[hibernate-dev] HHH-3225 is hitting Hibernate Search

[Emmanuel Bernard]

On  Mar 13, 2009, at 18:45, Sanne Grinovero wrote:

>> I don't quite understand why, the new EventListenerREgister will be  
>> bundled
>> with the IndexWork Listener always right? What backward compatible  
>> mode do
>> you have?
> So I'll discard compatibility with old configurations of Search? this
> means somebody
> just replacing the jar but "forgetting" to read the new docs and
> failing to update the
> configuration will have lots of new problems... was just worried  
> about this,
> but as you say they shouldn't be using it out of transaction anyway.
> What I have is just "do it immediately" as the current code in trunk  
> does, if
> we are out of transaction and I can't find the event listener to
> register the sync
> for later. I have to find it anyway, so it's not a big change.

ah I understand now, then you're right, if the listener is not  
registered raise a warning log

>
>
>> BTW, you should put a warning in the log when this event listener  
>> is used.
>> "Applying change to the full-text index before transaction  
>> completion.
>> Please use a Hibernate aware transaction (eg  
>> org.hibernate.Transaction,
>> javax.persistence.EntityTransaction, JTA transaction with the proper
>> TransactionFactory setting)"
>>
> so we're going to log that all usecases out of TX are evil?
> In this case wouldn't you prefer to avoid fixing it and just log the  
> message?

hum, here is my dilemma. Some people use it out of transaction and  
that's their problem.
Some people believe they do it within transaction but either they  
don't configure Spring properly or Spring does not work properly and  
that's what I am more concerned about and want the warning for.

I guess we can reduce it to INFO.

I am almost thinking we should have a flag in the configuration to  
explicitly allow out of tx indexing but I hate configurations, so  
we're not going to do that.