[
https://issues.apache.org/jira/browse/GEODE-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Bruce Schuchardt updated GEODE-2130:
------------------------------------
Description:
I hit this with a simple test using DataSerializer to write and read a java
Serializable object.
The problem I'm seeing is that there is a difference in the serialized form of
an object depending on whether DataSerializer is passed an ObjectOutputStream
or just a DataOutputStream.
With ObjectOutputStream DataSerializer.writeObject just calls writeObject on
the underlying ObjectOutputStream. With DataOutputStream, DataSerializer
constructs a new ObjectOutputStream.
The problem is that the constructor for ObjectOutputStream actually writes two
shorts to the stream. So the serialized form may or may not include those two
shorts.
DataSerializer.readObject expects to read those two shorts, because it
constructs a DSObjectInputStream unless the input is already a
DSObjectInputStream.
Because of that, just writing and reading an object and result in an
EOFException:
{noformat}
java.io.EOFException
at
java.io.ObjectInputStream$PeekInputStream.readFully(ObjectInputStream.java:2353)
at
java.io.ObjectInputStream$BlockDataInputStream.readShort(ObjectInputStream.java:2822)
at
java.io.ObjectInputStream.readStreamHeader(ObjectInputStream.java:804)
at java.io.ObjectInputStream.<init>(ObjectInputStream.java:301)
at
org.apache.geode.internal.InternalDataSerializer$DSObjectInputStream.<init>(InternalDataSerializer.java:3583)
at
org.apache.geode.internal.InternalDataSerializer.basicReadObject(InternalDataSerializer.java:2988)
at org.apache.geode.DataSerializer.readObject(DataSerializer.java:3281)
at DataSerializerJUnitTest.test(DataSerializerJUnitTest.java:20)
{noformat}
Here's a simple test of this issue:
{code}
import org.apache.geode.DataSerializer;
import org.junit.Test;
import java.io.ByteArrayInputStream;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import java.io.Serializable;
public class DataSerializerJUnitTest {
@Test
public void test() throws IOException, ClassNotFoundException {
final ByteArrayOutputStream byteArrayOutputStream = new
ByteArrayOutputStream();
ObjectOutputStream out = new ObjectOutputStream(byteArrayOutputStream);
DataSerializer.writeObject(new MyObject(), out);
out.close();
ObjectInputStream in = new ObjectInputStream(new
ByteArrayInputStream(byteArrayOutputStream.toByteArray()));
final Object object = DataSerializer.readObject(in);
}
private static class MyObject implements Serializable {
}
}
{code}
was:
I hit this with a simple test using DataSerializer to write and read a java
Serializable object.
The problem I'm seeing is that there is a difference in the serialized form of
an object depending on whether DataSerializer is passed an ObjectInputStream or
just a DataInputStream.
With ObjectInputStream DataSerializer.writeObject just calls writeObject on the
underlying ObjectInputStream. With DataInputStream, DataSerializer constructs a
new ObjectInputStream.
The problem is that the constructor for ObjectOutputStream actually writes two
shorts to the stream. So the serialized form may or may not include those two
shorts.
DataSerializer.readObject expects to read those two shorts, because it
constructs a DSObjectInputStream unless the input is already a
DSObjectInputStream.
