[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[gpsd-commit-watch] [SCM] GPSD annotated tag, release-2.96, created. rel

From: Eric S. Raymond
Subject: [gpsd-commit-watch] [SCM] GPSD annotated tag, release-2.96, created. release-2.96
Date: Mon, 03 Oct 2011 11:35:59 +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-2.96 has been created
        at  cfdd87cdf7f0dc5ae38060089c89abef6740824f (tag)
   tagging  abb336b69181ffb6b5d1f5087fd902c701d09c09 (commit)
  replaces  release-2.95
 tagged by  Eric S. Raymond
        on  Mon Mar 21 22:20:53 2011 -0400

- Log -----------------------------------------------------------------
Tagged for external release 2.96
Version: GnuPG v1.4.10 (GNU/Linux)


Andre Naujoks (5):
      Bug fixes and improvements for the NTRIP code.
      Simplify open_device(), eliminating an onnneeded table lookup.
      An additional condition for pass_rtcm to be build in drivers.c.
      initial device recognition is only done for non ntrip connections
      With this minor patch, working NTRIP is confirmed.

Beat Bolli (9):
      Pretty-print the nearest timestamp.
      Various small fixes to Python client mode.
      Allows NTRIP connection to ntrip://
      Add a Python port of Jason Hecker's
      Add Python output to
      gpsd.php: eliminate trailing whitespace use the HTML5 <canvas> instead of SVG
      gpsd.php: also show the current skyview in a table
      Don't let cgps leave the terminal in a bad state on timeout.

Bernd Zeimetz (33):
      Create pkg_config and prl files for Qgpsmm. Optimize libdir handling.
      Fix a few copy and paste errors. Nothing critical luckily.
      libgpsd needs to be linked with $(LIBUSB_LIBS)
      Spelling fix.
      Let the testregress target depend on geoid-regress.
      Do not wait for mountpoints in hotplug wrapper.
      Add regression tests for identation in Python files.
      Fixing several Python identation errors.
      Draw the sky view using Cairo instead of GDK
      Document usage of cairo.
      Add script to print Debian buildd regression test failures.
      Move to devtools.
      Be a bit more verbose and don't fail if a log was not available.
      Fixing a segfault in cgps.
      Add a wrapper around dchroot on Debian porter machines.
      libtool does not necessarily work with /bin/sh.
      Better clean for the various Python leftovers.
      Add missing linking for libbluetooth.
      Remove the libgps test binary during make clean.
      Documentain devtools/dchroot-exec.
      Spelling mistake fix.
      xgps/xgpsspeed: Check for a usable display before writing to it.
      Very very ugly workaround to depend ensure that gpsd.h exists.
      Fix out-of-src-directory builds for libQgpsmm.
      Fix out-of-src-directory builds for python.
      Party reverting the Python related changes from
      Check version after checking that Makkefile is present in
      Better workaroudn for autotools uglyness.
      Use distutils only, not setuptools.
      Add check_programs target to build test_*.
      Fix lat commit and make it portable thanks to autofoo.

Charles Curley (4):
      Expanded and improved the Troubleshooting Guide.
      Extensive polishing for the website.
      Minor tweaks for the libgps man page.
      Telnet addition to troubleshooting page

Chris Kuethe (47):
      this extra ]} is no longer needed
      fix the way skyview json is produced
      make the cgi point to the right hardware page
      Merge branch 'master' of ssh://
      make this work after esr's changes
      Some BU-353 receivers do support WAAS
      Somehow the geostar bit got lost
      over-zealous header removal
      Put back missing headers that break the build
      sys/types.h doesn't belong here...
      more header fixin'
      pacify cpp
      Subframe decoder will need PRN
      update inbufptr when skipping NMEA junk
      Wow, I totally baked the geostar import...
      Merge branch 'master' of ssh://
      missing headers for openbsd
      deheader broke the build again
      netinet/in.h goes before arpa/inet.h
      disable detection of comment packets
      let character devices default to RS232
      More fixes for the packet sniffer
      sizeof(long) doesn't always equal 4
      sizeof(long) is not portable
      use doublequotes when interpolating
      use select(2) rather than blindly reading fd
      explicitly initialize variable
      extraneous apostrophe
      Merge branch 'master' of ssh://
      scale M0 (mean anomaly) to match published almanac
      allow this to compile with pcc
      pull in sys/wait.h for WEXITSTATUS
      compose ERD with a loop rather than big snprintf
      compose sv health with loops rather than snprintf
      loop here too
      Merge branch 'master' of ssh://
      Revert "Try a tighter locking regime for serial devices."
      fix array index
      the subframe arrays are 1-index
      comment here too
      fix build on openbsd
      Add FAQ about whether unlisted receivers work
      unused variable

David Favro (2):
      Bring gps_read()'s return-code in sync with its manual entry.
      Ported libgpsmm to major-version 5.

