Messages

1

General Board / Putting States into Regions

« on: September 26, 2012, 11:17:13 pm »

Hi all,

can anyone tell me how to (reproducibly) put states into regions in a State Machine? For example, take the following State Machine:
 

I'd expect "State2" to be a member of region "a" and "State3" a member of region "b", but when exporting the XMI, both are put in a third region instead:

Code: [Select]

<subvertex xmi:type="uml:State" xmi:id="EAID_1FC64472_591A_4fa3_AF50_8441A75D99C1" name="State1" visibility="public" isSubmachineState="false">
    <region xmi:type="uml:Region" xmi:id="EAID_SP000025_591A_4fa3_AF50_8441A75D99C1" name="a" visibility="public"/>
    <region xmi:type="uml:Region" xmi:id="EAID_SP000054_591A_4fa3_AF50_8441A75D99C1" name="b" visibility="public"/>
    <region xmi:type="uml:Region" xmi:id="EAID_SR000009_57C6_4772_A2DE_A30F77B39290" name="EA_Region9" visibility="public">
        <subvertex xmi:type="uml:State" xmi:id="EAID_5641A017_130C_4b43_9FB1_C43FB395F020" name="State2" visibility="public" isSubmachineState="false"/>
        <subvertex xmi:type="uml:State" xmi:id="EAID_1B6F5722_D9E1_4411_9DD2_B794640437C0" name="State3" visibility="public" isSubmachineState="false"/>
    </region>
</subvertex>

Sometimes, dragging the substates around in the region they should belong to seems to fix this, sometimes moving the dashed line between the two regions up or down a few pixels seems to fix it, but sometimes there's nothing that seems to work. Is there any way to reliably assign a state to a region (preferrably without creating new states which would be very painful when changing an existing diagram)?

Thanks,
Rainer.

2

General Board / Re: Call/Signal Triggers with Arbitrary Specificat

« on: February 09, 2012, 11:39:38 pm »

Anyone?

3

General Board / Call/Signal Triggers with Arbitrary Specification

« on: February 01, 2012, 01:41:11 am »

Hi all,

please bear with me here, I do realize that my use case might not be the standard one . I'm using EA 9.0, if that should matter.

When adding a Trigger to a StateMachine Diagram [1], I can freely define its Specification if it is set to Type "Change" or "Time". If it is set to "Call" or "Signal", however, EA provides me with an Input window where I can select a method or a signal, respectively, that has to be defined somewhere in the model. Since my use case has EA only as a part of a bigger tool landscape, however, neither the caller nor the State Machine's class is specified in the model -- which means that there is nothing in the model that I could choose in this step. Is there some way that I can avoid this and simply enter the specification by hand?

I'd like to avoid (ab)using "Change" or "Time" triggers, since their intuitive meaning is contrary to what the triggers mean in the model's context. Furthermore, I guess I could somehow import the model's external specification from the other data sources we use in our tool landscape, but this doesn't sound too easy, and would additionally impose the overhead of keeping the imported interface in sync with the external spec on the designer.

Thanks and best regards,
Rainer.


[1] More precisely, I'm adding a Stereotype extending a Trigger (without adding/changing anything in the Stereotype), since I was unable to find Triggers themselves in the Toolbox.

4

Automation Interface, Add-Ins and Tools / Re: Constraint's attached links with shape scripts

« on: September 01, 2011, 06:26:56 pm »

Thank you very much for the workaround, it works nicely.

5

Automation Interface, Add-Ins and Tools / Re: Constraint's attached links with shape scripts

« on: August 23, 2011, 11:13:46 pm »

The problem persists with EA 9.0.907. Is there no workaround for this?

6

Automation Interface, Add-Ins and Tools / Re: Constraint's attached links with shape scripts

« on: August 04, 2011, 10:59:35 pm »

*bump* Anyone?

Is this a bug? Should I report it somewhere? Or is there some trick to get this done correctly?

7

Automation Interface, Add-Ins and Tools / Constraint's attached links with shape scripts

« on: June 21, 2011, 05:38:10 pm »

Hi again,

unfortunately I ran into another problem with shape scripts for Constraints: if I add a shape script to a stereotype of Constraint and apply the stereotype to a Constraint, its attached links are not drawn anymore. Checking via Right Click -> Advanced -> Set Attached Links they are still present, but the dashed line in the diagram is gone.

Which shape script is used does not seem to matter, I simplified it to

Code: [Select]

shape main {
    rectangle(0, 0, 100, 100);
}but the problem still occurs.

Again, thanks for your help,
Rainer.

8

Automation Interface, Add-Ins and Tools / Re: Shape Script for Stereotype of Constraint

« on: June 21, 2011, 05:10:55 pm »

Works great, thanks!

9

Automation Interface, Add-Ins and Tools / Shape Script for Stereotype of Constraint

« on: June 21, 2011, 12:53:39 am »

Hi all,

I'm trying to generate a shape script for stereotypes of Constraint. In this script, I can print tagged values of the stereotype (using print("#TAGNAME#")) but I was unable to print the contents of the constraint, ie. what the user can enter after doubleclicking it. Is there some magic #SOMETHING# to print this?

If this is not possible, I would also be fine with adding something below/above/... the default shape (basically, I just want to show some tagged values along with the constraint's content itself), but I was unable to scale what drawnativeshape() draws, ie. it always seems to draw to (100,100) and hence hiding my stuff. Drawing my additional stuff outside (0,0,100,100) works but seems to lead to strange effects with both selecting the item and drawing connectors from it.

Thanks,
Rainer.

10

Automation Interface, Add-Ins and Tools / Re: _image attribute ignored for UseCaseLink

« on: June 21, 2011, 12:59:39 am »

I had the same problem with Realizations, back then I noted that EA has different types of spelling Realization, ie. in the XML you posted it is written with an "s" instead of a "z" as in the Metamodel. Maybe this has something to do with it, though it would not account for your problem with the UseCaseLink.