Closed bjhargrave closed 17 years ago
Comment author: @pkriens
We are currently very much bound to 1.3 functionality. Today, generics, annotations, autoboxing, etc. are very useful. We must have a migration path for using, or allowing these features to be used, in our specifications.
Comment author: @bjhargrave
Jeff McAffer made an eloquent observation on this topic during the OSGi panel discussion at EclipseCon. Since developer should be writing POJOs and wiring them together with IoC extenders like DS and SpringOSGi, they wont be using any of the OSGi api directly. So there is no real need to waste effort on Java5-ing the OSGi api. I think I tend to agree.
Comment author: @pkriens
Not so fast ... I think we should first look what it means to the API because in certain cases things can become a lot simpler and there are people that write directly against the API. It might turn out that the gain is small, but we won't know until we looked at it.
Comment author: Jeff McAffer <jeff_mcaffer@ca.ibm.com>
I am fine with someone looking at it but this has to be a long hard look with at least some eyes focused on API evolution. in particular, the API and Java 5 presentation from EclipseCon http://eclipsezilla.eclipsecon.org/php/attachment.php?bugid=3626 may be useful in illuminating some of the challenges and pitfalls.
Comment author: Richard Hall <richard.s.hall@sun.com>
Might it also not be possible to add Java5 features to DS, e.g., annotations? May not end up being worthwhile, but it is worth investigating. I would shoot for as little as possible, though.
Comment author: @bjhargrave
CPEG call: Agreed to close until we get significant demand for a "new" api using Java 5 features.
Original bug ID: BZ#381 From: @pkriens Reported version: R4 V4.0.1