Re: [gpsd-dev] buzzkill is gone, please re-test

[Harlan Stenn]

"Gary E. Miller" writes:
> On Thu, 29 Jan 2015 19:25:39 -0500
> Greg Troxel <address@hidden> wrote:
>>> "Gary E. Miller" writes:
>>>> Everyone should be using PPS anyway...
>
>> Someone has to rewrite the gpsd PPS code to not depend on the
>> non-standard TIOCMIWAIT and just use RFC2783 calls.   I left some
>> comments explaining this, but haven't done it yet.
> 
> I have a general plan, but no point coding without a debug platform.

If/when I have time to solve the "remote access" problem to get folks
access to my dev box in Oregon that will get you access to it, too.

> Can't use just RFC2783 calls.  Some way to poll the RFC2783 needs to
> be used to replace the simple wait fot input that TIOCMIWAIT provides.
> 
> RFC2783 tells you what happened, but will not wake you up to tell you
> something happened.
> 
> Also, most linux kernels do not have RFC2783 enabled so TIOCMIWAIT
> must still be supported when available.

I'm not saying the following is a good idea.

Would libevent be a possible choice here?

I used it instead of a select() loop for sntp, as a proof-of-concept to
make sure it would be OK for use as the main-line IO handler for NTP.

And not that y'all care, but libevent supports Windows, too.

H