Re: [PATCHES] Automatic transactions in psql - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: [PATCHES] Automatic transactions in psql
Date
Msg-id 200202221609.g1MG9mT12566@candle.pha.pa.us
Whole thread Raw
In response to Re: [PATCHES] Automatic transactions in psql  (Thomas Lockhart <lockhart@fourpalms.org>)
List pgsql-hackers
Thomas Lockhart wrote:
> ...
> > > Secondly, it adds a "begin transaction" option that, when
> > > enabled, ensures that you are always inside a transaction
> > > while in psql, so you can always rollback. It does this
> > > by issuing a BEGIN at the appropriate times. This patch
> > > (if ever accepted) conflicts a bit with LO_RTANSACTION:
> > > psql now *does* have a way to know if it is in a
> > > transaction or not, so that part may need to get rewritten.
> 
> This part of the feature (corresponding to the Ingres "autocommit = off"
> feature) should be implemented in the backend rather than in psql. I've
> had a moderate interest in doing this but haven't gotten to it; if
> someone wants to pick it up I'm sure it would be well received...

Agreed.  I wondered whether we could use the psql status part of this
patch.  We currently cound parens and quotes, and show that in the psql
prompt.  Could we do that for transaction status?  Considering we
already track the parens/quotes, another level of status on the psql
display seems  a bit much, even if we could do it reliably.  Comments?

--  Bruce Momjian                        |  http://candle.pha.pa.us pgman@candle.pha.pa.us               |  (610)
853-3000+  If your life is a hard drive,     |  830 Blythe Avenue +  Christ can be your backup.        |  Drexel Hill,
Pennsylvania19026
 


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: CLUSTER TODO item
Next
From: "Zeugswetter Andreas SB SD"
Date:
Subject: Re: point in time recovery and moving datafiles online