Friday Nov 11, 2011

Attention Extension Developers: Your input wanted!

Your Input Wanted!

I've posted a lot of different topics throughout 2011, and would really like to provide info that is most important to you, the extension developer, as we head for 2012.

What are the most important areas that you want to learn more about?

Post your requests for examples and topics in the comments section. Let me know what you are struggling with, or something that you worked out, but it took way to long to figure out.  I'll take the list and do my best to provide samples over the coming months.

Please provide the version of JDeveloper that you want the topic to cover.


11gR1 = 11.1.1.x (e.g.

11gR2 = 11.1.2.x (e.g.

Thanks in advance for your comments and suggestions.  Let's get the JDev Extension community going in 2012!


John "JB" Brock
Oracle Product Manager - JDev ESDK

Tuesday Nov 08, 2011

Updating an existing file via the SDK

In my last post I showed how to add a file to an existing project, and have it show up in the project navigator. In this post, I'll give a very basic example of how to edit that file and have those changes show up in the editor right away.

This code is appropriate for both 11gR1 and 11gR2

Working with a files content in Java is usually done via the FileIO and buffers. When working with the content of a file that is already open in a JDev editor window, you want to do things a little bit differently.  Instead of modifying the file directly, we can use a TextNode and the acquireTextBuffer() method to get the TextBuffer that we will then modify. This will allow the IDE to handle the saving of the file to the filesystem, as well as updating the editor window so you see the changes as soon as they are made.


I've created a really simple xml file that I will use for this example.

I'm using a modified version of the FirstSample extension SDK sample project.

I've setup a context menu item in the extension.xml file that will only show in the Editor window when the context has an XML file being edited. Notice the "rule" setting. This is where the restriction is set.

I've done all of the work in the Controller classes handleEvent method. Here is what the try / catch block looks like

How it works

In the try / catch block above, we first make sure that we are dealing with a TextNode from the context that is passed to us from the context menu.

We then get a buffer that contains the current editors text.  We dump the entire buffer contents into a String variable and make the modifications to the string.

Once all of the modifications are made, we convert the String variable into a char array so that it can be used with the EditProcessor class.

The EditProcessor class is the magic of all this.  Instead of trying to do the buffer locks and beginEdit and endEdit methods ourselves, we can simply use the methods available to us from the EditProcessor and all of the undo state and locks will be handled for us.  It will even set the file in the editor to a dirty state so it shows that it has been modified.

The arguments that we pass into the doReplaceAll() method are:

-- char[] of the content that we want to be placed into the node.
-- the node that we want to change
-- true or false to set this change as undoable
-- a String that will show up in the Undo menu item.  It will read as "Undo <string message>"
-- the origin object.

You can read more about the other methods in the EditProcessor class from the JavaDocs


Don't use the classes to modify a file that is currently open in the IDE editor window.  Instead, use a TextNode and the TextBuffer associated with the current context to let the IDE do most of the work for you.  And remember to take advantage of the EditProcessor class instead of trying to do all of the undo and buffer locking yourself.

Friday Oct 28, 2011

Adding files to an existing project

I've run into this question a few times now in the Jdeveloper forums and thought it would be a good topic for this blog.

There are multiple ways that you can add a file to a project in JDeveloper, but only one that will really get the file properly setup in the IDE framework.

NOTE: The code shown below will work in both 11gR1 and 11gR2.

Let's start by showing the entire code block and then breaking it down.

The first thing we have to do, is get the current project, so that we know where to write this new file to.  I'm running this chunk of code from the controller class of a Context Menu item that I created to show up in the Navigator.  This provides me with a context to work with and I can get the project with a simple getProject() call.

The most important thing to note in all of the file manipulation, is that we are NOT using the standard methods.  We are using URLFileSystem and URLFactory to do all of the work.  This makes sure that the file that we create is being integrated into the IDE framework properly.

A quick lesson in Navigator view and Directory structures

The Navigator window in the IDE, is not a representation of what is on the filesystem.  This is really important to understand. That Navigator is a representation of what is in your Application (.jws) and Project (.jpr) files.  There are many things in the filesystem that do not show up in the Navigator.  One thing in particular is an empty directory is filtered out and not shown at all. 

The Navigator also does it's own grouping of source files into a folder called Application Sources and other files located in non-source-path directories are grouped under a folder called Resources.

By default any files under the /src directory will be shown in the Application Sources folder of the Navigator window.

Here is a quick look at the CreateStructure sample project and how things look on the filesystem and in the navigator window.



Notice how the directories that I created earlier (/src/test/name) do not show in the navigator window?

