MANDATORY_SCRIPT_VERIFY_FLAGS only exists to affect the banning behaviour
in AcceptToMemoryPool for transactions that fail to pass standard flags.
Since these tests are not about what kinds of script failures are
bannable / not bannable, they shouldn't care about what is in
MANDATORY_SCRIPT_VERIFY_FLAGS. (and such banning behaviour should in fact
be tested in the functional tests)
(cf: in Core, this flagset is only used in policy.h)