JAG BUG ? "Set Application Module Superclass to JhsApplicationModuleImpl"

Is it just me ? or ...
Whenever I set "Set Application Module Superclass to JhsApplicationModuleImpl" to false and run the Application Generation ... the base class of our ApplicationModuleImpl is set to JhsApplicationModuleImpl anyway (which it is not supposed to do !).
Our workaround is to go to our ApplicationModuleImpl and set the base class manually.
Is this a JHS bug ?

Hi Sandra,
I just made a testcase and tried to reproduce the "anomoly" I encountered. And ... whatever I try, I cannot reproduce it anymore.
My testcase:
==========
JHS 10.1.2.1 (build 27)
Jdev 10.1.2.1.0 (build 1913)
- created ADF baseclasses in new project : BaseJhsApplicationModuleImpl extends JhsApplicationModuleImpl
- created standard Emp business components
- open ApplicationModuleImpl : extend from BaseJhsApplicationModuleImpl
- enabled JHS on ViewController
- created new ApplicationStructure (view=JSP ; Set AppModImpl = false)
- run JAG
-- open ApplicationModuleImpl : class still extends BaseJhsApplicationModule iso JhsApplicationModuleImpl (--> which is the correct behaviour)
Whatever change I made to the ApplicationStructure ; the 'anomaly' did not return (fortunately).
There must have been a glitch in Jdev ... always easy to blame the IDE ; )
Thanks for your time,
Pascal Prins

Similar Messages

BUG? Application Module State Management

I have 2 read-only view objects VO1 & VO2 with a View Link that links them by a primary key field. VO1 (master) is on the first .jspx page and VO2 (detail) is on the 2nd .jspx page. VO1 has a SQL query, but is populated programmatically in an App Module Service method that sets the where clause parameters. VO2 is on a separate .jspx page. VO2 has one bind variable in its SQL query that is populated when the user clicks a commandLink on a row within VO1 - this takes the user to the 2nd .jspx page to view VO2 table results.
The issue is that when the user navigates back to the first .jspx page containing VO1, a duplicate row is added to the end the data table (or as the first record on the second page of results if there are multiple pages.) If I re-query VO1, re-sort the data using the column headings, or passivate the data (see below), the duplicate row goes away. Note I have tried using the same iterator on both pages and this same issue occurs.
Ok - after 5 days of troubleshooting, here is the potential bug. If Application Pooling is enabled on the Application Module, I experience the issue above. If I don't enale Application Pooling, and Passivate state for the App Module, this issue does not occur (see below):
Application Module Settings - Error Occurs if Pooling is enabled
(unchecked) Failover Transaction State Upon Managed Release
(unchecked) Disconnect Application Module Upon Release
(checked) Support Dynamic JDBC Credentials
(checked) Reset Non-Transcational State Upon Unmanaged Release
(checked) Enable Application Module Pooing
If I un-check Enable Application Module Pooling the error does NOT occur. Note I've experimented with many combinations, and its only the Application Module Pooling that affects the behavior.
Note both view objects are set to Passivate state (but not for transient values, though tested passivating those as well and there was no difference.)
Can someone please advise? This is a showstopper for implementing our application to production. See the threads below for more information. Finally note that I have not yet been able to replicate this in a simple test case, but there is some specific characteristics about my current application causing this behavior (and it appears others have experienced as well.)
Duplicate rows displayed in list
Need help - Duplicate Rows being added to View Object?
Thanks.
Here is the code on the AM service method to populate the rows for VO1 (no insert or create rows - only executing the query):
ViewObject vo1 = findViewObject("VO1");
vo1.setMaxFetchSize(250); // set max fetch size 250
vo1.setWhereClauseParam(0, x0);
vo1.setWhereClauseParam(1, x1);
vo1.setWhereClauseParam(2, x2);
vo1.setWhereClauseParam(3, x3);
vo1.setWhereClauseParam(4, x4);
vo1.setWhereClauseParam(5, x5);
vo1.setWhereClauseParam(6, x6);
vo1.setWhereClauseParam(7, x7);
vo1.setWhereClauseParam(8, x8);
vo1.setWhereClauseParam(9, x9);
vo1.setWhereClauseParam(10, x10);
vo1.executeQuery();
Message was edited by:
javaX

