the libegl error is interesting, its one thats probably going to keep cropping up if we don’t standardise on a clean raspbian stretch install… its always going to be tricky supporting a distro that someone has been installing stuff on, and so might get in an unknown state.
i guess ultimate goal would be to create a minimal building image… this is actually not that difficult to do, as theres a template project out there to do that.
(but its only really worth it, once we have a firm idea what to include - as it takes hours to build the image!)
anyway, that should not matter, as basically, it doesnt matter if the framebuffer initialisation fails (for whatever reason) as it nots (with code changes) going to do much useful for you - and matron still continues without it (and also going to fail for the encoders too)
one thing i found useful in debugging was to use oscdump to check that matron is sending out ready , and oscdump on crone to see if it was replying ( you can use oscsend to ‘fake’ the comms between them)
(if you see my post on the development thread, you can see i mentioned this)
a small note… which i dont think is the issue here, but useful for others
did you uninstall the old supercollider? just to make sure crone is picking up the correct one?
also when I switched, I needed to delete the norns build directory (?) , and rebuild it agains the new sc libs, and then run sc/install again to ensure I had the new extensions in place.
( I dont think these are your issues, as I seem to remember seeing errors, that made it obvious to me, which is why I forgot id done it 
)
