[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[gpsd-commit-watch] [SCM] GPSD annotated tag, release-3.10, created. rel
|
From: |
Eric S. Raymond |
|
Subject: |
[gpsd-commit-watch] [SCM] GPSD annotated tag, release-3.10, created. release-3.10 |
|
Date: |
Fri, 22 Nov 2013 13:40:37 +0000 |
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GPSD".
The annotated tag, release-3.10 has been created
at 989f35dfe3caf82d8ea489be0244b466ba496eb0 (tag)
tagging 2ac0cb7e1a03b761cd8c8bff0bb2d37341beb708 (commit)
replaces release-3.9
tagged by Eric S. Raymond
on Fri Nov 22 08:40:07 2013 -0500
- Log -----------------------------------------------------------------
Tagged for external release 3.10
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (GNU/Linux)
iD8DBQBSj163rfUW04Qh8RwRAntdAJ9LDu34u8BYPNtFsbxLRQOlTh2oTgCdFie7
XpVlGJDfG84/yBCX6a0LXHQ=
=+44x
-----END PGP SIGNATURE-----
Andrey Mazo (3):
Fixed incorrect string argument to dbus_message_get_args().
Crossbuild fix when using sysroots.
Added option to disable build of documentation.
Beat Bolli (29):
ntpshm.c: let the PPS init waiting loop terminate
webgps.py: fix import error when running directly from contrib/
webgps.py: set self.needsupdate correctly
webgps.py: fix errors when period is None
webgps.py: move the track drawing into class Track
webgps.py: simplify command line handling
webgps.py: add documentation
HOWTO: calculate the average GPS offset
Minor fixes.
Build: eliminate a lambda function.
leapsecond.py: set a socket timeout for slow servers.
leapsecond.py: make save_leapseconds() check all defined servers.
Define barrier() only if SHM export is enabled.
Typo fix.
gpsmon: fix the usage string
gpsmon: simplify the HAVE_BLUEZ conditional
Fix two splint annotations.
Typo fixes.
ppsthread: clarify the description.
gpsmon: log the file name instead of its FILE *
gpsmon: serialize output to the packetwin.
gpsmon: refactor gpsd_report().
gpsmon: unconditionally log the PPS reports to the packet window.
Fix compiler warnings with clang 4.2 on OS X 10.7.5
RTCM V2: use scons to determine the location of <endian.h>.
cgps: don't overwrite partial lines in the message dump.
Eliminate a redundant lambda function.
Build: fix an undefined variable.
Build: remove a redundant check.
Bernd Zeimetz (4):
Fix most of the undefined symbols in libgpsd.
Use UDP in regress-driver if ptys are not available.
Move gpsd_report from gpsctl.c into libgpsd_core.c.
Fix libdir in the libgps.pc file.
Brian Bieber (1):
Packaging fix for RHEL/Centos.
Christian Gagneraud (7):
Fix coveraging mode in SConstruct
www: Typo fixes
raw-regress: fix incantation
gps/fake.py: fix typo
SConstruct: fix indent
drivers.c: Remove dead reference to driver_{zodiac, ubx, sirf}_binary
drivers.c: move extern declaration of driver_nmea200 with other extern
declarations
Eeri Kask (1):
Build-system patches for MacOS 10.4.
Eric S. Raymond (721):
Commits are open again.
Documentation cleanup for gps2udp.
Note a missing feature.
Fix incorrect UNKNOWN value. Affects parsing of RTCM2 type 4 sentences.
Start on support for unpacking RTCM3 JSON.
Fix static-check warnings.
Two missing fields in RTCM3 1001 parsing.
Splint cleanup, spelling, and style fixes.
Move an assignment to pacify splint.
Suppress a compiler warning.
Unpack RTCM3 1002 JSON (untested)
Unpack JSON for RTCM3 R1007, R1008, and R1009.
Implement RTCM3 R1010 JSON unpacking.
Macro trickery that I might regret, but it makes the code more expressive.
Improve readability.
RTCM3 R104 JSON unoacking.
RTCM3 R1033 JSON unpacking.
Make the JSON unit tester more flexible.
Typo fix in AIDVM description.
Add a FAQ about almanac/ephemeris/pseudorange information.
Don't ignore a nonexistent binary.
Build and news housekeeping.
Minor plishing of GPS database.
Note Android usage.
Typo fix.
Fix type extraction for Evermore binary packts.
Disable Evermore binary trigger string so as to leave it in NMEA0183 mode.
Disable attempts to parse Evermore binary packets.
Nicer capability dumping on gpsd -l; merge with usage().
Remove dead code found by clang.
Another glitch caught by scan-build.
Add ex[perimental production for local scan-build.
Update the news.
Note the workaround for a Raspbian build problem.
A portability note.
Note the Raspberry Pi's issue.
Properly string-quote in ais.py output.
Cosmetic fix.
Minor fix to AIS type 1,2,3 radio field decoding.
Fix error in sixbit decoding.
Remove obsolete comparative.
Fix two minor errors in API description.
A note about threading.
Add record for Accutime Gold.
Increase GPS_PATH_MAX to correspond to practice on modern Linuxes.