More information on this issue:
After further testing, I discovered that the 'Disconnect Application Module Upon Release' also plays a factor with this issue.
If 'Enable Application Module Pooling' is unchecked the issue does not occur because the view object is passivated.
If 'Enable Application Module Pooling' is checked (jbo.ampool.doampooling=true), then the issue occurs UNLESS I also set 'Disconnect Application Module Upon Release' (jbo.doconnnectionpooling=true). If I set the latter to true, then I do not experience the duplicate row issue. I would rather not have to enable jbo.doconnectionpooling=true because of the performance impacts as recommended in the dev guide. Any Suggestions?
-- updated on 10/1/06 ----
Setting jbo.doconnectionpooling=true only fixes this issue when I use a ViewLink as the mechanism to pass the row Id from VO1 (master on .jspx page #1) to VO2 (detail on page #2.) If I use an actionListener to pass the value from page #1 to page #2, setting jbo.doconnectionpooling=true does not fix the issue.
<af:commandLink actionListener="#{bindings.setCurrentRowWithKey.execute}"
action="page2" text="#{row.Name}">
<af:setActionListener from="#{row.Id}" to="#{backing_bean_page2.rowid}"/>
</af:commandLink>
Message was edited by:
javaX

Guidelines for setting Application Module Pool Size Parameters?

Are there guidelines for setting the application module pool size parameters, such as initial pool size, maximum pool size, etc., based on the expected number of users or other factors? I've read the developer guide sections (ch 28-29), but still don't have a good feel for how to correctly set the optimal values for the pool configuration parameters? Even more importanty, how do I monitor the pool's efficiency during runtime to determine if the pooling parameters are configured correctly?
This will be critical to performance and scalability, so I'm looking for a way to get some visibility into how the pooling is working during production operation to assess whether there are bottlenecks/constraints/ineffeciencies?
Note I am using Tomcat as the java runtime container; ADF BC / JSF jdev 10.1.3.1
Thanks in advance and Merry Christmas!

KUBA - were you able to resolve these issues and if so are there any lessons learned you can share?
I'm hoping someone from the ADF team can answer our original question including guidelines for setting pool parameters and how to monitor the pool's performance while running in production.
thanks

Application Module's "query-on-commit" flag

Hi All,
Can you please guide how to set Application Module's "query-on-commit" flag as on/off.

You'll want to create a method in your application module class that will expose your reporting functionality. You can do this by:
1. Selecting your AppModule.
2. Opening AppModuleImpl.java
3. Creating a method to expose the necessary reporting services.
4. You'll then need to make this method accessible by opening AppModule editor and clicking on "Client Interface". Shuttle your method to the "Selected" side.
In the method you create in AppModule.java, pass the module (this) into your report class. You can then perform any actions you with such as querying VO's, updating, etc.
Finally, for your view/controller layer you'll want to call the newly added methods in your AppModule class instead of calling your reporting classes directly. This will make your code cleaner and easier to maintain as well and also control (encapulate) the reporting services you want to expose to others.
Hope that helps.

Application Module getting and using.

Hello all,
I am currently using this code:
DCDataControl dc = daContext.getBindingContext().findDataControl("AppModuleDataControl");
if (dc instanceof DCJboDataControl) {
ApplicationModule appModule = (ApplicationModule) dc.getDataProvider();
ApplicationModuleImpl myAm = (ApplicationModuleImpl)appModule;
It compiles well, but the problem is that the App Module that is returned is of type : WSApplicationModuleImpl.
I need to use the getDBTransaction() method from ApplicationModuleImpl, since when I cast my appModule, i get an error, I was wondering if there was something I was doing wrong.
I have searched high and low, and still haven't found what WSApplicationModuleImpl is for (i know that if I try to use some of the methods of ApplicationModuleImpl I get a moethod 'methodname' is not a working set Application Module method...)

