[
https://issues.jboss.org/browse/SOLDER-315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12662790#comment-12662790
]
Heiko Tappe commented on SOLDER-315:
------------------------------------
I'm not sure it's related, but can I somehow solve the following problem (with
3.1.0.Final):
I would like to catch some MyOwnValidationException. Later when rendering the
view I will be able to get some validation exception info and display what kind
of error occured. But right now also the rollback of the (jpa) transaction is
kind of swallowed. So everything persisted up to MyOwnValidationException is
commited to the database.
--Heiko
> Seam catch eats all exceptions silently
> ---------------------------------------
>
> Key: SOLDER-315
> URL: https://issues.jboss.org/browse/SOLDER-315
> Project: Solder
> Issue Type: Bug
> Components: Exception Handling
> Affects Versions: 3.1.0.CR1
> Environment: ANY
> Reporter: Mike Mosiewicz
> Labels: catch, exception_handler, exception_handling, exceptions
> Fix For: 3.1.0.Final
>
>
> I observed that seam catch eats all exceptions silently. Specifically it make
> my development stop for a few hours becouse we didn't know there was model
> update exception. Here is what I investigated:
> 1. Every exception in CatchExceptionFilter appear to be handled (ie.
> catchEvent.isHandled() returned always true)
> 2. It was becouse it is marked handled in
> ExceptionHandlerDispatch.executeHandlers switch statement. I.e. here:
> ...
> switch (breadthFirstEvent.getFlow()) {
> case HANDLED:
> eventException.setHandled(true);
> return;
> case MARK_HANDLED:
> eventException.setHandled(true);
> <<<<<<<<<<<<<<<<<<
> ....
> 3. breadthFirstEvent.getFlow() is MARK_HANDLED becouse it's the default value.
> The result is that it's enough there is a matching exception handler to mark
> the caught exception as handled and stop any exception display. And it's
> highly probable that you have seam-transaction in your library, so you have
> org.jboss.seam.transaction.SimpleTransactionExceptionHandler.markTransactionRollback
> handler active. That handler is intended to mark all the exceptional
> transactions to roll back. However it is irrelevant for this handler what
> kind of exception it handles. I wouldn't say that it actually handles
> exceptions. It just wants to know of all the exception that may cause
> transaction to be invalid and to rollback the transaction. But it definitely
> should be no mean to stop further processing of exception and it was probably
> not intended to hide exceptions.
> I think there should be (at least) two kind of handlers. Those that just
> listen to exceptions and do some actions but they are not supposed to be a
> final destination for exceptions. These could be transaction related
> handlers, logging handlers, etc. And there should also be handlers that are
> able to deal with specific exceptions ultimately.
> Right now there is a notion that exception is handled if there is a handler
> for the exception. If so - SimpleTransactionExceptionHandler should not be a
> handler.
> I think it would be better if MARK_HANDLED was renamed to UNHANDLED and
> wouldn't cause eventException.setHandled(true). I would then use some
> annotation to add to those handler that are ultimate and definite, or set
> handled property programatically. The whole difference between HANDLED and
> MARK_HANDLED states is artificial. It feels like there should be a difference
> when some handler says that it really handled the situation or did not.
>
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.jboss.org/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
_______________________________________________
seam-issues mailing list
[email protected]
https://lists.jboss.org/mailman/listinfo/seam-issues