News update.
API ninoer version needs a bump due to the GPS_PATH_MAX chhange.
Split Type 24 reporting, including the -s option to gpsdecode.
Fix for a typo spotted by Reinhard Arlt.
Document ITU-R 1371-4 interpretation of the vendorid field...
Explain the static library better.
Extract ITU-R 1371-4 model and serial fields from AIS Type24 messages.
Updare AIVDM regression tests for previous commit.
More about riverine extensions.
Wrap sixbit() calls in a macro to regularize them.
Documentation suggestion by Peter da Silva.
Fix up the ENDCHARS macro.
Another language tweak.
Document ITU-R 1371 change point.
cppcheck ckeanup. All regression tests pass.
Remove unused rate parameter.
More cppcheck cleanup.
splint/cppcheck cleanup.
cppcheck cleanup.
Original Trellisware patch.
Clean up the Trellisware patch by using event_identified.
Add a function to put a big-endian double. Not yet used.
Fix some minor bugs in debug reporting of the sat picture.
Restore some accidentally omitted code.
TSIP Double Precision XYZ needs to be passed, as it says in the comment.
Minor typo fix. All regression tests pass.
Typo fix.
Begin cleaning up some grubbiness in the packet getter.
More TSIP packet cleanup. All regression tests pass.
More TSIP packet cleanup. All regression tests pass.
Simplify TSIP satellite-picture packet recognition.
Remove some comments that belong in the revision history.
TSIP driver can only use one length of 0x5c packet.
More TSIP packet recognition cleanup. All regression tests pass.
More TSIP packet cleanup. All regression tests pass.
Deal with the Accutime Gold's known issues. All regression tests pass.
Pass through Trimble Accutime Gold bersion info.
Merge Igor Socec's initialization code for TSIP. All regression tests
pass.
Be more lenient about extraneous ? in a JSON command line.
TSIP packet 0x1c has to pass through.
Typo fixes.
Coverity cleanup.
Add gittip links and buttons.
Document use of manbuild.
Add some reassurance about NTP.
Add a warning by Andy Walls.
Note that NTP could be erroneous too.
More about time.
Caveat about PPS accuracy.
More about time constraints.
Keep adding detail about time problems.
Add link to "Notes on setting up reliable timekeeping".
Remove some thread debugging that stepped on other messages.
Eliminate a static global in favor of testing whether we're backgrounded.
Banish some compiler warnings.
Change gpsd_report to no longer use a global. All regression tests pass.
Remove duplicative code. Thee's only one copy of gpsd_report() now.
Properly label debug reports from different programs.
Cleanup after the big change.
Typo fix.
Use an initialization macro prperly. All regression tests pass.
Enhance gpsfake so it can be used to test TCP sources.
Cosmetic fixes.
Only run DEACTIVATE hook on close, not on hunt failure.
Increase the resilience of the packet getter.
splint and cppcheck cleanup. All regression tests pass.
Coverity false-positive suppression.
News update.
Better recovery in the longer Python example.
Partial Inland AIS coverage.
Inland AIS Number of Persons on Board description.
Describe all Inland AIS messages. These are not yet recognized by the
code.
Start a list of open questions.
More AIS open questions.
Axccept Persons On Board as either a type 6 or 8 message.
Note timezone ambiguity.
Remove some unused and erroneously included structure fields.
tablegen.py -t does more errorchecking.
tablegen.py conditionally includes -t.
Minor improvements to tablegen.py.
Support for Inland AIS ETA at lock/bridge/terminal.
Cosmetic fix.
More output improvements to tablegen.py.
Support for Inland AIS RTA at lock/bridge/terminal.
Support for Inland AISNumber of Persons On Board.
First step of disentangling text display from scaling.
Change JSON-dumper code generation to match my previous commit.
Fix up xgps to use _text names.
Dump preciptypes in the new manner.
Fix several small but fatal bugs in the new JSON generation.
The JSON parser now has an "ignore" type...
Rebuild the regression test loads so they dump in the new style.
Update documentation to capture the new behavior around _text attributes
News update about what has just been done.
splint cleanup.
Typo fix in some unused #defines.
New-style dumping (numeric and text both) for two more fields.
Track a previous name change.
Update the TODO.
Typo fix.
Driver code generation for boolean fields.
Derived-attribute generation for enumerated fields.
Fix a minor bug in boolean field generation.
Change a variable name that had become misleading.
TCX0 kit announces itself as a USB modem on ttyACM0.
Support Inland AIS ship static and voyage-related data.
When -t forces a driver type, show it in the gpsmon prompt.
Get the prompt placement right.
Generalize a logging function.
Redirect more data to the GPS through gpsd_write().
splint cleanup.
More splint cleanup.
We can now monitor low-level string sends through gpsmon.
Do things in a more natural order. All regressions pass.
Improve recognition of UBX chips with a slight change to the trigger
string.
More instrumentation and a name change.
uBlox 6 no longer sends model info in $GPTXT (observed on TCX0)...
Make binary driver selection sticky again.
Change the logic for managing the drive mode bit to report reality...
The trigger-string UBX detection was unnecessary; remove it.
Clarifying comment.
