Rare project developers

[Alexander Gall <>]

Hi

On Wed, 16 Mar 2022 14:05:17 +0100, mc36 <> said:

> On 3/16/22 13:37, Alexander Gall wrote:

>>> while (true); do
>>> /rtr/ttyLin.bin /dev/ttyS0 20021
>>> sleep 1
>>> done
>> 
>> I'm a bit reluctant to do this since the serial port is the last
>> resort for accessing the device. I suppose this can somehow co-exist
>> with the regular agetty process on /dev/ttyS0 that allows access to
>> the host system, but how can it be guaranteed that it will fall back
>> to a normal login shell if anything goes wrong with freerouter or this
>> "intercept" process?
>> 
> that's why the ttyLin should be an external process to guarantee that it 
> runs...
> then freerouter's side, it opens the lines (interfaces, etc) right after it
> parsed the -hw.txt and then it immediately sends out something, including
> all the boot defailts so i'm pretty sure it's the best we can do....
> let me show you how my freerouter only images boot....
> the syslinux commands are here: 
> http://sources.freertr.net/misc/image/boot.cfg
> then i use this as init: http://sources.freertr.net/misc/image/init.sys
> that's how i bring up udev: http://sources.freertr.net/misc/image/init.dev
> and finally the freerouter startup: 
> http://sources.freertr.net/misc/image/init.jvm

I don't want to go that far, at least not without consensus in the
group. This would be a fairly drastic change from the current design
where we essentially have a standard Debian system with RARE
implemented as a systemd service. How hard we want to try to hide the
system from the user is an open question, I think. My personal opinion
is that we should not try too hard, as long as the user understands
how to move between the system and freerouter.

-- 
Alex