[
https://issues.apache.org/jira/browse/WICKET-1445?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12581823#action_12581823
]
Niclas Hedhman commented on WICKET-1445:
----------------------------------------
Actually, not quite that fast...
IObjectStreamFactory has the weirdest construct....
What on earth's name is this for? And the close() of the patch above will now
instead be a NPE.
I think Eelco need to explain himself.
public ObjectOutputStream newObjectOutputStream(final OutputStream out) throws
IOException
{
final ObjectOutputStream oos = new ObjectOutputStream(out);
return new ObjectOutputStream()
{
protected void writeObjectOverride(final Object obj) throws IOException
{
try
{
oos.writeObject(obj);
}
catch (IOException e)
{
if (SerializableChecker.isAvailable())
{
// trigger serialization again, but this time gather
// some more info
new
SerializableChecker((NotSerializableException)e).writeObject(obj);
// if we get here, we didn't fail, while we
// should;
throw e;
}
throw e;
}
catch (RuntimeException e)
{
log.error("error writing object " + obj + ": " +
e.getMessage(), e);
throw e;
}
}
};
}
> StreamCorruptedException/PageStore/Serialization broken because
> ObjectOutputStream was not flushed
> --------------------------------------------------------------------------------------------------
>
> Key: WICKET-1445
> URL: https://issues.apache.org/jira/browse/WICKET-1445
> Project: Wicket
> Issue Type: Bug
> Components: wicket
> Affects Versions: 1.3.1, 1.3.2
> Reporter: Niclas Hedhman
> Assignee: Igor Vaynberg
> Fix For: 1.3.3
>
> Attachments: wicket-close-stream.patch
>
>
> The Objects.objectToByteArray() method incorrectly forgets to flush/close the
> ObjectOutputStream it uses. This can create corrupt object streams.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
