Coming from the other direction, I'd be very surprised if other JVM
languages don't begin offering a solution based on invokedynamic once Java
8 adoption takes off.

Given that Oracle will be optimising the VM for this pattern in particular,
the benefits in terms of size, performance, and interop are hard to ignore.



On 2 October 2012 10:10, Matthew Farwell <[email protected]> wrote:

> Is there any technical reason? Given that
>
> adding the option to create anonymous classes as well as invokedynamic
> would certainly make the compiler more complex;
> using invokedynamic will almost certainly be more efficient in terms of
> .class files generated & probably execution;
> and Oracle would probably want to push people to use the latest JVM (it is
> 99% backward compatible after all)
>
> I see this as being a reasonable choice. But obviously, I don't know
> anything :-) Of course, any particular compiler is free to implement
> lambdas however it wants.
>
> Matthew Farwell.
>
> 2012/10/2 Andreas Petersson <[email protected]>
>
>> Am 02.10.2012 10:36, schrieb Fabrizio Giudici:
>>
>>
>>> Definitely. Java is mostly backward compatible and I feel Java 8 will be
>>> the first exception.
>>>
>> Is there any particular reason a java 8 compiler woudl be unable to emit
>> bytecode (with a -target parameter) that is backwards-compatible? couldn't
>> the compiler just emit anonymous classes instead of invokedynamic?
>>
>>

-- 
You received this message because you are subscribed to the Google Groups "Java 
Posse" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/javaposse?hl=en.

Reply via email to