MDEV-34765: rpl.master_last_event_time_stmt fails with Result Length …

[bnestere]

trafficstars

…Mismatch

Starting this PR with only a regression patch to show the issue, to start a discussion about the direction for a fix.

When executing a Query_log_event that is a COMMIT query, gtid_slave_pos is updated before other replication status variables, so when an MTR test syncs a replica with primaries via GTID, there is a slight window where accessing status variables, e.g. via SHOW ALL SLAVES STATUS, results in "stale" values because gtid_slave_pos has been updated before inc_group_relay_log_pos() has been called, which is when gtid_slave_pos is typically updated.

Where we can easily fix the test to just use coordinates rather than GTID, it seems like unexpected behavior (and may also confuse users collecting info via scripts), so I wonder we should change the behavior to make things consistent.