Using BoostTest to run unit tests multiple times without quitting test runner

I want to use Boost Test to run unit tests multiple times without quitting
the test runner. Here is the pseudo-code to demonstrate what I want to do:

void main()
{
	for(int i = 0; i < 10; ++i)
	{
		char *arguments[] ={	"MyApplication.exe", 
					"--detect_memory_leak=0"
				};

		boost::unit_test::unit_test_main( &init_function, 2, arguments);
	}
}

The first time unit_test_main is called, all the boost tests run fine but in
the second iteration, boost test throws an exception.

I looked at the internal code and found that framework is initialized
everytime unit_test_main is called and if framework is already initialized,
boost test doesn't like initializing framework again.

Please let me know if there is any way to run units test multiple times
without quitting test runner.

--
View this message in context: http://boost.2283326.n4.nabble.com/Using-BoostTest-to-run-unit-tests-multiple-times-without-quitting-test-runner-tp4630095.html
Sent from the Boost - Testing mailing list archive at Nabble.com.

[cray] Failing almost every test?

I'm not sure whose running them but the Cray tests are failing every last 
Math lib test with an inscrutable internal error:

CC-1353 crayc++: INTERNAL File = 
/opt/gcc/4.4.4/snos/include/g++/x86_64-suse-linux/bits/c++config.h, Line = 
1431
  CIF Error

  #endif // _GLIBCXX_CXX_CONFIG_H
  ^

CC-2116 crayc++: INTERNAL
  "/opt/cray/cce/8.0.3/CC/x86-64/lib/ccfe" was terminated due to receipt of 
signal 013:  Segmentation fault.

See http://tinyurl.com/cusmk6g

Is this a compiler configuration error, or ??

Thanks, John. 

Trunk test failed, missing macro_default_test.cpp

FYI, this morning's trunk regression failed to run:

error: Unable to find file or target named
error:     'macro_default_test.cpp'
error: referred from project at
error:     '../libs/filesystem/test'

Upcoming failures

Hi,

My system(s) are moving to new hardware, os, etc.. and will likely  
adversely impact both the Sandia testers as well as the trunk and  
release reporting.  I'll attempt to stay on top of problems as they  
arise, just wanted to give you a heads up that problems may appear  
immediately and continue for the next week(s).

Sorry about the late notice.

-- Noel

NIghtly trunk tests broken

Hi,

Did someone forget to commit a file?

error: Unable to find file or target named
error:     'test_6498_pass.cpp'
error: referred from project at
error:     '../libs/ratio/test'

-- Noel

Sandia-clang-trunk - chrono link errors

_______________________________________________
Boost-Testing mailing list
Boost-Testing <at> lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-testing

Sandia-darwin-clang-trunk - chrono errors

_______________________________________________
Boost-Testing mailing list
Boost-Testing <at> lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-testing

Sandia-gcc-4.6.2_0x Fri, 30 Mar 2012

_______________________________________________
Boost-Testing mailing list
Boost-Testing <at> lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-testing

How to deactivate system error catching without using the commandline?

Dear all,

I need to deactivate system error catching without using the commandline.

I was looking for some possibility to do something like

get_execmonitor().p_catch_system_errors = false;

but I did not find any possibility, and no documentation how the exec
monitor interacts with init_unit_test_suite and main

My current test where I would like to deactivate system error catching
looks as follows (omitted application specific stuff):

...
 7 #define BOOST_TEST_DYN_LINK
 8 #define BOOST_TEST_MODULE "testDLVProcess"
 9 #include <boost/test/unit_test.hpp>
 10 #include <boost/thread.hpp>
...
 70 BOOST_AUTO_TEST_CASE ( testDLVProcess )
 71 {
...
 91 }

This is a cpp file, it is compiled into an executable and then executed.
I would like to set p_catch_system_errors=false within the cpp file, so
that I do not require a wrapper script.

After reading boost source code for about 1 hour I decided to ask on
the mailing list, because I think there must be an easy way. ;)

Best,
Peter

Reboot wowbagger

Hi,

My name is DongInn Kim, the sysadmin in the CREST.

We need to reboot the boost server (wowbagger.osl.iu.edu) to replace the
old RAM. The old one seems to act not normal and we will replace it with a
new one to fix any issues that may happen from the bad RAM.

The reboot time: 3:00PM-3:30PM (E.T.) today (Feb 16, 2012)

I am sorry for the very short notice.

Regards,

- DongInn

Why both tools_bb and tools_bjam?

Hi,

Anyone know the reason that "run.py" checks out the bjam sources twice?
One in "tools_bjam" and one in "tools_bb/engine"?

Snip of run.py output:

# Getting Boost.Build v2...
# Executing SVN command "svn --non-interactive co
http://svn.boost.org/svn/boost/trunk/tools/build/v2 tools_bb"
Checked out revision 77003.
# Getting Boost.Jam...
# Executing SVN command "svn --non-interactive co
http://svn.boost.org/svn/boost/trunk/tools/build/v2/engine tools_bjam"
Checked out revision 77003.

Thanks,
-Steve

_______________________________________________
Boost-Testing mailing list
Boost-Testing <at> lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-testing