By default the navigator shows the directories under /src as the package id.  You can adjust how this is displayed by adjusting the package level value in the navigator display settings

Writing the file

Ok, back to the actual code now. You will see that I created a URL variable called path that points to the directory that I want to place the new file in.  I used the URLFileSystem method mkdirs() to make sure the directory is created before I try and place the file in it.  The nice thing about mkdir*s* versus mkdir is that it will create the entire path, including parent directories, if they don't already exist.

In the sample, I've just hardcoded the file name. You can get this from a dialog or whatever means you want. I then created a new URL using both the path and filename variables.

I then created a new TextNode to represent our new file, and I called a method that I wrote called writeDefaultText(). This just adds a comment to the top of our new file so that it's not empty.  The editor handles files with a little something in them better. I believe most folks that take the time to write an extension that adds a file, is going to seed that file with some pre-generated code anyway.

Here is a look at the writeDefaultText() method

After we have some content in the file we do sav of the file in it's own try / catch block. 

Updating the Navigator

Finally we need to update the Navigator window so that it displays our new file as part of the project.


While you can use the standard classes to create the file on the filesystem, it's best to use the URLFileSystem methods provided by the IDE to do this.  It's really a pretty easy thing to do once you know the proper methods to use.

As always, don't hesitate to add your comments and experiences to the comments section so others can learn from your coding experiences as well.

Thursday Oct 13, 2011

Writing, Running, and Debugging Your First JDev Extension

It recently came to my attention, that while I've talked about it many times in the different posts on this blog, I've never written a post just about getting started with your first JDev extension.

Sorry about that!

So here I will finally correct that.  I'm going to quickly cover the basics of starting an extension project from scratch and then how to run it from the IDE, and how to debug it from the IDE.  I'll toss in NOTES when there is a difference between the R1 process and the R2 process.  Most of it is still the same between versions of the IDE

To keep this at the beginners level:  R1 refers to any version of the IDE that is 11.1.1.x  where R2 is referring to anything 11.1.2.x and higher.  You can read the post about migrating extensions from R1 to R2 to get some details about what is different in the newest release of the IDE.

Creating a new Extension Application and Project

There isn't a specific Extension Application selection when you go to the New Gallery and click on Applications. You will need to create a Generic (R1) or a Custom (R2) Application.

Double-Click on the application item to start the wizard.

Just enter the Application name and click on the Finish button at the bottom of the dialog.

The wizard will create a default project called "project1".  Right-click on the project name and select "Delete Project".
You should select the option to remove the project and all of it's contents so we don't have junk code laying around in your work folder.  We don't need this default project, as we will create our own extension project next.

Right-click in the empty space of the Application Navigator and select "New Project from the context menu.

Click on the Extension Development option under the Client Tier category, and then double-click on the "Extension Project" option.

Once the project creation wizard comes up, fill out the name of the Project with some thing that you like.

Give your project an identifier.  This is what will be used in your library path.

The rest will be filled in for you with the information that you gave in the two entries above.

After the project is created, your IDE screen should look similar to this

Running the Extension

Now that we have our project created we can start work doing all of the things that I've already written blog posts about.  For the sake of making this post only about Writing, Running, and Debugging, I'm going to just paste in some code and jump ahead to how to run the code from within the IDE.

After the project is saved and built(compiled) you can right-click on the project name in the Application Navigator and select "Run Extension".  This will start another instance of the IDE with your extension loaded so you can see how it works.

NOTE: In 11gR2 you must click on the "Deploy to Target Platform" menu item before you Run or Debug the project. This is required to setup the MANIFEST.MF file with any changes that you may have made to your code, and then package everything up properly.  This is not required for the 11gR1 version of the IDE.

Any error messages or warnings that occur during the running of the extension, will show up in the log window of the Parent IDE.  This includes any System.out.* commands that you may have put into your code.

Debugging the Extension

You can set breakpoints in the code just like you would for any other type of code. To debug your extension, save the project, then right-click on the project name in the Navigator and select the "Debug Extension" option. The new instance of the IDE will start up, and you can perform the functions of your extension in that test instance.

NOTE:  The same thing applies for Debugging as Running in 11gR2.  You must click on the "Deploy to Target Platform" option before you click on the Debug menu item each time.

When the debugger hits a breakpoint in the code, the Parent instance of the IDE will show all of the debug info.


As you can see, it's pretty simple to run and debug you extension directly from the IDE.  There isn't a need to save off the extension .jar file to the extension folder and restart the IDE like a lot of people have been doing.  I hope this helps speed up development for people, and makes extension development just a little bit smoother for everyone.

As always, please post comments or questions so all developers can learn from each other.

Friday Sep 16, 2011

Find your way Home

I often find myself working in one extension.xml file, while I have another extension.xml file, from a different application, open as a reference.  When working with multiple applications open, you can get in a state where the file you're looking at in the editor, is not really located in the application that is shown in the Navigator. 

It can be confusing and a bit of a pain to have to select the application from the pulldown in the navigator and switch back and forth.

