On 12/15/16 3:46 AM, Petr Jelinek wrote:
>>> Okay in case we decide it's the right way to go attached patch does
>>> that. I also added some more tests based on your feedback while I am at it.
>>
>> This looks mostly reasonable to me, but the location and xid output from
>> pg_logical_slot_get_changes() is not portable/repeatable, so it should
>> be omitted from the output.
>>
>
> Sigh, yes you are right.
committed
--
Peter Eisentraut http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services