Re: pg_recvlogical --endpos

On 22 November 2016 at 16:52, Okano, Naoki <okano(dot)naoki(at)jp(dot)fujitsu(dot)com> wrote:
> On Monday, November 21, 2016 1:08 PM Craig Ringer wrote:
>> I've updated the patch for this. It's already posted on the logical
>> decoding timeline following thread, so I'll avoid repeating it here.
>>
>> https://www.postgresql.org/message-id/CAMsr%2BYGd5dv3zPNch6BU4UXX49NJDC9m3-Y%3DV5q%3DTNcE9QgSaQ%40mail.gmail.com
> I checked the latest patch.
> I think that the error message shown below is a typo.
>
>> + if (endpos != InvalidXLogRecPtr && !do_start_slot)
>> + {
>> + fprintf(stderr,
>> + _("%s: cannot use --create-slot or --drop-slot together with --endpos\n"),
> The condition '!do_start_slot' is not reflected in the error message.

Would it be better rephrased as "--endpos can only be used with --start" ?

> The patch should allow --endpos to work with --create-slot.

How? It doesn't make sense with --create-slot .

> Also, the document explains as follows.
>> + specified LSN. If specified when not in <option>--start</option>
>> + mode, an error is raised.
> So, it is better to output an error message that matches the document when changing the error message.

Not sure I understand this.

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services