[JBoss JIRA] Created: (JBAS-3399) Transaction timeout fires too late
by tbech (JIRA)
Transaction timeout fires too late
----------------------------------
Key: JBAS-3399
URL: http://jira.jboss.com/jira/browse/JBAS-3399
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public (Everyone can see)
Components: Transaction Manager
Affects Versions: JBossAS-4.0.3 SP1
Reporter: tbech
Assigned To: Adrian Brock
Priority: Minor
Transaction timeout is fired in the 'commit' probably. It is far too late in order to really limit the transaction.
Scenario:
- do long job in transaction (CMT);
- job is finished;
- commit;
- timeout exception is thrown;
Doesn't matter what timeout is set, server does whole job (few hours even) and at the end throws exception. Shouldn't container:
a) throw an error after timeout
b) stop the called EJB method ???
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
19 years, 10 months
[JBoss JIRA] Created: (JBSEAM-308) DataModel instance changes even though the backing list did not when using PAGE scope
by Chris Rudd (JIRA)
DataModel instance changes even though the backing list did not when using PAGE scope
-------------------------------------------------------------------------------------
Key: JBSEAM-308
URL: http://jira.jboss.com/jira/browse/JBSEAM-308
Project: JBoss Seam
Issue Type: Bug
Components: Core
Affects Versions: 1.0.1, 1.0, 1.0 beta 2, 1.0 beta 1, 1.1
Reporter: Chris Rudd
The outjectDataModel code attempts to determine if the value changed and therefore needs to be re outjected.
When not using PAGE scope this works fine. When PAGE scope is used it always reports dirty, which causes the value to be outjected.
The issue is that if the backing value didnt change, it still gets re-wrapped and outjected instead of just re-outjected.
I noticed this when working with an @DataModel, the UIData components im using cache the DataModel to ensure that it doesnt change between the updateValues and the invokeApplication phases. So hwat ends up happening is that in the updateValues phase a property gets set on my component, causing the datamodel to be re-outjected, since the UIData has already cached the original value, it never picks up the "new" one. When I get to the invokeApplication phase and it calls an action method on the component the datamodel that the UIData updated the rowIndex on isnt the same one that seam pulls the rowIndex from to inject the @DataModelSelection.
I propose the following fix to the 1.0.0 GA code ( removed PAGE scope check from dirty assignment, added else clause to if(dirty) ):
private void outjectDataModelList(String name, Object list, ScopeType scope, DataBinder wrapper)
{
Context context = getDataModelContext(scope);
Object existingDataModel = context.get(name);
boolean dirty = existingDataModel == null ||
!wrapper.getWrappedData(existingDataModel).equals(list); //TODO: delegate to the wrapper to determine equality
if ( dirty )
{
if ( list != null )
{
context.set( name, wrapper.wrap(list) );
}
else
{
context.remove( name );
}
}
// If page scope, outject the existingDataModel even if there was no change
else if( scope==ScopeType.PAGE )
{
context.set( name, existingDataModel );
}
}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
19 years, 10 months