There is a new patch available.
I've addressed all your requests.
Luís Soares wrote:
> Hi Alfranio,
> Nice Work. Please find my review comments below. I share some
> of Zhenxing's concerns/comments and I am only putting in this
> review some of requests for clarification.
> Approved after consideration of suggestions and request for
> REQUIRED CHANGES
> R1. In this test case, why do we need the slave? Couldn't this
> test just use binlog and be filed under the 'binlog' suite?
> R2. If R1 is really needed, then why do we need to create
> tables separately on master and slave? Why not create on
> master and then sync_slave_with_master?
> S1. I would leave the check 'if(!empty())' in 'reset()'.