[Gpsd-commit-watch] [SCM] GPSD annotated tag, release-2.39, created. release-2.39

[Eric S. Raymond]

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-2.39 has been created
        at  938f2edc475e17ad72c4a1883719108eedf3f1e8 (tag)
   tagging  9e792156fd6ea20982e9038ed57297d920686456 (commit)
  replaces  release-2.38
 tagged by  Eric S. Raymond
        on  Wed Mar 18 17:17:30 2009 -0400

- Log -----------------------------------------------------------------
Tag for public release.

Bernd Zeimetz (2):
      Moving and updating Debian's etc_init.d_gpsd.
      Updating packaging/etc_default_gpsd.

Chris Kuethe (37):
      superstar2.h needs to be shipped
      Evade a core-dump bug.
      Prevent null-pionter bug rather than just evading it.
      enable the superstar driver.
      make this a little more find-n-replace friendly.
      repository beautification, part 1:
      repository beautification, part 2:
      make sure the right headers are shipped.
      repository beautification, part 2:
      little bit of post-rename cleanup
      fix for esr's previous file rename commit.
      Hook up the framework for a few more monitor types
      make sure that the Id keyword is present and expanded
      fix unbounded strcpy
      add some more useful flag bits
      Parse the satellite status
      remove extra space that messes up formatting
      the decoders should be "static void"
      Until configure knows how to look for the right python-config...
      fix warnings about missing initializers and sign/size mismatches
      all this time, we've only been counting half the SVs
      approaching useful.
      fix another bug in the italk decoder.
      fix decoding of used satellites.
      The itrax03 is a 12-channel receiver. so say all the spec sheets...
      move the NAV_SVINFO display up, just like in the italk display
      a better way to determine how many satellites were used
      fix the formatting of the satellite list
      squelch warning about sign comparison
      Space Vehicle (SV) is not right here.
      I have no idea what crack-addled rhesus monkey wrote this crap but it's 
wrong.
      don't return a value from a void function
      shrink down the NAV_FIX pane
      fix some sign warnings
      more ubx monitor machinery.
      rate switching for ubx
      oops. forgot to actually use the cycle time.

