Passing object with NIO

HI All
I am trying to figure out if we can pass objects when using NIO.
If anyone has any idea about this please let me know
Thanks and Regards
Pankaj Tiwari

Very tricky. If you are in non-blocking mode you will have to prepend a length word. Generally the idea is to serialize the object to a ByteArrayOutputStream, then write its length and its data to the channel. At the other end, read the length word, then read that many bytes into a byte array, wrap that in a ByteArrayInputStream, then deserialize from that.

Similar Messages

IllegalAccessError passing object with RMI

Hi,
I am playing with RMI at the moment.
I have a Remote class that implements a Remote interface
The interface has one method which the class implements and this is to get a TestSerial object.
If I run without using a Securitymanager then there are no problems.
When I do use the SecurityManager then there is an error.
This error remains even if set the permissions in the policy file to java.security.AllPermissions
Does anyone know why this is?
thanks in advance,
J
//Start the server....
java -Djava.security.policy=policy HelloServer
In main....
HelloImpl bound in registry....
//Run the client
java -Djava.security.policy=policy HelloClient
Exception in thread "main" java.lang.IllegalAccessError: tried to access TestSerial from class $Proxy0
at $Proxy.getTestSerial(Unknown Source)
     at HelloClient.main(HelloClient.jav:23)
* Remote interface
public interface Hello extends java.rmi.Remote {
    TestSerial getTestSerial() throws java.rmi.RemoteException;
* Remote class
public class HelloImpl implements Hello {
public HelloImpl() {
public TestSerial getTestSerial(){
    return new TestSerial();
* Server class
import java.io.*;
import java.rmi.*;
import java.rmi.server.*;
import java.rmi.registry.*;
public class HelloServer {
public HelloServer() {
public static void main(String args[]) {
    System.out.println("In main....");
    if (System.getSecurityManager() == null) {
      System.setSecurityManager(new SecurityManager());
    try {
      HelloImpl obj = new HelloImpl();
      Hello stub = (Hello) UnicastRemoteObject.exportObject(obj, 0);
      Registry registry = LocateRegistry.createRegistry(Registry.REGISTRY_PORT);
      registry.bind("Hello", stub);
      System.out.println("HelloImpl bound in registry....");
    catch (Exception e) {
      System.out.println("HelloImpl exception: " + e.getMessage());
      e.printStackTrace();
* Client class
import java.rmi.*;
import java.rmi.registry.*;
import java.io.*;
public class HelloClient {
public static void main(String args[]) {
   if (System.getSecurityManager() == null) {
     System.setSecurityManager(new SecurityManager());
   try {
      Registry registry = LocateRegistry.getRegistry();
      Hello obj = (Hello) registry.lookup("Hello");
      TestSerial x = obj.getTestSerial();
      System.out.println(x.getName());
    catch (Exception e) {
      System.out.println("HelloClient exception: " + e.printStackTrace();
* TestSerial class
class TestSerial implements java.io.Serializable {
public TestSerial(){
public String getName(){
    return "TestSerial";
* Policy file
grant codeBase "file:." {
permission java.net.SocketPermission "*:1024-", "accept,connect";
};

class TestSerial implements java.io.Serializable {
public class TestSerial implements java.io.Serializable {

Serializing objects with the NIO API

I try to use the new IO API to implement a non-blocking multi-threaded server that communicates with client applications by exchanging serialized objects. I started from the code in the JavaWorld article at http://www.javaworld.com/javaworld/jw-09-2001/jw-0907-merlin.html .
With the new features of J2SDK 1.4.0, you don't have to create a separate thread per connected client. That is a big improvement.
But when I read an object with ObjectInputStream, I get a java.nio.channels.IllegalBlockingModeException.
Has anybody successfully implemented such a server since the release of the new io API?
Thanks.
Jean-Robert

The ObjectStream code is basically incompatible with non blockin I/O since you must block on the stream until a whole Object is available. You could roll something like this yourself, by reading bytes until there are enough available to build the next object and then getting that Object from the buffer and shipping it to your client thread. All of this is far more trouble than just using the Stream oriented I/O.
NIO and mutilple threads are a nightmare, the whole idea of non-blocking I/O is to avoid needing multiple threads. If you do use multiple threads you will need queues between them. See the Taming the NIO Circus topic for lots of discussion of NIO and its problems. The system seems more stable in JDK 1.5, but most people haven't even started using 1.4 for production yet.
Personally I avoid ObjectStreams. They are only useful between Java applications, if I want that I just use RMI and let Java do ALL the hard work. When I write a Socket based app, its probably because the other end is a mainframe of a C program. For that reason I send data as bytes, packaged up in packets, and encode as XML or ASN.1 so the other end can interpret it.

