GPS offset time error reporting

[wehimwich]

@kpgosite requests that bad values of GPS offsets be reported and inverse video used to highlight them in the monitor window.

This is a little complicated. Technically, there are no incorrect GPS offset values (except maybe outside of ±1 second). What may make sense is to report an error when the GPS offset changes by more than some amount, say 1 µsecond, compared to the average of the previous, say 10, samples. The error could be reported at most once per minute.

It is not clear this would be well behaved in general, but two cases seem pretty straightforward:

• A one time jump of more than 1 µsecond would generate a few errors, so there would at least be some notice. The error condition would however not be persistent.
• The GPS offset becoming noise-like with a fairly large range (which is what the RDBEs at @kpgosite experienced) would persistently report the error once a minute until resolved.

For systems with RDBEs, the error could reported by multicast logging program rdtcn and inverse video used in the monit6 display window.

For systems that use gps-fmout (or fmout-gps) with GPIB, a clone of the hpib command that implements the error reporting could be used. It could also plant the offset in shared memory for use in the monit2 display window. This would create a builtin gps-fmout command. It could use a C language function to implement the new functionality.

Stations that use their own gps-fmout command, would either not have this feature or would if they utilize the C language function.

Maybe this functionality should be configured in a control file that includes:

• The size of change to flag (default: 1 µsecond).
• Range of acceptable values (default: 0 to 1 second, systems with RDBEs would need to change it to -0.5 to +0.5 seconds).
• Error interval or disable (default: 1 minute)?
• Others?