Steve,
I followed the instructions in your article, which gave me a better understanding of the situation but did not resolve my particular issue. I was confused by all the statements and insinuations in both JDev and various articles that configuring application modules to use interfaces is really to support a client that is not colocated with the business services. I did not see the relevance of the interfaces to getting the appropriate data type back from the data control (MyAppModule rather than WSApplicationModuleImpl, for example). Anyway, I followed the steps from that article that related to the application module. I did not create interfaces for the other business components (views, entities, etc). I continued to get back the wrong data type when retreiving the application module from the data control in an action. So I kept looking for other clues and finally came across the article about when to use batch mode (http://www.oracle.com/technology/products/jdev/tips/muench/batchmode/index.html). Reading that one of the benefits of batch mode is better enforcement of programming business calls to interfaces I thought I'd try changing the sync to immediate rather (from the default, batch). This did exactly what I was originally trying to do. It caused the data control to return a MyAppModuleImpl rather than a generic WSApplicationModuleImpl. Finally I could call methods on the business service from actions.
After learning about batch mode, I understand that my change is not advisable in the grand scheme of clean design. Unfortunately, even though I believe that I have all the components in place to support a clean design, the batch mode seems to be preventing me from getting data type back.
So I write this for the next person to run into ClassCastExceptions when attempting to retrieve their application module, since I haven't seen it suggested before: look into batch vs. immediate mode - at the very least it will help you diagnose the problem.
Jeff

Generated getter in Application Module impl overrides superclass getter

Hi,
I today came across this issue:
* Created an application module where the data model contains a view object named "Parent" (resulting form a parent-child relationship)
=> Application runs perfectly.
* Created an application module implementation class (just created it, no further code added)
=> Application throws java.lang.NullPointerException at oracle.adf.model.binding.DCIteratorBinding.executeQueryIfNeeded(DCIteratorBinding.java:2164)
After some debugging, it turned out that the issue is that JDeveloper created the following getter in the application module impl class:
public ViewObjectImpl getParent() {
return (ViewObjectImpl)findViewObject("Parent");
which overrides public NamedObjectImpl getParent() from NamedObjectImpl, which is a superclass of ApplicationModuleImpl, due to the method's covariant return type.
JDeveloper does not give me any warning or error, not while creating the impl class nor highlighted in the source code.
Actually this would mean that I must manually assure that I never ever name a view object the same as any getter method in any of the ApplicationModuleImpl super classes which have a return type which is covariant to ViewObjectImpl. getParent() is one example, another is getProperty(). Even if it might be unlikely, this adds a potential source for some subtle issues.
Is this known behavior of JDeveloper? Did I miss anything?
Thanks,
Andreas

Hi Frank,
I agree that naming conventions can avoid this type of issue. Nevertheless, as a matter of usability and reliability, I think it would be good if JDeveloper at least gives a warning during code generation when it generates methods which would override a method from a superclass.
In any case, I have logged a bug, lets see how it evolves...
Thanks,
Andreas

Table QBE-Filter BUG in combination with Application Module Pooling ?

Hi,
i use JDEVADF_11.1.1.1.0_GENERIC_090615.0017.5407, Java 1.6.0_14, ADF BC and ADF Faces.
I have one View Object, one Page with a panelCollection and a table with option filtering (create via drag-and-drop from datacontrol).
When I disable the application module pooling on the AM configuration and run the application, i can execute the query-by-example filter, but when I delete the filterCriteria and press enter (requerry for select all) the table shows the old data.
Image for setting on application module:
http://img265.imageshack.us/img265/1374/filterdoesnotwork1.png
Image for execute querry and requerry for "all rows" (deleted filterCriteria)
http://img140.imageshack.us/img140/1963/filterdoesnotwork2.png
When I enable application module pooling all works fine. Is this a issue?
kind regards

Hi Frank,
if I filter for employees which firstname starts with "D" and then for employees with "F" --> i see "No data to display.".
So uncheck "Enable Application Module Pooling" on application module doesn't work with QBE-Filter ?
I can send a testcase but it is so simple you can create it with jdeveloper in 2 minutes ;)
Martin

Set max fetch size on application module

Is there a way to set max fetch size on application module level, rather then setting it for every view object separately? I don't want to allow any of my view objects in application module to fetch more then 1000 rows.

