Re: [gpsd-dev] Updated docs on NTP segment management

[Miroslav Lichvar]

On Mon, Feb 23, 2015 at 12:28:07PM -0800, Gary E. Miller wrote:
> Miroslav Lichvar <address@hidden> wrote:
> > I'm following the discussion, but I don't have much to say. I'd prefer
> > sockets over shared memory and keep it as simple as possible to
> > minimize the amount of code in the clients that want to support it.
> 
> Yeah, sockets are nice.  But two concerns.  
> 
> The chronyd socket is binary which brings up issues of structure
> packing, versions, etcc.

Yeah, I've already found some fields that I'd like to add to the
protocol, but didn't do it to not break the compatibility.

> Would using JSON over sockets violate your 'simple as possible' 
> requirement?

Do we really need to use a context-free language for this? I
personally would prefer something that can be parsed with couple
strcmp/sscanf calls.

I think phk is planning to use a simple text protocol in the new ntpd
replacement (ntimed) as the main protocol for all refclocks, which
will run in separate processes. If gpsd spoke that, maybe it could be
connected directly without having to run a special driver for
conversion.

-- 
Miroslav Lichvar