Re: Nested transactions - Mailing list pgsql-patches

From Simon Riggs
Subject Re: Nested transactions
Date
Msg-id 1087425935.12015.1154.camel@stromboli
Whole thread Raw
In response to Nested transactions  (Alvaro Herrera <alvherre@dcc.uchile.cl>)
Responses Re: Nested transactions
Re: Nested transactions
List pgsql-patches
On Tue, 2004-06-08 at 23:23, Alvaro Herrera wrote:
> Hackers,
>
> Here is the latest installment of the nested transactions patch.
>
> What's in the current patch:
>

First of all, thank you for all your helpful comments recently.

The patch looks impressively technical, but overall I'm not exactly sure
what it does...I guess I'm just not clear why I would want it, except as
the main technical pre-work to later syntax changes. I'm sure some short
explanations would clear that up for me very quickly... :)

The Todo items were:
-Allow savepoints / nested transactions
-Use nested transactions to prevent syntax errors from aborting a
transaction

both of which I thought I understood:

The first one provides the SQL commands SAVEPOINT and ROLLBACK TO
SAVEPOINT as with Oracle/DB2, and also now ANSI SQL if I recall...

The second one again provides Oracle/DB2 support by conforming to their
interpretation of the ANSI transactional semantics definition. i.e. one
statement failure doesn't roll back the transaction, just the statement
that failed.

Being able to issue multiple BEGIN/END pairs isn't really (to me) the
same thing as the above, nor do I understand why I'd ever want to do
that - especially down to N levels....

Perhaps what I've just asked about is trivial icing on the cake you've
just baked, so forgive me, but could you explain the outward form of
your work and what that gives me? (or at least...what you think it gives
you...which I accept may be different)

Best regards, Simon Riggs


pgsql-patches by date:

Previous
From: Gavin Sherry
Date:
Subject: Re: Tablespaces
Next
From: Claudio Natoli
Date:
Subject: Re: pg_ctl service integration for WIN32