10.1.3 EA - AppModule Quirks in BC Diagram

Create an AppModule in a BC Diagram, drag a View Object with an attached View Link into the Module, then drag the View Link into the View Object Instance in the Module. What's supposed to happen is that the View Link brings with it the "other end", i.e. the other View Object that is related to this View Link. What happens is that two images of the trailing View Object Instance appear in the AppModule, one visibly linked to the first VOI (the one the VL was dragged into) and the other standing alone within the AppModule. They are two images of the same instance. If you change the name of one of them, the other's name is also changed. If you delete one of them, they both disappear from the AppModule.
Close the diagram and re-open it and you have what you expected in the first place -- two separate VOIs with a VL attaching them, no "mirrors".
Kinda cute, but weird.
Johnny Lee

Cool. You guys are SO on it! I can't wait for the production release. Some of these kinds of quirky things are just interesting, or mild nuisances, but there are other bugs that cause loss of work. We probably shouldn't be developing production applications with this release, but we are committed to working within an Oracle environment and the JSF/ADF combo is just too promising.
Anyway, thanks, Sioban, for your efforts monitoring and responding on this forum.
Johnny Lee

Similar Messages

10.1.3 EA - Visual Properties Quirks in BC Diagram

A couple of things I've noticed using the Visual Properties settings for Associations and View Links in the Business Components Diagrams:
1) After manually setting the "Show Association Name" and "Show Role Names" on all of the Associations in a diagram, then closing the diagram and JDeveloper and opening it all back up, several of the associations have their visual property settings reverted back to being unchecked. (I just set the global properties to default to being checking in Preferences. We'll see if that changes the behaviour.)
2) A uni-directional View Link still shows both accessor names on the diagram if the "Show Role Names" is checked, even though only one accessor is available. (The unavailable one shows greyed out in the Accessor Name field of the View Link Editor -- View Link Properties page.)
No biggies, just FYI.
Johnny Lee

The second one of these is intended, as you have asked for the role names to be shown, and the dialog option controls if these names get generated into the source code or not. Clearly you were expecting something else. Were you expecting the diagram only to show the accessors generated into the source code? If so we will file an ER for this feature.

Quirks and problems after a restore from Time Machine

I had need to restore from Time Machine. It did the job, but it wasn't as easy as restoring from SuperDuper. Apple iTunes Store passwords and iTunes registrations were all lost, requiring me to re-enter passwords for the various users' authorizations to play music (and burning a user on each of those accounts.)
As I used the restored system, there were several things that weren't quite right and had to be fixed as I encountered them. One persistent issue is that Safari cookies aren't working. For all the sites that had cookies to recognize me and log me in, the cookies no longer work. I have to log into the sites on each visit now. It will remember nothing.
Is there a general purpose restore tonic that will fix all these quirks?
I'm thinking of going back to SuperDuper and just set it up back up to a different volume each night for a week to have a full week's worth and skip this Time Machine mess.

In Safari, some of that password stuff is indeed handled by cookies from the various sites, but a lot of it is actually handled by the "autofill" function which fills in form fields (in this case name and password) with whatever you last used there -- independent of any cookies.
While running Safari, go to Safari / Preferences... Go to Security. Be sure that you are allowing cookies to be received in the first place.
Then go to Autofill. See if you have the item checked to allow recording and automatic filling of "User names and passwords". You can edit the list of them to see what's currently in there.
If this doesn't do it for you, you might want to try selecting Safari / Reset Safari... from the menu bar. The panel will allow you to select which data items to reset.

A number of quirks in iTunes 10.5.2

Don't get me wrong, I love iTunes. I use it everyday. I love perfecting my library and playlists, adding artwork, fiddling with the song quality, etc. But each week it seems like there's a new quirk with the program. I never had this many issues before iTunes 10 came along. It's still usable, but the annoyances are starting to add up. If anyone has some ideas why iTunes is doing the following, I'd greatly appreciate some insight:
1. When I double click on a brand new song file (say from the desktop) it'll open in iTunes, but the program will not go to the song (even though it's playing). It just sits there on the last song played. The song displays in cover view above the song list, but I have to type in the artist name to get it to the song.
2. When I'm in a song and I want to switch from the artist column to the album column it starts to go correctly, but then jumps down to the bottom of the album list. Same happens when I switch between other columns - just jumps to the bottom of the list.
3. I used to be able to choose cover view or list view for each playlist. Now whatever I have the main library on that's the only choice I have for my hundreds of playlists. I'd like the main library to be list only since it's so long, while I enjoy looking at album covers in individual playlists. It's a weird preference to take away Apple.
Thanks for any help!

