[type-annos-observers] Classfile format changes for more efficient processing

[Alex Buckley]

On 10/1/2012 7:54 PM, Srikanth S Adayapalam wrote:
> (1) There has been a flurry of mails about various proposals being
> accepted. Is there a refresh of the spec planned anytime soon ? Even
> an informal draft would be good enough for me.

I will leave this to Mike.

> (2) Could the spec leads coordinate with the lambda group so that
> TYPE_USE is available as a target in the lambda branch ? At the
> moment it is not and it is making life complicated for us alternating
> between 308 & 335 work. Per my understanding this is the only change
> to JRE to support 308 - Is this correct ? Are there any logistic
> issues in promoting this change alone to 335 branch that I am naively
> ignoring ?

I think you're looking for a JDK8 binary which supports lambda 
expressions in javac and the Runtime[In]VisibleTypeAnnotations attribute 
in ClassFiles.

This will happen in time. The OpenJDK Lambda and Type Annotations 
projects have their own forests which we don't intend to sync with each 
other directly. Lambda's forest is changing rapidly, whereas Type 
Annotations' forest is stable enough to undergo code review for a push 
into the JDK8 forest. (Specifically, type-annotations/type-annotations 
-> jdk8/tl -> jdk8/jdk8.) I will try to find out when Lambda expects to 
push into JDK8. Once both projects have pushed into JDK8, binaries will 
follow automatically.

Alex