Hi all: A while ago I mentioned that I had taken work on the surrogate container over to Github to try out both git and a Maven-based build.
It’s now reached an interesting milestone and I’d like to talk about bringing it back into the River project. Right now, you can startup the container with Reggie and Mahalo in it, plus whatever services you want to add. The container monitors the deploy directory and starts and stops services much like Tomcat or JBoss auto-deploys. You can start up a client program in a client container that manages the codebase server and security environment. As an example the Maven build includes the service browser. Check it out at https://github.com/trasukg/river-container Also there’s a ‘hello world’ example at https://github.com/trasukg/river-container-examples I’m thinking that this container could be a good entry point for those who wish to create services and clients that run in the Jini environment. Also, it will make a nice testbed for development of new deployment options, like annotation-based service authoring. I’d like to bring it back into River. It was derived from the surrogate code in skunk/surrogate. Obviously I’m under an ICLA, and all the files have Apache license headers, etc. Any opinions on whether it’s a reasonable River sub-project, or whether I should leave it at GitHub? And if we do bring it into River, would we like to have it in a git repository? I’m not sure what the state of git support is at Apache - there seems to be several projects using git (especially in the incubator), but the infra webpages still say it’s experimental. I would see the container as another top-level distribution artifact. And there’s plenty of development opportunities in making it more user-friendly, administrable, etc. In all, the goal is to make creation of Jini SOA as approachable as writing servlets. Cheers, Greg Trasuk.