Paul Brucker wrote:
I think this is simply a real step backward for the Mac developers.
I think it is simply an error/bug, not an intentional removal of the feature.r
I inadvertently got this upgrade
The only way to inadvertently get it is to log into developer.apple.com with your developer account and download it or download it from other sources which are illegally posting it.
Or how to download to an earlier version that works?
Download -> iTunes 10.5.1
Copy the most recent iTunes library.itl file in /Music/iTunes/Previous iTunes libraries/ to /Music/iTunes/.
Install and run 10.5.1.

Quirk, Open VI Reference and LLB

I found a strange quirk opening a VI Reference (on a PXI). The "normal" method of doing this is to Open Application Reference (to find the remote site) then to Open VI Reference with the path to the desired VI. Darren Nattinger pointed out (in 2006) that you can use just the name of the VI, provided the VI is on the same folder or LLB as the current VI.
Caveat -- this behavior is present in LabVIEW 7.0 (I'm maintaining some "legacy code"). I suspect it will be present in later versions, but haven't (yet) tested it.
Quirk 1 -- if the VI in question has been downloaded (or deployed) to the remote site, and hence is in memory, then the "path" information isn't used, only the VI name. Thus if you have a path "C:\ni-rt\startup\MyVI.vi" specified in the Open VI Reference, but have absolutely nothing in the startup folder on your PXI, but have "deployed" your Project, this call will work, even though the "path" part of the argument is not valid. Useful when testing.
Quirk 2 (and the "real" reason for this post) -- if the VI is embedded in an LLB, it depends on whether or not it is a "Top-level" VI or not. Suppose the path is C:\ni-rt\startup\MyLib.llb\MyVI.vi. Consider two VIs in this LLB -- TopLevel.vi and SubVI.vi (where TopLevel is at the top level of the LLB, whereas SubVI is not).
Case A -- the LLB has actually been downloaded to the remote site. Doing an Open VI Reference, with either the full path or just the VI name, works, whether or not you are calling TopLevel or SubVI. Nothing surprising here.
Case B -- you are taking advantage of Quirk 1, and have not actually put the file on the Remote site (i.e. there is nothing at the path specified). You have, however, downloaded the LLB to the Remote. Doing an Open VI Reference with SubVI, whether with full path or just VI name, works fine. However, trying to open TopLevel fails, with Error 7 (File Not Found), whether using full path or VI name.
Summary -- Quirk 2 is that Quirk 1 has a quirk if the VI being referenced is a top level VI in an LLB.
[My apologies if this is common knowledge -- I failed to find this behavior documented, but may have overlooked something].

Bob Schor wrote:
I found a strange quirk opening a VI Reference (on a PXI). The "normal" method of doing this is to Open Application Reference (to find the remote site) then to Open VI Reference with the path to the desired VI. Darren Nattinger pointed out (in 2006) that you can use just the name of the VI, provided the VI is on the same folder or LLB as the current VI.
Caveat -- this behavior is present in LabVIEW 7.0 (I'm maintaining some "legacy code"). I suspect it will be present in later versions, but haven't (yet) tested it.
Quirk 1 -- if the VI in question has been downloaded (or deployed) to the remote site, and hence is in memory, then the "path" information isn't used, only the VI name. Thus if you have a path "C:\ni-rt\startup\MyVI.vi" specified in the Open VI Reference, but have absolutely nothing in the startup folder on your PXI, but have "deployed" your Project, this call will work, even though the "path" part of the argument is not valid. Useful when testing.
Quirk 2 (and the "real" reason for this post) -- if the VI is embedded in an LLB, it depends on whether or not it is a "Top-level" VI or not. Suppose the path is C:\ni-rt\startup\MyLib.llb\MyVI.vi. Consider two VIs in this LLB -- TopLevel.vi and SubVI.vi (where TopLevel is at the top level of the LLB, whereas SubVI is not).
Case A -- the LLB has actually been downloaded to the remote site. Doing an Open VI Reference, with either the full path or just the VI name, works, whether or not you are calling TopLevel or SubVI. Nothing surprising here.
Case B -- you are taking advantage of Quirk 1, and have not actually put the file on the Remote site (i.e. there is nothing at the path specified). You have, however, downloaded the LLB to the Remote. Doing an Open VI Reference with SubVI, whether with full path or just VI name, works fine. However, trying to open TopLevel fails, with Error 7 (File Not Found), whether using full path or VI name.
Summary -- Quirk 2 is that Quirk 1 has a quirk if the VI being referenced is a top level VI in an LLB.
[My apologies if this is common knowledge -- I failed to find this behavior documented, but may have overlooked something].
This will likely not apply at all to >= 8.0. LabVIEWs Open VI Reference will now distinguish cleanly between absolute path and just the VI name itself. If you use just the VI name the VI needs to be in memory or for a built application included in the app. The quirk about VIs in the same LLB as the currently active VI sounds very strange to me but it may still work.
However if you specify an absolute path, Open VI Reference will bark with an error if the VI can not be found at that location or if the VI already loaded in memory with that name has not the same absolute path. This is supposedly to make it very hard to create cross linking even when using VI server.
Rolf Kalbermatter
Rolf Kalbermatter
CIT Engineering Netherlands
a division of Test & Measurement Solutions

