Hi Michiel,
Thanks for bring this up and driving it to a decision. Part of the
confusion probably comes from my not understanding what situations
this would be used in. Am I correct in understanding that the only
time this method will be used is when the user changes the notation
language (e.g. to C++ instead of of UML)? I think I was under the
impression that there places where it looked like it was designed to
be used when the owning UML element of a Fig changed (which, of
course, can't happen now).
How about instead of undeprecating this, we create a new method called
changeNotationProvider or something which makes it clear what its
redefined role is. That way we'll be able to easily find places that
are using the old method.
Currently the only two places that I can find where
initNotationProvider called is FigNodeModelElement.renderingChanged()
and FigEdgeModelElement.renderingChanged(). If my understanding above
is correct, perhaps we could move the invocation of the new method to
someplace specific to changing the notation language.
> protected void initNotationProviders(Object own) {
> if (notationProviderName != null) {
This is an actual NotationProvider, not just the name of one, right?
> if (Model.getFacade().isAModelElement(own)) {
Unrelated to the current question, but for best portability between
UML 1.4 & UML 2.x, things like this should use isAUMLElement()? There
are chunks of the UML 2 metamodel which are not subtypes of
ModelElement (and mistakes in the UML 1.4 metamodel which caused some
things to unintentionally not be subtypes either).
Tom
------------------------------------------------------
http://argouml.tigris.org/ds/viewMessage.do?dsForumId=450&dsMessageId=1003735
To unsubscribe from this discussion, e-mail:
[[email protected]].