Passing vector of object with multiple datatype

Hi
I want to pass vector of objects to stored procedure and I am facing some problem
please suggest is it possible or I need to follow some other way.
I have one class detail with two attributes i.e int ID and string course
In main I have created 1000 objects of different ID and course and pushed back the objects to avector of detail and passed to a function my function is
void bulkdatainsert(vector<detail>rec)
stmt = conn->createStatement ("begin storeBulk(:1);end;");
     stmt->setMaxIterations(rec.size());
     stmt->setMaxParamSize(1,1024);
     setVector(stmt,1,rec,"course"); // The error coming here
The error is due to the type rec .
I want to know can we pass objects of multiple data to stored procedure through OCCI ?
If not do we need to break every attribute and make vector of corresponding attributes and pass to stored procedure?
please suggest me
Thanks

I think Custom objects are not supported directly.
Following are the normal ones which you can use
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector< T > &vect,
const string &schemaName,
const string &typeName);
Intended for use on platforms
where partial ordering of function
templates is not supported, such
as Windows NT. Multibyte
support.
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<T* > &vect,
const string &schemaName,
const string &typeName);
Intended for use on platforms
where partial ordering of function
templates is supported. Multibyte
support.
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<BDouble> &vect
const string &sqltype);
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<Bfile> &vect,
const string &schemaName,
const string &typeName);
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<BFloat> &vect
const string &sqltype);
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<Blob> &vect,
const string &schemaName,
const string &typeName);
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<Blob> &vect,
const UString &schemaName,
const UString &typeName);
Sets a const Blob vector; UTF16
support.
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<Clob> &vect,
const string &schemaName,
const string &typeName);
Sets a const Clob vector;
multibyte support.
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<Clob> &vect,
const UString &schemaName,
const UString &typeName);
Sets a const Clob vector; UTF16
support.
void setVector(
Statement *stmt,
unsigned int paramIndex,
const vector<Date> &vect,
const string &schemaName,
const string &typeName);
Sets a const Date vector;
multibyte support.

Create an object with the name passed in as a string argument to a method

Hi,
I am sorry if it's too trivial for this forum , but I am stuck with the following requirements.
I have a method
void abc(String object_name, String key, String value)
method abc should first check if there exists in memory a hashmap object with the name passed in as argument object_name.
If yes ..just put the key and value there.
if not , then create the hashmap object called <object_name> and insert the key/value there.
Can anybody help me in the first step i.e, how to check if an object exists with the name passed in and if not then create it.
Will getInstance method be of any help?
Thanks in advance for your response.
-Sub-java

Dear Cotton.m,
Thanks for your suggesstion. I will remember that.
But somehow I have a strong belief that you still need to consult dictionary for exact meaning of the words like "upset" , "frustration" etc. Not knowing something in a language , that too as a beginner, does not yield frustration, but increases curiosity. And people like petes1234 are there to diminish that appetite.
To clarify above, let me compare jverd's reply to my doubt with petes1234's.
jverd said it cannot be done and suggested a work around (It was perfect and worked for me) While petes1234 , having no work in hand probably, started analysis of newbies mistakes.
jverd solved my problem by saying that it cannot be done. petes1234 acted as a worthless critic in my opinion and now joined cotton.m.
Finally, this is a java forum and I do not want to discuss human characteristics here for sure.
My apologies if I had a wrong concept or if I chose a wrong forum to ask, where people like petes1234 or Cotton.m show their geekdom by pointing out "shortfalls" rather than clearing that by perfect examples and words.
Again take it easy and Cotton.m , please do not use this forum to figure out others' frustration but be a little more focussed on solving others "Java related" problems :)
-Sub-java