Error when I call a method of AppModule from a backing bean.

I use Jdeveloper 10.1.3.0.3 EA.
In my application I have created ApplModule and UsrView(for my table USR).
I have a login page (login.jsp and the backing login.java) with 2 fields userName and password and an ADF Faces Core command button (cmdCheckPass) and the method cmdCheckPass _action() for the click event .
The problem is that I cant call method testPassword(userName, password) which is written in AppModule. (I get error:
javax.faces.FacesException: #{backing_login.cmdCheckPass_action}: javax.faces.el.EvaluationException: java.lang.NullPointerException) from line:
answer = app.testPass(st1,st2);
The code is as follows:
public String cmdCheckPass_action() {
String answer="";
String mes = "";
String st1=this.getInputText1().getValue().toString();
String st2=this.getInputText2().getValue().toString();
AppModuleImpl app = new AppModuleImpl();
answer = app.testPass(st1,st2);
if (answer == "OK") {
return "toMain";
} else {
answer = "";
mes="Invalid username, password !!!";
FacesContext.getCurrentInstance().addMessage(null,new FacesMessage(mes));
return answer;}
I dont know if theoretically is right to call, from a backing bean, a custom method of AppModule.
Can I have any help on this?
Thanks in advance,
Panos

Some quick ideas:
Is the clip the same NTSC/PAL standard as the project you are importing it to?
Have you tried to rename the offending clip (do NOT rename any clips inside the iMovie /Media folder!!).
How much free space do you have on the iMovie project volume -- what is the size of the imported clip?

Yoga 2 Pro Win 8.1 WiDi/Miracast quirk

So since Windows update offered a new Intel graphics driver 10.18.10.3379 I've had an odd quirk in connecting to my adapter.
I will connect to my PTV3000 fine, use it and disconnect all fine. The quirk comes in after I shutdown my PC and restart it. The next time I turn on the PC when I attempt to connect to the adapter, Charms>Devices>Project I am not given a list of adapter but the list of display options: Disconnect, Duplicate, Extend, Second Screen Only ???????????????
Why after the PC has been off (overnight) and freshly started would it think it's connected to an adapter already?
I have to perform a Restart in order for it to reset properly and show me a list of devices to connect to. This is pretty annoying since I just turned the PC on int eh first place. I will note that upon shutdown, I have checked that the PTV3000 is fully disconnected and the Project menu is showing the device as eligible to connect to.
Specs
Lenovo Yoga 2 Pro
Intel Dual-Band Wireless AC 7260 - 16.6.2.1
Intel Graphics 4400 - 10.18.10.3379
Ideas?
Solved!
Go to Solution.

Since the PTV is showing as a display on the Project menu that means the PC thinks it has successfully recoginized the device and thus a successful installation. Which it obviously does not.
Uninstall the widi software.
Go to Control Panel>Devies and Printers, delete the PTV device
Reinstall Wifi drivers
Reinstall Graphics drivers
Reboot (you probably will have to after reintalling those drivers anyway)
Then try to Add/connect to the PTV again.

Outlook Quirk With E-Mail Submission of XML and/or PDF

Hello all.
I am running into an annoying little Outlook quirk when testing e-mail submissions of PDF forms. When the submit button is clicked, Outlook will open a new message window as it should, but the
Send button will not trigger when clicked. The new message window doesn't react, it just sits there normally. Also,
Ctrl+Enter doesn't work either. But...if I save the message as a draft, and then send from Outlook's drafts folder, the e-mail miraculously goes its merry way.
I have tried this out with submitting the XML data by e-mail and with sending the completed PDF via e-mail (both by revising the
Submit By E-mail button's submit node to "PDF" in the XML and by using a regular button and changing the field's control type and submit info).
Does anyone have any ideas as to why Outlook is being finicky with this? I am running Office Outlook 2003 on Windows XP Pro and I use LiveCycle Designer 7.0 to create and edit the forms and Acrobat Pro 8.0 to test. My testers also use Outlook 2003/WinXP Pro and use Acrobat Pro 7.0 to fill in the forms.
Thanks for the time and any help.
M.L. Stone
[ a g d g ]

Try: Edit > Preferences > General > Reset All Warnings

Body Width Quirk I Need Possible Suggestions On

Hi,
I hope I can explain this clearly so someone can help me. I'm very familiar with the footer and page controls in Muse. I've made a quick and dirty mock-up of a site:
http://hissign.com/newdevelopment/footertest
The above mock-up is just something I threw together to show how if you use the body width of Muse and pull on those header and footer controls, you'll have a seamless fluid site that grows accordingly.
So I'm now working on re-creating a website for a client in Muse. The idea is to make all future changes to it in-house rather than go to the web design company and ask for changes and being charged an arm and a leg. Muse will obviously allow this to happen for the client, who is excited about this capability. I suggested a major re-design (I don't like the orange color scheme that's currently there) but got shot down. Maybe that will take place down the road.
Anyway, so far my re-design is going well as it looks just about as close as I can make it to the original. But there's a quirk in the design. Below is the link to my re-design. If you want to see the actual site for comparison and developed by the client's design company, just delete everything in front of the root of the URL:
http://hissign.com/newdevelopment
Here is my quirk. Notice in both sites (mine and the original) there's that orange body where the white text goes over it. In order for me to make my version look precisely like the original, I had to draw an orange rectangle so it matches up with the top of the header. The header contains the logo, that burgundy bar at the top, and there's also the top portion of the orange "body" background box. I then drew the second orange box and joined it with the header graphic's top portion of the orange box.
The reason why I'm doing it this way is because in order to make my version look like the original - and because that orange background in the original is inside of the header, I cannot take advantage of the body width in Muse - to just make the width the orange color and let it cascade down as the page grows vertically.
So on master pages I keep having to make different masters and stretching the orange box vertically either shorter or longer, depending on the content on the page. It seems to be working but it's getting to be a pain in the @@@ because you can imagine how I may have to keep making different masters of various vertical lenghs to fit the content for that page.
So I'm reaching out to the community to just see if there's an easier way to do this - or perhaps a trick that I'm unaware of to make this happen.
Again, I know this sounds like an odd and quirky thing to do. In a perfect world, and if I was designing from scratch, I'd definitely use the Muse body width like it should be used (and like I did when I threw together the quick and dirty mock up at the first link above). But for this one the only solution I've been able to come up with is the one described above.
Thanks,
Deaf Guy

OK here's some additional information. First here is a snapshot of the fake site I made (the link to it is in my first post) to demonstrate that as you keep adding content in it (like the long text boxes), the white BG continues to just spread down further:
In the site I'm making for the client, look at the image below. See how the header is joined to a rectangle orange box that becomes the "body" where all content appears over it? (I've split it in the image below so you can see what I mean)...
I have to do it this way in order for the Muse version of the site to look like the original design. If I simply fill the browser option in Muse with Orange, this is the result:
So that's why I have to draw these orange rectangles to match up with header, but then this solution is causing me issues with the vertical length of the site. For example, on Master 1 I made the orange rectangle go all the way down to 1440 pixels. One the home page it works fine. Here is that page:
http://hissign.com/newdevelopment/index.html
But because some of the other pages may have much more content on them, using Master 1 will not work because Master 1's orange box cuts off at 1440. Here is a page with Master 1 on it that has a lot more content on it, stretching it down further. As you'll see, because the content is much longer, this Master does not work because the orange box cuts off:
http://hissign.com/newdevelopment/request-interpreter.html
So what I had to do is create a 2nd master page and stretch the orange box much longer so it is behind the content.
What I'm just trying to find out is if there is some other (i.e., easier) way to do this than how I'm doing it.
Thanks, DG

Error: demo.model.AppModule: Application module demo.model.AppModule has a view instance EmpByEmail1 that references an invalid view object.

