David Jencks wrote:
On May 10, 2009, at 3:21 PM, Ate Douma wrote:
David Jencks wrote:
Also, does apache have an appropriately licensed copy of one or both
portlet xsds?
Not that I'm aware of.
However, why would we need them as we're not distributing them either?
why not :-) ?
In geronimo IIRC we ended up typing up the schemas without any
comments. This was unpleasant.... if there's a real copy we can have
in svn that would be great.
It might be useful I guess, but the porlet specs (pdf) already
contains them verbatim too.
Why would we need to "type up" these schema's?
At best sun probably provides cddl schemas. There has been some
argument about what license would apply to the result of running
xmlbeans or jaxb code generation on a cddl schema; Craig Russell thinks
it ought to be cddl.
Ugh, you're kidding me, right?
Are you saying the code generated by jaxb from the xsd "inherit" the license
from the xsd(s) itself?
I used the jaxb compiler initially to generated skeleton code based on the xsd, but it turned out so crappy I rewrote at least 90% of it by
hand... Only real "left-over" bits I kept are the generated javadoc headers in the sources as documentation.
See for example:
https://svn.eu.apache.org/repos/asf/portals/pluto/trunk/pluto-container/src/main/java/org/apache/pluto/container/om/portlet/impl/PortletAppType.java
In order to avoid any discussion of this having an
apache licensed copy of the parts of the schema that don't affect code
generation is valuable for anyone who wants to generate such code. As
far as we've been able to figure out, if we type up the schema without
any of the sun documentation annotations the result can be apache
licensed.
Jeez, how far does can this licensing madness go?
Does that mean if I simply delete the generated javadoc (based on those documentation annotations) from the code, we're fine just as well?
If so, I'm happy to do so right up... although I think its kind of nuts.
Regards,
Ate
thanks
david jencks