It seems to me that the most likely area of SRU regression would be an accidental functional change in 50_check_unity_support in the case of a user using a previous release. For SRU verification, in addition to the test case, please could you also check that the previous release correctly sets LIBGL_ALWAYS_SOFTWARE in the four cases of Unity vs. not-Unity and unity_support_test and !unity_support_test?
It seems to me that the most likely area of SRU regression would be an accidental functional change in 50_check_ unity_support in the case of a user using a previous release. For SRU verification, in addition to the test case, please could you also check that the previous release correctly sets LIBGL_ALWAYS_ SOFTWARE in the four cases of Unity vs. not-Unity and unity_support_test and !unity_ support_ test?