Hi team,
as we are moving forward w/ the releases, and we are close to have our 1.1.0.Final, I started to think about a proposal for a near-term future roadmap, and I'd like to get feedback, comments (or tomatos) on it.
- 1.1.0 -> very soon
- 1.1.x -> perhaps some needed bug fixes/improvements, in a short interval :-)
We have a release version in JIRA already.
Key features
- Keycloak isolation
- GCM 3 support (client and server)
- Improved docker support (e.g. tests/test suite -> Hopefully Travis supports 'docker run' by than ;-))
One could think that going back to JPA annotations is a key feature as well ;-)
There is no concrete release version for this, but we have a larger ups-future version in JIRA. The ups-future version/label has a few interesting things, that we may have to do right after 1.2.0
Key features
- APNs goes HTTP2 (a must)
- Quiete time for push (aka timezone awareness of devices)
- Scheduled pushes
- Proxy server support
In October (2015) the WildFly 10 should be released and I'd like to see us adapting this for the 2.0.0 series! Also for a possible release of our 2.0.0 in March 2016, I’d like to stop the 1.x series!
Note: We don't have a release version for JIRA here, but heck! this mail is asking for feedback ;-)
Key features / Core changes
- UPS based on (public) APIs that are available in WildFly-10
- looking at using WF feature packs (similar to what Keycloak did, e.g. for layered "product"
- looking at getting an UPS sub-system
- Java8 only (as well as making sure works w/ Java9)
- internal communication fully based on messaging (A-MQ / HornetQ)
- WildFly-Swarm launcher (aka fat JARs)
- helps with a more modular system:
- Modular system (e.g. different “webapps”, like "sender.war", "registration.war", "UI WAR" etc)
- allows us to play with different platforms for the different “web apps”
- for a 2.x.y we could see how a vertx microservice (e.g. for the device registration) will behave in the larger system
Other new features, e.g. based on needs and/or requests could be weaved into 2.0.0 or 2.0.x, based on timing. I think the core focus for such a 2.0.0 is more if a technical nature.
Please review this initial draft of a roadmap - once we are all in agreement, it's time to hammer the roadmap into stone, uhm... JIRA :-)
Greetings, Matthias