feat: add support for analyzing whether a type reference is generic or not (method isGeneric)

[monperrus]

No description provided.

[spoon-bot]

Revapi Analysis results

Old API: fr.inria.gforge.spoon:spoon-core:jar:6.0.0-20170926.224458-46

New API: fr.inria.gforge.spoon:spoon-core:jar:6.0.0-SNAPSHOT

Detected changes: 1.

Change 1

Name	Element
Old	none
New	method boolean spoon.reflect.reference.CtTypeReference::isGeneric()
Code	java.method.addedToInterface
Description	Method was added to an interface.
Breaking	binary: non_breaking,

[pvojtechovsky]

Hi Martin, there is already CtTypeInformation#isGenerics() with similar contract. May be we should implement your new method as CtTypeInformation#isGenerics(boolean recursive) ... or at least use more expressive name, because actually there are

• boolean isGenerics()
• boolean isGeneric()

with different behavior

[monperrus]

You're right thanks.

For now, I propose to keep this new behavior, and to reuse the existing signature.
This new behavior subsumes the older and also enables to know whether a type can be instantiated or not. In
particular, it makes it possible to test and enforce an interesting contract.

typeRef.isGenerics() == false <=> typeRef.getSignature()== typeRef.getDeclaration().getSignature()

WDYT?

[pvojtechovsky]

For now, I propose to keep this new behavior, and to reuse the existing signature.

do you mean to change behavior of isGenerics()?

I checked how this method is used in spoon and it is only needed in two tests.
It is OK for me to change this behavior.

What about adding another method isParametrized(), with this contract

Type	isParametrized()	isGenerics()
Integer	false	false
List<Integer>	true	false
List<T>	true	true

At the beginning I had problem to understand your change because I understood the behavior as isParametrized

[monperrus]

It is OK for me to change this behavior.

Cool, then we may merge.

What about adding another method isParametrized(), with this contract

Good idea! #1565