Because of that, just writing and reading an object and result in an
EOFException:
{noformat}
java.io.EOFException
at
java.io.ObjectInputStream$PeekInputStream.readFully(ObjectInputStream.java:2353)
at
java.io.ObjectInputStream$BlockDataInputStream.readShort(ObjectInputStream.java:2822)
at
java.io.ObjectInputStream.readStreamHeader(ObjectInputStream.java:804)
at java.io.ObjectInputStream.<init>(ObjectInputStream.java:301)
at
org.apache.geode.internal.InternalDataSerializer$DSObjectInputStream.<init>(InternalDataSerializer.java:3583)
at
org.apache.geode.internal.InternalDataSerializer.basicReadObject(InternalDataSerializer.java:2988)
at org.apache.geode.DataSerializer.readObject(DataSerializer.java:3281)
at DataSerializerJUnitTest.test(DataSerializerJUnitTest.java:20)
{noformat}
Here's a simple test of this issue:
{code}
import org.apache.geode.DataSerializer;
import org.junit.Test;
import java.io.ByteArrayInputStream;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import java.io.Serializable;
public class DataSerializerJUnitTest {
@Test
public void test() throws IOException, ClassNotFoundException {
final ByteArrayOutputStream byteArrayOutputStream = new
ByteArrayOutputStream();
ObjectOutputStream out = new ObjectOutputStream(byteArrayOutputStream);
DataSerializer.writeObject(new MyObject(), out);
out.close();
ObjectInputStream in = new ObjectInputStream(new
ByteArrayInputStream(byteArrayOutputStream.toByteArray()));
final Object object = DataSerializer.readObject(in);
}
private static class MyObject implements Serializable {
}
}
{code}
> Serialized data generated by DataSerializer.writeObject depends on the type
> of OutputStream
> -------------------------------------------------------------------------------------------
>
> Key: GEODE-2130
> URL: https://issues.apache.org/jira/browse/GEODE-2130
> Project: Geode
> Issue Type: Bug
> Components: serialization
> Reporter: Dan Smith
>
> I hit this with a simple test using DataSerializer to write and read a java
> Serializable object.
> The problem I'm seeing is that there is a difference in the serialized form
> of an object depending on whether DataSerializer is passed an
> ObjectOutputStream or just a DataOutputStream.
> With ObjectOutputStream DataSerializer.writeObject just calls writeObject on
> the underlying ObjectOutputStream. With DataOutputStream, DataSerializer
> constructs a new ObjectOutputStream.
> The problem is that the constructor for ObjectOutputStream actually writes
> two shorts to the stream. So the serialized form may or may not include those
> two shorts.
> DataSerializer.readObject expects to read those two shorts, because it
> constructs a DSObjectInputStream unless the input is already a
> DSObjectInputStream.
> Because of that, just writing and reading an object and result in an
> EOFException:
> {noformat}
> java.io.EOFException
> at
> java.io.ObjectInputStream$PeekInputStream.readFully(ObjectInputStream.java:2353)
> at
> java.io.ObjectInputStream$BlockDataInputStream.readShort(ObjectInputStream.java:2822)
> at
> java.io.ObjectInputStream.readStreamHeader(ObjectInputStream.java:804)
> at java.io.ObjectInputStream.<init>(ObjectInputStream.java:301)
> at
> org.apache.geode.internal.InternalDataSerializer$DSObjectInputStream.<init>(InternalDataSerializer.java:3583)
> at
> org.apache.geode.internal.InternalDataSerializer.basicReadObject(InternalDataSerializer.java:2988)
> at org.apache.geode.DataSerializer.readObject(DataSerializer.java:3281)
> at DataSerializerJUnitTest.test(DataSerializerJUnitTest.java:20)
> {noformat}
> Here's a simple test of this issue:
> {code}
> import org.apache.geode.DataSerializer;
> import org.junit.Test;
> import java.io.ByteArrayInputStream;
> import java.io.ByteArrayOutputStream;
> import java.io.IOException;
> import java.io.ObjectInputStream;
> import java.io.ObjectOutputStream;
> import java.io.Serializable;
> public class DataSerializerJUnitTest {
> @Test
> public void test() throws IOException, ClassNotFoundException {
> final ByteArrayOutputStream byteArrayOutputStream = new
> ByteArrayOutputStream();
> ObjectOutputStream out = new ObjectOutputStream(byteArrayOutputStream);
> DataSerializer.writeObject(new MyObject(), out);
> out.close();
> ObjectInputStream in = new ObjectInputStream(new
> ByteArrayInputStream(byteArrayOutputStream.toByteArray()));
> final Object object = DataSerializer.readObject(in);
> }
> private static class MyObject implements Serializable {
> }
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