Hi,
Tools-->... contains option for setting rows in batches of... , but doesn't contain option for setting max fetch size. I guess I'll have to do it in implementation class for view objects... Thanks anyway
Dragana

BUG: Can't create BC4J Application Module Configurations

Hello,
I'm using JDev 9.0.3.1 with Win2K and WinXP. There seems to be a bug in the BC4J Application Module Configuration Manager:
It seems to be impossible to create a new BC4J Application Module Configuration with the Configuration Manager of a BC4J Application Module. I can copy an existing AM configuration, but I can not rename it. Clicking on the OK Button in the Configuration Manager simply does nothing.
A work around seems to be to open bc4j.xcfg in any text editor, copy an AppModuleConfig element, save xcfg, in JDev remove the project form the workspace, readd it to the workspace (some caching issue of jdev?). Then the new AM config is there and can be modified.
In JDev 9.0.3 the Config Manager seems to work fine.
Regards
Stefan

The bug is mentioned in metalink bug 2849146.

Setting v$session.program property in application module config

How can I set the v$session.program property in the AM config, so that when a new connection is created, the v$session.program field is documented with the property value.
I added the following in JClient JCLoginDialog:
((Hashtable)connEnvironment).put("v$session.program", "my program name");
I traced the V$SESSION TABLE:
The added connection has "my program name" in the program field but if I open other application module instances, the new connections have an empty v$session.program field.
I suppose that a property must be added to all application module configurations.
Can somebody help me on this one?
The v$session.program field is helpful when sessions must be killed.
Thanks
Fred

I don't believe there is a built-in mechanism for that.
But you can easily do it yourself. In you AM class:
public void prepareSession( Session session ) {
    super.prepareSession( session );
    setDatabaseClientInfo( Bc4jConstants.APPLICATION_NAME );
    setDatabaseModuleInfo( moduleInfo );
}And then:
public void setDatabaseClientInfo( String clientInfo ) {
    try {
      this.getTransaction( ).executeCommand(
        "BEGIN dbms_application_info.set_client_info('" + clientInfo + "'); END;" );
    } catch( Exception e ) {
      // uncritical error
      log.warn( "Error setting the database client info: " + clientInfo, e );
}and
public void setDatabaseModuleInfo( String moduleInfo ) {
    if( moduleInfo != null ) {
      try {
        this.getTransaction( ).executeCommand( "BEGIN dbms_application_info.set_module('"
          + moduleInfo + "', NULL); END;" );
      } catch( Exception exception ) {
        // uncritical error
        log.warn( "Error setting the database module info: " + moduleInfo, exception );
}In a similar fashion you can set the client identifier. We use the moduleInfo to identify the module of the session. The client info helps us to identify the application.
Sascha

[BUG] TP3 duplicating imported application module data controls

Hi,
I'm using TP3.
I have an application project A which imports business components from an other project B. Whenever I open JDev (or the project A) the first time, the imported application modules from the other project B get an new entry in A's model.jpx file. This will lead to duplicate data controls in the Data Control panel.
There might be other possible ways to get duplicate entries, but this one I can reproduce. I'm not sure if there are any implications having duplicate entries (my be longer start up time in OC4J).
Can anybody confirm my findings?
Timo

Hi Frank,
right now I don't have a test case, but I'll try to set one up. This will take some time. Till then I can only describe in detail what I did.
I have an older application Module migrated from version 10.1.2.2 via 10.1.3.3 to 11TP3. The application is used to login a user to our application framework. User name and password are checked against a DB table and if correct some user information is compiled into a bean.
In all of our other applications this login module is used to authenticate a user.
I added the logingmodule.jar ro the classpath and the used 'file->import->business components' to import the application module into a new ADF Fusion Application (this was the method to import application modules from other projects in versions 10.1.2.1 and also worked in 10.1.3.3).
Afer importing the jar I can add the login module to the new application module in the 'Data Model-> Application Aodule Instances' shuffle component.
As far as I found out this triggers an entry in the model.jpx file: <AppContainee
    Name="ber"
    ObjectType="JboProject">
    <DesignTime>
      <Attr Name="_SourceName" Value="loginModel.jar"/>
    </DesignTime>
</AppContainee> This entry gets dulicated every time I open the project in JDev 11TP3 the first time.
I read in the new doc, chapter 31, that there is a new method to reuse application modules (or other objects). This I'm currently checking out. Nonetheless I try to set up a test case.
Timo

[svn] 3493: Fix bug setting focus into an A. 2 application that is in a dialog.

