BPMN 2.0 Processes: Saving and reloading doesn't preserve layout.

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Create a simple diagram with an open subprocess.
2. Save it and exit.
3. Reload it.

What is the expected output?

The exact same model that was saved.

What do you see instead?

While the model itself is intact, its layout is scrambled. This does only
happens if subprocesses are used in the diagram.

Please provide any additional information below.

See screenshots for effect. As additional information, JSON-serialization
of diagram before and after reloading is attached. 

Original issue reported on code.google.com by falko.ko...@googlemail.com on 31 May 2010 at 9:13

Attachments:

• subprocess_before.jpg
• subprocess_after.jpg
• before.json
• after.json

[GoogleCodeExporter]

I could reproduce this when pools are used as well.

Original comment by antoine....@gmail.com on 9 Jun 2010 at 3:14

[GoogleCodeExporter]

Dear Oryx Team! Can I ask you to force decision of this problem for BPMN 2.0. 
So we have plans to start using Oryx in Moscow Technical University for complex 
modeling (with pools) in 1-2 weeks. Thanks in advance.

Original comment by sergey.a...@gmail.com on 28 Oct 2010 at 11:22

[GoogleCodeExporter]

Hi, 
this is not a decision to make here. In fact, someone needs to dig into the 
code and find out what's going wrong. 
One solution has been proposed in our discussion group 
<http://groups.google.com/group/b3mn/browse_thread/thread/18a0f98d2291a13b#msg_c
852fdeddb1a53fb> that may or may not help you with this issue.

Original comment by mtku...@googlemail.com on 28 Oct 2010 at 4:42

[GoogleCodeExporter]

I digged a little due to some other layout issues and found this comment in 
docker.js

"TODO absoluteXY und absoluteCenterXY von einem Docker liefern falsche Werte!!!"

Which roughly translated means: "absoluteXY and absoluteCenterXY return 
incorrect values"

There seems to be a problem in the recursive calculation of these values, 
insofar as the contributing values of the edge to which the docker belongs are 
counted twice.
dividing the resulting values by two seems to result in correct values.

In fact, when importing structures in subprocesses (via importJSON) there seem 
to be layout inconsistencies which could be a result of this.

Original comment by falko.ko...@googlemail.com on 2 Nov 2010 at 10:21