Grzegorz Kossakowski skrev:
Ralph Goers pisze:
...
Version 2.3 should only be created if we have some new feature for it ..
like OSGi (and it might even be possible to do that on 2.2 since it is
leveraging Spring - but you'd have to ask the guys working on it). I am
not in favor of creating a new release branch just to increase the minimum
Java version. If that is done I doubt 2.2 will get any new features or
support so what is the point.
What about some less distant and less ambitious goals like further splitting of
cocoon core, less
number of dependencies on Avalon and new EL stuff as default for sitemap?
I would like to include in this list other goals like finished implementation
of postable source and
its infrastructure (like caching) but this is not going to be in 2.3 or even
2.4 of Cocoon core.
It's going to be feature of 1.1 version of servlet service fw.
I mean that we should be less ambitious about cocoon core when it comes to
feature-set for
particular release because most of the interesting stuff will be *outside* core.
When it comes to OSGi integration I keep my fingers crossed for this effort and
I even would like to
join into this camp eventually but I think OSGi should not block any release.
It must be desire and
labour of community to act as lifeblood not just our wishes.
The current plan is that Spring-OSGi will be released in december, so we
should most definitively not let Cocoon 2.2 wait for it. Right now,
I'm the only one working on Cocoon-OSGi so it must be more of a
community effort before we include it in any release.
OSGi does not work woth "split packages", i.e. that some of the classes
of a package is defined in one bundle and others in another bundle. So
some reorganization of package structure in Cocoon is needed. This
probably means that Cocoon-OSGi would need to be in 2.3 or maybe 3.0 for
the core blocks.
/Daniel
