I swear this is all true...
I am trying now to compile my old version of Asterisk with app_rpt and the necessary modules to enable the USB Radio Interface (URI) made by DMK Engineering to work on Arch, after a suggestion on a mailing list. That and the old adage "Not for beginners" tempted me.
I am needing to compile Zaptel, which, while out of date, is needed for various modules in order to get app_rpt to work. And I managed to get it to "make" - until it fell over whining about libnewt.
The biggest issue I have found with this, and I understand now why (more in a sec) is the fact the kernel sources are not provided properly for the OS to use. This is a vexed question, related to the fact that until five days ago, the GPU boot process was an industrial secret.
Hopefully now the boot process is open sourced, we can expect to see an open bootloader, and standard kernels being used, rather than these modifications. But in the meantime, my question is thus:
If Zaptel was compiling before, and not whining about the kernel sources, and the only thing I've changed from compilation to compilation is the presence of libnewt, what is actually going on here? Note, now, it bombs out straight away with the "no kernel source" message.
This is more a point of interest now, I'm actually waiting for the distros that will make use of standard kernels on the RPi, and I'll go from there. Hopefully the process for getting the kernel source, compiling Zaptel against it, and then onwards to Asterisk (with app_rpt) will be a simple one.
If anyone can correct me on the making of those distros, ie, will they actually be made (although I'm guessing that would be the first thing to do with the open sourced boot software), that would be great too. I'm not actually looking for a timeframe, as long as I can get this running before early October next year, thats the timeframe I'm looking at.
Onwards, due to the issues I found with the RPi and URI, I have invested some research into another device, but it still needs a computer running app_rpt (and chan_voter) to tie the network together. This'll allow me to have a full simulcast transmit PMR voting system, and the RPi is one of the lowest powered Linux computers I know. Also one of the cheapest. So getting the software, bereft of the URI itself, is now a priority, I don't want to run this "director" on a power hungry Atom system.
(And yes, it must be bad if I'm calling an Atom "power hungry" ^.~)
Thoughts on this matter of interest, and comments on the development of standard kernel systems is very much welcomed!