Eric S. Raymond (309):
      2.39 development commences.
      Ship gpsd.rules in the tarball.
      Splint cleanup.
      Fixed potential core dump in C client handling of "K" responses.
      Typo fix.
      Fix tracker bug #13605, hotplug breakage due to changes in udev.
      Include instructions for setting up hotplugging in INSTALL.
      Require Python.h for Python support, since gpspacket.c uses it.
      Update test report for Garmin GPS-16.
      Address tracker bug #14196.
      Remove a nonexistent method and use an existing one.
      Gary Miller reports good PVT on the GPS-18.
      Fix some stuttering.
      Note that the Garmin 18 returns skyview information.
      Address tracker bug #14776 by fixing a logic error in the raw-mode 
dumping.
      Thinko fix.
      This typo fix means I should probably be asleep.
      Don't bother trying to switch back to the Earthmate driver from Zodiac...
      All driver method tables are now const.
      Restore the logic for treating baud rate of 0 as "no change".
      Email address correction.
      Major and minor API revision numbers need to be returned by L command.
      Add some documentation on how to cope with the upcoming API change.
      Markup fix.
      Suppress attempt to copy CGI.
      Stop using dodgy PUT_ORIGIN feature.
      Refactoring step - begin separating out SiRF-specific logic.
      More factoring out of SiRF stuff.
      More refactoring.
      More refactoring.
      Yet more refactoring to separate out SiRF-dependent stuff.
      SiRF-dependent methods have been separated.
      sirfmon method table proof of concept.
      More refactoring.
      Begin move to libgpsd_core-based device handling.
      Another step towards using libgpsd_core.
      Typo fix.
      Documentation updates anticipating the v4 protocol change.
      sirfmon now fully uses the libgpsd_core.c hunt logic and packet getter.
      Presently, no way to handle stopbits other than 1.
      Now we can handle odd parity and stopbits settings.
      Coalesce two methods.
      More careful error checking on window allocation.
      Slightly more rational window layout.
      Carry stopbits in from the session structure, and note a limitation.
      Typo fix.  Be more explicit about what the speed-changer is doing.
      sirfmon now switches speed using the driver method.
      Driver control_send method is now used for 's' command...
      Capability table now declares how much screen space it will need.
      Note bugs and fix typos.
      Give the ubx driver a real control_send method.
      Change some method names to a more uniform scheme.
      Make tsip_write take a session argument rather than an fd;
      Refactor tsip_write() to assemble the entire control string before 
shipping it.
      Refactor ubx_write() to use an assembly buffer and a single write.
      Define TEXTUAL_PACKET_TYPE().
      Name changes in preparation for send-buffer snooping.
      Splint cleanup.
      Remove a fossil relic.
      Format driver sends through a common message buffer.
      splint cleanup.
      Mode switch without termination now works.
      sirfmon now has a driver-table architecture, like the daemon.
      Remove dead code.
      Elminate probe field of monitor method table.
      Introduce wrap method to deallocate windows.
      Simplify the repaint interface.
      More method simplification.
      First hack at supporting multiple devices in the monitor.
      First nontrivial NMEA monitoring -- accumulate the sentence set.
      Don't rely on sNMEA tags being a fixed size.
      Try to detect start-of-cycle based on timing.
      Set scrolling-region size correctly.
      Use subwindows in gpsnmon to decomplicate the refresh logic.
      Use subwindows to simplify layout.
      gsirmon now handles multiple types.  Time to rename it.
      sirfmon -> gpsmon name change.
      Switching from SiRF to NMEA with n1 works.
      Implemented and documented gpsmon 'i' command.
      We no longer need a separate hunt stage.
      Add GSV display.
      Add GPRMC display.
      Set validity masks on GSA sentences a bit more carefully.
      Add GSA display.
      Add GGA monitoring.
      Make fixframe public, fix bug in RMC display.
      Enable gpsmon to get at the raw fields from an NMEA parse.
      Comment corrections.
      Our NMEA field parser was silently ignoring the last (*-terminated) field.
      Dump VDOP and PDOP correctly.
      More GGA fields.
      Not dropping last field (an SNR) means we need to fix the GSV correctness 
check.
      Move some display locations.
      Move the NMEA monitor driver to its own module.
      splint cleanup.
      A control-send method for the Zodiac driver (untested).
      Create monitor_proto.c as a prototype for monitor objects.
      More on writing monitor objects.
      Simplify the packet-monitor driver interface.
      splint cleanup.
      More cleanup and documentation for monitor stuff.
      More documentation and tweaks on the monitor infrastructure.
      Zodiac has a send_control method now.
      Some gpsmon enhancements.
      Document our embarassing bug.
      Hack the gpscon comand set to resemble gpsctl switches.
