[infinispan-dev] PutForExternalRead and autoCommit

Mircea Markus mircea.markus at jboss.com
Tue Nov 22 14:17:52 EST 2011


On 22 Nov 2011, at 19:13, Slorg1 wrote:

> Hi,
> 
> I would like to take some time and make a proper answer to Galder,
> 
> However, for this quick remark please see below:
> 
> On Tue, Nov 22, 2011 at 14:07, Mircea Markus <mircea.markus at jboss.com> wrote:
>> 
>> On 17 Nov 2011, at 15:25, Galder Zamarreño wrote:
>> 
>>> For more background info, see: http://community.jboss.org/message/636797#636797
>> 
>> I also think this might be confusing for users.
>> What about:
>> a) don't inject a transaction for PFER. That's an easy change.
>> b) if autoCommit is off and PFER is not invoked in the scope of a tx we can either warn or throw an exception. Or just ignore.
> 
> I think that is the cache is set to TRANSACTIONAL, auto commit should
> be always off (is it not the point of transactions?) and as you point
> out using the cache outside of the scope of a transaction should be
> prohibited.
autoCommit was added mainly for backward compatibility, as there are user's accessing the cache in a mixed mode. 
It is also an API-sugar for single-write transactions, as one doesn't have to explicitly write then begin/commit.  
> 
>> The problem is that we inject the transaction twice.
> 
> I do not understand that statement, is it related to a or b or neither?
no necessarily - just me trying to get to the root of the problem :)
> 
> Regards,
> 
> Slorg1.
> 
> -- 
> Please consider the environment - do you really need to print this email ?
> 
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev




More information about the infinispan-dev mailing list