Discussion: FlushModeType.NONE ?

[motte1979]

Hi everybody,

I've many readonly - screens which do never need a flush ...
What about a new FlushModeType.NONE in
org.jboss.seam.annotations.FlushModeType

Comments & discussion are welcome ....

[christian.bauer]

Why not use FlushModeType.MANUAL and simply never call flush()?

[motte1979]

That is an option, yes. But this needs you to be sure that a flush is never called in your components ...
Just to ensure that a flush will be ignored ...

[denis-karpov]

FlushModeType.MANUAL does not help in every case.
I have tired to struggle with flushing. The only sure way now is to evict objects from persistent context.

Session session = (Session) getEntityManager().getDelegate();
session.evict(obj);

[christian.bauer]

FlushModeType.MANUAL does not help in every case.

And these cases are? I've been using it for 5 years and never had such a case.

[dreuzel]

THere are 2 FlushModeTypes



import org.hibernate.annotations.flushmodeType
import javax.persistence.FlushmodeType

i'm trying @begin(Flushmode=FlushmodeType.Manual
both imports are bad
expect second but manual not defined


Please assist

[fernando_jmt]

org.jboss.seam.annotations.FlushModeType

 @Begin(flushMode = FlushModeType.MANUAL)
 public String myaction() {
 ....
 }

[denis-karpov]

Sorry that I didn't confirm my words. Right now I do not have sample.

As I remember when I go through a decision with drools. Even in "invalid" case, when I simply return to the edit page, flushing occurs any way .

<decision name="validate">
 <handler class="org.jboss.seam.drools.DroolsDecisionHandler">
 <workingMemoryName>buyWorkingMemory</workingMemoryName>
 <assertObjects>
 <element>#{exchOrdHome.instance}</element>
 <element>#{facesMessages}</element>
 </assertObjects>
 </handler>
 <transition to="refresh">
 <action expression="#{exchOrdHome.merge}" />
 </transition>
 <transition name="invalid" to="edit"/>
 </decision>