[StephanS]

Hi there,

I discovered another problem with converging gateways. Maybe this one is not related to the converging function but maybe then a malfunction of using "DerivateCase" in order to hide the "next-user-mask" together with an following XOR. But let me show you what I discovered:

1. I`ve designet a little process, which spread to 3 parallel tasks and after these 3 tasks (Task 2,3,4) are completed goes to a final task (Task 5).
In every Task, I inserted a trigger with the PMFDerivateCase (in order to hide the next-user-mask) in the "Before Assignment" sectrion as described at the wiki

Code: Select all

PMFDerivateCase(@@APPLICATION, @%INDEX);
G::header("Location: casesListExtJsRedirector");
die();

The following picture shows, that everything works as designed, the "converging parallel gateway" wait for Task 3&4 to complete before

8e78d417-5c71-4e9b-a8c2-196d900fba25.png (54.45 KiB) Viewed 5254 times

2. Now, I want Task2 to decide, weather a message is send or not before going on. This is done by a XOR-Gateway after Task2 with the corresponding paths to the message-event and the converging parallel gateway. If we now complete Task2, Task 5 is immediately activated and the converging function is disregarded. This behaviour is shown regardless the XOR decides to send the message or not. The following screens shows, that Task2 is completed and Task5 is activated:

1f4a8eac-45a7-4012-b959-e45e467db8de.png (49.12 KiB) Viewed 5254 times

3. If I now remove the "PMFDerivateCase" from Task 2, everything is back to the expected behaviour. Regardless of the XOR-Gateway and its functions, the converging parallel Gateway is waiting for all previous Tasks to complete. See here:

02ff151a-c56b-4111-b59c-f99123501b4a.png (50.22 KiB) Viewed 5254 times

Conclusion: The PMFDerivateCase together with the XOR causes the converging parallel function to be disregarded !

Hopefully I did not designed anything far away from BPMN2. Is there another way to hide the "next-user-mask" instead of using the DerivateCase-Function ? Or is this in fact a malfunction/bug ? Maybe it is worth a note in the wiki, to be careful with the PMFDerivateCase ...
Possibly a solution in the future is a Designer-option to turn off the "next-user-mask" in Tasks instead of using trigger-code...

[amosbatto]

I wonder if killing the processing is causing some problem. What happens if you remove die() from your code?

[StephanS]

Hi,
same behaviour without the line "die();"

[zainab]

Hello StephanS,

As a workaround, you can change the information in the Routing Screen template.

The template for the Routing Screen can be found in the following directory:

Code: Select all

/opt/processmaker/workflow/engine/templates/cases/cases_ScreenDerivation.html

Change the information it displays accordingly.

Though what you said is weird, we will test and feedback you soon with the results.

Hope this helps.

Best Regards,
Zainab Sabunwala

[StephanS]

Hello Zainab,

thank you for pointing out the workaround.
I will wait for your feedback.
Have a nice weekend...

[trovetaira]

well done.

[StephanS]

Yes,

this behaviour is fixed with PM 3.2.
Just tested and now, derivate case with two following gateways does not disregard the converging-function any more.