Passing objects by reference in PL/SQL

Hi,
I have come across an unexpected problem using object types in PL/SQL that is causing me some grief. I'm from a Java background and am relatively new to Oracle Objects but what I'm trying to do is fairly trivial, I think. The code below illustrates the problem.
--- cut here ---
CREATE OR REPLACE TYPE test_obj_t AS OBJECT
num INTEGER,
CONSTRUCTOR FUNCTION test_obj_t RETURN SELF AS RESULT
CREATE OR REPLACE TYPE BODY test_obj_t IS
CONSTRUCTOR FUNCTION test_obj_t RETURN SELF AS RESULT IS
BEGIN
num := 0;
RETURN;
END;
END;
CREATE OR REPLACE PACKAGE test_obj_ref AS
PROCEDURE init(o IN test_obj_t);
PROCEDURE inc;
FUNCTION get_num RETURN INTEGER;
END;
CREATE OR REPLACE PACKAGE BODY test_obj_ref IS
obj test_obj_t;
PROCEDURE init(o IN test_obj_t) IS
BEGIN
obj := o;
END;
PROCEDURE inc IS
BEGIN
obj.num := obj.num + 1;
END;
FUNCTION get_num RETURN INTEGER IS
BEGIN
RETURN obj.num;
END;
END;
--- cut here ---
The object type test_obj_t holds a integer and the test_obj_ref package holds a 'reference' to an instance of the object.
To test the above code I run this PL/SQL block:
declare
obj test_obj_t;
begin
obj := test_obj_t;
test_obj_ref.init(obj);
dbms_output.put_line('obj.num='||obj.num);
dbms_output.put_line('test_obj_ref.get_num='||test_obj_ref.get_num);
test_obj_ref.inc;
dbms_output.put_line('obj.num='||obj.num);
dbms_output.put_line('test_obj_ref.get_num='||test_obj_ref.get_num);
test_obj_ref.inc;
dbms_output.put_line('obj.num='||obj.num);
dbms_output.put_line('test_obj_ref.get_num='||test_obj_ref.get_num);
end;
giving the output:
obj.num=0
test_obj_ref.get_num=0
obj.num=0
test_obj_ref.get_num=1
obj.num=0
test_obj_ref.get_num=2
It appears that the object held by the test_obj_ref package is being incremented as expected, but I would have expected the object declared in the PL/SQL block to be pointing to the same object and so should report the same incremented values.
I suspect that the object is copied in the call to test_obj_ref.init() so I end up with two object instances, one that is held by the test_obj_ref package and one in the anonymous block. Although, I thought that all IN parameters in PL/SQL are passed by reference and not copied!
Am I right?
Is passing objects by reference possible in PL/SQL, if so how?
I'm using Oracle 10.2.0.3.
Cheers,
Andy.

the object being passed to the test_obj_ref.init+ procedure is passed by reference; however, when you assign it to your package variable obj it is being copied to a new instance. you can pass object instances as parameters to procedures using the +IN OUT [NOCOPY]+ *calling mode, in which case modifications to the attributes of the passed object will be reflected in the calling scope's instance variable.
oracle's only other notion of an object reference is the +"REF <object-type>"+ datatype, which holds a reference to an object instance stored in an object table or constructed by an object view.
hope this helps...
gerard

