Typed Method handles

[John Nilsson]

What would be the problem? What exactly is planned for "true properties"?

The reason i'm persistent with this is that if I look at our current code
base one of the most obviois improvements would be to add statically
verified (refdrential integrity) propert references. Our current approach
using a string DSL and reflection makes it hard to do certain refactorings.

As it looks now I'm not sure the syntactic overhead of using lambdas will be
worth a change though. As discussed in another thread there needs to be
null, and sometimes exception, handling involved. To keep the code readable
there therfore needs to be a way to compose references with such handling
implicit. A methodchaining monad if you wish ;)

Br,
John

Sent from my phone
Den 14 jun 2011 05:21 skrev "Rémi Forax" <forax at univ-mlv.fr>:
> On 06/14/2011 04:07 AM, John Nilsson wrote:
>> On Tue, Jun 14, 2011 at 1:02 AM, Brian Goetz<brian.goetz at oracle.com>
wrote:
>>> - Four kinds of method references (static, instance, unbound instance,
>>> constructor)
>> Any thoughts about specific support for bean-property style
>> references? I could imagine something like #[customer.name] f.ex. (if
>> we include chaining and inference).
>
> If we do that, there is no way to add true property later.
>
>> BR,
>> John
>
> Rémi
>
>
>