Re: [HACKERS] Autocommit off and transaction isolation level

2003-01-02 Thread Stephan Szabo
On Thu, 2 Jan 2003, Tom Lane wrote: > "Michael Paesold" <[EMAIL PROTECTED]> writes: > > Now setting autocommit=off the set transaction isolation level command does > > not show any effect: > > > billing=# set autocommit to off; > > SET > > billing=# set transaction isolation level serializable; >

Re: [HACKERS] Autocommit off and transaction isolation level

2003-01-02 Thread Tom Lane
"Michael Paesold" <[EMAIL PROTECTED]> writes: > Now setting autocommit=off the set transaction isolation level command does > not show any effect: > billing=# set autocommit to off; > SET > billing=# set transaction isolation level serializable; > SET SET does not start a transaction block, so th

[HACKERS] Autocommit off and transaction isolation level

2003-01-02 Thread Michael Paesold
Hi all, I have come across some weird behavior in postgres concerning autocommit=off and setting the transaction isolation level. I have no explanation why things should work as they do, so I consider this a bug, no? With autocommit=on and normal begin; ... commit; block setting the transaction i