Re: SET TRANSACTION SNAPSHOT sounds like replicated environment but isn't - Mailing list pgsql-docs

From David G. Johnston
Subject Re: SET TRANSACTION SNAPSHOT sounds like replicated environment but isn't
Date
Msg-id CAKFQuwaxozDM-+P-nKPY1aBZ0feckRQbAQJ=d8HJtv8-E8TzQg@mail.gmail.com
Whole thread Raw
In response to SET TRANSACTION SNAPSHOT sounds like replicated environment but isn't  (PG Doc comments form <noreply@postgresql.org>)
List pgsql-docs
On Wed, May 11, 2022 at 7:33 AM PG Doc comments form <noreply@postgresql.org> wrote:
The following documentation comment has been logged on the website:

Page: https://www.postgresql.org/docs/14/sql-set-transaction.html
Description:

Hello,

the wording on the SET TRANSACTION SNAPSHOT left me a bit confused. It says
"To begin a new transaction with the same snapshot as an already existing
transaction" and it feels like basically taking over an existing
session/transaction or being able to replicate a transaction from that
snapshot on.

I'm not sure what a final documentation patch might look like but this section can rightfully assume some prior knowledge about how the system works with regards to snapshots.  In particular, a snapshot is effectively a single value that a session maintains that lets it compute whether a given transaction id is visible or not.  So, "whenever you give me data only do so up to this point-in-time".  Understanding that dynamic should lead one to conclude that just having a shared since of "point-in-time" doesn't have anything to do with permissions to interact with specific objects in the first place.  And, for temporary tables, to break the session isolation that it is assumed the reader is aware of.

David J.

pgsql-docs by date:

Previous
From: PG Doc comments form
Date:
Subject: SET TRANSACTION SNAPSHOT sounds like replicated environment but isn't
Next
From: PG Doc comments form
Date:
Subject: Reserved word: OWNER