Revision: 3493
Author: [email protected]
Date: 2008-10-06 13:40:22 -0700 (Mon, 06 Oct 2008)
Log Message:
Fix bug setting focus into an A.2 application that is in a dialog.
QE: YES
Doc:
Checkintests: Marshall Plan mustella tests
Reviewer: Carol
Bugs: SDK-17177
SystemManager.as
On mouse down, change the test from "the click was in a bridged application", to "the click is in a bridged application and it has the same focus manager as the document". A bridged application that is in a dialog will have a different focus manager than the document.
Ticket Links:
http://bugs.adobe.com/jira/browse/SDK-17177
Modified Paths:
flex/sdk/branches/3.0.x/frameworks/projects/framework/src/mx/managers/SystemManager.as

Thanks Jutapa.
When I ran the DPC Latency Checker, there were no significant hiccups, and nothing abnormal at all when the skips occurred. The PCI Latency Tool had no effect on the issue, either (I tried many different latency options, between 0 and 248). But maybe they'll help other people with this or similar problems!
Besides that, I tried logging huge amounts of data with the performance monitor and there's no consistent abnormalities at the times the skips occur as far as CPU usage or any other measurable variables go.
I'll check out http://musicxp.net! I'd tried fresh installs and disabling all unnecessary system services and startup programs before and it also had absolutely no effect, but that's probably not what MusicXP.net advises, so I'll take a look. Thanks!

Freeze JDeveloper10.1.3: Changing client interface on application module

Hello,
I could freeze JDeveloper 10.1.3.0.4.3673 build JDEVADF_10.1.3_NT_060125.0900.3673 3 times in a row with the following procedure:
1) You need HR schema from oracle examples
2) Create a new application workspace without any template (any name works)
3) Create the empty project (any works as well)
4) Set the project's technology scope to
4.1) ADF Business Component
4.2) Web Service
4.3) XML
4.4) Java will be added automatically
5) Generate business components from table
5.1) Generate component from table Employees only
5.2) Generate an updatable view for the component
5.3) Don't generate any read-only view object
5.4) Generate an app module
6) Add two methods to the application module class, here's are methods I'm using
public Element getCurrentEmployee()
ViewObject employeeView = getEmployeeView1();
Row current = employeeView.getCurrentRow();
if(current == null) {
current = employeeView.next();
Node employeeXml = current.writeXML(0, XMLInterface.XML_OPT_ALL_ROWS);
return (Element)employeeXml;
public void saveCurrentEmployee(Element xmlElement)
getEmployeeView1().readXML(xmlElement, 0);
getTransaction().commit();
7) Go to the application module properties and add the first method (getCurrentEmployee in my example) to the client interface
8) Click ok to confirm the change, some files will be generated
9) Reopen the application module properties, go to remote section and add a J2EE Web Service support
10) Click Ok, some files will be generated
11) Go to the web service properties in the state section and make the service stateful on session scope
12) Click Ok, some files are modified and a .deploy file is created
13) Save All
14) Go to application module properties and add the second method to the client interface
15) Click Ok, JDeveloper will freeze.
Anyone else can reproduce that or is it a strange bug from my environment?
Regards,
Simon Lessard

Simon,
I'm having a very similar problem except mine freezes at step 13. I click save all and jdeveloper freezes. Did you ever find a solution to this problem?
-brian

ADF panel executes all queries in the application module.

Hi all,
I have created an ADF Panel, which allows the user to run a few simple queries against an Oracle database done using ADF view objects and ADF view links and ADF application module.
As I add more functionality meaning more available queries meaning more ADF views and ADF view links, running the Application Modules becomes slower.
Let’s say the application module AM1 is made of 30 different view links. And application module AM2 is made of another 30 view links.
Now let’s say that ADF Panel1 references 8 view links from AM1, but it appears that all 30 view links that make up Application Module AM1 are executed before returning the result sets of the 8 view links referenced in ADF Panel1. Let me also underline the fact that each view link in Application Module AM1 can be executed independently, so the question is how I can make Jdev not to run all 30 queries in view links in the Application Module AM1 when ADF Panel1 is executed, but just the 8 that are referenced in the ADF panel1.
I hope I made my question understood.
Your inputs are greatly appreciated.
Thanks.
Bobby A.