Passing object as parameter in JSF using h:commandLink tag

Hi ,
I have a scenario in which i need to pass objects from one jsp to another using h:commandLink. i read balusC article "Communication in JSF" and found a basic idea of achieving it. Thanks to BalusC for the wonderful article. But i am not fully clear on the concepts and the code is giving some error. The code i have is
My JSP:
This commandlink is inside a <h:column> tag of <h:dataTable>
<h:commandLink id="ol3"action="ManageAccount" actionListener="#{admincontroller.action}" >
               <f:param id="ag" name="account" value="#{admincontroller.account}" />
               <h:outputText id="ot3" value="Manage Account" rendered="#{adminbean.productList!=null}" />
               </h:commandLink>
Also a binding in h:dataTable tag with the binding="#{admincontroller.dataTable}"
                  My Backing Bean:
public class CompanyAdminController {
          private HtmlDataTable dataTable;
          private Account account=new Account();
public HtmlDataTable getDataTable() {
        return dataTable;
public Account getAccount() {
         return this.account;
public void setAccount(Account account) {
          this.account = account;
      public void action(){
            account= (Account)this.getDataTable().getRowData();
       } faces-config.xml
<navigation-rule>
    <from-view-id>/compadmin.jsp</from-view-id>
    <navigation-case>
      <from-outcome>ManageAccount</from-outcome>
      <to-view-id>/manageAccount.jsp</to-view-id>
    </navigation-case>
</navigation-rule>
<managed-bean>
          <managed-bean-name>admincontroller</managed-bean-name>
          <managed-bean-class>com.forrester.AdminController</managed-bean-class>
          <managed-bean-scope>request</managed-bean-scope>
          <managed-property>
               <property-name>account</property-name>
               <property-class>com.model.Account</property-class>
               <value>#{param.account}</value>
         </managed-property>
     </managed-bean>My account object:
public class Account {
string name;
public String getName()
    return this.name;
public void setName(String name)
       this.name=name;
}I need to display #{admincontroller.account.name} in my forwarded jsp. But I am not sure whether the code i wrote in commandlink is correct. I get an error if i use <f:setActionListener> . No tag "setPropertyActionListener" defined in tag library imported with prefix "f"
Please advise.
Edited by: twisai on Oct 18, 2009 11:46 AM
Edited by: twisai on Oct 18, 2009 11:47 AM
Edited by: twisai on Oct 18, 2009 11:48 AM

Yes.. iam removed the managedproperty from faces-config. But still i get the null pointer exception on main jsp itself and i am taken to second jsp manageaccount.jsp. Did you find anything wrong in my navigation case in the action method??
                                 public String loadAccount(){
           Account account = (Account)this.getDataTable().getRowData();
            return "ManageAcct";And also can you please answer this question i have
The AdminController is the backing bean of 1st JSP and manageAccontController is the backing bean of forwarded JSP .
Can i put this action method and dataTable binding in a 2nd manageAccontController backing bean instead of AdminController. But if i do that way dataList binding in h:dataTable tag will be in first JSP backing bean and dataTable binding to 2nd JSP. Not sure how to deal with this.
{

RE: multiple named objects with the same name andinterface

David,
First I will start by saying that this can be done by using named anchored
objects and registering them yourself in the name service. There is
documentation on how to do this. And by default you will get most of the
behavior you desire. When you do a lookup in the name service (BindObject
method) it will first look in the local partition and see if there is a
local copy and give you that copy. By anchoring the object and manually
registering it in the name service you are programmatically creating your
own SO without defining it as such in the development environment. BTW in
response to your item number 1. This should be the case there as well. If
your "mobile" object is in the same partition where the service object he is
calling resides, you should get a handle to the local instance of the
service object.
Here is the catch, if you make a bind object call and there is no local copy
you will get a handle to a remote copy but you can not be sure which one!
It end ups as more or less a random selection. Off the top of my head and
without going to the doc, I am pretty sure that when you register an
anchored object you can not limit it's visibility to "User".
Sean
-----Original Message-----
From: [email protected]
[<a href="mailto:[email protected]">mailto:[email protected]]On</a> Behalf Of David Foote
Sent: Monday, June 22, 1998 4:51 PM
To: [email protected]
Subject: multiple named objects with the same name and interface
All,
More than once, I have wished that Forte allowed you to place named
objects with the same name in more than one partition. There are two
situations in which this seems desirable:
1) Objects that are not distributed, but are mobile (passed by value to
remote objects), cannot safely reference a Service Object unless it has
environment visibility, but this forces the overhead of a remote method
call when it might not otherwise be necessary. If it were possible to
place a copy of the same Service Object (with user visibility) in each
partition, the overhead of a remote method call could be avoided. This
would only be useful for a service object whose state could be safely
replicated.
2) My second scenario also involves mobile objects referencing a Service
Object, but this time I would like the behavior of the called Service
Object to differ with the partition from which it is called.
This could be accomplished by placing Service Objects with the same name
and the same interface in each partition, but varying the implementation
with the partition.
Does anyone have any thoughts about why this would be a good thing or a
bad thing?
David N. Foote
Consultant
Get Your Private, Free Email at <a href=
"http://www.hotmail.com">http://www.hotmail.com</a>
To unsubscribe, email '[email protected]' with
'unsubscribe forte-users' as the body of the message.
Searchable thread archive <URL:<a href=
"http://pinehurst.sageit.com/listarchive/">http://pinehurst.sageit.com/listarchive/</a>>
To unsubscribe, email '[email protected]' with
'unsubscribe forte-users' as the body of the message.
Searchable thread archive <URL:<a href=
"http://pinehurst.sageit.com/listarchive/">http://pinehurst.sageit.com/listarchive/</a>>

