Do you have a comment on this Henning? IMO this should be addressed for 2.3.2.
Scott Scott Eade wrote: >It looks like Henning's change in revision 221603 is the culprit. > >The commit message is: > > > >>Catch InvokationTargetExceptions and log them at error level. I was >>able to provoke this by rapidly firing and aborting HTTP requests for >>an action at a Tomcat server. Sooner or later, the container would get >>out of step and suddently the invoked methods would disappear from the >>class loader. >> >>I was also thinking about >> >> Throwable t = ite.getTargetException(); >> if (t instanceof Exception) >> { >> throw (Exception) t; >> } >> else >> { >> throw new TurbineException(t); >> } >> >>but decided against it, because this would be a spurious and hard to >>reproduce (and debug) error reported by the users. Discussion wanted. >> >> > >Another way of provoking this exception is to have an action method thus: > > public void doThrowexception(RunData data, Context context) > throws Exception > { > throw new MyException(".doThrowexception() - Goody!"); > } > >Scott > > Scott Eade wrote: >>Should an exception thrown in an action result in execution of the >>screen configured as "screen.error" in TurbineResources.properties? >> >>I had kind of assumed that it should, but it certainly does not - >>VelocityActionEvent simply logs the exception and then carries on as if >>all is well. This is not that great because it means the user receives >>no feedback that an exception has occurred. >> >>Scott >> > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]