OSGi
- seams to be an option I will check the specs
- spring osgi
- is m2 and therefore not an option
- has dependencies to other spring parts we would like to keep
out for now
- are the other options (ready to use and free implementations)
I started to investigate.
OSGi seams to solve:
- deployment
- module registry (checking for updates)
- lifecycle
It does not solve:
- update mechanism itself (we have to implement that anyway)
My concerns:
- not sure if it will be easy enough to create modules
Please correct me if I'm wrong!
I had a quick look at FELIX (apache), what kind of projects do you
recommend?
After an other hour of research I would make the following statement:
I like OSGi and think that magnolia can highly benefit of it. In case
we use spring it will be a very good idea to use springs OSGi. Since
OSGi would imply other architecture changes (services, ...) we
proceed by using our own implementation. We add/change only a few
things (update, module instantiation). We do also have additional
concepts like having a GUI, ...
The update mechanism is currently the last missing peace for
releasing 3.1 and I think we should not invest more than two weeks
for that. Definitely we should not change more apis as we already
have done and introducing OSGi would definitely need a lot more changes.
Never else I think we should not invest to much time into things
which are solved much nicer by OSGi (deployment, registry, ...).
When ever we will have enough courage and enough time to make such a
move to spring I will be happy ;-)
Philipp Bracher
----------------------------------------------------------------
for list details see
http://documentation.magnolia.info/docs/en/editor/stayupdated.html
----------------------------------------------------------------