Error: demo.model.AppModule: Application module demo.model.AppModule has a view instance EmpByEmail1 that references an invalid view object.
intially i created view object named EmpByEmail.........for that EmpByEmail1 is created in datacontrols....i deleted EmpByEmail in demo.model by mistake and i created again with same name EmpBYEmail another view object...again another EmpByEmail2 is created in datacontrols.....
while i am running i am getting following error:::::::
Error: demo.model.AppModule: Application module demo.model.AppModule has a view instance EmpByEmail1 that references an invalid view object.

Hi,
Please seasrch for the EmpByEmail1 file in ur project and remove all the details in the Page Def and UI and then try again.
Due to the EmpByEmail1 existance in page Def this may occur . It may help , still if it was not working then I am sorry for that.

How to use classes in AppModule?

How can I use custom classes (eg sqlj generated classes) in ApplicationModule?
I have some classes doing lots of db work, and I'd like to have a handle for it in my ApplicationModule.
Thx: Tib
null

Adding a class to your project is the same
regardless of project type.
You have code which calls this new class.
i.e. I assume you have methods in your AppModule which evenetually call this new class.
Add the class (Java, SQLJ, JSP, XSQL, etc) to
your project.
Add this class (if not done automatically) to
this project's deployment profile's list of files to include for deployment.
Then deploy and the class is available in the new deployment target.
I hope this is not too generic,
and can help you with your deployment.
-John
null

Form based security in WebLogic 7.0 - back button quirk

I have an application comprised of several JSPs that are protected via Form based
security and enforce an SSL connection via the appropriate declarations in the
web.xml. This aspect of the application seems to be working with the exception
of one small quirk.
If a user presses that back button until such time as the receive the container
provided login page once again, and subsequently provide a valid user id and password,
they are NOT successfully logged in. Rather, they receive the ugly 403 Forbidden
error that states that the server understood the request, but is refusing to fufill
it. This only seems to happen given the above course of events involving the
use of a back button in the browser (or selection of an item from the history
list). I suspect that this has something to do with the session id being cached
or something, but I'm not sure? Can anyone offer any assistance on this one?
Also, does anyone know of a way of preventing the user from bookmarking this container
provided login page as this also seems to be causing problems for users. If they
bookmark the first protected page of the application all is fine, but if they
bookmark the login page they receive the 403 error.
Thanks in advance!

The cure for the symtops described below was to simply add a welcome-file-list
element with appropriate welcome pages to the web.xml descriptor. It makes sense
now that I have worked it out.
Todd
"Todd Gould" <[email protected]> wrote:
>
I have an application comprised of several JSPs that are protected via
Form based
security and enforce an SSL connection via the appropriate declarations
in the
web.xml. This aspect of the application seems to be working with the
exception
of one small quirk.
If a user presses that back button until such time as the receive the
container
provided login page once again, and subsequently provide a valid user
id and password,
they are NOT successfully logged in. Rather, they receive the ugly 403
Forbidden
error that states that the server understood the request, but is refusing
to fufill
it. This only seems to happen given the above course of events involving
the
use of a back button in the browser (or selection of an item from the
history
list). I suspect that this has something to do with the session id being
cached
or something, but I'm not sure? Can anyone offer any assistance on this
one?
Also, does anyone know of a way of preventing the user from bookmarking
this container
provided login page as this also seems to be causing problems for users.
If they
bookmark the first protected page of the application all is fine, but
if they
bookmark the login page they receive the 403 error.
Thanks in advance!

SCC file quirks when viewed in Apple DVD Player

Greetings all,
I have a bunch of DVDs with closed captioning that I am implementing by importing scc files. Everything works fine and looks gereat on set top players and even most software players running Windows. However, on the Mac running Apple DVD Player 4.0 or higher (OS X), there are a bunch of quirks during CC playback. Among them are:
1. First and second lines of CC swap places (second line appears above first line)
2. First line doesn't appear at all, but second line does.
3. First line appear at top of screen while second line appears at the bottom.
These issues are showing up on multiple projects using burnt discs (not builds or in simulator). Again, the CC look fine on set top machines and on PCs running Windows.
Is this an issue with DVDSP? With Apple DVD Player? Or with the SCC files I'm working with? The files were delivered from the caption house and I have no way of opening them to look at the formatting.
Thanks in advance for any help.
Best,
Todd Nocera
G5 Dual 2.0 GHz, 3 GB RAM, DVDSP3 Mac OS X (10.3.9)

