+1 to handle older specs.

--
Uwe Barthel <bart...@x-reizend.de>



On June 11, 2015 5:22:32 PM "Jamie G." <jamie.goody...@gmail.com> wrote:

Long as the patched Karaf can handle the older specs too then all is good :)

Once the new spec is in place I'll talk to the SDN community about the
changes and help them migrate as well.

Cheers,
Jamie

On Thu, Jun 11, 2015 at 12:33 PM, Guillaume Nodet <gno...@apache.org> wrote:
> So I started experimenting a bit, and the second solution seems to be very
> easy to implement.
> I've been able to deploy an upgraded pax-web 4.2.3-SNAPSHOT with specific
> Karaf 4 features and install pax-http without any problems (but the change
> of a few imports in karaf to support pax-web 4.x).
> So unless there are objections, I'll go ahead and add degraded support for
> Karaf 4 features to Karaf 2.4 and 3.0 branches.
> Once those are released, we should be able to migrate downstream projects
> to leverage the new Karaf 4 features where it makes sense.
>
> 2015-06-11 9:47 GMT+02:00 Jean-Baptiste Onofré <j...@nanthrax.net>:
>
>> Hi Guillaume,
>>
>> I would do for the second one, I think it's easier and make sense for
>> backward compatibility.
>>
>> Regards
>> JB
>>
>>
>>
>> On 06/11/2015 08:51 AM, Guillaume Nodet wrote:
>>
>>> I want to work on a nice way to migrate to the new 1.3.0 schema for
>>> features for downstream projects (pax-web, cxf, activemq, etc…).
>>> I have two possible ways in mind which I'd like to discuss.
>>>
>>> The first one would be to write an additional mojo for the maven plugin
>>> which would translate the new 1.3.0 schema to older schemas, deleting
>>> unsupported stuff.  The mojo would thus generate an additional schema with
>>> a different classifier, either for the old schema, or for the new one.
>>>
>>> Another way would be to add this translation tool inside a bug-fix release
>>> of older branches, so that the old feature service could support the 1.3.0
>>> syntax.  The drawback is that this would not work on already existing
>>> releases obviously.
>>>
>>> Or we could do both.
>>>
>>> Fwiw, I haven't experimented yet on the translation, so for complex
>>> features definition using the new schema, I'm not sure yet if the
>>> translation will lead to usable results.  As a last resort, if it's not
>>> usable, the downstream projects can manually provide the two repositories.
>>>
>>> Thoughts ?
>>>
>>> Guillaume Nodet
>>>
>>>
>> --
>> Jean-Baptiste Onofré
>> jbono...@apache.org
>> http://blog.nanthrax.net
>> Talend - http://www.talend.com
>>


Reply via email to