[gpsd-dev] 48-hour 3.10 release warning

[Eric S. Raymond]

I think the codebase is now in as good a shape as it's going to get
without destabilizing changes and another development cycle.

The problems with SiRF IV and gpsctl speed changes being flaky in
client mode are annoying but minor in the larger scheme of things.

More importantly: Our tracker-bug inventory is down to one item and
all other pre-release issues appear to have been satisfactorily dealt
with.  Live-testing with SiRF-II, SiRF-III, and u-blox 6 looks good.
All regression tests pass.  The refactored PPS code is no longer
flaky, and delivers visible, moving time-delta estimates in gpsmon.

The code is passing inspection by splint, cppcheck, Coverity, and
pylint.  I will attempt scan-build cleanup in the next few hours, but
so far all of its current warnings look like false positives.

The Time Service HOWTO has a couple of minor FIXMEs pending in the
material on chronyd, but overall looks to be in good enough shape to
ship.

Gary, please do a last live-test with Garmin USB.  If anybody has 
any potential release blockers on their mind let's hear about it now.

My current plan is to ship some time Friday Nov 22. Until then I
expect to be working on documentation and the Time Service HOWTO;
the main issue will be deciding how much of Sanjeev Gupta's detail
on NTP calibration to merge.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>