Community
Participate
Working Groups
In the past, the Marketplace Client has been generalized into the P2 Discovery component and MPC has since built on P2 Discovery for its wizard implementation. With time, MPC has customized its major UI elements more and more to the point where they have little in common with the P2 Discovery implementation anymore. In fact, the use of the Discovery framework these days seems more of a hindrance, mostly just requiring us to adapt back and forth between our REST data model, the Discovery data model, and their and MPC's UI concepts. Also, many of MPC's UI performance problems are hard to solve due to the restrictions of the Discovery framework (e.g. the whole list refresh lifecycle, repeated destruction and construction of our item widgets etc.) I suggest to look closely at what we get from Discovery and if we can either remove its additional layer from our UI implementation.
Although I don't expect to actually get this done for Neon, I'm going to tentatively target our 1.5 release here, so we can get the discussion going. But I'd actually prefer to do this early in the next development cycle, when there is more time to get lots of early testing in more iterations done - and when there might be a chance for larger changes, maybe also in P2 Discovery.
Added bug 481165 as a subtask, since we could fix that with a better list control.
Moving to 1.6. I'll start the discussion about these plans on mpc-dev shortly after the Neon.0 release.
