[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[gpsd-dev] Now we know what breaks in GPSD on a leap-second transition
From: |
Eric S. Raymond |
Subject: |
[gpsd-dev] Now we know what breaks in GPSD on a leap-second transition |
Date: |
Sat, 11 Jul 2015 06:13:10 -0400 (EDT) |
New test in build.txt:
If you have had a leap-second transition, the following regression tests
will break:
bu303-climbing.log
bu303-moving.log
bu303-nofix.log
bu303-stillfix.log
bu303b-nofix.log
italk-binary.log
navcom.log
ublox-aek-4t.log
ublox-lea-4t.log
ublox-sirf1.log
There is no help for this other than a test rebuild. The problem is
that these devices rely on the build-time leap-second offset; you'll
see times one second off. Other GPSes either return
leap-second-corrected time or the test loads include a
leapsecond-offset report before any time is reported.
So our regression tests break, but the software itself does not.
I'll ship 3.16 sometime soon to clean this up. There haven't been any
substantive changes since 3.15.
--
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
The possession of arms by the people is the ultimate warrant
that government governs only with the consent of the governed.
-- Jeff Snyder
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [gpsd-dev] Now we know what breaks in GPSD on a leap-second transition,
Eric S. Raymond <=