[
https://issues.apache.org/activemq/browse/SM-992?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Freeman Fang updated SM-992:
----------------------------
Fix Version/s: (was: 3.1.2)
> Fix XSD schema locations for use with plain spring
> --------------------------------------------------
>
> Key: SM-992
> URL: https://issues.apache.org/activemq/browse/SM-992
> Project: ServiceMix
> Issue Type: Task
> Reporter: Guillaume Nodet
> Fix For: 3.2
>
>
> The spring application loaders use validation and requires a schema location.
> We need to do the same as ActiveMQ did:
> *
> https://svn.apache.org/repos/asf/activemq/trunk/activemq-core/src/main/resources/META-INF/spring.schemas
> * http://activemq.apache.org/xml-reference.html
> * http://activemq.apache.org/schemas/
> (12:10:25 PM) gnodet: jstrachan: what did you fix for the xbean / amq /
> spring problem ?
> (12:10:43 PM) jstrachan: the spring 2 stuff on the AMQ lists?
> (12:11:03 PM) gnodet: the schemas locations iirc
> (12:11:26 PM) gnodet: someone just reported the same problem on smx user list
> (12:11:32 PM) jstrachan: mostly it was to put the XSDs in a nice place and
> use a nice URI to the location of the schemas
> (12:11:32 PM) jstrachan: http://activemq.apache.org/schema/
> (12:11:41 PM) jstrachan: so the namespaces URIS are like:
> http://activemq.apache.org/schema/core/
> (12:11:47 PM) jstrachan: which points to all versions of the XSD
> (12:12:07 PM) jstrachan: then I've a cron script to copy 'em there from mvn 1
> repo
> (12:12:15 PM) gnodet: and the spring.schemas metadata, right ?
> (12:12:20 PM) jstrachan: then the other one - and this was the biggie I guess
> - was to create a spring.schemas
> (12:12:22 PM) jstrachan: yeah
> (12:12:32 PM) jstrachan: where it takes possible XSD locations (not namespace
> URI)
> (12:12:55 PM) jstrachan: this also means you have to in spring 2 to put the
> full XSD locations at the top of the XML file
> (12:13:11 PM) gnodet: yeah, because of strict validation
> (12:13:18 PM) jstrachan: http://activemq.apache.org/xml-reference.html
> (12:13:25 PM) jstrachan: yeah
> (12:13:32 PM) gnodet: ok, thx
> (12:13:43 PM) jstrachan: though there's no reason why an EntityResolver can't
> resolve an XSD purely on the namespace URI
> (12:14:03 PM) jstrachan: seems a bit bizarre that spring only supports
> namespace resolving on the XSD location (which is an optional thing)
> (12:14:26 PM) jstrachan: AFAIK an xml parser can still validate, so long as
> an EntityResolver can find the XSD location
> (12:14:38 PM) jstrachan: (I might be wrong on that - but I thought that was
> the case)
> (12:16:26 PM) jstrachan: here's the spring.schemas:
> https://svn.apache.org/repos/asf/activemq/trunk/activemq-core/src/main/resources/META-INF/spring.schemas
> (12:16:26 PM) jstrachan: notice how it points the XSD location to the local
> embedded xsd
> (12:17:32 PM) gnodet: yeah
> (12:17:40 PM) jstrachan: I did 2 kinds, like spring do - one XSD with no
> version number (which always resolves to the one on the classpath) and one
> using the actual version - only the first 2 lines are useful
> (12:18:01 PM) jstrachan: (I experimented adding the namespace URI in there
> but its never used)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
