Dave Taht is in my basement trying to use GPSD to set up NTP-independent time service on an WNDR3700 router, and having some problems. I’m upstairs teaching GPSD to emit a clock-drift message – both projects are because we’re trying to build a monitoring framework for accuracy-checking NTP. The following IRC exchange ensues:
[11:31] dtaht2 looks like I have an underconfigured gpsd, miscompiled gpsd or ntp [11:32] dtaht2 OR does gpsd not provide time until it gets a full fix? [11:32] esr That's correct. [11:33] dtaht2 yes, in terms of 'or' statements, the above evaluates to 'true'. However... which? [11:33] esr Some devices report time from one satellite but you can't count on that. Most won't report time without 3 sats in view and good enough SNR. [11:34] dtaht2 cgps does report the time, so this particular device is [11:35] esr OK, you have a problem somewhere else in the chain. And a learning experience just ahead of you. [11:35] dtaht2 and a dark tunnel ahead. There may be grues. [11:36] esr Take your flashlight. [11:36] dtaht2 w;w;w; [11:36] esr You see a rusty wand with a star on one end. [11:37] dtaht2 get wand; wave aimlessly [11:38] esr Nothing happens.
Actually, I went downstairs and said the last line to Dave rather than typing it. He then laughed immoderately.
If you failed to understand the above, you are probably a normal human being and not an unregenerate geek who spends too much time in basements. This is sad for you.