libsemanage/tests: make "make test" fail when a CUnit test fails

When modifications to libsemanage functions break the test cases tested
with the CUnit framework, "make test" currently succeeds, even though it
prints an output similar to:

    Suite: semanage_store
      Test: semanage_store_access_check ...passed
      Test: semanage_get_lock ...passed
      Test: semanage_nc_sort ...passed
    Suite: semanage_utilities
      Test: semanage_is_prefix ...passed
      Test: semanage_split_on_space ...FAILED
        1. test_utilities.c:150  - CU_ASSERT_STRING_EQUAL(temp,"baz")
      Test: semanage_split ...passed
      Test: semanage_list ...passed
      Test: semanage_str_count ...passed
      Test: semanage_rtrim ...passed
      Test: semanage_str_replace ...passed
      Test: semanage_findval ...passed
      Test: slurp_file_filter ...passed

Like commit 2489b50a9162 ("libsepol: make "make test" fails when a CUnit
test fails") did for libsepol tests, modify the logic of function
do_tests() to return an error value when there has been at least one
failure. This makes "make test" fail as expected.

Signed-off-by: Nicolas Iooss <nicolas.iooss@m4x.org>
1 file changed