JBoss Community

The bounds of a task should it be relative to its swimlane?

created by ooharak in jBPM - View the full discussion

I eventually come to an unexpected behavior

when planning to import/export a BPMN2 file to/from  the designer of drools-guvnor demo.

http://localhost:8080/drools-guvnor

Does anyone have any ideas on this issue?

 

If a swimlane is at (10,10) and its containing task (80,80) in a BPMN2 file,

then the web designer will display that task at (90,90) which is (10+80, 10+80), relative to the container lane.

 

However, by reading the specification of BPMN, I thought (80,80) would be correct.

 

* Rationale

The graphical diagram aspect of a process definition is serialized in the BPMN file using BPMN DI (BPMNShape etc.)

According to the Specification, The bounds of a BPMNShape are always relative to the origin point of its BPMNPlane.

> http://www.omg.org/spec/BPMN/2.0/PDF/

> 12.2.3.3 BPMNShape

 

 

* Expect:

As both swimlanes and tasks are depicted using BPMNShape,

the bounds of a task are't relative to the origin point of its container swimlane.

Instead should it just be relative to the origin point ofits owner plane, i.e. usually a process .

For instance, Figure 12.132 and Table 12.41 of the Specification

 

* Actual:

When I create an arbitrary BPMN file including a task within a lane and import it to drools-guvnor,

then the web designer displays my task relative to the lane.

This actual behavior is diffent from the expectation above.

 

* Note

I found a bug related to the issue above :

https://issues.jboss.org/browse/JBPM-3782

but it says node's x/y values in bpmn file are relative to Swimlane, which is different from my understanding.

Reply to this message by going to Community

Start a new discussion in jBPM at Community