Berin,
any email that mentions my name twice and promises
enlightenment for some coding is trigger enough.
I have a version of Fortress that I feed this config:
<test>
<component id="test4"
class="org.apache.excalibur.fortress.test.ComposerImpl"
role="org.apache.excalibur.fortress.test.Composer"
handler="org.apache.excalibur.fortress.handler.ThreadSafeComponentHandle
r"
logger="test">
<!-- no role mappings overridden -->
</component>
<component id="test1"
class="org.apache.excalibur.fortress.test.ComposerImpl"
role="org.apache.excalibur.fortress.test.Composer/Alternate"
handler="org.apache.excalibur.fortress.handler.ThreadSafeComponentHandle
r"
logger="test">
<provide role="org.apache.excalibur.fortress.test.Supplier"
via="org.apache.excalibur.fortress.test.Supplier/Alternate"/>
</component>
<component id="test2"
class="org.apache.excalibur.fortress.test.DefaultSupplier"
role="org.apache.excalibur.fortress.test.Supplier"
handler="org.apache.excalibur.fortress.handler.PoolableComponentHandler"
logger="test">
</component>
<component id="test3"
class="org.apache.excalibur.fortress.test.AlternateSupplier"
role="org.apache.excalibur.fortress.test.Supplier/Alternate"
handler="org.apache.excalibur.fortress.handler.PoolableComponentHandler"
logger="test">
</component>
</test>
The component with id="test1" gets the
org.apache.excalibur.fortress.test.Supplier/Alternate, and so on.
Performance penalty is negligible according to the profiling test case.
I am, however, incredibly unsure when it comes to the ComponentSelector
parts of Fortress, and I haven't touched the FortressServiceManager
pieces of the code.
So Berin, do you want me to committ the code back in for review, or
post the diffs here for review?
> From: Peter Donald [mailto:[EMAIL PROTECTED]]
>
> Luckily some Cocoon people (hi Leo!) have started to realize this
...
> I really hope Leo implements his ideas soon because I am sure that
> once he has implemented it
> he will find out how insane it it is to merge the notion of resource
> manager and directory service. I suspect he will also be wishing for
> separation of metainfo sooner rather than later.
/LS
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
