Agree.
My action plan:
1/ reopen KARAF-53 and set the validation "as optional" with a warning
2/ open the Jira in the dependent projects (Camel, AMQ, etc) to add the
namespace in the features descriptor
3/ when we have a new version of the projects with the namespace, I will
reset the features validation as required
Regards
JB
On 11/23/2011 05:02 PM, Guillaume Nodet wrote:
Well, not being able to install activemq and camel sounds like a problem to
me, even on trunk. Until we have releases of the satble branches that
reference the schema, i'd be inclined to support those older release. Even
then, people may want to deploy old versions, and I don't really see any
real valid reason to not allow them to do that. If the only real benefit
is to put something in the column when using feature:list, it may not be
worth it.
On Wed, Nov 23, 2011 at 16:52, Jean-Baptiste Onofré<[email protected]> wrote:
Does it mean that I reopen KARAF-53 to just raise a warn in the features
validation method ?
To be honest, I would prefer to leave the trunk with a "blocking"
validation and update the dependent projects.
WDYT ?
Regards
JB
On 11/23/2011 04:40 PM, Guillaume Nodet wrote:
I don't think it's about disabling the xml validation, but it's about
allowing old feature definition to work.
Afaik, both Camel and ActiveMQ can't work with the current trunk, so I'd
rather defer that change to 3.1, and make sure we have releases of stable
versions that will work with 3.0 before doing anything.
On Wed, Nov 23, 2011 at 16:23, Jean-Baptiste Onofré<[email protected]>
wrote:
Guillaume,
I was too quick in my previous answer. The Camel issue is not the same,
the name is well defined, but it doesn't define the namespace.
On trunk, the XML validation using the XSD is used for a long time now.
I don't want to disable it on trunk, and I think that the name definition
is fine.
WDYT ?
Regards
JB
On 11/23/2011 09:44 AM, Guillaume Nodet wrote:
Actually, I've just hit a problem where the latest camel stable release
2.8.3 can't be deployed anymore because it does not contain the
namespace.
I wonder if rejecting such features definition is a bit premature at
this
point.
On Sat, Nov 19, 2011 at 18:11, Jean-Baptiste Onofré<[email protected]>
wrote:
Hi Ioannis,
My comments:
1/ it's on trunk, I display just a warning on 2.2.x
2/ the main reason is all features without name will use the same
repository name (local-0 or so). It means that
features:listrepositories
is
not helpful at all. With a mandatory name, we have a clean repository
by
features.
3/ all projects (Camel, CXF, SMX, etc) define the name on the recent
versions. For instance, the last version without name is Camel 2.6.0.
That's why I changed it on trunk, as it makes sense for me.
Regards
JB
On 11/19/2011 02:02 PM, Ioannis Canellos wrote:
I came across this one today:
-<xs:attribute name="name" type="xs:string" />
+<xs:attribute name="name" type="xs:string" use="required" />
</xs:complexType>
Though I don't mind setting things strict, this will render some of
the
existing existing feature descriptors useless (I am refering to 3rd
party
descriptors).
This sounds kinda heavy and I do not see a good reason for it (what I
mean
is that in order to cover a trivial need we seriously break backwards
compatibility).
If we still have to do so, then we might need to bump the version of
the
xsd.
--
Jean-Baptiste Onofré
[email protected]
http://blog.nanthrax.net
Talend - http://www.talend.com
--
Jean-Baptiste Onofré
[email protected]
http://blog.nanthrax.net
Talend - http://www.talend.com
--
Jean-Baptiste Onofré
[email protected]
http://blog.nanthrax.net
Talend - http://www.talend.com
--
Jean-Baptiste Onofré
[email protected]
http://blog.nanthrax.net
Talend - http://www.talend.com
