Re: Relieving some ClassLoader contention in PreferredClassProvider

Bob Scheifler wrote:
>> Given the fact you are referring to the spec can you help me find where
>> it is stated that equality tests must be performed against the URL type?
>
> The spec does not have as much precision as one might like here, but to me
> it is readily implied by "the codebase loader is the loader in this
> PreferredClassProvider's internal table with the codebase URL path as the
> key's path of URLs and the current thread's context class loader as the
> key's
> parent class loader", "where each URL is parsed as with the URL(String)
> constructor".

I thought you would come up with this :-), but to me this left enough
room to interpret URL as the deprecated word to be replaced by URI
nowadays. If the type java.net.URL was meant I would have expected a
monospaced font in the specification as is common throughout all the
Jini specs.

But considering that the modifications would imply a specification
change, I believe I think such as change is justified. Not only because
of the performance penalty of URL.equals() but also because of its
behavior with regard to virtual hosting.
--
Mark

--------------------------------------------------------------------------
Getting Started:     http://www.jini.org/wiki/Category:Getting_Started
Community Web Site:  http://jini.org
jini-users Archive:  http://archives.java.sun.com/archives/jini-users.html
Unsubscribing:       email "signoff JINI-USERS"  to [email protected]