cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Berin Loritsch <blorit...@apache.org>
Subject Re: [C2] Conceptual Error in ComponentManager
Date Wed, 17 Jan 2001 18:33:02 GMT
Berin Loritsch wrote:

> This has to do with ThreadSafe Components.  The error is the assumption 
> that
> there will never be more than one Component that implements a class if it
> is ThreadSafe.  This is not true.  The case and point is that there are
> now three components whose role implement CocoonComponentSelector which is
> by definition ThreadSafe.  This means that the three needed instances with
> very different Configurations are never generated.  It is always the same
> instance.
> 
> If the "markup-languages" is processed first (as currently the case), you
> can find all of it's components.  If the "programming-languages" is 
> processed
> second (as is currently the case), you cannot find any of it's components.
> What's worse, the ComponentSelector is kind of a "namespace" for identical
> Components--if we came up with a markup-language named "java" and all 
> instances
> of ComponentSelector were merged, it would conflict with the 
> programming-language
> named "java".

This issue is corrected.  It also appeared in the ComponentSelector.  This
is a bad thing, because the ComponentSelector is used for DataSources.  By
definition the DataSourceComponents are ThreadSafe.  With this logic bomb,
you would only be able to have two different DataSources (with different
classes).

Now we can have as many as we need.


Mime
View raw message