| = Running Qpid C++ tests = |
| |
| General philosophy is that "make check" run all tests by default, but |
| developers can run tests selectively as explained below. |
| |
| == Valgrind == |
| |
| By default we run tests under valgrind to detect memory errors if valgrind |
| is present. ./configure --disable-valgrind will disable it. |
| |
| Default valgrind options are specified in .valgrindrc-default, which a |
| checked-in file. The actual options used are in .valgrindrc which is a |
| local file. Normally it is a copy of valgrindrc-default but you can |
| modify at will. |
| |
| Supressed errors are listed in .valgrind.supp. If you want to change |
| suppressions for local testing, just modify .valgrindrc to point to a |
| different file. Do NOT add suppressions to .valgrindrc.supp unless |
| they are known problems outside of Qpid that can't reasonably be |
| worked around in Qpid. |
| |
| |
| == Unit Tests == |
| Unit tests use the boost test framework, and are compiled to rogram |
| unit_test |
| There are several options to control how test results are displayed, |
| see |
| http://www.boost.org/doc/libs/1_35_0/libs/test/doc/components/utf/parameters/index.html |
| |
| NOTE: some unit tests are written as CppUnit plug-ins, we are moving away |
| from CppUnit so new tests should use the boost framework. |
| |
| CppUnit tests are run by the script run-unit-tests. |
| |
| == System Tests == |
| |
| System tests are self contained AMQP client executables or scripts. |
| They are listed in the TESTS make variable, which can be over-ridden. |
| |
| The ./start_broker "test" launches the broker, ./stop_broker" stops it. |
| Tests in between assume the broker is running. |
| |
| ./python_tests: runs ../python/run_tests. This is the main set of |
| system testss for the broker. |
| |
| Other C++ client test executables and scripts under client/test are |
| system tests for the client. |
| |
| By setting TESTS in a make command you can run a different subset of tests |
| against an already-running broker. |
| |
| |
| |
| |