gtest os.dll_address_to_function_and_library_name_vm fails

The following test fails in the JDK17 CI:

gtest/GTestWrapper.java

Here's a snippet from the log file:

[ RUN ] os.dll_address_to_function_and_library_name_vm
./open/test/hotspot/gtest/runtime/test_os.cpp:745: Failure
Expected: (::strstr(output, "(")) != ((char*)__null), actual: NULL vs NULL
./open/test/hotspot/gtest/runtime/test_os.cpp:745: Failure
Expected: (::strstr(output, "(")) != ((char*)__null), actual: NULL vs NULL
./open/test/hotspot/gtest/runtime/test_os.cpp:745: Failure
Expected: (::strstr(output, "(")) != ((char*)__null), actual: NULL vs NULL
./open/test/hotspot/gtest/runtime/test_os.cpp:745: Failure
Expected: (::strstr(output, "(")) != ((char*)__null), actual: NULL vs NULL
[ FAILED ] os.dll_address_to_function_and_library_name_vm (37 ms)

The failing test case, os.dll_address_to_function_and_library_name_vm,
was recently added by the following fix:

    JDK-8258606 os::print_signal_handlers() should resolve the function name of the handlers

This failure has happened in four of the last seven Tier1
job sets and only on macOSX (for now). I'm starting this
bug as a P2.