Eric S. Raymond (673):
      A slightly streamlined release procedure using the git tools.
      Bump maximum channel count to 32.
      Add the GeoS-!M (GLONASS, with 24 channels).
      With the higher channel ceiling, GeoS-1M no longer shows spurious errors.
      First round of changes for API major version 5:
      Move the buffer and waiting members from privdata to gpsdata, since
      splint cleanup.
      Grab some headroom for future expansion.
      Correct a vendor website.
      Carry through and document the poll() -> read() change.
      Fix and update the news.
      NMEA moves its subpages a lot, don't bother tring to track this.
      -f was replaced by -o, follow through.
      Add product-databse entry and regression test for Garmin 10x.
      Clean up the CGI a bit, factoring out host information.
      Switch TCP socket sources to nonblocking mode.
      Add regression test and product database info for Garmin Foretrex 201.
      Add -l option to keep logfiles.
      Add product database entry and regression test for uBlox-lea-5h.
      Enable much stricter compilation warnings
      splint cleanup.  All regression tests pass.
      Update to eflect current conditions.
      Typo fix.
      Most of the flocktest hosts are unusable for various reasons, alas.
      Accept AIVDO as well as AIVDM packets.
      Merge branch 'master' of ssh://
      More qualifications about the (possibly mythical) powerdown-on-close 
      Add a caveat about power management.
      Merge branch 'master' of ssh://
      Add an identity tag filter so tags won't vanish after the filter-branch 
      Merge branch 'master' of ssh://
      A production for deheader scans.
      Document a minor problem with AIVDO sentences and rebuild its regression 
      Typo fix.
      Unused-header removals by deheader. All regression tests pass.
      Minor code cleanup, banishing warnings.
      More deheaderizing.  All regression tests pass.
      Try to prevent deheader passes from doing mischief.
      Address bug #17510 gpsdecode reports AIVDM Type 9 aircraft height,speed 
      Address Berlios tracker bug #17485: problem with isotime in
      Use isfinite() rather than obsoilete finite().
      Resolve bug #17646: faq.html#bug-reporting does not mention the bug 
      Try to fix Gary's build with PPS.
      Address Berlios tracker bug #17638:
      Attempt to address #17564: gpspipe terminates because of non-blocking 
      Clean up the code base and add switches so deheader runs clean.
      Document the deheader production.
      Merge branch 'master' of ssh://
      Some ships in US waters transmit 8-bit MMSIs.
      Typo fix.
      Typo fix.
      Merge branch 'master' of ssh://
      Try to fix the script failure reported by Bruce Dale on 25 Sep.
      Partial splint cleanup. All regression tests pass.
      Old position odf struct sample was giving splint indigestion.
      More partial splint cleanup.
      Code splints clean except for endiabbess stuff in driver_garmin.c.
      Merge branch 'master' of ssh://
      Merge branch 'master' of ssh://
      Duplicate-inclusion and portability cleanup.  All regression tests pass.
      Merge branch 'master' of ssh://
      SuS conformance.
      Simplification through taking SuS seriously.
      For simplification on SuS assumptions.
      More header cleanup.
      More header simplification.  All regression tests pass.
      Header cleanup.
      More header rationalizations,
      Taking SuS seriously some more.
      Because we're taking SuS seriously we can simplify the configuration 
      More configure simplification.
      Still more simplification of configuration and header handling.
      We were checking for sys/modem.h and never needed it.
      Remove another unneeded header.
      More header cleanup.
      Remove second include of <sys/socket.h>.
      Another pass at removing unneeded headers.
      Typo fixes.
      Temporatily restore a header check removed in last commit.
      The difference between curses.h and ncurses shouldn't matter to gpsd.
      Not having setlocale(3) is so ten years ago...
      Guard some headers fropm deheader.
      Better approach to guarding some includes.
      Replace a FIONREAD call with select(2).
      Guarantee ~ expansion.
      Use our own macros for interpreting little-endian 16- and 32-byte 
      We can't use the Python -B option, it's not back portable to 2.5.
      Partial splint cleanup.  All regression tests pass.
      Partial splint cleanup. All regression tests pass.
      Lift some buried #includes.  All regression tests pass.
      Remove duplicate header.
      Splint cleanup.
      Trying to isolate a splint error.  No code changes.
      Work around a splint bug.
      Total splint cleanup.  All regressionm tests pass.
      cgps now handles resize signals.
      Something's still wonky in the Geostar driver.  Bulletproof xgps against 
      Typo fix.
      Fix for generated markup.
      NEWS update.
      Typo fix.
      Note some week-reollover problems and add a link to a GPS calendar.
      Add info to the Hacking Guyide on the GPS Week rollover problem.
      Add regression-terts for GP-330-FW illustrating the GPS week wraparound.
      Remove some unnecessary and broken code.
      Suppress spurious offline messages.
      Add explanatory comment.
      Version-stamp the SHA sum file.
      Noise reduction.
      Try to bulletproof xgps against missing satellite data.
      Refactor to turf all uses of gpsd_report() out of netlib.c
      Abolish gpsd_report.c.  No more hardwired logging from the client 
      Retore definition of TIOCMIWAIT so the PPS stuff will work.
      Factor out our implementation of daemon(3) and use it as a fallback...
      Typo fix.
      Merge in tthe -d and -f capabilties from Gleb Smirnoff's 
      Implemented Gleb Smirnoff's -m option for gpxlogger.
      Update to the new location of Diego Berge's Qt client.
      Apply Charles Curley's fix from <address@hidden>.
      Documentation update ansd typo fix.
      Attempt a better fix for libQgpsmm not linking properly.
      Document super-raw mode.
      Prevent JSON parsing from becoming confused by client locales other than 
      Remove a remnant reference to gpsd_report.c.
      Chase down and kill yet another remnant gpsd_report.c reference.
      Clients are no longer required to define a gpsd_report() hook.
      Use the century slot in the context structure.
      IERS has announced no leap second at end of December 2010.
      Add -i utility option to
      Use the system clock to get the current stury if its value is positive 
      Update context century from a greater ZDA century; handles century 
      Detect GPS week rollover in ZDA sentences.  All regression tests pass.
      More time computation in
      Last thing we need is two different GPS_EPOCH values.
      Clarify description of GPS epoch.
      Make the daemon startup time part of the context.
      Be more explicit about an assumption.  All regression tests pass.
      Insert GPS week rollover checks everywhere we rely on GPS time.
      Partial splint cleanup.
      Add a vendor abd engine type we're going to need shortly.
      Add a record for the RF Solutions GPS-320FW.
      Avoid generating Unicode into NMEA.html.
      Typo fixes and clarifications.
      Remove a few more unneeded headers.
      Use <time.h> a la SuS, rather than <sys/time.h>, for select(2) etc.
      Merge branch 'master' of ssh://
      Correct a previous commit changing <time.h> to <sys/time.h> eceywhere.
      Merge branch 'master' of ssh://
      Insert SuS headers required for portablity, as revealed by deheader.
      Remove a duplicated inclusion.
      Remove a -DBINARY_ENABLE that was screwing up deheader test compilations.
      More header portability auditing by deheader.
      More header portability audting with a new version of deheader.
      Be explicit about the type and signedness of subframe data.
      Reorder some declarations so as not to choke older compilers.
      Partial splint cleanup.
      More explicitness about type and signedness.
      Use bool where appropriate, this will be handy when we generate JSON.
      Experimentally revert ckuethe's change to not accept comment packets...
      Rebuild Geostar test as required by ckuethe's fix. All regression tests 
      Detect non-advancing GGA timestamps and latch mode to 1 when this happens.
