>> In this thread, we should resolve issue #1 using the patch provided in [5]. If we want to prevent
>> the newly added spec test triggering other issues (such as issue #3), revert the spec test to
>> the test cases provided in [6] may be a good choice.
>
>Absolutely correct.
If the spec test need to be reverted, the patch in [5] may be updated again.
>Sorry for confusing. I meant to say that issues #2 and #3 are the same, and there is a related
>discussion in [4].
Ok, I will investigate this issue again. And I will post my analysis in [4] if something new is found.