Yes, but I wasn't making a general comment, but specifically addressing your statement that inner classes "aren't real" because the there's no real specific JVM support for them.
It seems you were making a comment about the JVM in the guise of a comment about the language. Now that we've cleared that up -- I agree. But I'm not sure it's specific to the JVM. Not only do we have half- done closure handling and JVM support for inner classes, we have the even more incomplete generic types feature. In both cases, there's significant work lacking on the compiler side as well. I wonder if Sun's position as a hardware vendor had anything to do with it, and whether Oracle's purchase will have any impact on such matters. On Oct 27, 9:10 am, DanH <[email protected]> wrote: > By your definition, about 2/3rds of the JVM is "not needed". But what > you say I was "trying to get at" is what I was saying, if you go back > and read it again. > > There was, for some reason, great reluctance on the part of Sun to > even minimally extend the JVM for several years there, and then there > was a sudden burst of (often gratuitous) changes. I'm guessing there > were some personality conflicts involved. > > On Oct 26, 7:37 pm, Bob Kerns <[email protected]> wrote: > > > > > > > > > I used the word "need" as in "require to function", as opposed to > > "have a use for". I'm not about to argue that the JVM couldn't support > > them better with more information. Even if I couldn't think of a > > reason, I would still bet on more information allowing for better > > support. > > > Perhaps what you're really trying to get at is that the JVM (not the > > language) doesn't really support inner classes well. You could even > > take your original statement, and qualify it with "... in the JVM" > > rather than tying it to javac. > > > Or you could argue that they're tacked on as an afterthought. That > > would cover the 'final variable' issue as well. > > > On Oct 26, 11:43 am, DanH <[email protected]> wrote: > > > > The JVM certainly could make use of the knowledge. As it is, there > > > are some significant security holes, and the JVM must have some > > > "special" rules (loopholes) so that inner classes will work. And the > > > scheme used is slower and much more cumbersome than would be an > > > implementation where the JVM is in on the secret and could implement > > > the access checking and outer class access directly. > > > > On Oct 23, 2:57 pm, Bob Kerns <[email protected]> wrote: > > > > > That's because the JVM doesn't NEED any other knowledge. The JVM > > > > doesn't define the language -- the compiler does. > > > > > The JVM is the environment the compiler targets, but with a different > > > > compiler, you get, say, Scheme or AspectJ instead of Java. > > > > > Taking your argument to the logical conclusion, nothing about C++ is > > > > real, because the hardware doesn't know about it. Hmmm.... > > > > > Solipsism is fun, but not particularly useful. > > > > > On Oct 23, 5:53 am, DanH <[email protected]> wrote: > > > > > > > Inner classes certainly exist > > > > > > Only as a fiction of javac. The JVM has no knowledge of them, other > > > > > than the InnerClasses attribute that's really just for debugging and > > > > > reflections. > > > > > > On Oct 22, 11:17 pm, Bob Kerns <[email protected]> wrote: > > > > > > > There's always been Runnable, yes, but there hasn't always been the > > > > > > cultural emphasis on composition rather than derivation. In other > > > > > > words, I agree it was a mistake, but it was a mistake born of its > > > > > > time, and not one that would be as likely to be repeated today. > > > > > > > Inner classes certainly exist -- they're just broken (only capture > > > > > > final variables). The sense of unreality to which you refer is just > > > > > > an > > > > > > implementation technique, and only visible if you peek below the > > > > > > covers of reality. :) That's true of most any bit of programming > > > > > > semantics you choose to examine that closely. > > > > > > > On Oct 22, 5:15 pm, DanH <[email protected]> wrote: > > > > > > > > But there's always been a Runnable, and it's no more difficult to > > > > > > > subclass a Runnable than a Thread. Inner classes made it easier > > > > > > > (if > > > > > > > more obscure) to define your subclasses, but didn't change the > > > > > > > basic > > > > > > > nature of the beast. (In fact, in reality inner classes don't > > > > > > > exist.) > > > > > > > > On Oct 22, 7:04 pm, Bob Kerns <[email protected]> wrote: -- You received this message because you are subscribed to the Google Groups "Android Developers" 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/android-developers?hl=en