Well, you don't have to.  When you are in the editor for a specific file, just type Alt-Home and the Navigator will sync to whatever application and project that file is located in.  Quick and Easy!

 Hope this helps others as much as it has helped me.

Friday Sep 09, 2011

Cleaning up after yourself -- Deleting test extensions

When developing extensions, you often need to run the extension to test something, then make some changes, and run it again.  Often many times in a row.  If you're like me, you use the "Run Extension" option in the IDE a lot.  I'm constantly checking menu placements, and dialog look and feel, etc. while I'm developing the extension. Sometimes you run into issues where it just doesn't seem like the new changes are taking effect and you want to clean up the old extension files and start with a fresh environment.

NOTE:  These steps are for when you are doing a lot of start and stop testing.  You should not need to do this all of the time. Nor should you have to leave the caching setting, mentioned below, in place for normal development.

While JDeveloper does not officially support the uninstall of extensions (you can setup your extension to be disabled), there are ways to delete the extension while doing development so you aren't running into potential caching issues or just running into older versions of the jar files.

Cleaning things up

If you are doing everything in one jar file, it's really simple to clean things up and make sure you are not running into an older version of your extension. If your extensions is more complex, you will need to know what artifacts your extension is putting in what places, in order to clean them up.

One of the fastest ways to see what is being deployed, is to click on the "Deploy to Target Platform" in the context menu for your project.

This will open a Deployment Log window and you can see everything that is being deployed and to which locations.

Working with 11gR1

Once you know the location and name of the artifacts, close JDeveloper and open your favorite file manager to delete those files and/or directories.  After the files have been deleted, start JDeveloper back up.

Working with 11gR2

If you are working with the latest version of 11gR2, you will need to do the step above as well as a couple more steps to make sure you are not running into some potential caching issues.

First, when you add an extension via the Check for Updates feature in JDeveloper, information about that extension is added to an info file. This file is located at: <mw_home>/jdeveloper/configuration/  You will need to open this file in a text editor and remove the lines that pertain to your extension.  Be very careful that you don't remove lines that may be related to someone else's extension or an extension that you still want loaded. 

Second, when you right-click on the project and select "Run Extension" the IDE uses the default Run configuration for the project to startup the new IDE instance. To avoid potential caching issues when you are testing your extension, you will want to add a program argument to the default run configuration.

Double-click on the project name in the navigator and it will open the Project Properties dialog.  Select the Run/Debug/Profile option at the bottom left of the dialog

The Default run configuration should be selected, and you can just click on Edit to open it.

Look for the Program Arguments field and enter " -clean " in the field. With this argument you will have all of the caching cleared each time you use the Run Extension feature.  It does not remove any files or data.  It just cleans up the cache so that you are loading new versions of the extension libraries each time the command is run.

This is not something that you want to add to the actual IDE startup command, or have in place all of the time as it can remove some of the performance benefits that are new in 11gR2.


These couple of steps can save a lot of frustration at times.  I hope it helps others while doing the testing of your extensions.  I did mention above that you can setup your extensions to allow the end-user to disable them at runtime.  If you are interested in how to do that, please let a comment and if enough people are interested, I'll put up a separate post on how to do that in both R1 and R2.

As usual, comments are always welcome and encouraged!

Wednesday Aug 31, 2011

Running an audit from the command line with ojAudit

NOTE (UPDATE 9/30/2011):  This post currently does not work with 11gR2 ( because of a bug.  The bug has been fixed in 11gR2 Update 1 ( Please make sure you are running the latest release of R2.

 There is often a need to run an audit profile against a specific application, project, or even a file, and do it all from a script that may be part of a larger task.

JDeveloper provides a tool just for this task.  It's called ojAudit.

By default, you'll find this tool in the <oracle_home>/jdeveloper/jdev/bin directory.

If you run the tool without any arguments, you will get a really good usage guide, complete with examples.

The one thing that is needed by all of the examples, is a profile.  Let's look at how to create an audit profile so you can use it from the command line.

If you run JDeveloper and look under Tools >> Preferences >> Audit >> Profiles you will see a list of the profiles defined by default.

If you want to create your own profile, that is a subset of the existing rules, you can check and uncheck the various rules until you only have the ones that you want to run in your profile.  You will notice that the Profile name at the top of the dialog says that the currently selected profile has been (Modified).

Click on the Save As button and enter a new name for your profile.  You now have your own audit profile created and it can be used from the command line.  You don't need to know the location of the actual profile file, but by default, these audit profile files are saved at:

<UserDirectory>\Application Data\JDeveloper\system11.\o.jdeveloper\audit\profiles

In the example above, you see that I am using version of JDev.  In other words, 11gR1 PS4. Your systemXXXX will vary depending on what version of JDev you have installed. The rest of the path should be the same.

