Type Annotations RI update

[Werner Dietl]

- I diagnosed and fixed a problem with annotations in array component
and inner type positions: they ARRAY or INNER_TYPE tag was repeated
for each occurring annotation - leading to wrong type positions.
The fix is here:

http://hg.openjdk.java.net/type-annotations/type-annotations/langtools/rev/93c9048ce673

It would be great if SQE could add a combo-test for more syntactic locations.

I fixed two other small issues:

http://mail.openjdk.java.net/pipermail/type-annotations-dev/2013-January/000519.html
http://mail.openjdk.java.net/pipermail/type-annotations-dev/2013-January/000521.html

These can be reviewed for a push to TL.


- Last Wednesday I described my understanding of the way
AnnotatedTypes are constructed:

http://mail.openjdk.java.net/pipermail/type-annotations-dev/2013-January/000535.html

I am awaiting comments on how to proceed.


- I spent some time investigating why we are generating TypeCompounds
with UNKNOWN positions. I reverted my prototype, but it improved my
understanding and when we work on the above point we can resolve those
issues.

- I think it would be good to finalize javax.lang.model soon. My
understanding is that Joel is working on this. Is this right?
Did we decide whether we will introduce TypeKind.ANNOTATED? In code
that uses TypeKind.isPrimitive an annotated primitive type won't be
transparent. That is, asking "@TA int" whether it is primitive will
return false.

- To me a method reference receiver seems like a scoping mechanism,
not a type use. I tried a few examples and asked Mike for
clarifications.

cu, WMD.

-- 
http://www.google.com/profiles/wdietl