David,
First I will start by saying that this can be done by using named anchored
objects and registering them yourself in the name service. There is
documentation on how to do this. And by default you will get most of the
behavior you desire. When you do a lookup in the name service (BindObject
method) it will first look in the local partition and see if there is a
local copy and give you that copy. By anchoring the object and manually
registering it in the name service you are programmatically creating your
own SO without defining it as such in the development environment. BTW in
response to your item number 1. This should be the case there as well. If
your "mobile" object is in the same partition where the service object he is
calling resides, you should get a handle to the local instance of the
service object.
Here is the catch, if you make a bind object call and there is no local copy
you will get a handle to a remote copy but you can not be sure which one!
It end ups as more or less a random selection. Off the top of my head and
without going to the doc, I am pretty sure that when you register an
anchored object you can not limit it's visibility to "User".
Sean
-----Original Message-----
From: [email protected]
[<a href="mailto:[email protected]">mailto:[email protected]]On</a> Behalf Of David Foote
Sent: Monday, June 22, 1998 4:51 PM
To: [email protected]
Subject: multiple named objects with the same name and interface
All,
More than once, I have wished that Forte allowed you to place named
objects with the same name in more than one partition. There are two
situations in which this seems desirable:
1) Objects that are not distributed, but are mobile (passed by value to
remote objects), cannot safely reference a Service Object unless it has
environment visibility, but this forces the overhead of a remote method
call when it might not otherwise be necessary. If it were possible to
place a copy of the same Service Object (with user visibility) in each
partition, the overhead of a remote method call could be avoided. This
would only be useful for a service object whose state could be safely
replicated.
2) My second scenario also involves mobile objects referencing a Service
Object, but this time I would like the behavior of the called Service
Object to differ with the partition from which it is called.
This could be accomplished by placing Service Objects with the same name
and the same interface in each partition, but varying the implementation
with the partition.
Does anyone have any thoughts about why this would be a good thing or a
bad thing?
David N. Foote
Consultant
Get Your Private, Free Email at <a href=
"http://www.hotmail.com">http://www.hotmail.com</a>
To unsubscribe, email '[email protected]' with
'unsubscribe forte-users' as the body of the message.
Searchable thread archive <URL:<a href=
"http://pinehurst.sageit.com/listarchive/">http://pinehurst.sageit.com/listarchive/</a>>
To unsubscribe, email '[email protected]' with
'unsubscribe forte-users' as the body of the message.
Searchable thread archive <URL:<a href=
"http://pinehurst.sageit.com/listarchive/">http://pinehurst.sageit.com/listarchive/</a>>

View Object with User Defined Type input

I am trying to use a View Object with a query that requires a user defined object as an input parameter.
I have the query working with a PreparedStatement, but would like to use a View Object.
When I use the PreparedStatement, I prepare the user defined type data like this:
// get the data into an object array
Object[] wSRecObjArr = wSRec.getObjectArray();
// set up rec descriptor
StructDescriptor WSRecDescriptor = StructDescriptor.createDescriptor("WS_REC",conn);
// populate the record struct
STRUCT wSRecStruct = new STRUCT(WSRecDescriptor,conn,wSRecObjArr);
Then I can use this in the PreparedStatement like this:
OraclePreparedStatement stat = null;
ResultSet rs = null;
stat = (OraclePreparedStatement)conn.prepareStatement("Select test_pkg.test_function(?) FROM DUAL");
stat.setSTRUCT(1, wSRecStruct);
rs = stat.executeQuery();
I would like to do the same process with a View Object instead of the PreparedStatement.
My question is "How do I create the input objects"?
I obtain the View Object from the Application Module using findViewObject(). I don't actually have a connection object to pass into the StructDescriptor.createDescriptor method.
I have tried just using Java Object Arrays (Object[]) to pass the data, but that gave an error:
oracle.jbo.SQLStmtException: JBO-27122: SQL error during statement preparation.
Any help or pointers are greatly appreciated.
Thank you.
Edited by: 942120 on May 1, 2013 8:45 AM
Edited by: 942120 on May 1, 2013 8:46 AM
Edited by: 942120 on May 1, 2013 9:05 AM
Edited by: 942120 on May 1, 2013 9:06 AM

