[JBoss jBPM] - Nullable properties and Text control behaviour in the GPD
by koen.aers＠jboss.com
Hi all,
With this post I want to get some input from the community about two different topics.
The first of these is nullable properties in jPDL. As you might have seen in the latest release of the GPD (3.1.0.alpha4), I have provided a checkbox next to some of the text input fields. The checking/unchecking of this checkbox indicates that the corresponding property is not null or null. I introduced this because otherwise I had no means of differentiating a null value from the empty string.
Now with this included, and after the input of a number of users, I realize that some of you might find the result a bit awkward. I invite you to start a discussion on this. The alternative is to disallow empty strings for nullable properties. Would that be any better?
The second question I have is about the way properties should be set. Currently in most of the text fields I use the so-called default selection mechanism of Eclipse. That means that the property is not set unless you press the 'Enter' key. The property will explicitly not get set when you change the focus from the text field to some other control. In that case the value will revert to the original. The rationale behind this is that for consistency purposes I wanted to use the same mechanism as in the SWT tables, where the focuslost mechanism is also not allowed. Now afterwards, I realize there is not yet consistency because this default selection mechanism is not possible for multi line text controls. In that case I had to revert to the focuslost mechanism again.
The question I have is whether this is good or bad? Do you want me to implement the focuslost mechanism on all text fields again (leaving the incosistency with the SWT tables) or can you live with the current solution?
Regards,
Koen
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4043290#4043290
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4043290
19 years
[JBoss Seam] - Custom Injector - What interceptors should I go
by jbnewb
- I am using seam (no EJB), do use @In, @Out
- However, I need to inject domain specific objects (these cannot have default constructors), I have written a custom @Inject annotation
- I also have a dummy @Injector annotation at the class level, where I actually introspect all fields annotated with @Inject
- I am able to successfully inject the object, however, my interceptor is called a large number of times
- I have tried going around MethodContextInterceptor, BijectionInterceptor and some others. I see that the injection is called a large number of times (close to 50 for a page with about 5 fields)
Questions
- I use the "around" because I want the domain object to be injected before @Out comes into play
- Do I need to have the "around" in the interceptor class OR can I live without it, Besides pushing a flag in the eventcontext, is there any other way I can prevent such multiple calls?
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4043287#4043287
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4043287
19 years
[JBoss Seam] - Re: Pageflow Question & Switcher Issue
by Delphi's Ghost
Looking at this a bit more, it seems that join="true" might not be the best answer.
In my dummy app, I have meeting objects which has one or more person objects as members. When editing the meeting, you can select "add member" to go to the person search page where you select a person and the pageflow takes you back to the meeting edit with the person added as a member.
However, when on the person search page, you might want to invoke the add or edit person button to add or edit a person. The add/edit person page should have its own conversation and pageflow (right now, there is no pageflow). Bear in mind that add/edit person could be called from the search page (or a RESTful URL) without being part of a conversation, and therefore should be able to start it's own conversation.
If I start my person edit conversation with <begin-conversation nested="true"/> :
Edit meeting (start conversation A)
click add member->goto person search page
click add person->goto person edit page (start nested conversation B)
click save person -> (end conversation B) goto person search page
select my new person ->goto edit meeting page
save my meeting (end conversation A)
If I use join=true, which at the moment, the pageflow issue is forcing me to :
Edit meeting (start conversation A)
click add member->goto person search page
click add person->goto person edit page (join conversation A)
click save person -> (end conversation A) goto person search page
....I have no active conversation at this point!
My person editor just ended my original conversation.
So it looks like nested should be the best option when dealing with pageflows. However the problem there is that nested=true on the begin-conversation tag causes a new nested conversation with each page rendering. Again, it's this issue of repeating the conversation tag everytime the page is rendered.
It sounds like the I should go submit a JIRA, and link back to this thread, and see what the powers that be think.
OTOH, my little meet manager test app rocks with seam, I can add meetings go select people, and come back to the meeting editor, and the conversation stuff just makes it so painless and straightforward.
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4043286#4043286
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4043286
19 years
[Persistence, JBoss/CMP, Hibernate, Database] - Superclass of Entity getting nulled-out
by tglaess
I have an entity bean that extends another entity bean. All is fine until about 10 minutes of inactivity in the session.
So, I log in to our app and let the session sit idle for 10 minutes. At that time, the values in the superclass all become null. The values in the subclass have not changed.
What would cause this? The session timeout is set to 30 minutes, so there is no issue there. I've turned on the maximum logging I can and nothing is logged during that 10 minutes. Why would the superclass values become null?
thanks,
tglaess
Here are some code snippets from the two classes:
@Entity
@Table(name = "user")
@Inheritance(strategy=InheritanceType.JOINED)
public class UserBean {
@Id
@Column(name = "UserID")
protected BigDecimal userID;
@Column(name = "UserName", nullable = false)
protected String userName;
@Column(name = "Password", nullable = false)
protected String password;
...
@Entity
@Table(name = "dealeruser")
@PrimaryKeyJoinColumn(name="UserID",referencedColumnName="UserID")
public class DealerUserBean extends UserBean implements Serializable {
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4043280#4043280
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4043280
19 years
[JBoss Seam] - Re: Is seam-gen really a good idea to learn about seam?
by fperedo
I agree... Seam is simplest to use in an EE5 container... perhpas my problem is that I am used to other frameworks on which you learn the other way around:
1) How to do it without EE5
2) How to do it with EE5
I guess all this comes from the fact that I would love to use Seam over OC4J... but from what I have read over the net, OC4J is kindof a EE4 container with "full EJB3 support" (so... it is a an... EE4.5 container?) and since I liked Seam a lot a would like to use it, even if I have to "turn off" some of its features... but the thing is there is no guide to know what features will I miss if I have to do that (maybe I'll endup with a Seam so crippled that I would prefer to use some other framework)
I know it is not you fault... I would prefer to use JBoss or Glassfish, but OC4J is already deployed at my work, and there is some pressure to use it...
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4043278#4043278
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4043278
19 years