I have a VC++ project which has two build configurations, each causes a different 3rd-party library to be used. e.g Release_A & Release_B use libraries A & B, and wrap their functionality.
We use a set of boost.test unit tests so we can be sure both versions work the same, so we compile a test-suite app against both build_configurations to give test_A.exe & test_B.exe, and these tests emit test_A.xml & test_B.xml.
The problem is that the test XML does not distinguish in any way which version was tested, so when we pull these two XML files into a tool like Jenkins, it's not possible to tell which configuration is to blame for any failures.
Is there a neat way to do this in the C++ testsuite code somehow?