      Typo fixes.
      Add a pointer to the porterboxes.
      Rename the bits.h macros to be explicit about operand length.
      Refactor so almanac info is unpacked into a struct.
      Merge branch 'master' of ssh://
      More subframe unpacking into structs.
      We don't need yet another copy of the daemon() code in gpspipe.c.
      Turf out yet another instance of the daemonize code.
      Gary Miller explained about CNAV, and I was enlightened.
      Merge branch 'master' of ssh://
      Document what's going on with the source types.
      Conform to current udev naming conventions.
      Slight improvement in ckuethe's advice.
      Merge branch 'master' of ssh://
      Polish and better document the udev/hotplug code.
      Typo fixes suggested by Charles Curley.
      Turn gps_query in gpsctl into a send/expect engine.
      gpsctl device interpretation is working.
      Message cleanup following gem's notification-level change.
      Correct the test for GPS week rollover.
      Fix a bug that caused gpsctl reporting on the wrong device.
      More clarification about sensor behavior.
      Partial splint cleanup.
      Decoded subframe data is now part of the session structure.
      Quite down the JSON parser a bit.
      Another whack at making gpsctl work reliably.
      Attack Gary Miller's mysterious hang.
      Try to make gpsctl -s work again.
      Another attempted fix for changing baud rates.
      Salutary warning to future hackers.
      Fix Berlios tracker bug #17813.
      Fix the start-of-session race condition.
      Merge branch 'master' of ssh://
      Trivial fix for a segfault problem.
      Merge branch 'master' of ssh://
      Magic number elimination.  All regression tests pass.
      Suppress a compiler warning.  All regression tests pass.
      Reporting framework for subframes.  All regression tests pass.
      Good catch by Michael Cook.
      Update on the GPS rollover problem.
      Add some instrumentation for Roger Oberholzer.
      Merge branch 'master' of ssh://
      Enable gpspipe to start subframe reporting.
      WATCH_SUBFRAMES should not imply WATCH_JSON.
      Merge branch 'master' of ssh://
      Typo fixes.
      Back out the part of Beat Bolli's change that would spin on EAGAIN.
      Partial splint cleanup.  All regression tests pass.
      Typo fix.
      More splint cleanup.
      Rename a function before some substantial changes.  All regression tests 
      Markup fix.
      Typo fix.
      Clean up superstar2 initialization.
      Move the Navcom wakeuop event to the identify event.
      Oncore doesn't need to be woken up.  Move that send to the identify event.
      Don't send NMEA subtype strings until at least one NMEA packet has been 
      Cut down the number of independent driver_switch calls.
      Trigger strings that aren't NMEA packets won't work.
      Minor cosmetic change to logging.  All regression tests pass.
      Stop chattering about something no longer an error condition.
      We longer send wakeup probes to USB devices.
      Oncore VP needs a wakeup.  Fortunately it's not USB.
      Correct a comment.
      Under Linux, use /proc to avoid opening serial devices already open.
      Block other processes from using our serial devices while we have them 
      XHTML and style improvements by Charles Curley.
      Remove subframesflag so they're omitted uncinditionally.
      Correct a help string.
      Rebuilt regression tests to include subframe data.
      Comment typo fix.
      Allow use of Python 2.7.
      Followup top Paul Fertser's device-hook patch.
      Explain the uses of source typing better.
      Avoid incorrect ctype.h usage.
      Stop JSON hex escape emission from adding an extraneous backslash.
      It really needs to be %04x in JSON escaping, otherwise we don't get fixed 
      Try a tighter locking regime for serial devices.
      Re-commit the new device locking, previously removed by a finger error.
      Merge branch 'master' of ssh://
      Comment typo fixes.
      Regularize some function names.
      Partial splint cleanup.  All regression tests pass.
      It seems sys/wait.h gives splint fits.  All regression tests pass.
      deheader cleanup.
      More mud wrestling with autotools.
      More deheader testing.
      More deheader testing.  All regression tests pass.
      Warning cleanups.
      More deheader testing.  All regressions pass, codebase splints clean.
      Yet more header munging.
      Prevent some warnings under OpenBSD.
      Eliminate RAW_ENABLE, it's no longer needed.
      Quiet down the OpenBSD build.
      New first FAQ entry on troubleshooting.
      Break out the Troubleshooting Guide into a new document for expansion.
      Expand the Troubleshooting Guide.
      Update the Bluetooth instructions.
      Undo a previous commit to unbreak the DBUS build.
      First step towards sanity check by leap second.
      We can now generate C code for leap-second checking by year.
      Detect week-counter overruns using the leap second.
      GPS week rollover check extends to the rollover equivalent of now.
      Narrow a variable scope to eliminate a warning.
      Clean up warnings reported by Charles Curley.
      Refactor the rollover check.  All regression tests pass.
      Introduce GPS_TIME_VALID flag for GPS time validation.
      Rollover checking for NMEA.
      More refactoring of rollover checking.
      Refactor time resolution.  All regression tests pass.
      Make the Navcom driver use gpsd_resolve_time().
      Properly record when we get a valid leap second.
      Refactoring revealed a minor bug in Zodiac skyview timestamps.
      Cope with week counter rollovers while gpsd is running.
      Merge branch 'master' of ssh://
      Enable building even if we can't reach USNO leap-second history.
      Explanatory comments.
      Added information on GNS sentence.
      Remove the START_SUBFRAME optimization hack. All regression tests pass.
      splint cleanup.  All regression tests pass.
      Remove fossils related to START_SUBFRAME
      Comment and warning fixups.
      Keep leapcheck.i in the repo so we can rebuild without net.
      gemiller says the LEAP_SECONDS check is no longer needed...
      More magic-constant elimination.
      Default Evermore driver to NMEA mode so it gets leap-second correction.
      Stop making an assumption about the leap-second offset at startup.
      Update of release procedure.
      Update the release procedure.
      Refactor the leap-second fetcher.
      More refactoring, introduce stub for history graphing.
      Extract Unix times for leap-second history plot.
      Leap-second history plotting is *almost* right.
      Date in GNUPLOT ranges have to be quotes, it seems.
      Document how to get a leapsecond trend plot.
      Refactor leap-second fetching.
      Generate least-squares fit.
      Compute residuals.
      Generate coefficients for least-squares fit into the C code.
      Clean up the naming and data flow around generation of leapcheck.h.
      More cleanup of the machinery for wortking with leap-second history.
      Added a driver-type flags member to the driver method structure.
      Add explantory comments.
      Reduce dependency on the system clock.
      Unconditionall enable SiRF's PPS message, and watch for it.
      Go back to wiring in a default leap-second offset at start.
      Rebuilding of timebase constants is now completely automated.
      Further reduce dependencies onm the sytem clock.
      All the messy gpstime conversion stuff isin timebase.c now.
      splint cleanup.
      Document gpsd's assumption about the host system clock correctly.
      Standardize date format in all logfiles.  gpsfake will need this to set 
      More documentation of our assumptions about time sources.
      Update the history in the README.
      Refactor.  All regression tests pass.
      History update.
      Improve documentation comments.  All regression tests pass.
      Refactoring step.  All regression tests pass.
      Another refactoring step for the time logic. All regression tests pass.
      Add some suitable warning comments.
      Remove a dependency on the system clock.  All regression tests pass.
      Since we now disambiguate rollovers, remove a check no longer needed.
      Refactor time handling yet again.  All regression tests pass.
      Yet more time refactoring.  All regression tests pass.
