Got one, possibly two, interested parties at this weekend's BGD
meeting. Also got some coding done today.
I have a new LDraw model parser based on ANTLR4, and added in support for the BFC language extensions which I didn't have configured in my grammar for the Torque based parser. Also got more work done on the new preferences system to support loading of hierarchical configuration files from pure Python data structures. The next step is LDraw directory verification, then the caching, then model loading and rendering. I also have the necessary framebuffers set up for my inferred rendering pipeline, but nothing to do with them until I have LDraw models loaded with normal-group IDs. The infrastructure for the multiplexed console logging system is also functioning and in place.
Roadmap after that looks fuzzy but something like
* Functioning UI layer
* Networking + Physics
Also, I don't know if I mentioned this officially yet, but I'm now using Scala/Jython (and a smidgen of Java that is slowly being phased out), instead of C++/Boost+Python. I'm coding much more rapidly, and it should be more accessible for other people to help, though I'll have to update the GitHub accordingly. Also, JVM garbage collection is mature enough that well designed code shouldn't have to worry about stalling like it would in most other managed-memory environments.
At this point the only thing required for porting/cross platform support is the implementation of 5 or 6 C functions that I call through JNA to get information about platform-native folders and storage for applications to write to. The Windows and generic Unix code could actually have been done in pure JNA, but the "right way" to do it on OS X required calling Objective-C, which I didn't want to deal with through JNA, so I've separated all of that out into its own library (which I'm calling libcrane, as a some kind of pun about folders/origami). The OS X implementation is only 85 loc, and 6 functions, so this shouldn't be much of a burden on porting efforts.
Also, I'm trying to target OpenGL Core Profile 3.2, so if your card doesn't support that stuff....I'm sorry, but after the research I've done, I don't think there's much chance of having it performant anyway without the 3.2-era functionality I'm using.
Oh, and GitHub is updated if anyone wants to poke around.