      Note that the gpsmon command set has changed.
      Alphabetize command set.
      Make command parsing a bit more forgiving.
      Reserve lower-case letters for gpsmon generic commands.
      Markup fix.
      Reorganize.
      Further in the direction of only reporting raw data.
      More moves towards completely raw data.
      Various small cleanups and sanity checks.
      Typo fix.
      Add 't' command to gpsmon to force the monitor type.
      Compile gpsmon and friends correctly when various GPS drivers are 
conditioned
      Implement -l option analogous to gpsctl -l.
      Typo fix.
      Be clear about which SiRF fields are derived.
      Implement and document support for more NMEA devices.
      Add a to-do list.
      Add internals documentation.
      Documentation improvements.
      Relabel to make clear that more data is raw.
      Add support for mode switching on Ashtech chips.
      Respond to Gypsy.
      Typo fix.
      Link to my blog rant on GPSes.
      Documentation tweaks.
      Magic-number elimination and compiler hushing.
      Knock the rust off the D-Bus code.
      Simplify the internal D-Bus interface.
      Refactor in preparation for merging cgpxlogger.
      More refactoring.
      Merge cgpxlogger into gpxlogger.  Still needs some code merging in
      Code fully merged...
      As good a merge as we'll get, barring shipping mode data over D-Bus.
      Standardized parsing of server:port:device client arguments.
      Support device selection in gpspipe.
      Standard argument parsing for gpsmon, and related documentation fixes.
      ocumentation and a splint cleanup.
      Tweaked to avoid an echo problem.
      Fixed disordered rev numbers.
      Garmin dropped NMEA in 2004.
      Add SGPS version to $GPVID.
      Out-of order version history again.
      Set readonly properly for gpsctl -e.
      More tweaks to SGPS spec.
      Sentence limit to 255.
      Refactor channel allocation policy.
      New channel allocation policy: seek fix quality.
      Added FAQ entry about the possible effects of weird serial parameters.
      Typo fix.
      Updated and moved the Mac bug to Open User-Visible again.
      Documentation update.  Better description of the architecture.
      Paste-o fix.
      FAQ and HTML fixes.
      Create infrastructure for allowing changes of parity and stop bits.
      More steps towards parity/stopbit setting.
      Defensive programming.
      Full support for setting parity and stop bits.
      Document the fact that parity and stopbit setting is now supported but 
dicy.
      Process mode arguments a bit more forgivingly.
      Fix error message.
      Update driver guide for the recent extension.
      Alter setup.py to match the renames.
      Consolidate all the X code in one file to reduce clutter.
      Change name of the LCDproc client, regularize its options, document it.
      Regularize some options.
      Split up libps.c so the code overhead for users' clients will be lower.
      Introduce gpsdclient.h header, not shipped.
      File rename.
      Improved bounds checking.
      Fix wonky resizing of packet window.
      packetwin can no longer be directly accessible.
      Monitor API has changed, so clean out comments in the three stub monitors.
      Add TSIP support for setting parity/stop bits.
      Documentation fixes.
      Show currently selected driver type even if there's no monitor for it.
      Correct previous commit.
      Handle two more RTCM3 messages.
      One more RTCM3 message.
      Mention the Ohloh page.
      Have a single point of truth about SiRF monitor commands, on the manpage.
      Improve gpsmon help and fix a typo.
      Fix a misunderstanding about RTCM3.
      Two more RTCM3 messages.
      Handle two more RTCM3 messages.
      Another RTCM3 message (1013).
      Quiet down the compiler.
      An explict cast to match the format.
      Add warning about gps=true being needed in Kismet.
      splint cleanup.
      More splint cleanup.
      Cleanup some 12-channel-limit assumptions.
      List missing option.
      SiRF finally got around to documenting packet 0x1b...
      Link to latest SiRf binary reference.
      Reorganized.  Many documents move to Google Files.
      Housekeeping.
      Remove two things that don't need to be in the repo,
      Trial superstar speed-setting and mode-setting code, for review and test.
      control_send method is different from a straight write; be clear about 
this.
      Link fixes.
      Use portable memset() rather than non-portable bzero().
      Yikes!  Change to -x and 'x'.
      EverMore gets a rate-switcher method.
      Support changing cycle time with gpctl and gpsmon, if the device supports 
it.
      Add an -l option to gpsd that lists its compiled-in drivers and exits.
      Well, that was a bit embarassing...remove a remnant of previous code.
      Simplify the way ALLOW_RECONFIGURE works.
      Support --disable-controlsend
      Improved configure options listing.
      Compiler warning and partial splint cleanup.
      Partial splint cleanup, and a fix for a real bug splint caught...
      This finishes cleanup of the splint type warnings...
