Re: Multiplicity Questions
Tom Morris <[email protected]> Wed, 28 Mar 2012 11:32:51 -0400
| Newsgroups | gmane.comp.lang.uml.argouml.user |
|---|---|
| Message-ID | <CAE9vqEHaWBaV4nB-tTmCa2Vrd=xrY=9fY+BUuGfcd3Oneeb2CQ@mail.gmail.com> |
On Wed, Mar 28, 2012 at 11:06 AM, Mark Fortner <[email protected]> wrote: > Hi Linus, > I created the association again, clicked on the link in the Properties pa= nel > to go to the Department association end, and then clicked Private. =C2=A0= When I > looked at the source tab, it still showed up as public. =C2=A0Oddly enoug= h, on > the diagram we see "-employeeList" indicating that the association end is > private. I've lost track of the ends now, but my impression from the original message was that you were changing the visibility on the wrong association end. Remember that it's the *far* end that you need to change for things like visibility and navigability. > What's also odd is the fact that Employee also ends up with a public Vect= or > employeeList declaration. Is the association navigable in both directions? (the default) If so, you'll get data structures on both ends to allow you to get to the other end. Set isNavigable to false on the end opposite where you want the data structure to disappear from. > I suspect changing Vectors into generified Lists, and fixing the > Multiplicity=3D1 issue are probably trivial changes and I'll look into th= at > within the next few days. Support for UML Templates and Java generics is only partially implemented. You could probably special case the code generation for this specific use case without generics support, but you might need to take backward compatibility into account (e.g. provide a preference setting for what type to generate). Note also that ArrayList and Vector have slightly different semantics, particularly vis a vis synchronization, so the two types aren't completely interchangeable. Tom ------------------------------------------------------ http://argouml.tigris.org/ds/viewMessage.do?dsForumId=3D449&dsMessageId=3D2= 941640 To unsubscribe from this discussion, e-mail: [[email protected]= ris.org].