanks to Alon Diamant <diamant dot alon at gmail dot com> for supplying the patches.
This fixes bugzilla 3703.
* ace/Timer_Queue_Adapters.cpp
* ace/Timer_Queue_Adapters.h
* ace/Timer_Queue_Adapters.inl
Make it possible to use a customer event handler in ACE_Thread_Timer_Queue_Adapter.
Thanks to Alon Diamant <diamant dot alon at gmail dot com> for supplying the patches.
This fixes bugzilla 3614
* tests/tests.mpc:
* tests/Thread_Timer_Queue_Adapter_Test.cpp:
New test for testing custom event handlers in ACE_Thread_Timer_Queue_Adapter.
Thanks to Alon Diamant <diamant dot alon at gmail dot com> for creating
this new test
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* ace/config-g++-common.h:
PPU 4.1.1 for PPC doesn't have atomic builtin support
for datatypes of size 1 and 2
* ace/Atomic_Op.h:
* ace/Atomic_Op.inl:
Added defines to just disable short/unsigned short/bool specializations
* tests/run_test.lst:
Enable 2610 but mark it as not fixed
* ace/Process_Manager.cpp:
Layout change
* ace/WIN32_Asynch_IO.cpp:
Const changes
|
|
|
|
|
|
|
|
| |
* ace/Svc_Handler.cpp:
* tests/run_test.lst:
Fixed bug#2609 and enabled the test for it. Thanks to Milan
Cvetkovic <milan dot cvetkovic at mpathix dot com> for
contributing this.
|
|
|
|
|
| |
* tests/Atomic_Op_Test.cpp:
Added tests for short/unsigned short/bool
|
|
|
|
|
| |
* tests/Atomic_Op_Test.cpp:
Print the number of iterations
|
|
|
|
|
| |
* tests/Atomic_Op_Test.cpp:
Also measure postfix increment/decrement
|
|
|
|
|
| |
* tests/Atomic_Op_Test.cpp:
Added more checks to make sure Atomic_Op does work
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* tests/Atomic_Op_Test.cpp:
Rewritten this test to use a template method so that we can
test atomic op easily for various data types
* ace/Time_Value.cpp:
Layout changes and fixed compile problem with CB2010. This fixes
bugzilla 3835, thanks to Jan Kalin <jan dot kalin at zag dot si>
for reporting this.
* ace/config-win32-borland.h:
CB also has long timevalue mismatch
|
| |
|
| |
|
| |
|