Custom domains are the way to go.
When I try to pass custom domains that represent my user defined types - it works.
However, one of the functions requires a table of a user defined type be passed in.
I tried creating a domain of the table type. It forces me to add a field during creation (in JDEV), so I tried adding a field of type Array of Element of the domain representing the user defined type.
I populate the table by setting the field I created, but the table is empty in PL/SQL (TEST_TAB.COUNT = 0).
I also tried passing the oracle.jbo.domain.Array object, but that produced an error:
java.sql.SQLException: ORA-06553: PLS-306: wrong number or types of arguments in call
I also tried passing Object[], but that produced an error:
oracle.jbo.SQLStmtException: JBO-27122: SQL error during statement preparation.
How do I properly create, and pass an domain that represents a table of a user defined type?
When I use a OraclePreparedStatement, I can pass a oracle.sql.ARRAY using stat.setARRAY.
Thank you for the help you have provided, and any future advice.
JDEV 10.1.2.3
JDBC 10.2.0.5
Edited by: 942120 on May 13, 2013 7:13 AM
Edited by: 942120 on May 13, 2013 7:16 AM

Prob. Objects with same x,y,z coordinates appear at different positions

Hello,
I am having a problem where I have multiple objects with the same x,y,z coordinates but they appear at different positions on the screen. Could this have anything to do with scaling as I scale different types of objects differently. Also, I am using the simple Universe class and the objects that have the problem are added to the Universe when the program is running.
CB

I have encoutered this problem in two occaisions. I have ships that fire bullets. When the bullets are created they are not in the same position as the ship that fired them even though they have the same translation vector.
Here is the constructor for the Bullet--------------------------------
The transform3D passed in is the transform3D from the ship------------
public Bullet(Transform3D temp3d, BoundingSphere bounds) {
     // Create the root of the branch graph
     root = new BranchGroup();
// Create a Transformgroup to scale
tScale = new TransformGroup();
tempTrans = new Transform3D();
tempTrans.setScale(0.2);
tScale.setTransform(tempTrans);
     //create the main transform and attach the shape
transRoot = new TransformGroup();
transRoot.setCapability(TransformGroup.ALLOW_TRANSFORM_WRITE);
     transRoot.setCapability(TransformGroup.ALLOW_TRANSFORM_READ);
     transRoot.addChild(new ColorCube(0.4));
trans3d = new Transform3D(temp3d);
transRoot.setTransform(trans3d);
     //setup behaviour
BulletBehaviour myBehavior = new BulletBehaviour(transRoot, trans3d);
myBehavior.setSchedulingBounds(bounds);
transRoot.addChild(myBehavior);
     // Connect up graph
     tScale.addChild(transRoot);
     root.addChild(tScale);
     //root.addChild(transRoot);
root.compile();
} // end of Cube (constructor)
The other occaision when I find this problem is when comparing the position of two objects a target and a ship.
Check out this web page for pictures and debuging output.
http://www.sfu.ca/~ceden/java3d/snapshots.htm

[svn:bz-trunk] 16395: Bug: #2621264 [Regression] Small messages not working with NIO-HTTP endpoints in LCSD/trunk.

Revision: 16395
Revision: 16395
Author:   [email protected]
Date:     2010-06-02 05:00:56 -0700 (Wed, 02 Jun 2010)
Log Message:
Bug: #2621264 Small messages not working with NIO-HTTP endpoints in LCSD/trunk.
QA: Yes
Doc: No
Checkintests: Pass
Details: This is the BlazeDS side of the fix. This wasn't a regression, it probably never worked correctly. So, in the scenario where there's a Producer and a Consumer, and Producer sends a message for the Consumer, there are 2 messages from the server. The ACK message for Producer's send, and the actual message Consumer receives. I found that the ACK message was in small form, but the actual message was not in streaming messages. This was because we never really tried to convert streamed messages into small messages before.
Modified Paths:
    blazeds/trunk/modules/core/src/flex/messaging/endpoints/BaseStreamingHTTPEndpoint.java
    blazeds/trunk/modules/core/src/flex/messaging/endpoints/StreamingAMFEndpoint.java
    blazeds/trunk/modules/core/src/flex/messaging/endpoints/StreamingHTTPEndpoint.java