Slightly more sophisticated try at a synthetic config block.
Condition out the code that trues to synthesize ta CONFIG-PRT block.
splint cleanup. All regression tests pass.
Fix a minor linkage failure.
Implement handling of Inland AIS EMMA warning.
Fixes to generation of JSON dump code for arrays.
Fix a small slipup.
Express the end of the Inland Water Levels message as an array.
Inland AIS Water level support is done except for the JSON dumping.
ind -> i as a loop variable name.
Correctly generate JSON dumping code for trailing arrays.
Complete support for Inland AIS Water Levels message.
Tag layout tables so a program can easily search for them.
Table typo fixes.
Modified table parsing - a step towards parsing vocabulary tables.
Subtable parsing works.
Update a comment.
Generate structure name properly.
Make vocabulary references in dump code.
Correct some code generation.
Inland AIS full support (untested).
Variable renames, and remove an extraneous break statement.
Document some conventions.
Try another config synthesis. All regression tests pass.
Portability cleanup.
Some compilers barf if you name a struct member 'class'.
Remove overcomplicated type-switching logic.
Comment typo fixes.
We can finally see the UBX mode switcher being shipped.
Actually setting up a config block now.
gpsmon switching to UBX binary is now working.
Cosmetic changes.
Change the driver substructure to a struct from a union.
Cosmetic fixes.
Begin untangling reporting code from socket handing.
Another refactoring step. All regression tests pass.
Refactoring. All regression tests pass.
More refactoring. All regression tests pass.
Update a comment.
Cosmetic change. All regression tests pass.
Refactoring step. All regression tests pass.
Another small refactoring step.
More refactoring. All regression tests pass.
If NTRIP breaks, revert past here. Refactoring step, but not a trivial
one.
More separation of polling from fd set management.
Move some refactored code out of its original context.
Code transplant. No functional changes.
Give drivers a way to query the currewntly set stopbits and parity.
New, simpler UBX mode switching with synthesized config blocks works.
Switch UBX to binary mode when it's identified.
Introduce a mode option so gpscotl can be less intrusive than gpsd.
Partial splint cleanup.
Simplify signal handling.
splint cleanup.
Restrict the running of the device hook.
UBX mode-switching from gpsctl finally works.
Restore gpsctl's ID function and update docs.
Lower verbosity.
splint cleanup. All regression tests pass.
Fixed UBX baud-rate switching.
Cosmetic fix. All regression tests pass.
Make a 0 rewake time meaningful.
Update a comment.
Refactor gpsmon - moving towards using multipoll.
promptlen needs to ne gloal for monitor_complain to work.
Try for more usefyl sudden-death messages.
splint cleanup.
More informative error message.
Count subscribers more accurately.
Remove a mistaken optimization.
Deal with unsubscribed gpsctls more gracefully.
Yet another bug due to wacky CFG-PRT semantics.
Cut down on unnecessary sends during mode changes.
Add some new mechanism.
Make mode-switching from gpsctl work when going through the daemon.
Typo fix.
Minor command-handling fix.
Refactoring step.
Repaint less often.
Reduce amount of screen painting.
Refactoring step.
Fix a display glitch.
readpkt() no longer returns anything but the packet length; simplify.
Use gpsd_multipoll() for simpler code.
Handle error returns from gpsd_await_data() a bit more gracefully.
splint cleanup.
Narrow a variable's scope. All regression tests pass.
Refactoring step.
Renove an unused cursor move.
Fix a very minor bug reported by Trellisware.
Update some documentation.
Clarifying comment.
More clarifications.
Add a caveat.
Correct a comment.
Update release sequence. All regression tests pass. PPS on Macx-1 is
working.
Make gpsmon quit cleanly on 'q' again.
Having an generic NMEA mode switcher no longer fits the type transition
rules.
Enrich the packet type table.
Regularize driver names.
Move the MTK extension parset to where it belongs in the NMEA 0183 driver.
Move the AIVDM context into the main driver structure scratch area.
Compiler warning cleanup.
Add a driver property lister to test_packet.c.
Restore inadvertently removed .splintrc
splint cleanup.
Remove some driver methods that should not have been included.
Add the driver method consistency check to normal tests.
Enrich the driver method listing. Shorten an overly long name.
Don't block more packet display while accepting commands.
Update news and fix typo.
More news.
Observer-driver reversion logic. This is a simple but major change.
Deal with the Macx-1's quirk of not being able to turn off ASCII.
Create driver slot and scons option for a visualizer driver method.
Second step in the visualization code.
Make all hex-dumping truly thread-safe.
splint cleanup.
Add a note about misformatted Type 5s.
Typo fixes.
Be more relaxed about AIS Type 5 message lengths.
Add the link to the Maritec decoder.
OK, visualize as a method was a bad idea. All regression tests pass.
Fix a bug in MT3301 handling.
Eliminate a swich_driver() call that is never used. All regression tests
pass.
Remove dead code and clean up. All regression tests pass.
Clean up an archaism in the prototype driver.
More elimination of code rendered dead by packet-type dispatching.
The OnCore driver had nonfunctional stub control methods. Remove them.
The old Delorme Earthmate control_send method is never used...