      Compensate for century rollover in NMEA.
      Add a FAQ about NMEA 2000.
      Comment fixup.
      Reduce the amount of padding time in the regression tests.
      Make regression-test behavior independent of the century the daemon 
started in.
      Text polishing by Charles Curley.
      Be explicit about requiring C99.
      Update with Python example.
      The raw_hook mwember and set_raw_hook are gone; client API cleanup is 
      Fix typos pointed out in Berlios tracker bug #17858.
      Fix tracker bug #17832.
      Distinguish LOG_SHOUT from LOG_ERROR. Reduce verbosity.
      We no longer rely on long long to be 64 bits.
      Remove unneeded configuration checks.  All regression tests pass.
      Restore the ability to build under Python 2.4.
      Remove more stuff made unnecessary by ANSI/POSIX assumptions.
      More excision of unneeded configure tests.
      NAN is in the SuS <math.h>.
      More advice on Bluetooth and gpsd.
      Add required Cflags to *.pc files.
      Added a FAQ on interference with other USB devices.
      Update a manual page to reflect reality.
      Minor update.
      Reorganize stream argument flags, and create one for timing data.
      Add mozilla/firefox to our client list.
      Update an upload prduction.
      Fix tracker bug #17810: ?WATCH filter isn't work[ing].
      More sophisticated version extraction for SiRF. All regression tests pass.
      In the RPM packaging, allow udev to do its thing.
      Implement ?devices control-socket command for Bluetooth discovery.
      Splint cleanup. All regression tests pass.
      Most of the way to out-of-directory builds.
      Another step towards out-of-directory builds.
      Address bug #17887: Infinity gps device not recognized by udev.
      Minor gpsmon cleanup.
      splint cleanup.
      Build timebase.h correctly out of directory.
      Address bug #17532: gpsd 2.96~dev fails to build from separate directory.
      Address bug #13345: python module installed in wrong location on 64 bit 
      Typo fix.
      Merge updates from IMO 289, communicated by Kurt Schwehr.
      Fix minor bug in 'clean' mode.
      Catch and display regression errors from malformed JSON.
      Document the latest round of API changes.
      Timestamps in TPV and SKY go from float seconds since Unix epoch to 
      Comment improvement.
      Markup fixes and minor clarifications.
      More clarifications.
      First moves in the direction of supporting IMO 289 subtypes of AIS 
message 8.
      Describe some message 6 and 8 subtypes.
      Describe Tidal Window subtype of message 6.
      Document another message 8 subtype.
      Clarify the release history.
      Be explicit about message preambles.
      In pseudo-NMEA generation, emit ZDA at top of cycle.
      Cleanup after the NOISE change.  All regression tests pass.
      Change Noise sentence to use ISO8601 timestamp.
      Typo fix.
      More name changes related to GST.
      Make generated ZDAs report a UTC timezome, since UTC is what they emit.
      Some build chaing won't take "-I" without argument.
      Explain why we emit ZDA in pseudo-nmea without timezone.
      Remove a Subversionism.
      Make sure version is always present for
      Modernize the autotools initialization.
      Add a test log containing GPGST sentences.
      Remove a dependency on the leapseconds table being correct.
      Guard a potentially fatal call better.  All regression tests pass.
      Cosmetic fix.
      A working test for the C++ binding.
      Simplify the C++ lib, noo point in supporting the old API now.
      Rebuild the C++ interface to use RAII.
      Refactor some documentation.
      Update a capability table.
      Fix a bug in the LEA-5Q entry.
      Magic-number elimination.
      Be clearer about which quantities are not available.
      Fix a screen-layout multiplier.
      Exile noise statistics from main display of xgps.
      Remove a dependency on the number of rows.
      Tweak a field size.
      Correct an error pointed out by Danny Backx.
      Garbage-collect the client-side *SET flags.
      Compute satellites_visible in the client-side JSON parser.
      Partial splint and warning cleanup.  All regression tests pass.
      Fix a code-generation issue.
      Attempt to ensure that mask constant expressions won't sign-extend.
      splint cleanup
      splint simplification and cleanup.
      Cosmetic fix.
      Typo fix.
      Remove unneeded splint options.
      Portable version of kukabu's fix for the regression-test driver.
      Correct two wrong bit widths.
      Describe IMO236 Met/Hydro information message.
      Minor corrections.
      Add IMO236 Dangerous Cargo Indication.
      Typo fix.
      Straighten out the mess with Tidal Window variants.
      Reorder some descriptions.
      All IMO236 and IMO289 type 6 special messages are now described.
      Added Fairway Closed message.
      Add another message type.
      Field length correction.
      Fixed up VTS sentence.
      Add semantics descriptions for several message types.
      Correct some special type tables.
      Message length correction.
      Prevent xgps from fooing up on JSON with no device label.
      Add a gpsfake option to set the listening port.
      Typo fix.
      Add IMO289 met/hydro message, correct the IMO236 version.
      Add to the repo.
      Added Beaufort scale.
      Add tables for MARPOL codes.
      Correct some message numbers.
      Add IMO289 Extended ship static and voyage-related data (broadcast).
      Number of persons on board.
      Fix an arithmetic error.
      Add IMO289 Clearance Time to Enter Port.
      Add Marine Traffic Signal message.
      Add IMO289 Berthing Data message.
      Add Route Information (broadcast).
      Rebuild check logs to include genearated $GPALM.
      SUBFRAME JSON object gets a device tag.
      Data hiding.
      Bug fix: AIS JSON objects get a device field.
      Bug fix: RTCM2 JSON dumps get a device field.
      Documentation update.
      Splint cleanup.  All regression tests pass.
      Properly conditionalize gpsdecode in case AIDVM support is not compiled 
      Refactoring step, no logic changes.  All regression tests pass.
      Refactoring step.  All regression tests pass.
      More layer separation.  All regression tests pass.
      More layer separation.  All regression tests pass.
      Cosmetic tweak.
      Drastic simplification of gpsdecode using recent refactorings in libgpsd.
      splint cleanup.  All regression tests pass.
      gpsdecode can now dump NMEA and GPS binary data presented on standard 
      Reduce verbosity.  All regression tests pass.
      Improved building of the hardware page.
      Added Lyconsys GPSMOD1
      Cosmetic fixes.
      Typo fixes.
      Minor patches for AIVDM.txt from Kurt Schwehr.
      Improved error and progress logging.
      One variant of weather observation report.
      Check the WMO variant of the weather report.
      Bump MAXCHANNELS to 72 in order to avoid a future ABI break.
      Greg Troxel says CLOSE_DELAY needed to go to 0.2 on his sytem to avoid 
      Regression-log rebuild of two tests is required by the maximum-channel 
      Merge branch 'master' of ssh://
      Bulletproof a formatting function against C library variations.
      Make gpsmon's logfile option more useful.
      Merge branch 'master' of ssh://
      A log message needed to move.  No logic changes.
      splint cleanup.
      More splint cleanup.
      NTRIP fixes from Andre Naujoks.
      Typo fixes.
      Typo fixes.
      Move a warning about speeds < 38400 that triggered too often.
      Assume any 'satellites' attribute in JSON is a JSON array.
      More uniform error messages.  All regression tests pass.
      Simplify year computation in NMEA. There are no out-of-bound years.
      Fix the negative-date problem in gpsmon.
      Comment fix.
      Eliminate some spurious messages.
      Prevent emission of empty device tags.
      Somewhat more effective asciifying in the Python client.
      Typo fix.
      Add -t option to gpsmon.
      Log mode/speed/rate switches.
      Re-enable mode switching from gpsmon.
      Make gpsmon aware of the SHOUT log level.
      Fix some timeout, and explain why they're needed.
      Splint cleanup.
      Binary and NMEA-mode switching work again.
      strcpy() elimination.
      Remove a mo-op initializer from the driver type list.
      splint cleanup.
      Address Berlios tracker bug #17951 FTBFS with disabled nmea.
      Fully enable compilation withe NMEA_ENABLE off.
      Remove all support for RTCM2 Sager dump format.
      Fold the rtcm-104(5) manual page into the gpsd(8), now that it's 
      Documentation update.
      Documentation editing.
      Add and document gpsdecode -v option.
      Refactor device opening.  No logic changes. All regression tests pass.
      Vertical window size computation had to be redone to fix a display bug.
      Avoid any possibility of rejecting a $PASHR due to bad checksum.
      Guard probes on whether the device being activated is serial.
      Move NTRIP per-connection state from global context to private driver 
      Eliminate the dsock member in the context structure.
      Eliminate the context sentdgps member in favor of a per-session state 
      Eliminate the context->netgnss_service member.
      Add an explanatory comment.
      The NTRIP connection state musn't be zeroed on gpsd_activate().
      Abolish device->is_serial.  Replaced by device->servicetype.
      Make xgps indifferent to whether it gets float or string time.
      Exit without trailing debris on Ctl-C.
      Merge branch 'master' of ssh://
      Split an excessively long function. No logic changes. All regression 
tests pass.
      Minor bug fix: packet echo shouldn't be pinned to stdout.
      -j is supposed to be the default.
      Make the SiRF driver obey context.readonly.
      Faster implementation of strlcpy(), thanks to Michael Cook.
      Fix a bug in computation of satellites_visible.
      Fix an explanatory comment.
      Fix a problem noted by Charles Curley.
      Try to fix Greg Troxel's format mismatches.
      Eliminate the ntrip_state global.
      splint cleanup.
      Alas, PRI64 macros choke splint.
      Make the length of ISGPS packets accessible.
      Add instrumentation. All regression tests pass.
      RTCM[23] are single-packet formats, so they have reliable cycle detection.
      Make sure ISGPS packets are available raw in the output buffer.
      Reverse an earlier error, and harden RTCM2 decoding a bit.
      Optimized strlcat() to match the optimized strlcpy().
      Remove a probe for which we do not actually recognize any response 
      Don't pretent to have methods we have not implemented.
      Add Ande Naujoks's Ntrip sample to the test suite.
      Remove pass_rtcm(),
      Try to re-enable recognition of the MediaTek 33301.
      splint cleanup, includubng an error in the last commit caught by it.
      Update product database, adding the MTK-3329.
      Comment fixes.
      Hide something that doesn't have to be globally visible.
      Fix a bug in some mode-restoration code.  All regression tests pass.
      Verbosity reduction when context.readonly is off.
      Enable gpsdecode to dump subframe data.
      Abandon the attempt to de-Unicodify JSON as it comes in.
      More about the rollover problem.
      splint cleanup.
      Improve some progrss messages,
      Typo fix.
      Verbosity suppression. All regression tests pass.
      The old Sager dump function is gone.
      Inline a header that is only used once. No code changes.
      Implement and document RTCM2.3 message type 14. All regression tests pass.
      Implement and document RTCM 2.3 message type 13.
      Reserve space for a feature.
      News update.
      Reserve a flag bit.
      Typo fix.
      Simplify the encoding side of gpsdecode.
      splint and cosmetic cleanup.
      Replace the crappy old RTCM3 dumper with one that emits JSON.
      Simplify the RTCM3 analyzer.  All regression tests pass.
      Preprocessor abuse for a better world.  Or simpler code, anyway.
      More creative macro abuse.
      A step towards decoding RTCM3.
      RTCM3 test capTures.  Not ready to be part of the regulae regeressions 
      Suppress a spurious error message.  All regression tests pass.
      Tell the compiler that gos_errstr() ia an observer.
      Add explanatory comments.
      Have the Python packet-module interface export the character counter.
      Include a stream counter in the gpscat -t output.
      Documentation tweaks.
      Add a feature and a minor fix to the packet-sniffer tester.
      Sigh, seems a flush is required.
      Bug fixes for RTCM3 decoding.  All regression tests pass.
      Typo fix.
      Limit precision based on the units of measure.
      Code simplification.
      Code simplification.  All regression tests pass.
      Sniffer wasn't going to ground state properly afer an RTCM3 packet.
      Don't crap out on garbage data.
      Add RTCM3 regression-test load.
      Be more specific in two cites.
      News update.
      Make extra sure Ntrip correction state is correctly initialized.
      splint workaround.
      Merge branch 'master' of ssh://
      Dead-code removal.  All regression tests pass.
      Dead code removal. (This has been replaced by the -e option.)
      Remove a silly #ifdef.
      Simplify and regularize #ifdefs.  All regression tests pass.
      Fossil removal.
      More fossil #ifdef removal.  All regression tests pass.
      Regularize #ifdefs.  All regression tests pass.
      Revert "Fossil removal."
      Merge branch 'master' of ssh://
      Un-botch the #ifdef CLIMB removal.
      Android directions.
      Update documentation about gps_read() and gps_waiting().
      Typo fix.
      Give gps_waiting() a timeout argument. Use it systematically in test 
      POLL_NONBLOCK is no more.
      More POLL_NONBLOCK elimination and an example update.
      Documentation update.
      waiting method in the Python binding gets a timeout argument.
      Describe how to be carefuln about wait timeouts.
      Wait method needs an argument.
      Make the Qt implementation honor timeout.
      Reduce an unnecessary double to time_t.
      The recent change to use GST as a time check for end of cycle required 
      Rearrange and update option listings.
      Give timestamps their own typedef. No logic changes.
      More timestamp_t conversions.  No logic changes.
      Yet more double->timestamp_t moves.
      And still more double -> timestamp_t moves.
      Library time parsing is now backward-compatible with 2.95.
      Clear errno at an appropruaye place in gps_read().
      TODO and NEWS updated for release.
      Only send fixes via DBUS when you actually have sat lock.
      Build system simplification.
      Merge branch 'master' of ssh://
      Use $(LIBM) consistently ratcher than -lm.
      Repair 'make dist' so build from tarball works again.
      Remove outdated comment.
      Comment typo fix.
      More entries for the WMO observation-report table.
      Finished WMO weather-report variant.
      Specify coverage of the AIVDM document.
      Change a structure name to avoid a future API break.
      Add Globalsat ND-1005 to product database and regression tests.
      Small markup fixes.
      Some build chains *still* don't like -I wuthout an argument.
      issuedata -> iod in RTCM2
      rangerr -> prc, rangerate -> rrc in RTCM2
      Dump contents of of RTCM2.3 type 31 as a JSON object rather than hex 
      Documentation update.
      Minor splint fixes.
      Add examples to documentation.
      Take gps_poll() off the near-term agenda for the C binding.
      News update.
      More doc updates.
      An upstream-bug complaint can be nuked because libgps:gpsd_report() is no 
      Leave plenty of room for GGA timestamps with subsecond parts.
      Document that the command parser is indifferent to the order of JSON 
      Protocol documentation improvements.
      More dicumentation improvements.
      Revised section on time accuracy.
      The overlength-GGA fix requires rebuild of two tests.
      Improve an explanatory comment.
      Attribute typo fix, and bring the manual in sync.
      And fix the only other place "timestamp" was mentioned.  In a comment 
      Typo fix.
      Minor fixup to driver prototype example.
      In the client API, take the data buffer out of the gpsdata_t structure.
      Sync the Python and C++ bindings to the C one.  Document it properly.
      Banish compiler warnings in the Qt support.
      Moinor correction to previous commit; buffer must be static.
      Move a declaration so it doesn't follow executable code.
      Version bump and news update for 2.96 release.

Gary E. Miller (187):
      Embryonic support for linux kernel PPS.
      Use the proper include file for kernel PPS.
      Grr, get the right PPS header to use.
      Another baby step to kernel PPS.
      Finally have the right header and can get a valid PPS handle.
      Init kernel PPS.
      Kernel PPS compiles, mostly there, but it does not work yet.
      Can now init linux kernel PPS, but only by hardcoding a magic dev path.
      Now getting real useable data from linux kernel PPS.
      Linux kernel PPS now drives ntpd.
      Hide the kernel PPS code when timepps.h not found.
      Good news and bad news about kernel PPS.
      A note on how to find the magic path for the kernel PPS interface.
      Fix bad comment.  Add more work todo for kernel PPS
      Refactor kernel PPS a bit.
      Kernel PPS comment tweaks.
      First pass at reading /sys/class/pps to find /dev/pps? matching our 
      Reading the contents of /sys/class/pps/pps?/path now to find right 
      Final tweak for a clean read of /sys/class/pps/pps?/path
      More kernel_pps.  Fine tune the config for another required header.
      First cut at automatic edge detetion for kernel pps.
      Tweak the kernel PPS edge selection logic.
      Another small step at automatically finding the proper kernel pps device.
      The kernel PPS init now automatically finds the /dev/pps?
      Compute kernel PPS cycle and duration.
      Fix a small typo in last kernel pps change.
      First bit at adding chrony SOCK driver compatibility.
      Open the chrony socket for nSec resolution kernel pps.
      Now can write chrony SOCK style time.
      Add comments on chrony SOCK behavior
      Don't write to chrony if crony socket not open.
      The kernel PPS now seems to be talking to chrony SOCK driver.
      tv_usec != tv_nsec, make adjustment for chrony SOCK.
      GPZDA reported broken by aquarat, and others.
      Add regressions for HP (SYmmetricom) 58534A GPS.
      Add some more error logging for GPGSV.
      Fix endianess issues in Garmin binary driver.
      Another tweak to find the wandering endian.h
      Another attempt at getting portable endianness in Garmin driver.
      Yet another try at endianness portability.
      Yet another attempt and getting rid of endian.h stuff.
      Revert "Yet another attempt and getting rid of endian.h stuff."
      Revert "Yet another try at endianness portability."
      Revert "Another attempt at getting portable endianness in Garmin driver."
      Yet another try at portable endian macros.
      Fix the order of include in lcdgps.c
      Looks for timepps.h in the right place.
      Use a better path for chrony.sock
      Allow for multiple chrony sockets.
      Improve conversion of timespec to timeval
      More macro madness to clean up timeval/timespec handling.
      Cleanup timespec and timeval normalization code with macros.
      Move more time variables from timeval to timespec.
      A half step moving from gettimeofday() to clock_gettime()
      When PPS is built, try to use, but do not require, librt.
      Fix stupid error in timeval/timespec rounding macros.
      Improve comments on the 50bps data.
      Further cleanup of logging of 50bps data.
      Decode a bunch of 50bps data, output to the log.
      Decode a tad more 50bps data.
      Fix the formatting of some 50bps data messages.
      And yet anopther try at counting 50bps parameters.
      Remove endian.h, it was not portable.
      Remove splint override of potential errors that are not there.
      Tweak some PPS logging.
      Tweaks to PPS logging.
      Missing newline in PPS logging.
      More logging tweaks for PPS
      Enable SiRF subframe (50bps) data on USB as well as high speed serial.
      Enable SiRF subframe (50bps) data on USB as well as high speed serial.
      Enable SiRF subframe (50bps) data on USB as well as high speed serial.
      More logging tweaks to subframe decoder.
      The 64 bit header on all subframes is now decoded.
      TOW is really TOW17 (truncated TOW)
      Subframe 1 data is now completely decoded to log output.
      Merge branch 'master' of ssh://
      All of subframe 2 data now decoded to logs.
      Subframe 3 is now fully decoded to the logs.
      Tweak to the subframe log format.
      Almanac data in subframes 4 and 5 is now decoded to the logs.
      Subframe 5, page 25 is now decoded.
      Comment undocumented Subframe 5, page 51.
      Subframe 5 page 25 is sent as page 51.
      Wow, finally nail down the real subframe page numbers.
      Subframe 4, page 18 now decoded.
      Placeholders for Subfram 4, pages 13 and 25.
      Fix Subframe 4 page 17 log.
      Subframe 4, page 25 is now decoded.
      Subframe 4, page 13 is now decoded to the logs.
      Subframe almanac data is now properly unsigned/signed.
      Subframe 1 data is now properly unsigned/signed.
      Subframe 2 data is now properly unsigned/signed.
      Subframe 3 data is now properly unsigned/signed.
      Subframe 4 page 13 is not properly unsigned/signed
      Merge branch 'master' of ssh://
      Subframe 4, page 18 now properly unsigned/signed.
      With luck all the subframe data is properly signed or unsigned.
      unint2int() macro was backwards
      Revert SUB test for seeing if a SiRF is fast engough to send subframes.
      Log eccentricity (e) in human readable form.
      Subframe comment inprovement.
      50bps data comments, sqrtA is now in final units and verified.
      Omega dot now in proper units and commented.
      Scale and commment IODE and AODO in subframe data.
      Comment and scale af0, af1 and af2 in the subframe data.
      IODC and Cic commented and scaled in subframe data.
      Comment and scale Oega0 and omega in subframe data
      toa, Tgd, and M0 commmented and scaled in subframe data.
      The almanac data is now all commented and scaled
      Subframe 1 data all commented and scaled.
      Subframe 2 data is now commented and scaled */
      Change uint2int() macro so as not to offend ckuethe.   More comments.
      Subframe 3 data is all commented and scaled.
      Obfuscate and over-complicate the Subframe 3 code like esr wants.
      Document scale factors in subframes 1 and 2 to match subframe 3.
      More Subframe 4 page 13 into big-ass structure, document data.
      Merge branch 'master' of ssh://
      Structurize and document Subframe 4, page 25
      Structurize and document subframe 4, page 17.
      Subframe 4, page 18, structurized, commented and scaled.
      Subframe 5, page 25, structurized, commented and scaled.
      Move the last data items into subframe structure with comments
      Fix preamble data size in subframe.
      Warn users if their SiRF link is too slow.
      Add generic warning on slow link speed.
      Move the speed warning so each autobaud test does not give the warning.
      Add errlevel to gpsd_report() output.
      Remove some old cruft that libusb replaces.
      Fix globalsat URL.  Thanks to Charles Curley.
      Even if ALLOW_RECONFIGURE is not set still report SiRF firmware version.
      Allow reading of SiRF nav parameters, even if ALLOW_RECONFIGURE not set.
      Tweak a SiRF log message
      Merge branch 'master' of ssh://
      Tweak a SiRF log message.
      Make gpsctl debug output look like gpsd debug output.
      A lot of subframes are invalid, only SUBFRAME_IS on good ones.
      Subframe 5 almanc part output to JSON.
      More changes to gpspipe for subframes, no joy, yet.
      Output some JSON for all frames.  Lot's more to come.
      Add TOW17 to JSON SUBFRAME
      Dump subframe 4 almanac to JSON too.
      Add two more significant digits to sqrtA.
      Merge branch 'master' of ssh://
      Subframes 1,2, and 3 (the ephemeris) now output to JSON.
      Handle Subframe 4, page 55, the system message to JSON
      Decode subframe 4, page 56 to JSON.
      Subframe 5, page 25 to JSON
      The last subframe decoded to JSON.
      Fix the JSON for subframe system message.
      Organize the Subframe JSON a bit for easy decoding.
      Add unscaled versions of subframe almanac and ephem3.
      Subframe EPHEM1 and EPHEM2 now have unscaled output too.
      All the subframe data now has scaled/unscaled.
      Fix JSON to properly escape.  Escape subframe system message.
      Merge branch 'master' of ssh://
      really fix JSON strings this time.
      Ooops, forgot to JSON Subframe 4, page 25.
      Fix subframe JSON to not emit integers with leading 0's
      Make more subframe JSON pretty and easy to decode.
      Fix JSON strings to handle non-ascii.
      Make the JSON subframe system message the proper size.
      Another tweak to JSON strings
      Yet another cut at JSON strings.
      Another JSON string tweak.  Leave enough room in the buffer.
      Don't truncate the escaped system message.
      Fix visibilize() for chars with the top bit set.
      Merge branch 'master' of ssh://
      Fix typo in comment.
      Merge branch 'master' of ssh://
      Fix a small, but fatal error in the HEALTH JSON.
      Sort the sat list in xgps, Thanks to Thomas Sprinkmeier
      Merge branch 'master' of ssh://
      Fix a loss of precision in KPPS,  By: Miroslav Lichvar
      Mention the WATCH_SCALED also applies to Subframe data.
      Document Subframe data in the gpsd man page.
      Revert "Remove a mo-op initializer from the driver type list."
      Revert the removal if GPS driver initializers.
      If misdetection is going to occur, better to fail to Garmin Serial
      Merge branch 'master' of ssh://
      Revert esr's breakage of chrony interface.
      Tweak a KPPS log message that is not really an error.
      Patch to allow uBlox to send time to ntpd/chrony
      Remove the root only interface to chronyd.  Document that interface.
      Whoops, get the chronyd socket file name right.
      address@hidden pointed out that ntpd SHM needs to be volatile.
      Silence compiler warning about inconsistent use of volatile in ntp SHM.
      Shut up a compiler warning on unused var if not defined USE_QT

