Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Uffe

Pages: 1 ... 52 53 [54] 55 56 ... 76
General Board / Anyone for Ada?
« on: October 17, 2007, 08:14:19 pm »
Hello people,

Just wondering if anyone's got a Profile for Ada tucked up their sleeve somewhere. Or better yet, a full-on Add-In with reverse engineering capabilities.

Is there an echo in here?


General Board / Re: How to Generate Code
« on: October 17, 2007, 07:43:42 pm »
Oh. Ah. Yes, that's exactly what it is. There's no code generation from the Desktop Edition, you need at least the Professional (next step up) for that.


General Board / Re: How to Generate Code
« on: October 17, 2007, 07:12:35 pm »
Hey there,

You need to select one or more classes or packages. I recommend right-clicking in the project browser; the Code Engineering submenu is the one you want, and in there Generate Source Code.
In the dialog which opens you will normally want to tick the "Include all Child Packages" box, otherwise code will only be generated for the classes immediately below the package you selected.



General Board / Re: How to persist edits of code generation templa
« on: October 03, 2007, 03:54:15 pm »
On this topic, I've had problems getting the Code Generation Templates to work after importing them as reference data. I've traced this to the absence of types defined for my language.

In other words, if there is no language X defined in the "Programming Languages Datatypes" dialog (Settings - Code Datatypes...), language X does not appear in the Language dropbox the Code Template Editor (top left).

Defining a single type for language X resolves the issue. The Code Generation Templates do not need to be reimported.

Has anyone else had this problem? I'm using 7.0.815.

Bugs and Issues / Re: Traceability between swimlanes and classfiers
« on: December 02, 2015, 09:20:37 pm »
So the traceability can not be used since it relies on connectors.
This is incorrect. The traceability window keeps track of other types of relationships as well, not just those expressed with connectors.

The classifier/instance relationship is also shown in Traceability, and it actually works even if the instance is not a proper element (like a Partition is) but just a diagram object, like a Boundary.

Ctrl-U / Find in Diagrams also finds instances of the selected element, and that includes boundary instances.

My point was that swimlanes are not even diagram objects, so Traceability doesn't work for them. And neither does Ctrl-U. Swimlanes are magical, but boundaries can be used as a workaround.


Bugs and Issues / Re: Traceability between swimlanes and classfiers
« on: December 02, 2015, 01:28:14 am »
Hi Ursula,

The reason the swimlane classifier doesn't count as a reference in either the Traceability window or the Elements Usage (Find in all Diagrams) dialog is that swimlanes aren't diagram objects.

Exactly how the reference is stored I don't know, but it does work in the sense that the swimlane labels is updated when the classifier's name is changed.

So the potential is there, it's just not implemented. You could post a suggestion and see if anyone likes it.

As a workaround, you could use boundaries or (in activity diagrams) partitions. You can set classifiers on both these, and since a boundary is a diagram object and a partition is a fully-fledged element, the backwards references work.



Bugs and Issues / Re: Decision tables in EA11
« on: November 17, 2015, 07:25:28 pm »
Aaaah - right. I'm on a Corporate installation here, I'll check it out on my Ultimate instead.

Thanks, KP!


Bugs and Issues / Decision tables in EA11
« on: November 17, 2015, 01:13:28 am »
Hi all,

In the View menu in EA 11.1, there's a disabled item Decision Table. Is it supposed to work, and if so how do I open it?

The help file says it should enable when I select an activity in a diagram. It doesn't.
The help file also says there should be a Decision Table item in the diagram context menu for activities, as well as a Decision Table toolbar. I haven't been able to locate either.
I am using the Complete command set.

According to the version history, the function was introduced in EA 11.0.1105, not touched in the remaining 11.0 or 11.1 releases, then in 1207 it was added to a portal, and activities became renderable as decision tables.

But I can't find any way to add them in the first place. Is this a case of the help file and bits of the GUI being shipped before the function was in place, or am I just doing something wrong?



Bugs and Issues / Re: Showing Realized Interfaces - broken
« on: November 12, 2015, 09:25:51 pm »
I eat red herrings for breakfast, me. Yummy!

But here's my problem.

It allows you to quickly set up a summary of what interfaces are provided, by what components, to service what dependents.
This would be extremely useful. Except that in order to do that, I must now draw an additional connector. I've drawn a realization and a dependency, from one component each, to the same interface. If I must now also draw a dependency from one component to the other, I introduce a direct dependency between them -- precisely what I wanted to avoid by using an interface in the first place.

