As part of the work for bugs 236064
and 236711,
I have been trying to figure out a
simple and consistent naming convention for EclipseLink's schemas.
- rename the EclipseLink native deployment XML schema to eclipselink_persistence_map_1.0.xsd
- rename the EclipseLink sessions XML schema to eclipselink_sessions_1.0.xsd
- remove mention of Oracle in the namespace for native deployment XML:
Index: NamespaceResolvableProject.java
===================================================================
--- NamespaceResolvableProject.java (revision 1661)
+++ NamespaceResolvableProject.java (working copy)
@@ -25,7 +25,7 @@
public static final String ECLIPSELINK_PREFIX =
"eclipselink";
public static final String ECLIPSELINK_NAMESPACE =
- "http://xmlns.oracle.com/ias/xsds/eclipselink";
^^^^^^^^^^^^^^
In (the renamed) eclipselink_persistence_map_1.0.xsd
file, the namespace is declared as
<!-- Eclipse Persistence Service Project
:: Map Schema file for ORM/OXM/EIS -->
<xsd:schema
targetNamespace="http://xmlns.eclipse.org/xsds/eclipselink"
^^^^^^
I discussed this with Blaise - it is common practice to make your
namespace URI match up to an actual
URL so that people can download your schemas. The website "http://xmlns.eclipse.org" doesn't exist,
nor do I think the Eclipse Foundation would want us to 'take over'
something like that all by ourselves.
PeterK mentioned that we have complete control of everything under "http://www.eclipse.org/eclipselink",
so how about:
native deployment XML schema namespace : "http://www.eclipse.org/eclipselink/xsds/persistence"
new EclipseLink ORM XML schema namespace : "http://www.eclipse.org/eclipselink/xsds/persistence/orm"
The schema document itself has version information in the name eclipselink_persistence_map_1.0.xsd,
eclipselink_orm_1_0.xsd, etc) so
I don't think we need a version number in the namespace URI.
Comments?
Note 1: the changes are not backwards compatible so once
they are 'set' for the 1.0 Release, we will have to live with
them for a long time!
Note 2: once these changes get checked in, old 'legacy'
files will no longer work since by default schema validation is
turned on. I can scan the SVN repository for any and all files, but it
is unlikely that every place will be found on the
first pass - failures will point the way to where things were missed.
--
Mike Norman | Principal Software Designer | 613.288.4638
Oracle Server Technologies | EclipseLink Product
45 O'Connor Street, Suite 400 | Ottawa, ON K1P 1A4 | (fax) 613.238.2818
|