review: feature: ParentFunction

[pvojtechovsky]

Implements new mapping functions, which visits all the parents of current element.
It is needed by to LocalVariableScopeFunction to detect scope of local variable, by searching for a parent which matches some rules... will be used in next PR, which fixes LocalVariableScopeFunction for variables used in CtLoop expressions and CtTryWithResource.

Note: I have one question to the scanning of all parents. It is "Which parent is the top parent?"
If I visit parents until the getParent()==null, then I get to helper annonymous class based on CtElementImpl, which is declared in CtModelImpl$CtRootPackage$1. I guess it is not correct ... Your opinion please?
The test actually fails because the scanning does not end on RootPackage, but one level below...

[surli]

If I visit parents until the getParent()==null,

why don't you use getParent() == factory.getModel().getRootPackage() as condition to stop?

[pvojtechovsky]

I know several ways how to do it different/better. But I did not know what is the correct behavior from the spoon design point of view. So I waited for a feedback of an experienced spooner.
So, please confirm that parent visiting should stop on factory.getModel().getRootPackage() - including that node. And I will do so. Or discuss it and let me know what is correct. Thank You!

[tdurieux]

Note: I have one question to the scanning of all parents. It is "Which parent is the top parent?"
If I visit parents until the getParent()==null, then I get to helper annonymous class based on CtElementImpl, which is declared in CtModelImpl$CtRootPackage$1. I guess it is not correct ... Your opinion please?
The test actually fails because the scanning does not end on RootPackage, but one level below...

If I correctly remember, we only created the CtRootPackage to "store" all elements of a model. It does not represent any Java concept.
It is probably better to include CtRootPackage in order to match the behavior of getParent(Filter<E> filter).
Even though there is no interest to visit it.

[msteinbeck]

If I correctly remember, we only created the CtRootPackage to "store" all elements of a model. It does not represent any Java concept.

Isn't it the package of files without any package declaration?

[tdurieux]

Isn't it the package of files without any package declaration?

Yes, and also for the parent of declared packages.
Before CtRootPackage, we identified the default package by comparing its name to TOP_LEVEL_PACKAGE_NAME

[pvojtechovsky]

Thanks to everybody for good feedback. I modified the code and test so it expects that last visited parent element is CtModelImpl$CtRootPackage.

[spoon-bot]

Revapi Analysis results

Old API: fr.inria.gforge.spoon:spoon-core:jar:5.6.0-20170130.234600-20

New API: fr.inria.gforge.spoon:spoon-core:jar:5.6.0-SNAPSHOT

Detected changes: 0.

[surli]

Ok it seems good to me.

[spoon-bot]

Revapi Analysis results

Old API: fr.inria.gforge.spoon:spoon-core:jar:5.6.0-20170131.234706-21

New API: fr.inria.gforge.spoon:spoon-core:jar:5.6.0-SNAPSHOT

Detected changes: 0.

[surli]

Test revapi.