Comment 10 for bug 1420606

Upstream has it's own fix for bug 1331586.
It passes our test case. I am going to revert our fix.

Author: Joao Gramacho <email address hidden>
Date: Mon Oct 20 08:24:05 2014 +0100

    Bug#18885916 RELAY LOG WITHOUT XID_LOG_EVENT MAY CASE PARALLEL
                 REPLICATION HANG

    Problem:
    =======

    With MTS, GTIDs and auto positioning enabled, when a worker applies a
    partial transaction left on relaylog by an IO thread reconnection, it
    will wait for the XID log event to commit the transaction.

    Unfortunately, the SQL thread coordinator will reach the master's
    ROTATE event on the next relaylog file and will wait for all workers
    to finish their tasks before applying the ROTATE.

    Analysis:
    ========

    As the whole transaction is retrieved again by the IO thread after the
    reconnection, the slave must rollback the partial transaction once
    noticing this ROTATE from the master.

    This bug reports the same issue already fixed by BUG#17326020, and the
    reported issue is not reproducible anymore. So, this patch is just
    adding a new test case.