Re: cursors outside transactions - Mailing list pgsql-hackers

From Tom Lane
Subject Re: cursors outside transactions
Date
Msg-id 19750.1048131493@sss.pgh.pa.us
Whole thread Raw
In response to Re: cursors outside transactions  (Bruce Momjian <pgman@candle.pha.pa.us>)
List pgsql-hackers
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Hiroshi Inoue wrote:
>> AccessShare table locks are only needed.
>> What is wrong with it ?

> But that is going to block VACUUM, right?

No.  It's a matter of bookkeeping more than anything else.  Right now,
transaction end releases all the locks a backend holds.  You'd need to
figure out which locks are associated with cross-transaction cursors
and keep those.  This overlaps to some extent with bookkeeping that we'd
need to add for nested transactions --- but we haven't got a plan for
that, either.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Nested transactions: low level stuff
Next
From: "Shridhar Daithankar"
Date:
Subject: Re: [pgsql-advocacy] PostgreSQL flamage on Slashdot