[JBossCache] - Re: Please vote for your favorite issues -- mine :-)
by manik.surtani@jboss.com
Re: 841
Since with 2.0.0, the node is now a top-level construct, I can see the value in adding comparators to the dataset. I'mnot averse to adding stuff to the API provided it is useful and cohesive to the core offering of JBoss Cache. Again, +1 on the idea.
Functors is a nice genericisation for applying transformations to the data in the cache, and is a useful hook for allowing people to do stuff to the data that we *cannot* foresee or is sufficiently domain-specific, but I think sorting is (a) fairly fundamental to many use-cases and (b) fairly core to any system that organises and stores data.
So +1 on the API.
Again, re: the implementation, I don't think JBoss Retro will support JDK6 -> 4 weaving for a while. And there is no way I'm baselining on 6 for some time to come.
Also, I need to be careful of feature-creep in 2.0.0, so I may defer this to 2.1.0. (Adding an API method at that stage will not break the existing 2.0.0 API)
Unless, that is, this feature (complete with comprehensive test suites and full regression against the existing test suite) is contributed in time for 2.0.0. ;-)
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3990803#3990803
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3990803
18 years, 1 month
[JBossCache] - Re: Please vote for your favorite issues -- mine :-)
by manik.surtani@jboss.com
Re: 880:
No, it's not that I am against this idea, just that I think that this should be a custom extenstion (something a user may implement) rather than a part of the core JBoss Cache product.
Then again, there has been some talk about this on the forums and it is useful for some of our core use cases (inc. http session repl).
Regarding implementation, Elias' impl using an 'expiry' attribute in the node's data map is a little hacky and I don't like the fact that it is not typed so there is no way of checking this, but it is easily customised and extended as a result - plus it has very little impact on the overall arch.
So I'm leaning to a +1 for the feature, and a +-1 on the impl. I'd like to hear more from others (especially users on this forum) before I make a clear decision on this though.
Thanks for providing a prototype on this though, Elias, I do appreciate that.
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3990802#3990802
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3990802
18 years, 1 month