[JBoss jBPM] - Re: Exception-Handler an Transitions
by kukeltje
ok, for the 'intimi': I stil had to respond BEFORE going to sleep.
the
anonymous wrote : No, it is clear (to me ;-))
is 'vague' because I'm a non-native English speaker.. ahh no, it was just vage.. sorry. I combined the answer to the first and second question...
now the real response
| (01:11:56) Me, Myself and I: one thing... I do not understand your response about exceptionhandling
| (01:12:16) bradatamentra: i was saying i didnt get your response
| (01:12:19) bradatamentra: either
| (01:13:11) Me, Myself and I: The action in the exception-handler is the exception..... meaning that in THAT kind of exceptions you cannot put the token anywhere. I should have mentioned that if that is what you mean
| (01:13:55) Me, Myself and I: and yes... nested exceptions between the java code and the engine are not possible
| (01:14:11) bradatamentra: so the action handler in my exception-handler
| (01:14:15) bradatamentra: why cant that move the process?
| (01:14:21) bradatamentra: or it can?
| (01:14:33) bradatamentra: gosh, i dont get it... i have read that section of docs many times
| (01:14:39) Me, Myself and I: no it can't afaik... because you are on a different level
| (01:14:43)] bradatamentra: just trying it for the first time this week on client side
| (01:15:26) Me, Myself and I: compare it to starting a program from a shell script. If the program throws an error, you are back in your shell script and cannot influence the program anymore since it 'crashed'
| (01:16:23) bradatamentra: i just think the doc is a bit contradictory. it says you cant at first, but you can.... i dont understand why you couldnt since you could get the root token and move it to a given node
| (01:17:11) bradatamentra: i figured what they were saying is in jbpm, the exeception handling doesnt mean the actual token is moving into the global exception handler... but just that the actions would be applied from the global exception handler to the token at the point the exception occurred
| (01:17:38) bradatamentra: and in that sense, the flow of control is not working like it does in java.... where you can use the scope of the exception to determine flow control
| (01:18:18) bradatamentra: instead they are saying you could use an action to determine flow control, but otherwise no flow control changes are made for handled exceptions
| (01:19:29) Me, Myself and I: Yes in that sense it sounds contradictory... but your analysis is good
And yes, that was 1.19:29 local time... 1 hour before the chat ended and even then it took me 20 minutes to cut/paste this response... should have started from scratch
If it is still not clear, please comment
View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4220790#4220790
Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4220790
17 years
[JBoss jBPM] - Re: Exception-Handler an Transitions
by bradsdavis
Ronald, maybe I am wrong, but it is talking about your Action within the <exception-handler> moving the exception to an arbitrary node. But before that, it states:
anonymous wrote :
| In the case of jBPM, control flow cannot be changed by the jBPM exception handling mechanism.
|
I guess its saying that in Java, you can have nested exceptions, and the thread execution will move to whatever point the exception is caught. In other words, JBPM does not move its token automatically to like a global exception handler... it just executes a script or action within a given exception handler. It is the responsibility of the action within the exception handler to actually move the process to another node.
anonymous wrote :
| No, it is clear (to me ;-))
|
What are you saying no about? Clearing it up?
View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4220777#4220777
Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4220777
17 years