I would argue against putting finder methods in the generated
entities, as I think it is a separate concern and you can easily end
up with tightly coupled code which is difficult to refactor.
The "Rich Domain" model can end up becoming the "Blob" anti-pattern.
I don't think stateful entities objects should be be mixed up with
static finder methods.
regards Malcolm Edgar
On Wed, Mar 31, 2010 at 6:10 PM, Andrus Adamchik <andru..bjectstyle.org> wrote:
> Also I suspect I am not using a classic version of DAO. My DAO's are morphed
> with the "business layer" and can have interactions with other subsystems,
> not just Cayenne (e.g. caching, data access via web services, etc.). Also
> they are not auto-generated, as any new methods are created when a need in
> them arises in the application. The idea is to keep it tight and avoid
> having any unused and unneeded methods.
>
> Andrus
>
> On Mar 31, 2010, at 10:02 AM, Andrus Adamchik wrote:
>
>> I am using a form of DAOs (with injection) with the main reason being that
>> I can provide alternative implementations in a context-sensitive manner
>> without changing the DAO interface (something that can't be done with the
>> static methods), which is very handy in numerous situations.
>>
>> Andrus
>>
>>
>> On Mar 30, 2010, at 11:51 PM, MGargan..scholar.com wrote:
>>>
>>> When using Cayenne are DAO's still the pattern used for findAll,
>>> findByProperty, and findById query methods or is there some better way to
>>> do this in Cayenne? For example in the pet store project I see DAO's
>>> being implemented, but I know that example was taken from other projects
>>> that use DAO's to begin with so I'm not sure if it was just retrofitted
>>> to
>>> use DAO's with Cayenne. I tried to look at some other examples from the
>>> website, but most of those are either not complex enough to warrant that
>>> kind of structure or the links are dead. I hope this is clear. :) Also
>>> if DAO's are the recommended pattern to use is there a tool (i.e. eclipse
>>> plugin) that will generate them?
>>>
>>> Thanks.
>>> -Mike
>>
>>
>
>
This archive was generated by hypermail 2.0.0 : Wed Mar 31 2010 - 23:32:44 EDT