Also save a driver type as last controller if it's NMEA but not generic.
First cut at salience test.
This works, but it's inefficient.
More efficiebt sticky-flagging.
Zip file production for binpress release.
When making a zip for binpress, must do directory recursion.
Revert driver whenever it's non-sticky.
Simplify the drive transition test.
Typo fix.
Clarifying name change. All regression tests pass.
Under the new transition rules, event_identified fired too often. Fix
this.
Narrow a variable scope. All regression tests pass.
Prevent a race between KPPS initialization and privilege dropping.
Clean up the KPPS initialization wait.
Some clarifying comment.
A volatile that might be helpful, and explanatory comments.
Allow gpsdctl to run as non-root.
Don't spam the user with device notifications.
Be certain the new logic filters out packets with bad checksums.
Remove an FR, because the JSON passthrough driver exists.
Egg-on-my-face build and compiler warning fixes.
Typo fixes.
Gary says the table of NTP offsets for old devices is junk. Removing it.
First, very early cut at a time service HOWTO.
Transplant some sections of gpsd(8) to the new HOWTO.
Fix a strange typo that crept into ntpshm.c.
Trivial markup fix.
More for the time service HOWTO.
Describe strata.
Avoid croking asciidoc with a mysterious bad-style message.
Regularize some markup.
Use correct plus-or-minus.
HOWTO improvements.
Add "Choice of Hardware".
Minor corrections.
Minor HOWTO additions and corrections.
Typo fix.
Address Savannah tracker issue #40297: AIS type 19 decoding bug.
Address Savannah bug #40298: AIS type 27 too strict size check
Check various test sentences against the Maritec decoder.
Thread-lock the code for writing JSON to clients...
Typo fix.
Remove a reference to a table now nonexistent.
Attempt to arrange reentrancy. All regression tests pass.
Catch the documentation up with Gary's change to PPS.
Reinsert the language about known latency in Sratum-00>Stratun-1
connections.
Chimers explained. Also, how time service is decentralized.
Add RFC-1305 as a reference.
Minor polishing.
Documentation, comment, and HOWTO updates.
GPSD assumes 1PPS before ins-stream report. Document this assumption.
Rework the HOWTO section on recommended hardware.
More references and accuracy figures.
More about accuracy, and some sources.
On providing public NTP service.
Minor polishing.
Add project links and explanation.
HOWTO polishing by HÃ¥kan Johansson <address@hidden>.
Document gpsdctl's return value.
Man page typo fix.
Start new section on NTP peerformance tuning.
More for the Time Service HOWTO.
More about strata and the NTP pool.
Heavy revision of the permissions stuff.
We've put to bed one set of HOWTOS.
Ignore (but log) "Secure SIS" channel C messages.
Address Savannah bug #40331: invalid empty AIS channel -> Warning instead
of error
Beat Bolli's stuff on performance tuning.
Acknowledge Beat Bolli.
eorganize a but of HOWTO for clarity.
Update the HOWTO todo list.
Fold in Beat Bolli's ntpoffset script.
Add Hal Murray's link and some clarifications.
Revert some strata language. More about using the pool.
"server" is overloaded.
Not just the USNO...
Changes suggested by Miroslav Lichtvar.
Clean up remaining HOWTO to-do items.
Correct an error message.
Time service issues with single-core systems.
Update the hardware database; add a regression test for the GR601-W.
Minor correction of number.
Time Service HOWTO polishing.
Polishing and a bit of rearrangement on the HOWTO.
No longer a draft; now nearly final and being polished.
Polishing changes. Good analogy for wobble from Sanjeev Gupta.
Time Service HOWTO is nearly ready to be shipped.
Remove a redundant test, and explain the redundancy.
Try for a better account of NTP time error.
Many polishing changes to the Time Service HOWTO, mostly from Sanjeev
Gupta.
The word "hopefully" is a kiss of death to decent writing.
First minor tweaks to setup instructions from the live test.
Move npshm.c into the libgpsd service library. All regression tests pass.
Add a clarifying comment.
starg.h is needed before maentioning va_list. (NetBSD port patch)
Remove ntpd-specific language from a section that applies also to chrony.
Replace a non-portable endianness check.
Remove an endianness assumption.
Eliminate a really annoying build glitch. All regression tests pass.
Cleanup fixes. There was a typo in the last change to the Zodiac driver.
Fix duplicative section title.
Build system symplifications fixing minor bugs exposed by 13.10.
Add contrib/ppscheck utility to test for 1PPS without a lot of code
complexity.
Build instructions from Hal Murray.
Begin the painful process of factoring the thread processing.
Attempting to disentangle the thread monitor from ntpd and chrony
reporting.
More factoring out of chrony code.
Last commit needed this, too.
Remove fossil duplicative section title.
Move some code and remove unneeded deoendency.
More code rearrangement. No logic changes.
Comment polishing.
Add watch for DSR.
Session clearing wasn't being done quite right. All regression tests
pass.
HOWTO markup fixes.
Polish ppscheck.
asciidoc markup fix.
Documentation typo fixes.
Another factoring step in ntpshm.c.
Unconditionally enable PPS at startup.
Eliminate shmTimePPS, no longer used.
A big step - ntpd/chrony-specific code becomes method ponters.
Now we can handle a PPS thread per session. Tested with live PPS and
works.
ntpshm.c splits up; thread management now lives in ppsthread.c
Separate policy from mechanism.
Cleanup. No logic changes.
Compile correctly with pps off.
Either a non-null session or context hook lets the PPS thread run.
Fix for a nasty typo.
Remove unused or redundant #includes.
splint cleanup. Regression tests pass and PPS live-tests correctly.
Fix the autobaud hang on SiRF-IIIs. See comment after line 1208.
Re-enable compilation with ntpshm=no.
Force pps=no if TIOCMIWAIT is not available.
Separate ntpd support from chrony support, step 1.
Fix a minor layerring violation. All regression tests pass.
All four combinations of ntpshm=[on|off] and chrony=[on|off] now build.
Revert "All four combinations of ntpshm=[on|off] and chrony=[on|off] now
build."
Revert "Separate ntpd support from chrony support, step 1."
Fix build when pps=off.
Better fix for pps=no build. All regression tests pass, PPS is live
PPS on a pty doesn't work.
Document good support for MTK3329 as reported by Andy Walls.
Typo fix.
Move NTP link activation from device stash time to device open time.
Eliminate the race in KPPS thread startup.
Contain the assumption that time hinting is done to NTPD.
Fix build error when HAVE_SYS_TIMEPPS_H is not available.
Update some instructions.
Add some (hidden) instructions.
Cease ignoring an obsolete file.
Fix hang bug in the SHM client code. Thanks to Michael Tatarinov.
Repair outdated comment.
Another whack at getting the endianness check right.
-I/usr/include/sys blows up on Gentoo and *BSD.
Trying to get the endian.h location right.
Clarify an obscurity.
gpsprof can collect timeplot data, but not yet plot it.
Generic replotting code.
gpsprof can now make plots of time deta from PPS.
Finish decoupling the PPS watcher code from NTP shared-segment management.
Yet more cohesion-busting.
Namespace cleanup. All regression tests pass, PPS is live.
Eliminate an enable flag that always gets set true.
Yet more decoupling.
Cleaned-up version of Beat's patch to use the PPS loop in gpsmon.
There is no longer any reason that > 1 PPS threads can't run.
Pretty up the PPS display in gpmon.
gpsmon display of PPS events is now a public feature.
GPSD is fully up on Mac OS 10.7 - regression tests pass, PPS is live.
This appears to nuke the carnivorous config bug.
Exile a grotty endianness check to the only pace it's used.
Add autoconf-like distclean...
Last commit was too soon, this revision actually works.
Cleanup.
Show PPS offset in the NMEA monitor.
Get the size of the UBX monitor window right.
Remove embarassing incorrect apostrophe.
Support reading PPS messages in libgps.
Fix up gpsprof's dime difference calculation and add a mtching C macro.
Use timedrift_t for all hook arguments.
Change two variable names for clarity.
Eliminate a variable simplifying data management.
A fine point from Andy Walls.
FakeTCP source don't work yet in gpsfake, but this is a step less broken.
Improve gpsfake debugging.
Restore "Transport: UDP" functionality to gpsfake.
Explanation from Hal Murray.
Another steo towards getting fake TCP sources to work.
TCP fake sources now work but can't be reused yet.
TCP fake sources are now fully working.
Create a regression test for the TCP fake source capability.
Clean up the new HOWTO section slightly.
Document some magic comments.
Note that asciidoc is needed for building the website.
Move some installation instructions to a better location.
Delay-Cookie feature in test loads almost works.
Remove redundant variable.
Savannah tracker bug #36409 is now reproducible inside gpsfake.
Both the SiRF-III live test and the bug #36409 pathological case work.
Typo fix.
Solve the driver-flapping problem. Comments after line 1211 explain it.
Add PPS offset display to the SiRF monitor.
Insert memory barriers where requested, since we actually have them.
More intelligent disabling of the progess vaton.
Memory barriers are required for safe NTPSHM.
Test-machinery improvements. All regression tests pass. PPS is live.
Test machinery improvements. raw-regress is now only partially broken.
Suppress JSON notifications to clients who haven't enabled them.
Fix a small but fatal typo.All regression tests pass.
A third wack at the recent autobaud hangs.
Partial splint cleanup. All regression tests pass. PPS is live.
More splint fixups.
Full splint cleanup. Partial cppcheck cleanup.
Use a clever test for big-endianness.
Make the barrier() definition fail quietly, as it did in 3.9.
More cppcheck cleanup. All regression tests pass. PPS is live.
Update for the existence of the barrier() function.
Clarifying name change. All regression test pass. PPS is live.
Correct an NMEA sentence name.
splint cleanup.
Show the repo version, if possible, when running scons check.
Terminological fix.
Terminological adjustment.
Unify time-difference computations. All regression tests pass, PPS is
live.
Partial cppcheck and Coverity cleanup.
cppcheck and Coverity cleanup (not yet complete).
Remove code which Coverity correctly calls out as statically dead.
Code is fully cppcheck, splint, and Coverity clean.
Revert "Use a clever test for big-endianness."
Name change.
Another name change.
Polish the JSON passthrough code, and add a regression test.
Another whack at in-compiler endianness setup.
Add -j option to striplog.
Regretfully, give up on the raw-regress test.
Update the client applications list.
Partial code for setting MTK pulse width to a good value.
Configure and explain pulse width black magic.
More clarification.
Better solution to the flapping problem.
Improve the UBX sentence mix in NMEA mode.
Typo fixes. All regression test pass. PPS is live.
Improved cleanup procedure - clears up webpages.
More work on the cleanup productions.
Add material about PPS to the gpsmon manual page.
Merge branch 'master' of git.sv.gnu.org:/srv/git/gpsd
The real packaging fix for RHEL.Centos. Previous commit was a finger
error.
Minor improvements for the UBX monitor.
uBlox -> u-blox; the company does not biicapitalize its name.
Clean up manual pages.
Show PPS edge (when the hook fires) at LOG_INF.
Typo fix.
Typo fixes.
Fix inexplicably inserted whitespace.
Documentation fixes.
Make gpsmon properly clear screen at startup.
Have gpsmon do the right thing for NMEA packets shipped by a sticky
driver.
Clear the NMEA monitor's sentince list when the driver type switches.
u-blox name cleanup. All regression tests pass, PPS is live.
PPS events now show in gpsmon runb through the daemon.
Teach the client library about ppsbar...
A regression test had to be rebuilt to track the ppsbar library change.
Explain that gpsmon doesn't autoprobe.
Eliminate spurious input character echoes in gpsmon.
Typo fix.
Make interpretation of the split24 flag propperly per-subscriber.
Update a comment with a caveat.
Cosmetic fix to screen sisplay.
More gpsmon display cleanup.
Add hardware entry for BU-353-S4 based on Gary's 2012 report.
Add a hardware entry for the EVK 6H.
Reports on delays needed in the test framework.
Typo fixes.
Fix the stutter problem with gpsmon input.
Fix up command echoing.
Catch an illegal-argument error thrown by Fedora Linux...
It would be helpful to get the exception class name right.
Trying to narrow down the serial-parameter error.
Stop trying to set serial parity in the test loads.
Don't slow regression tests down more than we have to.
Add a warning about sporadic build failures and the delay constants.
More field reports on test delays.
More results from mike_t.
A cleaner custom test for the endiannsss check.
Correct a delay figure.
gpsmon i command now works.
Cosmetic fix.
Deal more gracefully with a zero timeout.
Explain some code.
Fix a truly stupid u-blox driver bug. Make baud rate change effective.
Nuke short timeout on gpsctl sync, it was producing spurious failure
messages.
In DEVICES, don't display serial parameters for a device with speed 0.
Correctly mark deactivated devices, and do a better job of dumping them.
Polish the modified gpsmon timeout logic.
I don't know how the previous commit compiled...
Fix more build breakage. Partial splint cleanup.
Full splint cleanup.
We can now change baud rates on the GR601-W.
Documentation fix.
sys/timepps.h needs sys/time.h on NetBSD, perhaps FreeBSD
Fix PPS offset display bug by properly thread-locking some accesses.
SPLIT24 and PPSBAR flags belong in the Python interface, too.
Fix buggy enable-flag-to-JSON mapping in the Python interface.
Typo fix.
Fix pps=no build.
Virtuous code removal.
Slightly improved thread0locking in gpsmon.
Fix minor bugs in JSON packet passthrough.
Documentation and TODO updates. All regression tests pass, PPS is live.
splint cleanup.
Rename and rationalize driver table instances. No logic changes.
Reduce gpsd log verbosity a lot.
More consistent ise of LOG_PROG.
In gpsmon, more informative error messages.
Another error message upgrade.
Update the TODO list.
Address Savannah bug #40461: leapsecond.py and http proxies
Document the leapfetch=no workaround.
Documentation update.
Update build and packaging docs.
More hints for packagers.
Teach gpsctl to extract subtype information.
gpsctl ID now gets basic version info from u-blox receivers.
Address Savannah tracker bug #40335: Accept overlong AIS messages?
Relax some more AIS lebgth checks. All regression tests pass. PPS is
live.
Fix a misplaced degree symbol.
Explain a permissions issue better.
In libgps, incorrect defaults resulted in garbage values for empty fields.
In gpsctl, make client mode ID operation report the device subtype.
Typo fix.
No response from David Ludlow. Give up on the Windows port.
Clean up an obsolete comment and redundant #ifdef.
ppsbar -> pps.
Add -P option to gpspipe for including PPS drift JSON in NMEA or raw
dumps.
Change the PPS bar over to a JSON drift message.
Tweak SConstruct so 'scons install' does no builds after 'scons build'.
A to-do item just got done.
More information on MMSIs for the AIVDM document.
Add gittip link.
gpsmon refactoring step. Live-tests correct.
Rarrange gpsmon code for clarity. No logic changes.
More gpsmon refactoring.
More isolation of curses code.
More gpsmon refactoring.
More curses code isolation. All regression tests, PPS is live.
