Re: locks

From: HansF <Fuzzy.Greybeard_at_gmail.com>
Date: Fri, 11 Aug 2006 01:50:17 GMT
Message-Id: <pan.2006.08.11.01.49.51.891553@gmail.com>

On Thu, 10 Aug 2006 22:39:26 +0100, Andy Hassall wrote:

> On Thu, 10 Aug 2006 20:59:16 GMT, HansF <Fuzzy.Greybeard_at_gmail.com> wrote:
> 

>>On Thu, 10 Aug 2006 19:37:09 +0200, Sybrand Bakker wrote:
>>
>>> exiting *normally* will end the database connection and will issue a
>>> commit.
>>
>>I always thought 'exiting *normally* without explicitly issuing a commit'
>>will issue a rollback, and the 'commit on exit' syndrome was a 'feature'
>>of the calling environment (such as SQLPlus).

> 
>  The OCI function OCISessionEnd has always committed the current transaction.
> No idea about the v7 and earlier equivalent - ologoff or something wasn't it?

Thanks - dug into it further and ologof() also did an explicit commit. The general rule (I vaguely recall) was to explicitly commit or call a function like ologof() that includes an explicit commit.

I do clearly remember "Don't depend on the environment to commit on disconnect - different environments may behave differently."

-- 
Hans Forbrich   (mailto: Fuzzy.GreyBeard_at_gmail.com)   
*** Feel free to correct me when I'm wrong!
*** Top posting [replies] guarantees I won't respond.

Received on Thu Aug 10 2006 - 20:50:17 CDT