| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
The numbers look better now.
|
|
|
|
|
|
|
|
| |
Try to do it with forward and backward compatibility, which is
challenging with current miscojson.
Sometimes epe was used to 2D estimated erro. Sometimes for 3D error.
So make it explicit eph is 2D, and sep is 3D.
|
|
|
|
| |
Odd that it was never there before...
|
|
|
|
| |
Replace zero times with blanks.
|
| |
|
| |
|
|
|
|
|
| |
This released a flood of GPGBS. But the GPGBS are wrong format
according to NMEA 3.0...
|
|
|
|
|
|
| |
The broken tests are the ones where we have to do our own conversion from
week/second to UTC because the device either doesn't supply UTC or its
reporting is broken.
|
|
|
|
|
| |
Also update the affected regression files. gpsd had been
throwing away the DGPS status.
|
|
|
|
| |
Explain what breaks on a leap-second transition, and why, in build.txt.
|
|
|
|
|
| |
A concerted effort to reduce all tests to below 10K in volume each while
preserving all significant test features.
|
|
|
|
|
| |
Turned up a bug in where a counter was incremented un the Navcom driver;
this required one test rebuild.
|
|
|
|
|
|
|
| |
Instead, set the used member in the satellites array directly where possible.
The NMEA0183 and TSIP drivers still need a local equivalent.
This changes pseudo-NMEA GSA output in several binary-protocol tests.
|
|
|
|
|
|
| |
Affects only GSA emission in pseudo-NMEA mode, fixing a lingering bug
whee the last sat in the listing was sometimes duplicated. Required
one test build.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
...from a set of parallel arrays. This change flushed out a
longstanding bug in the computation of DOPs for estimated error bars.
Some test-load rebuilds were required:
geostar-geos1m-binary.log.chk: With this change error
estimates are computed and reported.
trimble-lassen_iq-3dfix.log, trimble-lassen_iq-3dfix.log: the
change revealed a bug in the computation of satellite-seen bits.
Error estimates did not change.
navcom.log: Error estimates changed.
With these rebuilds, all regression tests pass.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Um. What was I thinking when I wrote that guard? It zeroed out
reports of satellites used.
Required a rebuild of the following (binary-orotocol) test logs.
test/daemon/bu303-climbing.log.chk
test/daemon/bu303-moving.log.chk
test/daemon/bu303-stillfix.log.chk
test/daemon/italk-binary.log.chk
test/daemon/navcom.log.chk
test/daemon/superstar2.log.chk
test/daemon/trimble-lassen_iq-3dfix.log.chk
test/daemon/trimble-lassen_iq.log.chk
test/daemon/ublox-aek-4t.log.chk
test/daemon/ublox-lea-4t.log.chk
test/daemon/ublox-sirf1.log.chk
test/daemon/zodiac.log.chk
After that, all regressiion tests pass.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Required a regression-test rebuild, of course. The field is still set by
the TSIP and SiRF drivers; the SiRF driver actually uses it. It may be
possible to eliminate the TSIP uses, but so far attempting this has
produced odd regression-test failures.
|
| |
|
| |
|
|
|
|
| |
Required a regression-test rebuild.
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Protocol version number is bumped. Python and C test clients are known
to work; interfaces of the C and Python client bindings are
unchanged. Third-party client-side bindings which rely on naively
copying JSON members will break (implementers have been repeatedly
warned not to do this).
|
|
|
|
|
| |
I thought I could avoid this, but it turns out SiRF chips before firmware
rev 2.3.2 don't reliably get a leap-second report either. Sigh...
|
|
|
|
|
|
|
|
|
|
|
|
| |
Gains: the stored leap-second offset we used for this could go stale,
breaking our regression tests in the process - itr's just dumb luck
that it hasn't since done so since 2008.
Losses: If the receiver doesn't have leap-second cached in NVRAM (e.g,
between cold boot and the next subframe message) time will be inaccurate
by a few more seconds.
10 regression test outputs of 66 had to be rebuilt.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This solves the disappearing epx/epy bug on SiRFs, but it was actually a
systemic problem that affected several drivers. Navigation solution messages
were clearing DOPs, making it impossible for the error modeller to compute
estimates. New logic: Clear DOPs only when we get a skyview report. They'll
be regenerated by our visibility-matrix calculation when the skyview sentence
is analyzed.
If a sentence from the device supplies a DOP between the time the visibility
calculation is done and when the next fix is reported, it will override
our computed value. This might change later!
This change required a regression-test rebuild.
|
|
|
|
|
|
| |
The guard controllong DOP computation wasn't right. The result was
that x/y error estimates were computed much less often than they
should have been.
|
|
|
|
|
|
|
|
|
| |
Change the data management slightly so the fixbuffer private member
doesn't have to be exposed outside libgpsd_core.c any more. Oddly,
this accidentally fixed a minor bug, shown by 5 regression tests in
which the pseudo-NMEA code should have been generating $GPGBS (because
ex/ey/ev show up in the TPV output) but weren't actually doing so.
Other than these virtuous additions, all regression tests pass.
|
|
|
|
|
|
|
|
| |
Rather than computing the value the WGS84 ellipsoid semi-minor axis in
every call to ecef_to_wgs84fix, and having this computation in a few
different source files, use a predefined value. This slightly changes
the output value of ecef_to_wgs84fix - in the least-significant digit,
or about 0.1mm.
|
|
...to something more descriptive.
|