[Better Comment Initiative]: FAMIXNamedEntity

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[Better Comment Initiative]: FAMIXNamedEntity

Stéphane Ducasse
We have
-------

I'm an abstract class, root of the hierarchy modeling  source code entities.

Any of my subclasses must define the meaning of the belongsTo, an abstract property that provides polymorphic traversal. For example,  FAMIXClass defines belongsTo as being the container, while the FAMIXMethod defines belongsTo to point to the parentType.


I propose
--------
I'm an abstract class, root of the hierarchy modeling  source code entities.

Any of my subclasses must define the meaning of the belongsTo property, an abstract property that provides polymorphic traversal. For example,  FAMIXClass defines belongsTo as being the container, while the FAMIXMethod defines belongsTo to point to the parentType.

I support the tagging of entity as stub. Marking an entity as a stub means that the
Stubs should only be marked by importers and extractors. Stubbed entities represent the border of the modeled world as extracted. Stub tagging should not be recursive. Use other tags for such a purpose: a class may have an instance variable tagged as a stub because an importer got a memory leak or a bug. Still the class is not a stub, but extra analysis can mark it as incomplete.

When applicable, I also point to my containing package accessible via parentPackage and modifiers.

I also can return the list of invocations performed on this entity (considered as the receiver) (receivingInvocations).

Finally, I provide simple name for the entity I represent. The pattern is that name returns the local name of the entity.
_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev