Resolving generic type bounds to a j.l.r.Type, not a j.l.Class?

[pholser]

Hello, thanks again for your magnificent work.

I'm using generics-resolver 3.0.0 in junit-quickcheck, and am trying to resolve an issue with it that a user raised.

Given this class:

    @RunWith(JUnitQuickcheck.class)
    public static class GenericMethodVars {
        @Property
        public <X extends HashMap<Integer, Character>> void x(
            Callable<X> callable) {

            System.out.println(callable);
        }
    }

junit-quickcheck will do something like this to resolve the variables to some type:

        MethodGenericsContext generics =
            GenericsResolver.resolve(GenericMethodVars.class)
                .method(GenericMethodVars.class.getMethod("x", Callable.class));
        Map<String, Type> typeVariables = generics.genericsMap();
        typeVariables.putAll(generics.methodGenericsMap());

I was hoping that type variable X would be associated with a parameterized type HashMap<Character, Integer>. Instead it's associated with the java.lang.Class object for java.util.HashMap.

Am I doing something incorrect? Can I accomplish what I want using generics-resolver?

[xvik]

Hello!

It's my bad.. I did not face such cases myself, and so this part remains not updated for a long time.. I indeed loose type information there. The same problem with constructor contexts.

I was preparing new 3.1.0 version with a lot of changes.. but it's not ready now for publication (I did not check everything). So instead, I will release 3.0.1 hotfix (hopefully today).

[xvik]

New version is released, but only in jcenter because central's nexus is dying since yesterday. New version will appear in central as soon as it will become responsible.

[xvik]

Finally, central accept publication. Version will appear soon.