If the connector (I suppose an assembly would be most correct, but I just plain don't like those bulky little buggers) was ghosted in the same way as the lollipops, I'd love this feature. Love it.

And, just to pile on for no better reason than it's raining outside, here's how it works in practice.

1) I set my model up the normal way with an interface, a component which realizes it (A), and a component which depends on it (B).
2) I drop the components in a diagram and select to display realized / dependent interfaces. Lollipops appear - yay!
3) I draw a new dependency from B to the lollipop decoration on A. The dependency sticks to the lollipop on A. Good.
4) But what's this? There's a new dependency lollipop on B, called "A"!
5) I go to A and deselect presentation of realized interfaces.
6) The dependency end now hovers in the air.
7) I move "A" around in the diagram. The dependency sort of twitches a bit, but the end stays where it is.
8) I switch A's realized interfaces back on. The dependency sorts itself out.

I should point out that if you close and re-open the diagram after 6) it all works correctly. But 4) makes the function less than intuitive, and having to draw the extra connector myself rules it out altogether -- not because I'm such a lazy sod, which I totally am, but because it requires me to break my precious model.

So IF you're going to have another look at this feature, please consider changing it to auto-draw connectors between the ghosted lollipops. Preferably headless ones, like associations or connectors (ie the ones between ports, which are actually named Connector).

Also, IF you do take a look, the whole thing would be much more useful if you could select which side of the element each type of lollipop appeared on. That way, connectors or no connectors, I could at least set them up roughly facing each other.

Cheerie-bye then,


Bugs and Issues / Re: Showing Realized Interfaces - broken
« on: October 27, 2015, 07:29:54 pm »
Hey Paolo,

I think all you're going to get from that is a correction in the help file.

I've never seen it work that way, the "shown" interfaces (whose names are in italics, not regular font as the help file image claims -- further proof that the help file is wrong) have always1 been unselectable ghosts.

It's a pretty useless feature IMO. In order for it to be useful, the connectors to/from the "shown" interfaces (the ones from Class1 in the help file image) would also need to be drawn in the same ghostly manner, but they're not.


[size=9]1 For "always", please read "since version 7.0".[/size]

Bugs and Issues / Re: Wrong use of extends
« on: November 04, 2015, 11:31:38 pm »
My research over the last decade has convinced me that the more rigorous the relationship between the modelling syntax (especially including [highlight]naming of objects[/highlight] according to their metatype) and the semantics, the better and more insightful the model.
You should read Wittgenstein. Things are only what their name is - and vice versa.
True, but then Wittgenstein was a beery swine who was just as sloshed as Schlegel.


Bugs and Issues / Re: EA crashes when I try to use ShowInProjectView
« on: October 28, 2015, 06:16:06 pm »
Hi Aaron,

What you're doing there is selecting the element that triggered the event. I think the OP is trying to select a different element from the one that triggered the event.

I'm thinking that might shove EA into a tailspin, because it changes the context item while the context item change event is being processed.

But I might be woofing up the inappropriate woody plant.


Bugs and Issues / Re: EA crashes when I try to use ShowInProjectView
« on: October 28, 2015, 12:18:09 am »
Well, by calling ShowInProjectView() you are in fact changing the context item -- right in the middle of EA's processing of a context item change event.

I wouldn't be entirely surprised if EA finds this difficult to deal with.


Bugs and Issues / Re: EA crashes when I try to use ShowInProjectView
« on: October 27, 2015, 09:20:21 pm »
Hi Javier,

If you return true out of EA_OnContextItemDoubleClicked(), you're telling EA that you have taken over handling of the event and EA should now stop processing it. If you return false, you allow EA to continue processing the event, which in most cases means opening the properties dialog.

It is probably a good idea to return false in this case.

As to why EA crashes, before you call Repository.ShowInprojectView(), make sure that element isn't null. If for instance you have deleted and recreated the use case, element ID 13 will never be reused and so Repository.GetElementByID() will return null.



Bugs and Issues / Re: Using multiple keywords
« on: October 12, 2015, 07:36:54 pm »
Hi semo,

I can confirm that the behaviour is the same in 11.1.

It looks like the "One Of" filter condition is strictly one of -- not one-or-more-of. Might be worth logging this as a bug, but I have a feeling it might be a case of unexpected design consequence rather than implementation error.



Pages: 1 ... 52 53 [54] 55 56 ... 76