And where does merging your changes in with pd-extended 0.43 fit into all this?
Not sure. I've submitted at least half-dozen patches into the sourceforge already and many more via mailing list and only a fraction of them have been looked at, and even less merged. Granted, some of them are somewhat controversial (e.g. revamping the scrolling algorithm), yet with such a low response rate one certainly feels discouraged in contributing further, particularly considering just how time-consuming fragmenting improvements into sub-patches can be. OTOH, I do understand just how hard it can be for one to maintain code when there are a bunch of patches trickling in at all times--it's a full-time job in and of itself, particularly in respect to regressions. Yet, having spent good two weeks chasing exactly such regressions and IMHO improving editor experience to the point where both show-stopping as well as usability bugs have been by and large squashed, I certainly hope they will find their way eventually into the core Pd. The code produced so far has been clean and (apart from fprintf's for debugging purposes that are currently commented out awaiting further potential development) should be rather easy to merge into the main trunk. The real question is whether Hans, or perhaps more importantly Miller will find doing so to be of their interest.
All that said, I think I'll continue to maintain a L2Ork variation until either its feature-set becomes synonymous with the core Pd package or there is no more reason to maintain it (and FWIW as of right now there are plenty, so I don't see me stopping the support anytime soon).
Cheers!
Ico