If you still can't locate the file, do a search on the name.  If you entered spaces in the profile name when you did the Save As in the dialog, the spaces will be replaced with " - ".  In the example that I used above, I named my profile to "My Audit Profile".  The resulting file is "my-audit-profile.xml". 

Once you have your profile created, you can use it from the command line like this:

ojaudit -profile 'My Audit Profile' -workspace <path-to-workspace(jws)-file>

If you forget exactly how you spelled your profile name, or just can't remember it because you created it a while ago, you can run:

ojaudit -profilehelp

and it will list all of the available profiles that can be used.

You can audit an entire workspace, a project, or a specific file.  If you point to a specific file, you will need to set  -classpath and -sourcepath for the files that you want to audit.

The report will output to stdout by default, so unless you are ok reading a bunch of xml from the command window, I would also add: -output report.xml

The final syntax would look something like:

ojaudit -profile 'My Audit Profile' -workspace <path-to-workspace(jws)-file> -output report.xml

Thursday Aug 25, 2011

Adding Exteneral Libraries to an Extension in 11gR2

I mentioned this recently in a comment response, but it has come up enough times now, that it deserves it's own topic.


If this is a new project in 11gR2, you should already see a MANIFEST.MF file in your project in the same location as the extension.xml.

If you don't see it, add the file in the META-INF folder along with the extension.xml file.
Add the default 3 lines to the file of:
Manifest-Version: 1.0
Bundle-ClassPath: .
<blank line>

Adding an External Library


Using the default above, and this example line, the Bundle-ClassPath would look like.

Bundle-Classpath: ., external:$ORACLE_HOME$/jdeveloper/mydir/mylib/myjar.jar

You can verify that your external library is being added to the bundle by right-clicking on the project in the navigator panel, and selecting "Deploy To Target Platform".  In the Deployment tab of the Log window, click on the link to the deployed package and open the file to make sure everything is listed as you want it to.

If you find that the MANIFEST.MF is not being merged after you do the "Deploy to Target Platform", do the following to force the merge.

Go to project properties and click on the Deployment section. Select the "Extension" deployment profile and click on Edit.

Add the path to your new MANIFEST.MF file to the "Add additional manifest files..." section.

You usually do not need to do the step above, but I have had people tell me that the merge isn't happening for some reason, and this fixes it in those cases.

Tuesday Aug 16, 2011

Don't fear the Audit -- Part 2

In Part One of this topic, I covered the basics of creating an extension that will implement the audit framework and find a specific value for the FetchSize property of a ViewObject.  Once we found the value, we sent a Warning to the Editor if the value was greater then 100.

In this part, I'll cover how to create a Transform, or fix, that the end user can select to reset the value to the recommended value of 100, automatically.

An updated .zip of the sample project, with the transform code added, can be downloaded from here.

Just as with Part One, we will work with two separate areas of code.

The declarative entries in extension.xml and the code itself in the Transform class and the Transaction class.


Previously I setup the <audit-hook> element in the extension.xml file, and I added a <rule-definition> to define what I wanted to report, when I found it. 

Similarly, I will add a <transform-definition> element to the audit-hook to define what the transform will be called, and where the transform class is located.

As you can see, the <transform-definition has one parameter of "id".  This must be a unique value.  As with the <rule-definition> I've set the label and description for this element in the resource file, instead of directly in the extension.xml.

Inside of the transform-definition is the <transform-class>.  I've added the entry that points to the class that I will be using to extend Transform.

Once you have the transform defined, you have to bind the rule to the transform, so that the IDE know which transform to display in the Editor.

Inside of the <rule-definition>, I've added a <transform-binding> element and the <transform> element which points to the ID of the transform definition that we set above.

That's it for the extension.xml.  Let's take a look at the Transform and Transaction classes now.

Transform class

The transform class must be a subclass of oracle.jdeveloper.audit.transform.Transform .

I'm going to be transforming an xml document so I'll make an explicit superclass constructor invocation of XmlTransformAdapter (for Java I would use JavaTransformAdapter).

To make sure all of this is done in a transaction, that can be undone by the end user, I've created the class FixFetchSize and called teh run() method to make everything happen.

Transaction class

The transaction class is a subclass of oracle.bali.xml.model.task.FixedNameTransactionTask.

I've overwritten the performTask() method to do the actual value replacement.

The Final Look

Once you have everything completed, if you run the extension from within the IDE (remember for 11gR2 you have to: Build, Deploy to Target Platform, then Run Extension)  you should find the audit available in the Tools >> Preferences >> Audits >> Profiles and if you run the audit against something like the Storefront demo, you will see something like this


Setting up the extension.xml to point at and use a transform is pretty simple. Getting the proper values from the file and making the modifications is where the real work is at.  In this particular example it took a lot more work to understand how to work with the XML document model and what to look for and replace, then it did to add the few lines in the extension.xml file. Hopefully this example will provide enough information to get you pointed in the right direction.