Non-ncurses mode of gpsmon (for diagnostics) is closer to working.
Cope better if something tries to strip binaries.
Code simplification.
In gpsmon, further decouple data handling from display.
In gpsmon, more simplification. All regression test pass, PPS is live.
Add a necessary 'volatile'.
In gpsmon, better signal protection.
Remove a stray (but harmless) unlock.
Don't condition out thread locks when PPS disabled, gpsmon will need them.
Partial pylint cleanup. All regression tests pass, xgps works.
More pylint cleanup in some build helpers. Build succeeds.
Insert required calls to Python Eceptions base class.
Ritualistic pacification of pylint. All regression tests pass, xgps runs.
Give up on pychecker. pylint is better, add it to the audit sequence.
Revert "Don't condition out thread locks when PPS disabled..."
More splint and pylint cleanups.
Update shipping metainformation.
Adapt to new shipper CLI.
-a mode now stops on input char, but needs some fiddling yet.
-a mode now works except for initial character echo.
And that fixes the echoing.
gpsmon logging improvements
Reduce noice when debugging PPS.
Introduce new LOG_CLIENT looging level.
Drastic simplification of gpsmon initialization sequence.
Fix kexer to be able to see JSON after NMEA.
PPS JSON messages are now visible in client-mode JSON.
Suppress path rewrites in one important special case.
PPs drift is now visible in client-mode JSON...
Fix up the pps=no build. All regression tests pass.
Try to ensure that threads don't keep running after gpsd is signalled.
Stomp a sign-extension bug.
Guarantee that PPS threads are reaped when GPSD is signaled.
Another whack at display flapping. Never reselect the monitor on PPS
packets.
Document gpsmon -a
PPS offset field now updates properly in gpsmon client mode.
Be explicit when PPS is not available because the build used pps=no.
Stall SiRF writes the minimum necessary to guarantee a 4-second cooldown.
spint annotation fixes. No logic changes.
Polish the Time Servuce HOWTO.
Cope better if asciidoc is absent.
dia is better and producing SVG than PNG.
We attempt to count ACKs to avoid pproblems on the SiRF IV.
Structure change went missing from my last commit.
Minor documentation updates.
Coverity annotations.
Ignore some new generated files.
Check in the SVG generated from the cycle diagram...
Alas, last commit wasn't enough. cycle.dia has to be invisible to scons.
Add a warning when a SiRF message send fails.
Add a caveat and a TO-DO.
A to-do got done.
Document the SiRF IV issue.
Add timeout on wait-for-ACK to SiRF driver.
Trivial fix for Hal Murray's 8N1N1 artifact.
Upgun the gpsmon prompt for Hal Murray's benefit.
Skip configuration when cleaning,
Remove an unused check.
Skip configuration when just emitting help.
Simplify the build. No logic changes.
Build simplification.
Code is now static-checker clean with splint, cppcheck, and Coverity.
Time Service HOWTO improvements - a quick start sequence.
Add FIXMEs to the Time Service HOWTO.
Resync the Python interface's list of various magic constants.
More about chrony vs. ntpd.
Typo fix.
Delete somebody's stray character.
Restore inadvertently deleted code.
Speculative patch (code conditioned out) to prevent buzzing on fast
devices.
Markup fix.
Documentation polishing.
Edit down the news to the hight points.
FIXMEs resolved thanks to mlichtvar.
Fail more gracefully on parallel builds.
Avoid DESTDIR pollution in the build paths.
Address Savannah sr #108406: Trimble Thunderbolt reporting wrong time.
Final pre-release changes to Time Service HOWTO.
Version bump for 3.10.
Fulup.ArFoll (1):
Added gps2udp a gpsdclient to dispatch AIS to AISHUB, MarineTraffic, ...
Gary E. Miller (59):
Not all systems use group dialout for tty devices. Gentoo, Fedora,
Whoops, fix a typo to previous fix.
Fix typos in help text
Don't gratuitously switch drivers, if we already
Make clear -n is not really optional for ntp/chrony usage and that
fix typo #definr -> #define
Fix compilation with no IPV6. Can imagine why anyone would
There is a race between opening /dev/pps0 and dropping root. Make it
Fix a bug where KPPS values were getting stepped on.
Rename a KPPS variable so all KPPS vars end in _kpps.
Add a sanity check for KPPS. KPPS in usb may return all zeros.
Fix missing semilcolon and a style nit. Build was broken.
Fix wrong comments. Build still broken.
More updates to the NTP howto.
pps_hook (PPS in JSON)now correctly reports the PPS time and local clock
ntpd now can handle nSec. So all PPS hangling moving from uSec to
More tweaks to the time keeping howto.
Whoops, I patched gpsd.h, not gpsd.h-tail. Now fixed.
Flesh out the time howto with a full ntpd.conf example.
Add more detail to what expected ntpq data should look like.
Leave root permissions for a tad long to be sure KPPS initialized.
Tweak PPS reporting a bit.
Missing step to getting contrib/ntpoffset to run.
Add a FIXME of a potential, never reported, problem.
Tweak the description of ntpd pools. Thanks to Beat Bolli.
Tweak Beat Bolli's sample output per his requesst.
