As you're probably aware, an Apache Chemistry F2F was held this week in Munich 
with the aim of merging the two code bases of Chemistry and OpenCMIS.

The attendees of this meeting were Jens Hübel, Florian Müller, Florent 
Guillaume, Nick Burch (mentor), Stephan Klevenz, Paul Goetz, Thierry Delprat 
and David Caruana. To ensure decisions were not made in isolation, details of 
discussions and outcomes were posted to the chemistry-dev mailing list 
throughout the week.

Here we will summarise the outcome. For details, see the mails posted this week.

The good news: there will be one Apache Chemistry Java code base.

As proposed by Florent, we started with the OpenCMIS code base and improved 
upon it by merging elements of Chemistry into it. This has resulted in several 
refactorings of the OpenCMIS code base, which has reduced the number of classes 
considerably, as well as clarified the client api. The current output of this 
work can be found at 
http://svn.apache.org/viewvc/incubator/chemistry/opencmis/trunk/.

There is still further work to do on this branch to complete the merge (as 
resprented by JIRA tasks CMIS-193 to CMIS-199) which will take place over the 
next couple of weeks. After completing these tasks and ensuring all is stable, 
this branch will be considered the formal trunk that all of the community can 
start using.

What's going to happen to the current Java code base? As there are still some 
clients and server implementations relying on this code, it will be moved to a 
branch and deprecated. The aim is to re-host these implementations on the 
/chemistry/opencmis/trunk/ code line after it's stabilized. At this point, it 
would seem good timing to provide the first formal incubator release of Apache 
Chemistry OpenCMIS subproject. This will also require writing new documentation 
and for the Apache Chemistry web site content to be updated.

There was also discussion on how to structure the various subprojects of Apache 
Chemistry. You may notice that OpenCMIS has moved directly under /chemistry in 
svn. A suggestion is to move all other subprojects (e.g. cmislib, jsclient, 
abderaclient, tck, shell etc) to /chemistry so they can support their own 
development and release cycles. 

Hopefully that reflects things accurately, that's all for now

Reply via email to