Mailing-List: contact dev-help@manifoldcf.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@manifoldcf.apache.org
Message-ID: <53B3EBC6.3030606@apache.org>
Date: Wed, 02 Jul 2014 13:23:50 +0200
From: Rafa Haro <rharo@apache.org>
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9;
 rv:24.0) Gecko/20100101 Thunderbird/24.6.0
MIME-Version: 1.0
To: dev@manifoldcf.apache.org
Subject: Re: Configuration Management at Transformation Connectors
References: <53B2EA33.4010202@apache.org>
	<CALUFAGAXc_OzLxeniSCrnbn+yFrhoBvbvR6Ovxy8HtWThSc3dA@mail.gmail.com>
	<53B3C00D.4090409@apache.org>
	<CALUFAGBBvEgBRVoZc0DHo8aFZA0gm2YoP4J7767YCDW-D=ixGQ@mail.gmail.com>
	<53B3DE8E.5090908@apache.org>
	<CALUFAGCC1iXOcMCH2a7N9QOL49zrCvKXeGALMNXUu+1jo9anBA@mail.gmail.com>
	<53B3E294.8040907@apache.org>
 <CALUFAGB8KcZEMcOhHHzHhMOodnRsv5kRLBxeMtRSLG+CzmgZBA@mail.gmail.com>
In-Reply-To: 
 <CALUFAGB8KcZEMcOhHHzHhMOodnRsv5kRLBxeMtRSLG+CzmgZBA@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

Hi Karl,

El 02/07/14 13:17, Karl Wright escribió:
> Hi Rafa,
>
> Without a concrete example, you won't get a concrete recommendation.  But
> in general please understand that the architecture is what it is, for very
> good reasons, and if you want your connector to work within it, you need to
> play by the rules.

Of course, I just wanted to be sure about the options. I appreciate a 
lot you help and your clarifications and I hope that also the uses cases 
of the developers integrating ManifoldCF can help the community to 
improve it by addressing (or not :->) the proposed issues. From my side, 
I'm going to further investigate how the framework works at code level 
to see if I can extend the Transformation connector workflow for 
allowing resources initialization from Connectors Configuration and 
Output Specificiation.

Thanks,
Rafa
>
> Thanks,
> Karl
>
>
>
> On Wed, Jul 2, 2014 at 6:44 AM, Rafa Haro <rharo@apache.org> wrote:
>
>> Hi Karl,
>>
>> El 02/07/14 12:37, Karl Wright escribió:
>>
>>> Hi Rafa,
>>>
>>> If there are a limited number of read-only models, then you can build a
>>> static hashmap between model data and the file name, and populate that on
>>> demand.
>>>
>> It is a possibility, although the models stuff was just an example. The
>> generic question is how to reuse resources in a Transformation Connector?
>> or, which is the proper place in the implementation to initialize a
>> resource that has to be reused during a job running?
>>
>> I'm still missing this because I still don't know where to access the
>> Specification object apart from the getPipelineDescription method. So, if
>> there is no other point in the architecture, I'm afraid I have to check if
>> the initialization has been done or not and even synchronize it by myself.
>>
>> Thanks for your help so far Karl
>>
>>