[hibernate-issues] [Hibernate-JIRA] Commented: (HHH-2455) "Could not close a JDBC result set" output very often

[Steve Ebersole (JIRA)]

    [ http://opensource.atlassian.com/projects/hibernate/browse/HHH-2455?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=45025#comment-45025 ] 

Steve Ebersole commented on HHH-2455:
-------------------------------------

Why does this have to be so difficult?  Try it.  Does it still happen in 4.0?

The code has significantly changed as I already said and as you said you have already seen.  If you had dug a little deeper you would have seen that Hibernate proxies all JDBC resources.  Which lets it do many things, specifically 2 of which can be seen in this case:
# when a {{ResultSet}} is opened from a {{Statement}}, register that {{ResultSet}} against that {{Statement}}.  
# when {{Statement}} is closed, because of the above, we know all the {{ResultSet}} that need to be closed.

So please, when we ask you to verify something please just try to verify it.  You sitting here arguing about how its "probably still a problem" serves no purpose.

See {{org.hibernate.engine.jdbc.spi.JdbcResourceRegistry}} and the {{org.hibernate.engine.jdbc.internal.proxy}} package if you are interested in the details of how this all works

> "Could not close a JDBC result set" output very often
> -----------------------------------------------------
>
>                 Key: HHH-2455
>                 URL: http://opensource.atlassian.com/projects/hibernate/browse/HHH-2455
>             Project: Hibernate Core
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 3.2.2
>            Reporter: Dirk Feufel
>            Priority: Minor
>         Attachments: patch-for-3.6.4.patch
>
>   Original Estimate: 1h
>  Remaining Estimate: 1h
>
> If you call this type of code (like the DbTimestampType class does), the AbstractBatcher outputs a warning "Could not close a JDBC result set". 
> The problem should be that closing the prepared statement internally also closes the associated result sets and the AbstractBatcher still has a reference to this result set.  
> One possible solution might be to provide an additional method 
> 	public void closeStatement(PreparedStatement ps, ResultSet rs); 
> (as already present for closeQueryStatement) in the AbstractBatcher allowing to close both in the right order. 
> PreparedStatement ps = null;
> try {
> 	ps = session.getBatcher().prepareStatement( timestampSelectString );
> 	ResultSet rs = session.getBatcher().getResultSet( ps );
> 	....
> } finally {
> 	if ( ps != null ) {
> 		session.getBatcher().closeStatement( ps );
> 	}
> }

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira