openjpa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Craig L Russell <Craig.Russ...@Sun.COM>
Subject Re: FetchGroup implementation
Date Thu, 28 Aug 2008 22:40:21 GMT
Hi Kevin,

On Aug 28, 2008, at 2:45 PM, Kevin Sutter wrote:

> Craig,
> Comments below...
>
> On Thu, Aug 28, 2008 at 4:21 PM, Craig L Russell <Craig.Russell@sun.com 
> >wrote:
>
>>
>> On Aug 27, 2008, at 4:05 PM, Craig L Russell wrote:
>>
>> The JDO expert group has adopted a user-level API for configuring
>>> FetchGroups at runtime. There is a factory for FetchGroup in the
>>> PersistenceManager and PersistenceManagerFactory, corresponding to
>>> EntityManager and EntityManagerFactory.
>>>
>>> I'd like to implement this concept for OpenJPA, and have a few  
>>> questions
>>> before I start.
>>>
>>> Currently there is an Annotation  
>>> org.apache.openjpa.persistence.FetchGroup
>>> that allows an annotation to be defined for a class or field.  
>>> There are also
>>> FetchPlan.java and FetchPlanImpl.java in that same package. The  
>>> FetchPlan is
>>> the interface (not standard) and the FetchPlanImpl is the  
>>> implementation for
>>> the FetchPlan interface.
>>>
>>> So if we want to have an interface representing the FetchGroup how  
>>> do we
>>> avoid the name conflict between the annotation name and the  
>>> interface name
>>> for FetchGroup?
>>>
>>
>> This name conflict is making me crazy. Why did we put annotations  
>> into the
>> same package as interfaces and implementations?
>>
>> FetchPlan is an interface but it might also be an annotation in  
>> future,
>> along with FetchPlans. These annotations would allow you to define  
>> named
>> fetch plans in annotations that could be used for static definition  
>> of fetch
>> plans for queries, etc. without needing an API. Sort of the inverse  
>> of what
>> we found useful in FetchGroup.
>>
>> What if we moved all the annotations currently in
>> org.apache.openjpa.persistence to org.apache.openjpa.annotations,  
>> deprecate
>> the current annotation definitions in org.apache.openjpa.persistence?
>
>
> Although I can understand the frustration with the current  
> organization,
> moving and deprecating the current annotations would be a major
> compatibility issue as we move forward.  Besides the native OpenJPA
> applications, we have many other packagers of OpenJPA (IBM, BEA,  
> others?)
> that have documented the use of OpenJPA annotations.  Changing these  
> now
> would be disruptive.
>
> Since the use of the annotation would be more common than the  
> interface, I
> would prefer to separate out the interfaces into a new directory.   
> But, even
> that, I'm not thrilled with.

I understand and thought that throwing out a proposal would stimulate  
some thought.

I really hate all of these options:

creating a new directory in which we put only FetchGroup (the  
interface) and FetchGroupImpl (the implementation of the interface)

deprecating and breaking existing applications

keeping both interfaces and annotations in the same place

calling FetchGroup something entirely different, although calling it  
OpenJPAFetchGroup, or JDOFetchGroup might not be a bad idea

Of all the unattractive options, I'd probably vote for the latter  
(lattest?) (JDOFetchGroup).

Craig
>
>
> Kevin
>
>
>>
>> Craig
>>
>>
>>>
>>> Craig
>>>
>>> Craig L Russell
>>> Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
>>> 408 276-5638 mailto:Craig.Russell@sun.com
>>> P.S. A good JDO? O, Gasp!
>>>
>>>
>> Craig L Russell
>> Architect, Sun Java Enterprise System http://java.sun.com/products/ 
>> jdo
>> 408 276-5638 mailto:Craig.Russell@sun.com
>> P.S. A good JDO? O, Gasp!
>>
>>

Craig L Russell
Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
408 276-5638 mailto:Craig.Russell@sun.com
P.S. A good JDO? O, Gasp!


Mime
View raw message