Hi all,
I am using Wicket, and I manage to detach the session with this :
package com.netzep.common.frontend.application;
import org.apache.wicket.model.IModel;
import org.qi4j.api.entity.Identity;
import org.qi4j.api.unitofwork.UnitOfWork;
import org.qi4j.api.unitofwork.UnitOfWorkCallback;
import org.qi4j.api.unitofwork.UnitOfWorkCompletionException;
public class EntityIModel<T extends Identity> implements IModel<T>,
UnitOfWorkCallback {
@Override
public boolean equals(Object obj) {
if (obj != null && this.identity != null
&& obj instanceof EntityIModel<?>) {
EntityIModel<?> eim = (EntityIModel<?>) obj;
return eim.getIdentity() != null
&&
eim.entityClass.equals(this.entityClass)
&& this.identity.equals(eim.identity);
}
return false;
}
String identity;
Class<T> entityClass;
transient private T entity;
public EntityIModel(T entity, Class<T> entityClass) {
this(entity.identity().get(), entityClass);
}
public EntityIModel(String identity, Class<T> entityClass) {
super();
this.identity = identity;
this.entityClass = entityClass;
}
private UnitOfWork getUnitOfWork() {
final UnitOfWork currentUnitOfWork = ((Qi4jWebApplication)
Qi4jWebApplication
.get()).module.unitOfWorkFactory().currentUnitOfWork();
currentUnitOfWork.addUnitOfWorkCallback(this);
return currentUnitOfWork;
}
@Override
public T getObject() {
if (this.entity == null)
this.entity = getUnitOfWork().get(entityClass,
identity);
return this.entity;
}
@Override
public void setObject(T object) {
assert false;
}
@Override
public void detach() {
this.entity = null;
}
public String getIdentity() {
return this.identity;
}
@Override
public void afterCompletion(UnitOfWorkStatus status) {
this.detach();
}
@Override
public void beforeCompletion() throws UnitOfWorkCompletionException {
}
}
It is far from perfect especially since the entity identity string ends up
in the session (a security layer may be needed here but I don't rely on
anything coming from the web tier anyway, the security aspects are in my
business layer)
Regarding the unitofwork management :
package com.netzep.common.frontend.application;
import org.apache.wicket.Page;
import org.apache.wicket.Response;
import org.apache.wicket.protocol.http.WebApplication;
import org.apache.wicket.protocol.http.WebRequest;
import org.apache.wicket.protocol.http.WebRequestCycle;
import org.qi4j.api.structure.Module;
import org.qi4j.api.unitofwork.ConcurrentEntityModificationException;
import org.qi4j.api.unitofwork.UnitOfWork;
import org.qi4j.api.unitofwork.UnitOfWorkCallback;
import org.qi4j.api.unitofwork.UnitOfWorkCompletionException;
import org.qi4j.api.unitofwork.UnitOfWorkFactory;
import org.qi4j.api.unitofwork.UnitOfWorkCallback.UnitOfWorkStatus;
public class UoWRequestCycle extends WebRequestCycle {
UnitOfWorkFactory uowf;
public UoWRequestCycle(WebApplication application, WebRequest request,
Response response, Module module) {
super(application, request, response);
this.uowf = module.unitOfWorkFactory();
}
@Override
protected void onBeginRequest() {
super.onBeginRequest();
if (this.uowf.currentUnitOfWork() == null) {
this.uowf.newUnitOfWork();
}
}
@Override
protected void onEndRequest() {
final UnitOfWork currentUnitOfWork =
this.uowf.currentUnitOfWork();
if (currentUnitOfWork != null) {
try {
currentUnitOfWork.complete();
} catch (ConcurrentEntityModificationException e) {
e.refreshEntities(currentUnitOfWork);
try {
currentUnitOfWork.complete();
} catch (ConcurrentEntityModificationException
e1) {
e1.printStackTrace();
} catch (UnitOfWorkCompletionException e1) {
e1.printStackTrace();
}
} catch (UnitOfWorkCompletionException e) {
e.printStackTrace();
}
}
super.onEndRequest();
}
@Override
public Page onRuntimeException(Page page, RuntimeException e) {
this.uowf.currentUnitOfWork().discard();
return super.onRuntimeException(page, e);
}
}
Any comment and improvement welcome !
Phil
Niclas Hedhman wrote:
>
> Adding to Rickard's reply...
>
> The mentioned web framework wanted to leverage Wicket, but there are
> problems around the serialization of sessions. Exactly how this should be
> resolved isn't clear.
>
> There is also work done with Struts2 integration, which I think currently
> sits in qi4j-sandbox. I have not looked at that myself, and perhaps Edward
> has other Richard Wallace.
>
> Lastly, I have been thinking a lot on how this *should* be done with the
> Qi4j toolkit, but have so far failed to formulate a structure that I am
> happy with. GutFeeling is that it should turn out really tight and neat,
> thanks to contextual fragments and use of entities for state and
> valuecomposites for everything else.
>
> Cheers
> Niclas
>
> On Jul 18, 2009 10:36 PM, "Robert Krüger" <[email protected]> wrote:
>
>
> Hi,
>
> I'm currently beginning to evaluate Qi4j and have stumbled across mentions
> of a web framework that's also being developed by the Qi4j team but have
> been unable to find any further information on that topic on the site or
> in
> the list archives.
>
> Could anyone point me in the right direction or is it simply too early to
> ask?
>
> Thanks in advance,
>
> Robert
>
>
>
> _______________________________________________
> qi4j-dev mailing list
> [email protected]
> http://lists.ops4j.org/mailman/listinfo/qi4j-dev
>
> _______________________________________________
> qi4j-dev mailing list
> [email protected]
> http://lists.ops4j.org/mailman/listinfo/qi4j-dev
>
>
--
View this message in context:
http://www.nabble.com/Web-framework-tp24548505p24555233.html
Sent from the Qi4j-dev mailing list archive at Nabble.com.
_______________________________________________
qi4j-dev mailing list
[email protected]
http://lists.ops4j.org/mailman/listinfo/qi4j-dev
