jni boilerplate: c -> cpp

[pazos]

It isn't intended to go in for a while as it is untested.

pros:

• It is easier to read for a ramdom lurker.
• C/JNI boilerplate should be banned from existence.

cons:

• It is c++

but we already use c++ for the rest of boilerplate, so why not here too? :)

---

This change is 

[Frenzie]

No real opinion. I suppose "real" objects might have some advantages over pointers but I thought it was really just a slightly different syntax for the exact same thing.

[pazos]

No real opinion. I suppose "real" objects might have some advantages over pointers but I thought it was really just a slightly different syntax for the exact same thing.

They do the same thing. The c version also showcases what the cpp does under the hood.

From a reading perspective the cpp version is way easier. Modern "source readers" are used to deal with instances of things and are self aware, so the construct:
(vm->AttachCurrentThread(&env, NULL) == 0)) is easier to grasp than
((*vm)->AttachCurrentThread(vm, &env, NULL) == 0)))

They are also easier to think and write. Hence cpp elsewhere :)

But, yeah: really just a slightly different syntax for the exact same thing. is 100% correct :)

[Frenzie]

Not convinced after all?