regardless, it seems like a good thing to test carefully in contrib before moving into core. also, by implementing as an extension first, we get to test out and ensure that our hooks for this sort of stuff work. if it turns out to be as great as you say, i'm sure we'll all discuss the possible core-ness of your contrib plugin in the future.


i think one thing at a time here is best. xstream is definitely not a 1.0 feature (we have to ship this darn thing and the date keeps sliding as people find more problems!). maybe contrib in 1.1. and then if that works, it could be considered for core in 1.2. please note the word "considered". we'd have to come to consensus. but the best way to get that consensus is surely implementing an xstream plugin in contrib that knocks our socks off.

Gili wrote:

        <sigh> I give up :)

        XStream is not "walking on the wild side". JSX, its predecessor
has been around for *many* years and is a proven technology. XStream
does the same thing, only it's free (the former was commercial). Both
of these products simply hook into underlying JDK mechanisms to get at
private fields, methods, etc. There is no native code or any magic at
play here. Performance is very good.

        So fine, if all you want is a non-default contrib module then
maybe we can start with that but honestly the kinds of arguments I'm
hearing here from Chris:

1) It's not good to add another JAR because it'll increase the size of
Wicket
2) Serialization requires "a certain discipline" which would be lost if
we allowed XStream

        go smack in the face of Wicket's goals of focusing on ease of
use before performance. No "discipline" is lost whether you use
standard serialization or XStream. The same persistance rules apply
(transient keyword, etc) whether you use one or the other. All we're
doing is making it *easier* to serialize and if users want to come back
and fine-tune the process further they can still do so. I think you
guys are assuming (incorrectly) that XStream is "unstable magic" and
that its performance is bad. Neither is true.

        I have compared XStream vs Serialization to Java vs C before. I
honestly think it's that sort of thing.

        And Eelco, I'll work on the plugin as soon as I get around to
working with Wicket again. Right now I'm knee-deep migrating to
Hibernate3.

Gili

On Tue, 08 Mar 2005 09:00:04 -0800, Jonathan Locke wrote:



yeah. i'm +1 on that too. i'd also like versioning to be hookable enough that you could replace cloning code with a contrib'ed version mgr that uses xstream. best of all worlds this way. wicket would default to the normal Serialization model, but then you could replace that using contrib'ed xstream code if you felt like taking a walk on the wild side...

Christopher Turner wrote:



My personal view is that we should not be including this as a standard feature of the core. It just adds extra jar dependencies, increases the size of the web application and so on. I'm personally more than happy with using the standard serialization mechanism provided by my application server vendor.

Serialzation is a messy thing to deal with and I think that a certain discipline and design ethic is forced on you by having to implement Serializable and deal with the differences between transient and persistent fields. If you want to design efficient web applications then you have to be aware of the issues involved in this. Wicket can hide it for simpler applications but as soon as you start growing to servers supporting session clustering and recovery then you have to understand a bit more about what is going on and make some important design decisions.

However, I would be +1 for providing XStreamPageState and other Xstream support as an optional implementation in the contrib package.

Regards,
Chris





if xstream doesn't have downsides and is more efficient and we don't care about vm requirements, we could use xstream for undo cloning operations. then if we provided XStreamPageState, you wouldn't have to make anything serializable. (uh... i think). xml seems like a terribly verbose format though. i'm concerned about memory impact...

Gili wrote:



On Tue, 08 Mar 2005 08:32:18 -0800, Jonathan Locke wrote:





if gili wants to go test our new PageState mechanism by


making a nice


subclass of PageState using xstream (XStreamPageState?),


that might make


a really nice contrib or extensions class...




Before I do I want to make sure I have a better


understanding of how


far I can go to serialize as little as possible with the web


container


and handling the rest with XStream. My main concern is for
user-classes: users should not have to worry about serialization or
no-arg constructors and I simply want to give them the


option of doing


so. So it would be nice if say we'd still have to persist


Wicket using


the web container but everything else could be handled "in-house" by
Wicket using XStream.

Gili



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real
users. Discover which products truly live up to the hype.


Start reading


now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

<http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click>


_______________________________________________
Wicket-develop mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/wicket-develop





-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from
real users. Discover which products truly live up to the
hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396


<http://ads.osdn.com/?ad_id=6595&alloc_id=14396>> &op=click


_______________________________________________

Wicket-develop mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/wicket-develop




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Wicket-develop mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/wicket-develop







-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Wicket-develop mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/wicket-develop





-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Wicket-develop mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/wicket-develop

Reply via email to