If you are going to be working with a Java file instead of XML, please take a look at the existing Extension SDK sample project called AuditRefactor for an example on how to work with java files.  It shows how to implement a transform by using the RefactoringManager, which can be very useful.

As with all my posts, if you have any comments, experiences, or questions, please add them to the post so everyone can learn from each other.

Monday Aug 15, 2011

Don't fear the Audit -- Part 1

Have you ever looked at the JDeveloper IDE, while you're writing code, and seen the warnings or errors that come up in the left and right gutters? Have you thought, "wouldn't it be great if the editor caught XYZ in my code?"

Well, you can write your own Audit extension to have the IDE look for just about anything.

The Audit Framework that comes with JDeveloper is very powerful, and really not that hard to extend.  It can help with maintaining company standards, or enforcing efficiency and protocol standards that may be set by your team.

You can even go as far as to write transforms (fixes) that will make the changes to the code for you when it finds something that you pre-defined.

For this topic, I'm going to use a project that I wrote specifically for this example.  You can download the entire project as a .zip file and place it into your work directory if you like.  I'm using JDeveloper 11gR2 for this example project. If you don't have this latest release, you can still follow along and I will try to point out the differences between the R1 syntax and the newer R2 syntax.

UPDATE (02/17/2012): There has been a nice write up about this topic, specifically done for 11gR1 over on Arvinder Singh's blog If you are using 11gR1 (11.1.1.x) then I would head over there as well.

In Part One of this topic, I'll cover the writing of the analyzer, to find the suspect code and send a report to the IDE.  In Part Two, I'll cover how to write a transform to automate the modification of what you report in the analyzer.

For most audits setting up the analyzer and having it report something to the IDE is all you need to do.

This is broken down into two parts.

  • Declarative entries in the extension.xml file
  • Analyzer class

The Extension.xml 

Let's start in the extension.xml file and work from there.

As you can see, we are going to add an <audit-hook> element to the file.  If you are using JDev 11gR1, this will go into the <hooks> section, if you are using 11gR2, this goes into the <triggers> section. Otherwise, the syntax is the same.

Let's take a closer look at each of the elements in the example above.

The <audit-hook> ID parameter must be a unique id.

The <category-definition> is what tells the IDE where to display the rules when you look at the Tools >> Preferences >> Audits >> Profiles dialog.
There is a label, description, and message parameter for the category-definition element. I've set these in a Resource file instead of putting them directly in the extension.xml file.  I've done the same thing for the <rule-definition> element. Here is what that resource file looks like.

The <rule-definition> element is really where everything is setup.

The ID parameter has to be a unique string.

The category is set to the category that you defined above, or an existing category that you know the ID of.

Enabled is set to true or false.  This is very important.  If this is set to true, the IDE will run your audit as soon as the file is loaded in the editor.  It's best to set this to False until you know what impact your audit is going to have on performance of the editor.  If you do something really time intensive in the analyzer class, it could bring the editor to a crawl.

The severity element can be set to one of four options.

  • Error
  • Warning
  • Incomplete
  • Advisory
  • Assist

All of these have a different way of displaying in the IDE when they are encountered by the audit framework. The Error value will stop the compiler from completing until the issue is resolved.

The parameter element is optional.  This is what you will use to pass any values back from the audit.  You may want to pass a value back to be used in the message that is displayed in the editor when you mouse over the issue.  Take a look at the Resource file example that I included above, and you will see how I am using the "currentsize" parameter in my message.

Multiple Rules can be setup in an audit-hook element.  However only one analyzer class can be referenced per audit-hook so it's best to group rules into common analyzers.

The <analyzer-definition> element is set to point to the analyzer class that the rules will be implemented in.

The Analyzer Class

The analyzer class is where all of the real work is done.

An analyzer class must be a subclass of oracle.jdeveloper.audit.analyzer.Analyzer.

Before Audit begins each traversal, it creates new Analyzers instances, creates and configures new Rule instances, and injects the rules into the analyzers. The @ExtensionResource annotation on the INVALID_FETCH_SIZE field instructs Audit to inject the instance of the Invalid Fetch Size rule into the field. The analyzer will need the instance stored in that field to report an issue.

In this class you can implement validation code for the workspace (application), project, document, and element.  In the traversal, the audit framework will visit the objects in the model as follows:

  • Enter Root
  • Enter Workspace
  • Enter Document
  • Enter Element
  • Exit Element
  • Exit Document
  • Exit Workspace
  • Exit Root

For this example, I don't have any validating to do in the workspace or project methods, so I'll skip down the document method and start there.

In this method I check to see if the first node in the document is "ViewObject" and if the file name ends with VO.xml. If this criteria is not met, then we set the setEnabled() method to false, disabling the audit for this traversal.