Memory leak problem while passing Object to stored procedure from C++ code

Hi,
I am facing memory leak problem while passing object to oracle stored procedure from C++ code.Here I am writing brief description of the code :
1) created objects in oracle with the help of "create or replace type as objects"
2) generated C++ classes corresponding to oracle objects with the help of OTT utility.
3) Instantiating classes in C++ code and assigning values.
4) calling oracle stored procedure and setting object in statement with the help of setObject function.
5) deleted objects.
this is all I am doing ,and getting memory leak , if you need the sample code then please write your e-mail id , so that I can attach files in reply.
TIA
Jagendra

just to correct my previous reply , adding delete statement
Hi,
I am using oracle 10.2.0.1 and compiling the code with Sun Studio 11, following is the brief dicription of my code :
1) create oracle object :
create or replace type TEST_OBJECT as object
( field1 number(10),
field2 number(10),
field3 number(10) )
2) create table :
create table TEST_TABLE (
f1 number(10),f2 number (10),f3 number (10))
3) create procedure :
CREATE OR REPLACE PROCEDURE testProc
data IN test_object)
IS
BEGIN
insert into TEST_TABLE( f1,f2,f3) values ( data.field1,data.field2,data.field3);
commit;
end;
4) generate C++ classes along with map file for database object TEST_OBJECT by using Oracle OTT Utility
5) C++ code :
// include OTT generate files here and other required header files
int main()
int x = 0;
int y = 0;
int z =0;
Environment *env = Environment::createEnvironment(Environment::DEFAULT);
Connection* const pConn =
env->createConnection"stmprf","stmprf","spwtrgt3nms");
const string sqlStmt("BEGIN testProc(:1) END;");
Statement * pStmt = pConn->createStatement(sqlStmt);
while(1)
TEST_OBJECT* pObj = new TEST_OBJECT();
pObj->field1 = x++;
pObj->field2 = y++;
pObj->field3 = z++;
pStmt->setObject(1,pObj);
pStmt->executeUpdate();
pConn->commit();
delete pObj;
}

Hoe to pass object as a parameter thru webSerive

I need to pass Object as a parameter to web service:
example:
public calss returnCodes{
int ret_code;
String ret_string;
public retrunCodes(int ret_code, String ret_string){
this.ret_code = ret_code;
this.ret_string = ret_string;
public int getret_code()
return this.ret_code;
public String getret_string {
return this,ret_string
and other repc holder class to hold returnCodes
public class retrunCodesHolder implents javax.xml.rpc.Holder
returnCode value;
pubic returnCodesHolder(){
public returnCodesHolder(returnCodes value) {
this.value = value;
In Webservice I use returnCodesHolder as parameter.
WebService getReturnCodesWs I have
returnCodes retCodes = new returnCodes(04,"Good return code)
returnCodesHolder retholder = new returnCodesHolder(retCodes)
return myString;
When I am calling above webService in servlet I have code:
org.test.mytest.returnCodesHolder retHolder = new org.test.mytest.returnCodesHolder();
String result = port.getReturnCodesWs(retHolder);
How to I access /get values of returnCodes class in my calling servlet?
I am using Tomcat v 6.0

I believe (although have never used, so if I'm wrong,
correct me please!) you can do it with implicit
dynamic cursors, such as:
For rec in 'select col1, col2 from table 1 '||lstr
loop
Not really:
michaels> declare
   lstr   varchar2 (30) := ' where empno = 7788';
begin
   for rec in 'select * from emp e ' || lstr
   loop
      null;
   end loop;
end;
Error at line 1
ORA-06550: line 5, column 4:
PLS-00103: Encountered the symbol "LOOP" when expecting one of the following:
   . ( * @ % & - + / at mod remainder rem .. <an exponent (**)>
   ||

Passing Objects to Stored Procedures in SQL Queries: UNREF(REF(o))

