On 4/22/2013 11:24 PM, Peter Levart wrote:
On 04/23/2013 05:27 AM, Mandy Chung wrote:
On 4/21/2013 6:39 AM, Peter Levart wrote:
On 04/21/2013 04:52 AM, Mandy Chung wrote:
Hi Peter,
I want to give it some more thought and discuss with you next
week. As for the zero number of interface, I think it's a bug and
it should throw IAE if the given interfaces array is empty.
Thanks for finding it and I'll file a separate bug for that since
it requires spec update/clarification.
I think it's a feature. It's useful, since it forwards Object
methods to InvocationHandler (equals, hashCode, ...). Sometimes
that's all you need.
Do you have specific use case or know any existing applications doing
that? What's the reason one would prefer to create a proxy class
with InvocationHandler rather than defining its own class that
implements equals, hashCode, or toString() method?
I really don't have a use-case here. I can only think of a
hypothetical case where one would already have an InvocationHandler
capable of serving proxies for various interfaces, including
equals/hashCode methods. Now to be able to re-use that logic for
constructing keys for Maps, one could simply request a proxy with no
interfaces. The fact is that currently this is allowed and although
there might not be any usages, it doesn't hurt to allow it in the
future and there's no performance cost supporting it. I think that not
putting artificial constraints on API usage without a reason is a good
design.
I can't say for sure allowing to define a proxy class with no interface
is a good design and thus I propose to file a separate issue to follow
up and determine if there is any issue with and without such constraint
(interfaces.length() > 0 or not). Yes - I am in a position not to allow
creating a proxy class with no interface since that was not the original
use case (for RMI). But I see your point and we need to investigate
before we decide one way or the other. If we determine no need to
constraint that, we can close that bug as will not fix.
It's similar in a way to some things in language, like for example the
support for empty arrays. Why would anyone want to have an empty array?
I don't think the Proxy API can be compared with empty array language
support :)
Regards, Peter
P.S. What do you think of the latest webrev? I did some further
simplifications to code (removed checks for reverseMap != null) and
re-worded some javadocs. I can publish another webrev together with
possible further changes when you review it.
Sorry I didn't get the chance to look at your past 2 webrevs (I was
distracted with higher priority tasks).
Mandy
Mandy
