Thanks for the feedback Kevan!

We understand from your comments that you are positive with regards to having the changes integrated with Geronimo, but that they are too significant to be applied without further discussion. We have no problem understanding this.

Our goals with the project has been to 1. add some new features (i.e. spec changes), 2. upgrade to supported versions of 3rd party libraries (for instance CXF), 3. upgrade to latest version of 3rd party libraries (SLF4J, Plexus and so on) and 4. better understand the inner workings of Geronimo. We have not concerned ourselves with TCK compliance, though this is obviously important from the community perspective.

We have done the project on the 2.2 branch because 3.0 was too unstable (as in changed too fast) for us to work with. Some of the changes are backports of your work on 3.0, others should in our opinion be applied to 3.0 as well (ActiveMQ, CXF), and some are 2.2-specific (GShell).

For the community to better understand the implications of the changes, we suggest the following approach:

1. We try to split the changes into smaller parts with specific areas of concern, and create JIRA issues for each of them.
2. The community decides whether or not to accept the issues.
3. We commit and support the accepted issues.

There will be quite a bit of work involved in splitting the patch, and some changes are dependent on others. Before starting this we should come to agreement on which specs upgrades are wanted. This will in most cases decide which major 3rd party frameworks should be upgraded. We should also decide if the "latest and greatest" approach to minor 3rd party libraries (typically commons-libraries) is feasible for Geronimo. There might not be a single answer to this.


(we'll email the corporate CLA to

On Thu, Nov 17, 2011 at 12:47 AM, Kevan Miller <> wrote:
Hi Trygve
Thanks for the email!

If anybody else struggles with Trygve's name, you might find this helpful: (how could it be pronounced any other way!)

On Nov 16, 2011, at 4:37 PM, Trygve Sanne Hardersen wrote:

> Hello,
> Our company has done some work lately on upgrading the dependencies used by Geronimo 2.2. The most notable changes are:
> As you can see the Axis2 integration is still quite buggy, and will have to be improved (we're not using it internally).
> Community
> Are you interested in having these changes merged back into the Geronimo source repository? We would be very happy to see them accepted in some way.

Those are some fairly significant changes. I haven't reviewed them in detail… In addition to the integration test issues, there are likely to be TCK issues to contend with.

In the rest of my note, I'm *not* going to be commenting on the details of these patches. There may be specific technical/philosophical problems with some of the changes. These problems would need to be resolved through community discussion. Note that this means there may be some give-and-take. There may be some parts of your changes that the community feels are wrong/incorrect. So, some parts of the patches may require change.

However, let's assume we can resolve any of these issues (I'd certainly hope so).

> We are aware that the first 3.0 beta has just been released (congrats!), and don't know how this fits into your 2.2 release policy. We have previously done internal Geronimo releases using patches, but as the changeset has grown, this approach is less and less practical. Our feeling is that we either have to get the work back into Geronimo in some way, or do a fork. One possibility is to branch 2.2, and apply our changes there.

There has been a lot of community focus on 3.0. However, this does not mean our 2.1/2.2 branches are closed down… I am expecting to see additional releases out of these branches… However, the scope of your changes may go beyond a normal 2.2.x service release. I wouldn't be a big fan of a 2.3 release. However, it's not impossible either...

So, my personal opinion -- I'd be interested in seeing your contributions make their way into Geronimo SVN. I would certainly hope that we can avoid a fork… I am assuming, however, that additional people (e.g. you) would be helping the community with integrating, testing, releasing, and *supporting* these changes. We're quite open to new contributors. And I would expect that it would not take long to become a committer on the project.

One final note -- I would not be in favor of *permanent* development on separate Geronimo branches (2.x vs. 3.x). So, I am assuming that given time -- the community would be working towards common goals.
Given the scope of these changes, I think a software grant would be required (either or ).

I'm sure there will be additional comments from other community members.