To me it seems that passing an object (type) to stored procedure in a SQL query is quite laborious. See:
I created the following simple type:
TYPE LAST AS OBJECT (
id NUMBER
defined a function:
FUNCTION "CHECKLAST" (l LAST)
RETURN NUMBER
AS LANGUAGE JAVA
NAME 'Checker.checkLast(Last) return int';
created a typed table:
CREATE TABLE lastTable OF Last;
inserted some objects:
INSERT INTO lastTable VALUES (1003);
NOW! If want to use the StoredFunction I have to use the following syntax:
SELECT checkLast(DEREF(REF(o)))
FROM lasttable o;
That's not very elegant.
I would expect the follwing to be working:
SELECT checkLast(o)
FROM lasttable o;
Did I do something wrong? Or did Oracle forget to implement the simple syntax?
Even worse, the DEREF(REF(o)) trick does not work for member functions...
Regards,
André

I did some experiments and can now (partly) answer my own question:
Indeed, it is possible to declare a function with a parameter passed as reference:
CREATE OR REPLACE TYPE Point AS OBJECT
ID INT,
MEMBER FUNCTION distance(q IN REF Point) RETURN NUMBER
Unfortunately, it is not possible to use the REF in PL/SQL directly:
CREATE OR REPLACE TYPE BODY Point is
MEMBER FUNCTION distance(q IN REF Point) RETURN NUMBER
AS
BEGIN
RETURN q.ID - SELF.ID; --Error(5,14): PLS-00536: No navigation through REF
END;
end;
So, for testing purposes I decided to implement a very simple PL/SQL member function:
CREATE OR REPLACE TYPE BODY Point is
MEMBER FUNCTION distance(q IN REF Point) RETURN NUMBER
AS
BEGIN
RETURN -1;
END;
end;
BUT: The SQL query without VALUE is still not possible. The query
SELECT *
FROM PointTable d, PointTable e
WHERE d.distance(e) > 0
renders ORA-00904: "E": invalid identifier
Somehow the 'e' seems to be misinterpreted completely.
Illogically, the following query works fine:
SELECT *
FROM PointTable d, PointTable e
WHERE d.distance(REF(e)) > 0
The illogical thing is that, now 'E' seems to refer to a VALUE and not to REF (because we can use the REF function), whereas in Gaverill's post 'O' seems to refer to a REF (because we can use the VALUE function).
My intent is to avoid both the VALUE(o) from Gaverill's example and the REF(e) from my example, because both is not intuitive for end users...
Maybe I have to define a VIEW holding the object values... But then the end user has to use it in the FROM part.
All in all, there seems to be no elegant solution for passing objects as parameters to functions...
Regards,
André

BC4J Objects with PKs based on two columns using multiple sequences

I have implemented a BC4J object that is based on a DB table that has two columns as the primary key. For example Table A's primary key is Group and ID and there is a DB sequence for each Group.
I use the BC4J object using servlets and JSPs and I have been able to implement an Update form.
I now want to implement a Create (New) form. I've read forum postings and info on the help regarding using the SequenceImpl class to override the EntityImpl object's create method as shown below:
protected void create(AttributeList attributeList)
super.create(attributeList);
SequenceImpl mySeq = new SequenceImpl("MY_SEQ",getDBTransaction());
setMyObjectId(mySeq.getSequenceNumber());
But as you can see I need to know the value of the group attribute before I can get the next value sequence:
For new EO belonging in Group1 I need sequence from "GROUP1_SEQ", Group 2 from "GROUP2_SEQ", etc.
I would like to put this code in the Entity Object where it belongs. I guess my problem is simply how to create a new row in the View Object with a passed parameter which is the value of Group.
Thanks in advance!

Angelo:
How is the EO suppose to receive the group id? Is it from the view row? If so, the VO can pass that group id to the EO through the attribute list and calling the
public Row createAndInitRow(AttributeList nvp);
API.
Example code would be:
ViewObject myVO;
// myVO is initialized
AttributeList nvp = new oracle.jbo.NameValuePairs();
// GroupID is the attr name of the Group ID
nvp.setAttribute("GroupID", <some-value>);
Row row = myVO.createAndInitRow(nvp);
If you do that the create(AttributeList attributeList) method of the EO will receive the GroupID value coming from the VO.
Then, you can use that for your sequence.
Thanks.
Sung

Passing object with NIO

Similar Messages

Maybe you are looking for