Gleb Smirnoff (2):
      Enhance the earth-distance() function to compute initial and final 
      make earth_distance() calculate initial and final bearings.

Greg Troxel (9):
      Initialize gpsdata->raw_hook to avoid use-before-initalized.
      Find {,sys/}endian.h via autoconf.
      Add sys/socket.h, following POSIX.
      Pass MAKE variable to
      Link test_bits with bits.o instead of libgpsd.
      Use PRIx64 to print int64_t in hex.
      Use PRIx64 for uint64_t.
      Don't assume python scripts in source tree are directly runnable.
      Don't use  to run binaries (fix previous).

Haakan Johansson (4):
      Combine calls to ntpshm_pps. Use chrony socket if KPPS is in use or not.
      Allow oncore devices in postion hold mode to deliver PPS data to ntpshm.
      Take programmed oncore PPS offset into account. Name correction delay -> 
      Do not modify TRAIM on startup, instead poll it.

Jon Schlueter (77):
      gpsutil.c: fill_dop bug if sats < 4
      marking quark and nog as back online for the flocktest
      Add -h option to regress-driver
      Fixed warning in regress-driver script
      monitor_oncore.c: fix some old bugs and add survey to the status line
      monitor_ubx.c: convert usages of mvwprintw to display for consistancy
      Rework diagnostic in driver_aivdm to print usable info
      trailing whitespace cleanup
      trailing whitespace removal in driver_aivdm
      Spelling corrections in faq.html
      whitespace cleanup in regress-driver
      summarize which regression tests fail
      add m4 to the .gitignore list
      splint cleanup make context,devices, subscribers static scope
      whitespace cleanup in gpsd.c
      change gpsd_init to take a const char * device (name) param
      change add_device,find_device all take const char *
      notify_watchers take a const char * sentence
      update explan_gpsd with notes on open_device
      remove unneeded splint directives around open_device/add_device
      spelling error in comment for driver_nmea
      add return codes when parsing partial GSV packets
      Match debug levels passed into gpsd_hexdump_wrapper with gpsd_report() 
      NULL-guard device in open_device.
      Convert from int to natural size_t/ssize_t
      NULL-guard device_name in find_device() function call.
      warning hunting const char * in libgps_core.c
      Hide vim .swp tmp files
      make sure regress-driver calls gpsfake explictly
      fix broken build of lcdgps
      spelling fix in explan_serial.c.xml
      whitespace cleanup in gpsd.xml
      spellchecking and catching all of the same spelling mistakes in other 
      spellchecking in gpsctl.xml
      whitespace cleanup in gpsctl.xml
      spellchecking in gpsmon.xml
      whitespace cleanup in gpsmon.xml
      spellchecking in writing-a-driver.xml
      whitespace cleanup writing-a-driver.xml
      spellchecking in explan_packet.c.xml
      spellcheck and cleanup related spelling issues
      whitespace cleanup explan_libgps_core
      spellchecking in explan_libgps_core
      whitespace cleanup gpspipe.xml
      spellchecking in gpscat.xml
      spellchecking in gpsfake.xml
      spellchecking in gpsprof.xml
      spellchecking in gps.xml
      whitespace cleanup in libgpsd.xml
      spellchecking in libgps.xml
      whitespace cleanup explanation.xml
      spellcheck explan_driver_nema.c.xml
      spellchecking in explan_geoid.c.xml
      spellchecking of explan_gpsd.c.xml
      spellchecking in explan_libgpsd_core.c.xml
      spellchecking in explan_netlib.c.xml
      spellchecking in explan_ntpshm.c.xml
      spellchecking in driver_tsip/geostar
      spellchecking in explan_serial.c.xml
      whitespace cleanup in faq.html
      spellchecking in faq.html
      Corrected bad reference to RS3232 in gpsprof
      spellchecking and bad tag fix in upstream-bugs.html
      spellchecking on bt.html
      whitespace cleanup on bt.html
      whitespace cleanup in client-howto.txt
      spellchecking in client-howto.txt
      spellchecking in excellence.html
      spellchecking gps-hacking.html
      spellchecking of gypsy.html
      spellchecking hall-of-shame.html
      spellchecking in protocol-evolution.txt
      spellchecking in protocol-transition.txt
      spellchecking in references.html
      spellchecking in troubleshooting.html
      expand logging for error condition of too many satellites reported

