lucene-solr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ryan McKinley <ryan...@gmail.com>
Subject Re: Initializing - break init() API compatibility?
Date Thu, 15 Nov 2007 16:39:39 GMT
> 
> Instead of changing 'init' (& keeping the 1.2 version), we can either make
> classes take a solrSystem as (first) constructor parameter (but I know this
> is not the current nor preferred usage) 

Are you saying 'if a constructor exists that takes a solrSystem, it will 
use that - otherwise, just the standard newInstance()'?

that seems ok.  But it would mean that a request handler that needs the 
core for initialization would need to hold on to the core until the 
'init( NamedList )' method though.


> or have the proper solrSystem
> instance available through a static method with guaranteed call scope
> availability in the constructor & in init.

ThreadLocal?

> 
> I might be able to publish this experimentation later today.
> Any "stop" sign ?
> 

I'm all for experiments!

ryan

Mime
View raw message