Re: [gpsd-dev] gpsd rolls back to 1999

[Greg Troxel]

Hal Murray <address@hidden> writes:

> Using the system time at startup might work for gpsd.  It seems counter to 
> ntpd's assumptions.

I think that's too dangerous (system time and gpsd).  Time issues can
become circular.

We are basically working around broken equipment (it's quite possible
for the GPS receiver to remember the last date in flash and have a
manual epoch setting, and basically get things right).  But it seems
reasonably safe to add 1024 weeks to reports from hardware that are
multiple years before some compiled-in-constant.