Hi,
if this is what you see then you should file a bug because certainly its not the way it is designed to work. Note that in JDeveloper 11g - you don't specify the release you are on - the iterators by default are set to refresh deferred. Only iterators that belong to a PageDef file are executed when the page loads (you can suppress this by setting a refresh condition). Other queries are not executed
Frank

ADF BC: Optimal Application Module Runtime Configuration

Have implemented an intranet application for 1000 users using the following technologies implemented on Oracle JDeveloper 10.1.3.0.4:
ADF BC for my Services layer
JSF for my Controller layer
JSF and ADF for my View layer
ADF Model for my Data layer
The application is entirely designed by .jspx and .jsp pages (for reports), handling 106 application modules, with several entity and view objects and without any step-by-step stateful scenario. At each page, there appears a commit button to commit the end-user transaction, and a rollback button to return to the previous page.
The application is database-backed from an Oracle Database 10g, deployed on three application servers Oracle 10.1.3.1 on three different hosts on Suse Linux OS, used data-source connection and AM pooling, and load-balanced by one webcacher in front of the 3 servers.
Under stress test scenario, on a heavy system load in production environment, several ADF State Management errors were recognized that pushed the webacher and application server with the maximum load to automatic restarts due to either demand for JVM increased heap size or errors from the database that the TNS listener complained it could not handle any more connections.
Thus, we started a safety/reliability tess for all our application module components (disabling "Enable application module pooling"), and a stress test using a third-party stress tool at work, but we're still faced with several questions.
(a) What is the best strategy for AM configuration, to quit State Management at all since our application doesn't need it, and if so can it be done globally or declaratively and not individually and programmatically?
(b) Testing on application module runtime configuration, we observed that not to "hang-on" on a connection -default functionality- but to release it at the end of each HTTP request, as each deployment is concentrated on a single data-source connection pool/single user, brought less open JDBC connections on A.S.? Is this the optimal run-time configuration than the default for our case?
(c) If State Management errors are resolved, and using a software balancer before 3 application servers, regulate the traffice by 1/3 to each, is it possible that a system crash will be avoided next time?
So, our problems related to ADF tunning and State Management Errors, but we need a good advice on this topic as the overall documentation is case-specific.

Chapter 29 "Understanding Application Module Pooling" in the ADF Developer's Guide for Forms/4GL Developers on the ADF Learning Center at http://www.oracle.com/technology/products/adf/learnadf.html is the best place to start. There is no "magic" optimal configuration that I (or anyone here, really) can suggest to you. You need to understand the various "knobs and dials" that control the pool sizing and pool cleanup behavior, and then apply application-specific information to this information to derive what the optimal configuration for your particular application will be. The chapter explains how the application module pool and database connection pool interact, and in which situations it is best to change the default setting of the jbo.doconnectionpooling property to true.
Chapter 28 "Application Module State Management" describes the functionality of that facility of the framework, as well as how to release the application module in stateless, unmanaged mode if you need to.
By following the tips in the section "28.3.2.4 Setting Release Level in an Custom ADF PageLifecycle" (which builds on information introduced in section "10.5.4.1 Globally Customizing the ADF Page Lifecycle"), you can generically impose a different release level in your application without "touching" every page.
As a general rule, if you have:
(*) Lots of distinct application modules
(*) Each of which is used as a separate data control (rather than being nested inside a containing application module as described in section "8.9.3 Root Application Modules Versus Nested Application Module Usages" of the dev guide)
(*) Each of which uses the same database connection
Then as described in section "29.8 How Database and Application Module Pools Cooperate" in the guide, your application can decrease the total overall number of connections used by using the non-default setting of jbo.doconnectionpooling=true.
Applications using many distinct data controls in the span of a single user's session need to be aware of Bug# 4566186 ("PERF: ADFBINDINGFILTER CHECKS OUT ALL ADF BC DATA CONTROLS USED AT LEAST ONCE"). This issue has been fixed in JDev/ADF 11g, but apparently the fix was too complex to accommodate a 10.1.X backport. I'm trying to document a workaround and publish it on my blog, but I'm not finished with it yet.

JAG BUG ? "Set Application Module Superclass to JhsApplicationModuleImpl"

Similar Messages

Maybe you are looking for