geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Jencks <david_jen...@yahoo.com>
Subject Re: de-gbeaning
Date Mon, 02 May 2011 22:44:32 GMT

On May 2, 2011, at 3:18 PM, Kevan Miller wrote:

> 
> On May 2, 2011, at 5:54 PM, David Jencks wrote:
> 
>> I'm making progress on the 3.0-osgi branch and hope to have the j2ee plugin building
shortly at which point I plan to look at the tomcat plugin.  (connector is next in the build,
but I have a different approach in my sandbox txmanager.  I'd like to get something less radical
working before looking at integrating that code).
> 
> Cool. Just pulled the changes in, I think -- still getting used to git -- and kicked
off a build.
> 
>> 
>> Once j2ee is in it should be possible to work on openejb, tomcat, jetty, and probably
most of the other builder plugins more in parallel.   Meanwhile a few tasks I didn't get to,
in case anyone has time on their hands :-)
>> 
>> -- the karaf-assembly packaging doesn't include legal files and probably not resources.
 I'd like it to include the ones generated by maven-remote-resource-plugin.  I haven't figured
out why they aren't included yet.
> 
> What exactly are you generating (trying to generate) using the maven-remote-resource-plugin?

Well, we need legal files from somewhere in the assemblies, and the default source is usually
m-r-r-p.  I don't really care how they get in there or from where.... and the solution needs
to be in karaf too.

> 
>> 
>> - I haven't actually checked that I've removed all use of ConfigurationManager and
DependencyManager and related classes by removing them and trying to build framework.
> 
> Can give that a try, once my initial build is done.
> 
>> 
>> -- I've been rewriting our deployment plan processing to use jaxb rather than xmlbeans.
 As part of this I'm removing the jsr88 DConfigBeans and related stuff.  I don't think there
are any tck tests for this.  If I'm wrong we may have to reimplement the jsr88 beans around
the jaxb tree.  Reviewing the code in framework to make sure I've been thorough might be useful.
> 
> Just to be explicit, are you proposing this as a potentially permanent condition? Or
temporary?
Permanent.  I think the jaxb trees are a much easier way to deal with manipulating plans.

thanks
david jencks

> 
> --kevan
> 


Mime
View raw message