[JBoss JIRA] (SRAMP-246) Delete derived artifact are still available
by Kurt Stam (JIRA)
Kurt Stam created SRAMP-246:
-------------------------------
Summary: Delete derived artifact are still available
Key: SRAMP-246
URL: https://issues.jboss.org/browse/SRAMP-246
Project: S-RAMP
Issue Type: Bug
Security Level: Public (Everyone can see)
Components: Persistence
Reporter: Kurt Stam
Assignee: Kurt Stam
Fix For: 0.3.1 - jBPM - bugfix 1
When a primary artifact is deleted the JCR node is moved from being rooted in '/s-ramp' to '/s-ramp-trash'. It's derived nodes are children and are moved as well. The JCR filter by '/s-ramp' still brings back the derived nodes, so they don't appear as being deleted. Opened a discussion with the ModeShape guys about this: https://community.jboss.org/message/840238#840238
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
11 years, 3 months
[JBoss JIRA] (RTGOV-289) Improve mechanism for obtainng core components where injection not possible
by Gary Brown (JIRA)
[ https://issues.jboss.org/browse/RTGOV-289?page=com.atlassian.jira.plugin.... ]
Gary Brown commented on RTGOV-289:
----------------------------------
Issues:
- need registration of epn, etc to be performed before server up, so blocking may be necessary unless another approach can be found
- could use a shared state approach for the 'singleton' components, so multiple copies could be injected in different archives, rather than relying on a single singleton being created by the container and having to obtain a reference to it
- however this approach means that EPN, ACS, IP (etc) archives would need to contain all of the relevant dependencies to resolve the beans - so for reducing dependencies/clutter/platform specific etc, it would be better if (for example) the EPN loader could obtain a reference to a previously initialized EventProcessorManager
- current use of Singleton annotation has a problem, in that if BeanManager is used to obtain bean, then using the create method appears to create a new bean instance, regardless of the singleton annotation
> Improve mechanism for obtainng core components where injection not possible
> ---------------------------------------------------------------------------
>
> Key: RTGOV-289
> URL: https://issues.jboss.org/browse/RTGOV-289
> Project: RTGov (Run Time Governance)
> Issue Type: Enhancement
> Reporter: Gary Brown
> Assignee: Gary Brown
> Fix For: 2.0.0.M1
>
>
> Currently components within the overlord-rtgov.war can use CDI injection to resolve references.
> However there are other places where the core components need to be accessed, that are outside this war, and therefore cannot benefit from CDI.
> The current approach is to gets these components to register themselves with an 'accessor' class that manages a singleton reference. The problem is that due to startup initialisation order ambiguity, if the singleton has not been set, the caller must be blocked for a period of time.
> Rather than block the deployer, it would be better to find an asynchronous approach to resolve these references.
> Previously EJBs were being used to provide access to these components - and with the injection of those EJB resource, the dependencies were automatically managed by the container. However use of EJBs resulted in client proxies, which affected performance.
> Need to rethink how these references are resolved in a way that is generally applicable.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
11 years, 3 months
[JBoss JIRA] (RTGOV-289) Improve mechanism for obtainng core components where injection not possible
by Gary Brown (JIRA)
Gary Brown created RTGOV-289:
--------------------------------
Summary: Improve mechanism for obtainng core components where injection not possible
Key: RTGOV-289
URL: https://issues.jboss.org/browse/RTGOV-289
Project: RTGov (Run Time Governance)
Issue Type: Enhancement
Reporter: Gary Brown
Assignee: Gary Brown
Fix For: 2.0.0.M1
Currently components within the overlord-rtgov.war can use CDI injection to resolve references.
However there are other places where the core components need to be accessed, that are outside this war, and therefore cannot benefit from CDI.
The current approach is to gets these components to register themselves with an 'accessor' class that manages a singleton reference. The problem is that due to startup initialisation order ambiguity, if the singleton has not been set, the caller must be blocked for a period of time.
Rather than block the deployer, it would be better to find an asynchronous approach to resolve these references.
Previously EJBs were being used to provide access to these components - and with the injection of those EJB resource, the dependencies were automatically managed by the container. However use of EJBs resulted in client proxies, which affected performance.
Need to rethink how these references are resolved in a way that is generally applicable.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
11 years, 3 months