On Mon, Oct 17, 2022 at 7:11 PM shiy.f...@fujitsu.com <shiy.f...@fujitsu.com> wrote: > ... > > Thanks for your patch. Here are some comments. > > In Example 2, the returned slot_name should be "myslot". > > +test_pub=# SELECT * FROM pg_create_logical_replication_slot('myslot', > 'pgoutput'); > + slot_name | lsn > +-----------+----------- > + sub1 | 0/19059A0 > +(1 row) >
Oops. Sorry for my cut/paste error. Fixed in patch v6. > Besides, I am thinking is it possible to slightly simplify the example. For > example, merge example 1 and 2, keep the steps of example 2 and in the step of > creating slot, mention what should we do if slot_name is not specified when > creating subscription. > Sure, it might be a bit shorter to combine the examples, but I thought it’s just simpler not to do it that way because the combined example will then need additional bullets/notes to say – “if there is no slot_name do this…” and “if there is a slot_name do that…”. It’s really only the activation part which is identical for both. ----- Kind Regards, Peter Smith. Fujitsu Australia