When the setEnabled() method is set to False, it stops the traversal from going any deeper into the process.  It will immediately start the return trip in the traversal and call the appropriate Exit methods.

If this method passes, then it continues down into the Enter method for the Element. You do not need to do a setEnabled(true).

Here we check to make sure that the FetchSize element exists and if so, and if it does, we check to see if a value has been set. We do our test to see if the value is greater then 100 and if so, we return the report, adding our parameter that we set in the extension.xml.

If we had created multiple rules in the extension.xml file, and they all defined some kind of audit on a ViewObject, we could do multiple validations here in this method. 

You would need to add an additional @ExtensionRegistry and "private Rule" definition for each rule at the top of the class, and then set the report variable to send out the report to the different rules as needed.

To run your audit, you can click on whichever object you want to run the audit against (Workspace, Project, file) and then click on the Build menu in the main menu bar. At the very bottom of the menu list, you will see "Audit <object>".

You'll notice in the example above, I have the Project selected in the Navigator window, and in the Build menu, it is showing "Audit Audit-Sample.jpr" which will run my audit against all of the files in the project.

To make sure your audit rule will be run, look at the Tools >> Preferences >> Audits >> Profiles dialog and make sure your rule is checked.

Getting Audit Extensions to load in 11gR2

If you are working with 11gR2, you will find that the audit extension will not fully load when you run it.  Looking at Help >> About >> Extensions you will see that your extension has "Triggers loaded" but it never gets to "Fully loaded". This is a result of the lazy loading feature introduced with 11gR2.  To get the extension to load when it's first initialized, we add the following element to the <audit-hook> element.


The value for <technology> is a technologyKey and will cause the extension to be loaded when a project is initialized that uses this technology.  For this example, I've set it to ADFbc since I am doing an audit against an ADF application with the ADFbc feature enabled.


In this post I covered the basics of creating an audit extension that audits and XML file. In Part Two I add a transform (fix) to this same project.

A very special Thank You! to Louis Tribble (Audit Framework master) and Jun Shozaki who spent countless hours traveling the internal roads of the Audit framework, and thankfully documenting them. Without these two developers, this topic would not be written.

As usual...  Comments are always encouraged.  Toss out your experiences and questions for others to learn from. All are welcome.

Wednesday Jun 15, 2011

Migrating an Existing Extension to JDeveloper 11gR2 -- Part Two

We covered the basics of how to migrate an extension that does not use the Addin.initialize() method, in Part One of this topic. If you haven't taken a look at that yet, I strongly recommend that you do.

Quick review

With the release of JDeveloper 11g R2, the extension framework is now based on OSGi.  Because of this, we no longer load all of the extensions at the same time, at the startup of the IDE.  We now use something called Lazy Loading which only loads extensions when they are needed. Because Addin.initialize() is no longer called at IDE startup, you have to think a little more carefully about of how your extension integrates into the IDE. You can't just dump everything out there and have it ready to go.

I'm going to use the Extension SDK sample project, "OpenNodes" for this migration example.

A different point of view

The first thing we need to think about, is what user interface element can be used to trigger the loading of the rest of the extension. When we open the OpenNodesAddin class, which extends Addin, we see that there are two different things going on in the initialize() method.

A new factory object is being created and a call to a method called installViewMenu, which does exactly that.  It installs a menu item into the View menu. This is exactly what we are looking for.

If we move the functionality of this method to be done declaratively in the extension.xml file's trigger-hooks section, we will be able to show the menu item in the View menu, without actually loading any of the other code in the extension.  It will not get completely initialized until the menu item is selected. Go ahead and comment out the call to the installViewMenu() method and save the file.

The existing extension.xml file also contains an action element which will need to be migrated.  You can refer back to Part One for information on how to modify the action element and add the now required controller element as well.

We'll focus on the installViewMenu() method and how we will migrate it's functionality over to a menu element in the extension.xml file.  Here is what the existing method looks like

This is a pretty simple method that adds a menu item to the View menu on the IDE main menubar. It calls the only action that we have in the extension.xml file already. It doesn't have anything in the method that defines where the menu item is supposed to be placed within the View menu itself.  We can just do the same thing and let the IDE place the menu item in the default location if we like.

Let's move over to the extension.xml file now and add our menu element to the triggers section of the file. For this example, I have already gone ahead and done the migration of the action element, as well as the changes to the reuired-bundles element.

I'm going to go ahead and add the <menu-hook> element to the bottom of the <triggers> element, right after the new <controllers> element that we added.

The <menu-hook> element has one required argument of "xmlns".  We'll set this to ""

Once you have the menu-hook element started, you can use the xml schema tool in the Structure window, just as we described in Part One of this topic. We will need to add the following elements to finish the menu-hook section.

