> My guess was how I was setting "rf.valid_bands" based on the '144/440'
> model or '144/220' model. Hopefully this will 'fix' it.

I can't apply a patch like that since I already applied your first one,
just FYI. However, the problem was that now get_features() was assuming
self._mmap was always loaded for band selection, which isn't the case
when we're just inspecting the driver. So, I just pushed a change that
should have retained your intended behavior as well as fixed the case
where self._mmap isn't loaded.

Please give it a look and make sure it's good.

Thanks!

--
Dan Smith
www.danplanet.com
KK7DS