M Joonas Pihlaja (7):
      xgps: Fix setting colours by name.
      xgps: Use colour components in [0,1] rather than [0,255].
      xgps: Remove gratuitous saves and restores of cairo context.
      xgps: Grid fit coordinates to sharpen up lines.
      xgps: Unclutter horizon circle computations.
      xgps: Make sure everything in SkyView agrees on the grid fit horizon 
      xgps: Typos and comments.

Michael Tatarinov (9):
      Cosmetic improvements for SiRF monitor screen.
      More cosmetic tweaks for the SirF monitor.
      More autotools modernization.
      Cosmetic fixes for the SiRF monitor.
      In gpxlogger, filter for device on the server rather than client side,
      In gpxlogger, added <sat>, <hdop>, <vdop>, <pdop> elements.
      Some tweaks for the GPS monitor.
      Bugfixes in the GPS monitor.
      Generate NMEA GPALM in pseudo-NMEA when almanic information is available.

Miroslav Lichvar (3):
      Make the hotplug wrapper work when USBAUTO is not set.
      Don't crash in gpscat when started without arguments.
      Update gpscat usage text.

Nirgal Vourgère (1):
      Fixes SEGV on reception of short aivdm type 26 message

Paul Fertser (1):
      Add device activation/deactivation hook

Thomas Sprinkmeier (3):
      Process GPGST into noise statistics.  All regression tests pass.
      Typo fix.
      Enable GSTs to update time properly.

Timo Juhani Lindfors (3):
      Bug gpsd:17494 fix: cgps don't exit if interrupted
      Bug #17637: /dev/ttySAC1 is always opened in read-only mode
      Bug #17638: control socket & command can return OK even if write fails

Viktar Palstsiuk (1):
      Driver for the GeoS-1M GPS/Glonass module

Wolfram Sang (1):
      Two options needed to enabled once, now they need to be disabled.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]