If you are not sure what values to use for the <menubar> id and <menu> id, you can refer back to a blog post I did early in 2011 about how to work with declarative menus in jdeveloper extensions.

The new trigger-hooks section is going to look like this when it's done.

We should be able to test the new version of the extension now, by doing the steps of:

-- Build
-- Deploy to target platform
-- Run Extension

The "Deploy to Target Platform" step is required in 11gR2 before you try the "Run Extension" feature.  This builds the OSGi bundle so that it can be deployed properly.

Cleaning things up

Once you run the new version, you should come across two things.

1) There is a warning telling you that "registerDockableFactory" has been deprecated, after you do the compile (I'm doing a rebuild when I see this)

2) When you actually run the extension, a Warning will be thrown telling you that the <feature> hook doesn't have a corresponding <feature-member>.

While everything will run fine with these warnings in place, lets go ahead and clean things up.

To get rid of the runtime warning, let's add a <feature-member> element just before we start the <trigger-hooks> section.

The registerDockableFactory() method is being replaced with a trigger-hook.  You can add this to the triggers element using the schema editor in the Structures window.  The class name and id will be required.

The result will look like this.

The last thing to do, is go into the OpenNodesDockableFactory class and comment out the method that does the existing registration.  Leave the ID variable declaration though.  That's needed by other parts of the code.

Now if we rebuild, deploy, and run the extension, everything should work without any warnings or errors.


By moving the GUI elements out of the Addin.initialize() method, we can display the connections that our extension requires for the user to enable the full functionality and start usng the extension. Once the trigger is hit, the IDE will call the Addin.initialize() method and do the rest of the loading of the extension.

There may still be cases where an extension needs to load, but doesn't have a GUI element to use as a trigger.  In those cases, there is usually something else that is being added to the IDE by the extension, that can be used as a trigger.  Take a look at the Trigger Hooks Wiki for a list of trigger-hooks provided by the IDE, and some more information about each of them.

As usual...  Comments, Comments, Comments.  Toss out your experiences and questions for others to learn from. All are welcome.

Wednesday May 11, 2011

Letting the Usage Reporting feature show you the way

Some times when you're looking to start an extension project, you're just not quite sure where to start to accomplish the task that you have in mind.  Maybe it's something that is already possible in JDeveloper, but you just want to modify it in some way, or do things a little bit differently then the current functionality.

One way to look at what is going on behind the scenes, is to enable the Usage Reporting feature in JDeveloper and look at the usage log that is generated after you do a certain process in the JDeveloper UI.

prefereces setting for Usage Reporting feature

This report will not show you the code that you have to write, but it will show you the extension classes being called when you perform certain steps in the IDE, and that could be enough to give you something to look up in the javadocs and get going in the right direction 

Of course this will only happen if you have opt-in to the usage-reporting feature in the first place. Which you can do on the Preferences panel as shown above.

What the Usage Reporting feature does, is create an xml file after each IDE session. It records some of the API usages during that session. A session is defined as the time in the IDE from startup to close.

By default the usage log is saved to the location shown below on a windows machine. This install was done with the Local Users option selected during the installation. The location could vary depending on how you did your installation.

C:\Documents and Settings\<username>\Application Data\JDeveloper\system11.\o.ide.usages-tracking

The systemXXX number will vary depending on which version of JDeveloper you have installed.  The one above, shows the version number for JDeveloper (the newest version as of this post)

The Trick...

There is one trick to being able to actually look at the usage log, before it's sent off to Oracle. 
To stop the file from just being sent and erased as soon as you close JDeveloper,  you will need to set some bogus proxy info in the Tools --> Preferences --> Web Browser and Proxy panel.

proxy settings in preferences

The file will stay in that location until JDeveloper can connect again. If you don't want the file to be sent, just go ahead and delete it after you are done, or move it to another location.

The data that is gathered is used internally by the JDeveloper product management and engineering teams to help make more educated decisions on feature updates and enhancements, as well as adding new features down the road.

Here is the contents of an actual report that was created as the result of opening an extension.xml file and then opening a .java file.

