model validation problem sequence diagram [message #1060544] |
Sun, 26 May 2013 15:04 |
Joost Kraaijeveld Messages: 273 Registered: July 2009 |
Senior Member |
|
|
Hi,
I validated my model and Papyrus complains (multiple time) about:
"The required feature 'action' of '<Action Execution Specification>
DisplayInitialiseAction' must be set"
where "DisplayInitialiseAction" is different each time, depending on the
concrete action.
It happens whenever I have a syncCall in a sequence diagram that ends in
a "Action Execution Specification" and I do not specify the "Action",
i.e. it says "<Undefined>" in the UML properties.
I can make it disappear by letting it end in a "Behaviour Execution
Specification" but what should I do to make it disappear by maintaining
the "Action Execution Specification", because it is actually an action,
and not a behaviour?
TIA
Joost
Cheers,
Joost
|
|
|
Re: model validation problem sequence diagram [message #1060567 is a reply to message #1060544] |
Mon, 27 May 2013 05:52 |
Marc-Florian Wendland Messages: 83 Registered: January 2013 |
Member |
|
|
Hi Joost,
without providing an explicit Action, you cannot do anything to let the
validation problem disappear. This is not an arbitrary validation rule
specified by Papyrus, but a syntactic obligation of the UML metamodel. An
ActionExecutionSpecification requires an action to be valid. Riddle solved!
Regards,
Marc-Florian
"Joost Kraaijeveld" wrote in message news:knt8a9$rld$1@xxxxxxxxe.org...
Hi,
I validated my model and Papyrus complains (multiple time) about:
"The required feature 'action' of '<Action Execution Specification>
DisplayInitialiseAction' must be set"
where "DisplayInitialiseAction" is different each time, depending on the
concrete action.
It happens whenever I have a syncCall in a sequence diagram that ends in
a "Action Execution Specification" and I do not specify the "Action",
i.e. it says "<Undefined>" in the UML properties.
I can make it disappear by letting it end in a "Behaviour Execution
Specification" but what should I do to make it disappear by maintaining
the "Action Execution Specification", because it is actually an action,
and not a behaviour?
TIA
Joost
|
|
|
|
Re: model validation problem sequence diagram [message #1060605 is a reply to message #1060567] |
Mon, 27 May 2013 09:38 |
Joost Kraaijeveld Messages: 273 Registered: July 2009 |
Senior Member |
|
|
Hi Marc-Florian,
Thanks for the answer.
On 27-05-13 07:52, Marc-Florian Wendland wrote:
> Hi Joost,
>
> without providing an explicit Action, you cannot do anything to let the
> validation problem disappear. This is not an arbitrary validation rule
> specified by Papyrus, but a syntactic obligation of the UML metamodel.
> An ActionExecutionSpecification requires an action to be valid. Riddle
> solved!
I should have asked the right question ;-).
The practical problem at hand hand is that if I select the "+" just
right after the action I can choose between ~41 actions. If I select
CallBehaviorAction or CallOperationAction (and many other actions) I
have to choose "a parent element of the new object". I have no idea what
to choose from the options that are given. Most things are greyed out,
but some become active if I click on them. But I have no idea which
should or could be the right one. Should the interaction (which is
displayed by the sequence diagram) that contains the lifeline with the
action execution specification be the parent? And if so, what does this
mean?
TIA
Joost
Cheers,
Joost
|
|
|
Powered by
FUDForum. Page generated in 0.02765 seconds