      Some things splint just has to be told to shut up about.
      Other things can actually be annotated sanely.
      We're relying on the compiler not to evaluate function actuals in 
parallel...
      A command name changed.
      Framework for UBX rate switcher, everything except the field value 
calculation.
      Handle RTCM3 message 1014.
      Remove 'cycle' member from device structure.
      Housekeeping.
      More housekeeping.
      .cycle_chars is gone, replaced by min_cycle.
      Might fix an odd bug reported by Greg Troxel.
      Documentation comment tweak.
      Patch from Michael R. Davis.
      Add RPM-compatible init script from Michael R. Davis.
      On NMEA packets always insert \n, since the trailing \n gets hexdumped.
      New packaging/ directory.
      Asserts will allow us to test that bit widths are correct with a test log.
      Command name fix.
      Don't use hand-rolled hex-packing code when we have a shared function to 
do it.
      Improve error detection from gpsd_hexpack().
      Hush a compiler warning.
      Add X command to gpsmon.
      Typo fix.
      The shell of an AIVDM driver.
      Now that the driver structure includes its packet-type information...
      Document the simplification in the driver interface.
      More driver-interface simplification.
      Second stage of AIVDM decoding:
      First cut at describing AIVDM.
      Report Claude Lapointe's bug.
      Packaging changes from Michael R. Davis.
      We now have complete descriptions for AIS message-types 1-5.
      Minor corrections to AIVDM document.
      Minor corrections.
      Checkpoint before going to ubits-based decoding.
      All AIVDM type 1-3 fields except latitude and turn rate now decode 
correctly.
      Remove a comment that is now misleading.
      Dump Type 1 fields in the order they occur.
      Sign-extend negative long longs correctly.
      Typo fixes.
      Full decoding of AIVDM Type 1 packets.
      Full decoding for AIVDM Type 4 messages.
      Store unscaled AIDVM values so we never throw away information.
      Documentation fix.
      Turn off debugging.
      Proof of concept for AIVDM packet decoding in the rtcmdecode filter.
      rtcmdecode -> gpsdecode.
      More work on the AIVDM driver and spec.
      Document -l and -u options.
      Pass comments through.
      Correct checksum computation on AIVDM packets.
      More minor corrections to AIVDM.txt.
      Tweak some logic for verbosity control.
      Remove obsolete paragraph.
      Prevent spurious echo of data on bad packets.
      Dump ship type in text.
      Document length limit on bitfield processing.
      Remove a word-width dependency.
      Sixbit decoding was slightly off.
      Check in AIVDM regression tests for sentence types 1-4.
      Explain the Repeat Indicator.  Att Kurt Schwehr's types 25 and 26.
      Correct a make production.
      Back down to where all regression tests, including RTCM, still pass.
      Refactor the AIVDM driver to remove dependencies on the session structure.
      More refactoring of the AIVDM code.
      AIVDM tests for packet types 1-4 are stable.
      Correctly unscramble and document the wacky six-bit message format.
      Type 5 decoding is stable.
      Type 9 decoding is stable.
      Tests for Type 18 are stable.
      Fix infinite-looping configuariuon bug reported by Andrew Mileski.
      New version of AIDVM document.
      Get rid of stupid "smart quotes".
      libgps -u is now a unit test for unpack.
      Fix and unit test for the watcher+raw coredump.
      Enable passing of RTCM data to rtcm_writer() methods.
      Fix code for RTCM distribution.
      Rearrange a visible struct.
      Try fixing the RPM build.
      Parse, and describe, the RAIM bit.
      Typo fix.
      Bump the version, as we have significant new info.
      Fix the spec template to handle gpsdclient.h properly.
      Add mrd's RPM-building production.
      Regression machinery cleanup.
      Attempt to cope with locales where the decimal separator is not dot.
      Typo fix.
      splint cleanup.
      Note that Martin Schaller's bug is fixed.
      Typo fixes.
      Prepare to ship 2.39.

Greg Troxel (13):
      add plan for v6 support
      add pointer to overly complex configure.ac example
      Move gpsfix decl out of ifdef DBUS.
      Use %d to print difference of pointers, rather than %ld.
      Add manpages (not makefiles as I confusedly said in my mail) to clean 
target.
      Note that libxslt and docbook-xsl packages are needed to build man pages.
      Label the satellite column PRN instead of SV, following cgps and xgps.
      Remove leftover usleep of 100 ms.
      Spiff up error message about 'garmin_gps not active'...
      Change cycle_chars to min_cycle so my third-to-last revision compiles.
      Totally rewrite comment about what the driver is for.
      Add a couple of comments as I try to understand this; no code change.
      In dist-rpm target, use tab instead of 7 spaces.

Jeff Francis (1):
      Added man page for lcdgps.

-----------------------------------------------------------------------


hooks/post-receive
-- 
GPSD