<?xml version = '1.0' encoding = 'UTF-8'?>
<usages xmlns="">
   <hash n="system-info">
      <value n="build-label" v="JDEVADF_11."/>
      <value n="dev-build" v="false"/>
      <value n="guid" v="07dd49ee-012e-1000-8001-0a9a27d99b84"/>
      <value n="jdk-version" v="1.6.0_24"/>
      <value n="operating-system" v="Windows XP"/>
      <value n="product-edition" v=", oracle.j2ee, oracle.jdeveloper"/>
      <value n="product-name" v="Oracle JDeveloper 11g Release 1"/>
      <value n="product-version" v=""/>
      <value n="session-end-time" v="1305043200531"/>
      <value n="session-id" v="1305043105550"/>
      <value n="session-start-time" v="1305043105550"/>
      <value n="user-role" v="<none>"/>
    <hash n="usage-data">
      <list n="activities">
           <value n="extension-product-id" v="oracle.jdevimpl.extensiondt.editor.ExtensionManifestEditor"/>
           <value n="property-id" v="oracle.jdeveloper.extensiondt.model.ExtensionManifestNode"/>
           <value n="time-stamp" v="1305043182769"/>
           <value n="usage-type" v="OPEN_EDITOR"/>
         <value n="extension-product-id" v="oracle.jdevimpl.extensiondt.editor.ExtensionManifestEditor"/>
         <value n="property-id" v="oracle.jdeveloper.extensiondt.model.ExtensionManifestNode"/>
         <value n="time-stamp" v="1305043182832"/>
         <value n="usage-type" v="ACTIVATE_EDITOR"/>
         <value n="extension-product-id" v="oracle.ide.ceditor.CodeEditor"/>
         <value n="property-id" v="oracle.jdeveloper.model.JavaSourceNode"/>
         <value n="time-stamp" v="1305043190096"/>
         <value n="usage-type" v="OPEN_EDITOR"/>
         <value n="extension-product-id" v="oracle.ide.ceditor.CodeEditor"/>
         <value n="property-id" v="oracle.jdeveloper.model.JavaSourceNode"/>
         <value n="time-stamp" v="1305043190767"/>
         <value n="usage-type" v="ACTIVATE_EDITOR"/>

As usual, please feel free to post any questions or personal experiences in the comments.

Monday Sep 06, 2010

Declarative Menus in JDeveloper Extensions – Part Three

Part Three: Creating Gallery menus (File / New) using a Wizard

If you came to this post directly, you may want to take a look at the previous posts first.
Part One: Introduction and Creating an Action
Part Two: Creating Context Menus

Using the JDeveloper SDK Sample project, "Create Dialog" we can take a look at how the extension.xml is setup for using a Gallery item.



The nice thing about setting up a Gallery item is that it can be done with a simple Wizard.  From the File menu, select New and look for the "Gallery Item (Wizard)" in the Extension Developer folder under Client Tier.


If you don't see the Client Tier category when the dialog first opens, make sure you are on the "All Technologies" tag and not the "Current Project Technologies" tag.

Double click on the Gallery Item option and you will see a dialog like the one below.


Enter the name of your Gallery item.  You can select the Category and Folder from the pulldown list, or type in your own. You can also select the icon that will be displayed with your menu item at this time.

NOTE:  Take a careful look at the "Wizard class" name that is being created for you.  Specifically make sure that the package name in front of the actual class name is what you want it to be.

The checkbox for adding this to the Tools Menu is something that I haven't used that much. It adds the menu item using a MenuSpec object in the code.  I'll cover how to work with the main MenuBar and ToolBar menus declaratively in Part Four of this post.

Click OK once you have all of your data entered and the wizard will add the basic lines to the extension.xml file and create a new class file based on the wizard class name that you set in the dialog.


If we look at the java class created by the wizard, you'll see two methods that manage the execution of the menu item.  

The "isAvailable" method is where you add any code that will tell the IDE when to actually show your menu item.  In the example above, you can see that this menu item is only going to be shown when there is a valid Project in the application.

The "invoke" method is where you will define the UI for your wizard or perform the action for the menu item.  In the example above, a call is made to the runDialog() method in another class that does all of the UI work. 

If we take a look at what the wizard added to the extension.xml file, we'll see that  the basic elements are added in for us.  "name" (which is pulled from the generated class), "category", and "folder".


The result of this is what you see below.


Notice however, that this only shows up in the "All Technologies" tab.  If we look at the "Current Project Technologies" tab, our menu item isn't there.


The Current Project Technologies tab is a filtered list of items. It only shows items that are related to the Technology Scopes that you are using in the current project.

If you know that your extension is working with projects that have a specific Technology Scope, Java for instance, you can set the menu item to show up in the Current Project Technologies tab when that Tech Scope is present by using the "technologyKey" element, right after the "folder" element, in the extension.xml file.  If your extension is working with any project, no matter what the tech scope may be, you can still set the technologykey to "General" and it will show up for any project.



To determine what the correct value is for the different technology scopes, just click on the "selected technologies" link at the top of the dialog and you'll see a complete list.

As usual, please post up any questions or personal experiences in the comments.

Previous parts in this topic:

Part One: Introduction and Creating an Action
Part Two: Creating Context Menus

Next Steps:

-- Part Four: Creating  Toolbar and Menubar menus


profile image
My name is John 'JB' Brock.
This Blog will focus on tips and tricks for working with the JDeveloper Extension SDK.
I hope to bring clarity to some of the mysteries around developing extensions for JDeveloper.


« April 2014