2nd try at Beat Bolli's patch. I got confused between minus from diff
Tweaks to the time howto. Added more chrony detail.
Harlan Stenn <address@hidden> correctly point out this needs to reference
Flesh out the chrony part of the howto. More to do.
Add some more error logging to chrony socket opening
Anoter chrony socket error reporting tweak.
Another step at using nSec instead of uSec in ntpshm.c
Remove broken and unused ONCORE PPS code
Make chrony_send() and ntpshm_pps() take the same arguments in the
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/gpsd
Prevent double reporting of the same PPS second
Add new ntpd nSec fields to SHM interface
Fix a rounding problem in last patch.
Fix typo. Thanks to Mick Durkin <address@hidden>
Be a little more paranoid about shmTime.
Fix a PPS debug report nit.
Use the new timedrift_t to simplify PPS.
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/gpsd
Get some ntpshm coments to align with new reality
Fix a bad precision in a debug report
Stamp out timeval in gpsd_ppsmonitor()
Nit in the time howto.
C variables are case insensitive, but still nice to have matching case.
Whoops, fix ppsthread.c. Can't fit float nSec in a long.
Start to decode new SiRF IV messages.
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/gpsd
A bit more SiRF IV decoding.
Merge branch 'master' of ssh://git.sv.gnu.org/srv/git/gpsd
Collect all the SiRF settling time constants into one define.
Disable MID 64 in SiRF as it floods the stream at 4800
Document en-/dis-able SiRF subframe. remove redundancies.
More consistent use of SiRF settling time.
SiRF now recommends at least 57600 to use nav data.
Greg Troxel (2):
Fix endian detection on BSD.
Detect endianness on OS X.
Igor Socec (1):
Fixing a bug in the function writing a big endian float number.
Michael Tatarinov (13):
Some GPSes use the USB ACM driver (/dev/ttyACM*, major number 166) on
Linux.
The gpsd doesn't compile if PPS disabled.
Lost separator.
Cleanup after refactoring AIVDM driver.
In gpsdctl: handle the unknown action.
Fixed the choice of method in gpxlogger.
Use GCC built-in function for memory barrier.
ntpshm_pps() is gone.
Simplify the PPS hook.
Sets precision inside ntpshm_put().
Fix bug in array bounds and replace exit() with
exit(EXIT_{SUCCESS,FAILURE}).
Cleanup.
Remove more duplicative logging.
Mick Durkin (1):
Fix an incorrect byte offset.
Mike Frysinger (1):
fix typo in drivers define
Miroslav Lichvar (1):
Improve chrony examples in time howto.
Pavel Kirienko (1):
Minor correction to use UBX error estimstes.
Reinhard Arlt (35):
Allow sat report for pseudonmea and nmea2000
Add gps2udp to .gitignore.
Add support for 8devices usb2can module.
Groundbreaking for AIVDM messages in nmea mode.
Encode a few more AIVDM messages.
More work on AIVDM messages in nmea mode.
Add AIVDM message 24 Part A and Part B.
Pacify splint.
Decode AIS channel info from nmea2000 messages.
Merge branch 'master' of git://git.sv.gnu.org/gpsd
Improve decode_ais_channel_info().
Add missing mesages 21 and 27 to pseudoais.c.
Fixed ugly copy and paste error.
Pacify splint.
Allow more splint checks for ais_binary_encode().
Let gpsmon -n survive AIVDM messages.
Merge branch 'master' of git://git.sv.gnu.org/gpsd
Remove flicker when receiving AIVDM messages in -n mode.
Accept a few more NMEA2000 PGN's.
Merge branch 'master' of git://git.sv.gnu.org/gpsd
Comments for PGN methods.
Merge branch 'master' of git://git.sv.gnu.org/gpsd
A few cleanups.
Fix bug in (still unused) handling of addressed PGN's.
Add PGN 127258 skeleton GNSS Magnetic Variation
Merge branch 'master' of git://git.sv.gnu.org/gpsd
Add skeleton for PGN 129283 Cross Track Error, PGN 129284 Navigation
Data, and PGN 129285 Navigation - Route/WP Information.
Decode PGN 127250 and PGN 128267.
Fix crash due to ais->type24.part.
Merge branch 'master' of git://git.sv.gnu.org/gpsd
Honor split24 flag in ais msg24 part a.
Honor split24 flag in ais msg24 part b.
Add -2 flag to set split24 flag.
Isolate the access to split24 flag.
New AIS type24 reporting.
Sanjeev Gupta (11):
Fix Makefile, was targetting wrong file
Hal Murray's explanations of how ntpd drift works
Initial draft; how to estimate time1 offset
Typo fix only
Terminological fixes
Proof read only
Proof read only
Correct attribution email
Extensive review by Jaap Winius
Update SConstruct to build new docs added in 3.10
Fix typo in SConstruct
-----------------------------------------------------------------------
hooks/post-receive
--
GPSD
| [Prev in Thread] |
Current Thread |
[Next in Thread] |
- [gpsd-commit-watch] [SCM] GPSD annotated tag, release-3.10, created. release-3.10,
Eric S. Raymond <=