Remove now-redundant test result check
Since 349eadc58faff851ee8dcbb43c702f0ddecfcbb8, test_fail() reports
the first failure. So it's safe to call test_fail() again to report a
cleanup failure when we don't want to potentially erase information
about an earlier failure.
The behavior of mbedtls_test_helper_is_psa_pristine() changes if
test_info.result was neither TEST_RESULT_SUCCESS nor
TEST_RESULT_FAILED, but this should not matter since a skipped test
should not cause mbedtls_test_helper_is_psa_pristine() to fail.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
diff --git a/tests/include/test/psa_crypto_helpers.h b/tests/include/test/psa_crypto_helpers.h
index 9263a93..e62bc04 100644
--- a/tests/include/test/psa_crypto_helpers.h
+++ b/tests/include/test/psa_crypto_helpers.h
@@ -46,10 +46,7 @@
msg = "Some slots are still marked as locked.";
}
- /* If the test has already failed, don't overwrite the failure
- * information. Do keep the stats lookup above, because it can be
- * convenient to break on it when debugging a failure. */
- if( msg != NULL && test_info.result == TEST_RESULT_SUCCESS )
+ if( msg != NULL )
test_fail( msg, line, file );
return( msg == NULL );