Hi Todd,
I haven't seen this abberation but I am in PAL land. It doesn't surprise me though. DVD Player has numerous issues with adhering to the spec. If you are happy with the playback on all but a Mac I would proceed. If you are really worried when you go to replication you can have this looked at by the engineers ie through MEI or their in house testing systems.
Your caption supplier will also be able to examine the integrity of your VIDEO_TS build.
For future reference most captioning houses can supply you with the encoded mpeg rather than relying on DVDSP.
Good Luck

Creating master-detail VO in custom AppModule method

Hi.
I need to create master-detail view objects and save them in DB in my custom AppModule Service Interface method.
I have the next master-detail view objects:
RegistryVO (1) -- (*) RegistryItemVO
Also I have a view link for this relationship.
I wrote custom method with next signature:
Integer createRegistry(RegistryVORowImpl registry) {
Then, I added this method to my AppModule Service Interface.
Method works fine, but only master row is created and inserted to DB (only Registry but not RegistryItem). I.e. when I testing this method in EM I fill full RegistryVO with nested RegistryItemVO data but in my method RegistryVO comes without nested RegistryItemVOs.
Then, I added SERVICE_PROCESS_CHILDREN custom property to view link (RegistryVO-RegistryItemVO) as it described in chapter 11.2.5 of Fusion Developer 's Guide for Oracle Application Development.
After that nested RegisttryVO in my method contains nested RegistryItemVOs.
In my custom method i need to set primary key for RegistryVO. So I set it in registryVO to PK attribute and set it in nested registryItems to FK attribute.
But when I doing commit I have an exception:
Exception while executing the business logic: JBO-26048: Constraint "PP_PST_RG_IT_PP_PST_FK" is violated during post operation "Insert" using SQL stat
ement " *INSERT INTO PP_PERSON_ST_RG_ITEM* (PP_PERSON_ST_RG_ITEM_ID,IN_DOCUMENT_ID,POL,FULLNAME,NAME,FULLOT,TIN,D_ROG_DT,D_REEST
_DT,PROCESSING_DT,PROCESSING_STATE_CD,RECEIVE_YEAR) VALUES (:1,:2,:3,:4,:5,:6,:7,:8,:9,:10,:11,:12)".: ORA-02291: integrity
constraint (PP_PST_RG_IT_PP_PST_FK) violated - parent key not found
It seems that only registryItems are commited in DB. I'm not seeing INSERT for Registry in ServieInterface log...
But if I setting primary key for RegistryVO when I invoking method (if I setting id for registry in test console of EM, not in method) - it's work fine .
What I'm doing wrong ?
How can I create custom method for AppModule Service Interface to create master-detail view objects and set PK for master and FK for details within method ?
Edited by: user2975420 on 19.10.2010 4:46
Edited by: user2975420 on 19.10.2010 4:49

Thanks for reply, John.
I have already added master and detail view objects usages to my AppModule Data Model. And yes, I add detail VO as child of master.
It seems that the reason of my problem is another...
I can't create master/detail when I'm setting id of master in custom method. If I set id of master in input parameter (RegisterVORowImpl) then method invokes successfull.
I noted one moment - if I'm not setting SERVICE_PROCESS_CHILDREN custom property for view link, I can set id of master VO inside method. But in this case my master VO haven't got nested RegisterItemVOs.
Edited by: user2975420 on 19.10.2010 10:23
Edited by: user2975420 on 20.10.2010 1:31

How to add User defined return type in service interface of AppModule

HI All,
I am creating service interface using AppModule. Added custom method with Java primitives and working fine. Now i want to added custom method with return type is EMP object means user defined object. How can i do so?
Thanks in advance.

I looked it up in the mean time ...
The docs state that you only can use simple data type or java.util.list or AttributeList (which is a wrapper for any viewRow). So I guess you have to somehow flatten your complex data type or create a custom VO which holds it.
http://download.oracle.com/docs/cd/E12839_01/web.1111/b31974/bcextservices.htm#CJAEHFJD
Scroll down to 11.2.3
Timo

10.1.3 EA - AppModule Quirks in BC Diagram

Similar Messages

Maybe you are looking for