Re: unlogged sequences - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: unlogged sequences
Date
Msg-id b3323839-8899-badc-9067-2b2207af4a4f@enterprisedb.com
Whole thread Raw
In response to Re: unlogged sequences  ("David G. Johnston" <david.g.johnston@gmail.com>)
List pgsql-hackers
On 04.04.22 01:58, David G. Johnston wrote:
> "Because pg_dump is used to transfer data to newer versions of 
> PostgreSQL, the output of pg_dump can be expected to load into 
> PostgreSQL server versions newer than pg_dump's version." [1]
> 
> That is what I'm getting on about when talking about migrations.  So a 
> v14 SQL backup produced by a v14 pg_dump restored by a v15 psql.

It has always been the case that if you want the best upgrade 
experience, you need to use the pg_dump that is >= server version.

The above quote is a corollary to that we don't want to gratuitously 
break SQL syntax compatibility.  But I don't think that implies that the 
behavior of those commands cannot change at all.  Otherwise we could 
never add new behavior with new defaults.




pgsql-hackers by date:

Previous
From: Andrei Zubkov
Date:
Subject: Re: [PATCH] Tracking statements entry timestamp in pg_stat_statements
Next
From: Julien Rouhaud
Date:
Subject: Re: [PATCH] Tracking statements entry timestamp in pg_stat_statements