I've thought about this some more and have some major reservations
about this approach.
- The generated code wont have apache headers, so we can't actually
check in into svn. There might be a way to fix this with a script of
some kind.
- More seriously, this will eliminate traceability to the source
schema. Sun has in the past fixed errors in the schema without
changing anything that can be recognized from the generated source.
(I think they changed a date in a comment). This is a major problem
IMNSHO.
So, I think we need to keep the source schemas in svn. One way to do
this would be to move this module into the tck project and publish
the binary jar from there. If we found a way to change the headers
we might be able to publish a source jar too.
So, I'm -0.999 on this proposal at the moment... I think we need to
find a way to get the source schemas into svn.
thanks
david jencks
On Aug 19, 2006, at 9:49 AM, David Jencks wrote:
GERONIMO-2332 didn't show up in the list of RTC issues sent out
today, and I don't see the "Begin RTC Review" button.... any ideas?
Anyway, please vote so we can get this taken care of.
thanks
david jencks
On Aug 18, 2006, at 4:15 PM, David Jencks wrote:
I've implemented the proposal to not ship any sun schemas or dtds
to avoid licensing questions and to check in the xmlbeans
generated files.
Please vote promptly.
I've provided patches for trunk and branches/1.1. I expect that
the 1.1 patch should work for 1.1.1 also, I'm hoping someone else
will check for me.
Before releasing 1.1.1 including this work we'd have to release
the proposed new specs jar with the generated xmlbeans classes in
it: this would involve changing the version number to 1.0 in
several places.
http://issues.apache.org/jira/browse/GERONIMO-2332
thanks
david jencks
