Community
Participate
Working Groups
I recognize that this might be tough to do but it would be great to somehow avoid that first download of content.jar when looking for new software to install.
Simon, are you thinking that the platform would ship this way, or the packages built from EPP? For the platform to ship this way (at week +0), won't it need access to all the bundles / features from the later weeks (for version number and timestamps)?
Yep, it's not easy to do much for this with the raw platform. As I downloaded content.jar for twntieth time today I was just thinking that it might be worthwhile to do something especially since it really does impact people's first impression. I know I was sick of waiting... Something like EPP is probably the path of least resistance although I haven't really thought about what's involved.
I think EPP is way to go because it is the most popular download as well as the last thing being built. To ensure success at this we would have to guarantee stability of the cache location over VM.
We could also ship an SDK at +4 (or whenever EPP builds). It would be the original platform (build at week 0) with the final Gaileo content.jar included. Not sure how much work this would be, or if it would be worth it. Pascal, I'm not sure what you mean by stability of the cache location over VM?
I'm referring to the CacheManager code that computes the location of the cache on disk. From memory I believe it uses some hashing. As for the idea of modifying the SDK later I would vote -10000 :)
This would be nice, but I can't see it working in practice. Every release we've had some 11th hour tweak to the remote repositories, which would invalidate any local cache we have shipped with the build. I agree though that it's impossible for the plain SDK to ship galileo caches because we shut down two weeks before everyone else. If there is any hope for this, it would be in the EPP packages I think.
.