XTAL error calc not using NAV-TIMEGPS

[pshep]

Should this: https://github.com/Lora-net/lora_gateway/blob/a955619271b5d0a46d32e08150acfbc1eed183b7/libloragw/src/loragw_gps.c#L678 Not use gps_time, not utc for calculating the 'slope'?

On a separate but related topic, I understand the NAV-TIMEGPS is used to avoid leap seconds in the NMEA stream, but why are they an issue? I assume it's got something to do with the possibility of time going backwards?

3rdly... right there we're measuring the time between PPS pulses. Should this not be expected to be exactly 1 second (within 30ns or so)? Why even measure the period over hard-coding to 1 second?