David Winslow ha scritto:
> Justin Deoliveira wrote:
>>> I was thinking of reusing as much as possible for the XStream 
>>> backend.  If the current backend is going to use XStream for 
>>> persistence anyway then that doesn't mean much :)  But yes, I had 
>>> 'reusable for alternative implementations' in mind there, as opposed 
>>> to 'reusable by other subsystems,' which I was taking for granted as 
>>> a design goal.
>> OK, still not sure I 100% understand but yes :). I do think that we 
>> should structure it in a way that we don't add any dependencies from 
>> the core model on xstream, so that we could swap out other persistence 
>> backends easily.
> Ah, got it now.  I had been thinking persistence mechanisms would 
> require reimplementing the entire catalog but it could probably be done 
> in most cases with just loaders and change listeners, leaving the 
> in-memory objects to be shared across implementations.  I wonder, 
> though, whether a 'store everything on shutdown' policy is 
> necessary/useful if we are saving all changes that are meant to be saved 
> along the way.  

I share the same concern. We do want to save along the way since we
cannot be sure there will be a clean shutdown (jvm segfaults anybody?),
so eventually on shutdown we'd just be saving whatever was already
saved the last time the config did change?

Cheers
Andrea


-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Geoserver-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

Reply via email to