On 08/16/2014 12:39 AM, John Rose wrote:
Do you have a use case in mind that could be expressed as a more tightly
focused API?
My use-case is run-time class generation and avoiding unnecessary
retention of classes which are referred to by the generated class.
I do not need constant pool patching. It would be interesting to inject
static final values of arbitrary types, but I could use a thread-local
variable for that if necessary. Being able to pass parameters to
<clinit> would be a much cleaner way to achieve that, but it doesn't
seem strictly necessary.
> Or, most or all of dAC could be simulated using regular class
> loading, into a single-use ClassLoader object. The nominal bytecodes
> would have to be rewritten to use invokedynamic to manage the
> linking, at least to host-class names.
Why is that? Because you cannot directly reference to a class by name
if you don't use the class loader which was used to define the class to
look up the name?
--
Florian Weimer / Red Hat Product Security
_______________________________________________
mlvm-dev mailing list
mlvm-dev@openjdk.java.net
http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev
