From alex.buckley at oracle.com  Tue Jun 18 16:51:11 2013
From: alex.buckley at oracle.com (Alex Buckley)
Date: Tue, 18 Jun 2013 16:51:11 -0700
Subject: Metadata about parameter names
Message-ID: <51C0F26F.3020908@oracle.com>

Prompted by discussions on jdk8-dev [1], I have specified an 
isNamePresent() method in java.lang.reflect.Parameter. It exposes 
whether the parameter has a name _according to the MethodParameters 
attribute_. Non-Java clients can use it to determine whether the 
parameter "really" has a name, because for the sake of Java clients, a 
parameter always has a name _according to getName()_ even if we have to 
synthesize it.

The mini-spec for this method is at [2] and the enhanced metadata spec 
incorporates it for completeness [3].

Alex

[1] http://mail.openjdk.java.net/pipermail/jdk8-dev/2013-June/002658.html
[2] http://bugs.sun.com/view_bug.do?bug_id=8016285
[3] http://cr.openjdk.java.net/~abuckley/8misc.pdf

From alex.buckley at oracle.com  Fri Jun 28 15:33:19 2013
From: alex.buckley at oracle.com (Alex Buckley)
Date: Fri, 28 Jun 2013 15:33:19 -0700
Subject: Repeatable non-inheritable annotation types
Message-ID: <51CE0F2F.5080600@oracle.com>

Deven You raised the following point at [1]. Assume the declarations:

   @Foo(1) @Foo(2) class A {}
                   class B extends A {}

where Foo is _not_ inheritable. Foo is repeatable with FooContainer as 
its containing annotation type, and FooContainer _is_ inheritable.

B.class.getAnnotationsByType(FooContainer.class) will of course expose 
the inherited @FooContainer annotation. But, 
B.class.getAnnotationsByType(Foo.class) should not expose @Foo(1) 
@Foo(2), as Foo is not inheritable.

The definitions of "present" and "associated" annotations in the spec 
are correct, but this particular scenario - with a _non_-inheritable Foo 
- was not covered in the examples in section 1.2.

I have extended example 1.2-1 to cover it, and updated the spec PDF at:

   http://cr.openjdk.java.net/~abuckley/8misc.pdf

Alex

[1] 
http://mail.openjdk.java.net/pipermail/core-libs-dev/2013-June/017675.html