archiva-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brett Porter <br...@apache.org>
Subject Re: from plexus to spring...
Date Mon, 25 Feb 2008 14:50:54 GMT

On 26/02/2008, at 1:46 AM, nicolas de loof wrote:

> On plexus XML descriptors, the <field-name> seems to be optional...  (
> CacheFailuresTransferTest.xml)

Yes, I think it will attempt to guess based on the interface type if  
there is only one.

>
> This makes things more complex on the spring side. Is this a common  
> use case
> or can we live without this feature ?

If that's not doable, can we throw an unsupported exception if it's  
not given?

>
>
> I've found the bug in archiva-proxy CacheFailuresTransferTest :  
> there was
> some plexus to spring mistakes, but the final issue is the  
> requirement on
> RepositoryContentFactory for the Contextualizable interface, and  
> direct
> access to the PlexusContainer.
>
> I plan to create a PlexusContainerAdapter to expose Spring beans
>
> Nicolas.
>
> 2008/2/25, Brett Porter <brett@apache.org>:
>>
>>
>> On 25/02/2008, at 10:23 PM, nicolas de loof wrote:
>>
>>> Can someone confirm the plexus behavior for IoC :
>>>
>>> When the target field is a Map or a Collection and no role-hint is
>>> specified, inject all component implementations
>>>
>>> When the target field is the component interface and no role-hint is
>>> specified, inject the no-hint component or the one with rol-hint
>>> "default".
>>
>>
>> Correct.
>>
>>
>>>
>>>
>>> Is there a special handling for "mock" role-hint ? I have some
>>> issues with
>>> component initialisation in archiva-proxy CacheFailureTransfertTest
>>> that
>>> seems to be related to MockArchivaConfiguration...
>>
>>
>> No - are you reading the CacheFailureTransferTest.xml file as a  
>> plexus
>> descriptor?
>>
>> - Brett
>>
>>
>> --
>> Brett Porter
>> brett@apache.org
>> http://blogs.exist.com/bporter/
>>
>>

--
Brett Porter
brett@apache.org
http://blogs.exist.com/bporter/


Mime
View raw message