In SCM8.5, after getInterMergeCandidates(), the merge candidate bi-pred restriction, i.e. xRestrictBipredMergeCand(), is done before the merge candidate rounding, i.e. roundMergeCandidates().
But in the Spec, the merge candidate list generation follows the below steps, where the bi-pred restriction should be done after the merge candidate rounding. This caused the inconsistence between SCM and Spec.

Step 1: Derive the merge candidates (step 1-8 in spec 8.5.3.2.2, and SCM function is getInterMergeCandidates())
Step 2: Round the merge candidates if needed (step 9 in spec 8.5.3.2.2, and SCM function is roundMergeCandidates())
Step 3: Disallow 8x4 and 4x8 bi-pred merge candidates(step 10 in spec 8.5.3.2.2, and the SCM function is xRestrictBipredMergeCand())
Step 4: Disallow bi-pred merge candidates for some 8x8 CUs (the below part in spec 8.5.3.2.1, and the SCM function is xRestrictBipredMergeCand())

When all of the following conditions are true, refIdxL1 is set equal to −1 and predFlagL1 is set equal to 0.
– predFlagL0 is equal to 1.
– predFlagL1 is equal to 1.
– nPbSw is equal to 8.
– nPbSh is equal to 8.
– TwoVersionsOfCurrDecPicFlag is equal to 1.
– noIntegerMvFlag is equal to 1.
– identicalMvs is equal to 0.