Eclipse Community Forums: Papyrus » Corrupt Ordering of Interaction Fragments in Sequence Diagram

Help

Home

Home » Modeling » Papyrus » Corrupt Ordering of Interaction Fragments in Sequence Diagram

Show: Today's Messages :: Show Polls :: Message Navigator

Corrupt Ordering of Interaction Fragments in Sequence Diagram [message #469839]

Tue, 21 April 2009 07:09

Eclipse User

Originally posted by: florianwendland.freenet.de

Hello @Papyrus,

yesterday, i've explored a bit the capabiliy of Papyrus for modelling
sequence diagrams. I found out, that it causes some trouble, if you want
to insert an instance of InteractionFragment (InteractionUse,
CombinedFragment) into an existing sequence diagram as the first
interaction fragment od the sequence. Graphically it is possible, but
the corresponding xmi is corrupted. The ordering of the interaction
fragments is established by their appearance in the xmi and every
additional, non-chronological placement of an interaction fragment will
be add as the last element in the interaction's xmi.

If one create a sequence diagram straight forward without changes in the
structure of the sd, you will have a correct xmi file. But i don't
think, that this is desired, because the refinement or restructuring of
models is common behavior, i think.

Are there plans, when the behavior of sequence diagram and their
corresponding interaction xmi will be more stable?

Thanks anyway
Marc-Florian Wendland

Report message to a moderator

Re: Corrupt Ordering of Interaction Fragments in Sequence Diagram [message #469842 is a reply to message #469839]

Thu, 30 April 2009 09:16

Patrick Tessier

Messages: 341
Registered: July 2009
Location: Paris Saclay, France

Senior Member

This comment is for papyrus I (www.papyrusuml.org) ?

-----
Patrick

"Marc-Florian Wendland" <florianwendland@freenet.de> a

Report message to a moderator

Re: Corrupt Ordering of Interaction Fragments in Sequence Diagram [message #469844 is a reply to message #469842]

Thu, 30 April 2009 12:51

Eclipse User

Originally posted by: florianwendland.freenet.de

Patrick,

yes, i'm using Papyrus 1.11.0. The sequence diagramming feature has a
lot of bugs. Perhaps, when you resize the heihgt of a lifeline, all
events on that line, will be moved.

I've tried several hours, but i wasn't able to create a very simple
sequence diagram with a combined fragment and some interaction operands,
that produce valid non-corrupting UML model structure...

I will raise several bug reports in the future.

Thanks
Marc-Florian

Patrick Tessier schrieb:
> This comment is for papyrus I (www.papyrusuml.org) ?
>
> -----
> Patrick
>
> "Marc-Florian Wendland" <florianwendland@freenet.de> a écrit dans le message
> de news: gsjr86$qkb$1@build.eclipse.org...
>> Hello @Papyrus,
>>
>> yesterday, i've explored a bit the capabiliy of Papyrus for modelling
>> sequence diagrams. I found out, that it causes some trouble, if you want
>> to insert an instance of InteractionFragment (InteractionUse,
>> CombinedFragment) into an existing sequence diagram as the first
>> interaction fragment od the sequence. Graphically it is possible, but the
>> corresponding xmi is corrupted. The ordering of the interaction fragments
>> is established by their appearance in the xmi and every additional,
>> non-chronological placement of an interaction fragment will be add as the
>> last element in the interaction's xmi.
>>
>> If one create a sequence diagram straight forward without changes in the
>> structure of the sd, you will have a correct xmi file. But i don't think,
>> that this is desired, because the refinement or restructuring of models is
>> common behavior, i think.
>>
>> Are there plans, when the behavior of sequence diagram and their
>> corresponding interaction xmi will be more stable?
>>
>> Thanks anyway
>> Marc-Florian Wendland
>
>

Report message to a moderator