[Gpsd-commit-watch] [SCM] GPSD annotated tag, release-2.91, created. release-2.91

[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-2.91 has been created
        at  71c1f77e812c213275bda612a56dff6da8441d7a (tag)
   tagging  d9d8640d7afe84ef708c3aa2b2a6cd883df6a5af (commit)
  replaces  release-2.90
 tagged by  Eric S. Raymond
        on  Mon Mar 1 23:03:12 2010 -0500

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

Bernd Zeimetz (6):
      Fix a wrong test in gpsd.hotplug.wrapper.
      Make gpsd.rules work again. We need ATTRS, *not* ATTR.
      Document another model one of our udev rules matches on.
      Replace tabs by spaces where they were used for identation.
      Support Pharos GPS-500 in udev rules.
      Revert "Support Pharos GPS-500 in udev rules."

Chris Kuethe (3):
      pacify cppcheck
      Use the Vincenty formula rather than the Haversine formula for 
earth_distance.
      Regenerate regression tests after changing to the Vincenty formula.

Eric S. Raymond (251):
      Version bumped to 2.91dev.  Conmmits are open again.
      Update website documentation for the new release and new wire protocol.
      Fix a glitch in the capability database.
      Typo fixes.
      Update the list of known regression-test problems.
      Documentation polish.
      Typo fix.
      Add a new supported device, and a USB chip ID to the hotplug rules.
      Documentation fix.
      Add a test log for the CH-4711.
      Markup fix.
      Make Subversion ignore some generated Python.
      Update TODO.
      Ship a DEVICE notification on every driver change
      Add non-blocking check for data waiting from the daemon to the C and C++ 
APIs.
      Comment fix.  All regression tests pass.
      A gpsfake testload can now include magic comments inducing delays.
      Proper conditionalization in case CLIENTDEBUG_ENABLE is off.
      Introduce new argunent WATCH_DEVICE to the stream-option-control entry 
point...
      Refactor gpspipe to use gps_stream() and gps_open().
      Make use of -D uniform for enabling debug/progress messages, and document 
it.
      Extend -D convention to gpscat.
      Fix a truly obscure bug in the packet getter.
      Make a list of client project contacts in devtools.
      Add gpsdrive to the projects list.
      Update the transition plan.
      Notes about C++ and package systems.
      rtcm2_unpack() is no longer part of the core client library.
      Markup fix.
      Disable unnecessary logging, ix a typo.
      Cruft removal.
      Possible bug fix for gps_waiting().
      Cosmetic fix.
      DGPSIP and NTRIP device opens were being done in a bad way.
      Add the master list of NTRIP casters, we may use it later.
      gpsd can now accept as a command-line argument a URI like 
ais://data.aishub.com
      Merge Peter Stoyanov's description.
      Correction by mledford.
      Add a list of AIS feeds and caveats.
      Minor bug fixes for Python client binding.
      Fixed JSON dumping of AIS type 15.
      Fix up JSON dumping of the (rare) AIS message 23
      Fix decoding glitch in AIS type 22 and update docs.
      Give gpxlogger the -D option, abolish the last remnant of cgpxlogger.
      Implement line buffering in the client library...
      Change the client read logic in gps_poll()...
      Document the new async-read behavior.
      Add POLL_NONBLOCK flag to the C API.  Document all flags.
      Improved auditing of gps.h structure sizes.
      Refactor the Python poll() meethod.
      Make gpsfake.py use PACKET_SET properly.
      Remove unused method with a dangerous lockstep assumption.
      Dead code removal.
      Give up on interpreting Sager RTCM2 dump format in the Python poll() 
method.
      Better Python code for handling async writes.
      Documentation update; we can't get multiple responses per poll any more.
      Improve handling of buffered input from asynchronous writes.
      Fix a minor error in dumping of type 21 AIS messages.
      Begin a replacement for xgps using pygtk.
      Add toolbar to scratch xgps2 code.  It doesn't work yet.
      xgps need to run in rare mode, not raw which will freeze the display.
      Small step forward for xgps2.
      Checkpoint the evolving xgps2.
      Another xgps2 checkpoint.
      Successful setup of daemon momitoring in xgps2.  Verify by doinf xgps2 -D 
1
      Keep gps.py from accumulating satellites forever...
      Satellite list is live in xgps.  Now we need to make the font fixed-width.
      Raw-data display is live in xgps2.
      Initialize satelliste list as a ListView.
      Refactor satellite-view code for simpler display.
      xgps2 has a live skyview.
      Fix buggy used-flag display.
      Make xgps2 tootips work.
      Change to new gtk tooltip API.
      Preliminary layout of gps display fields in xgps2.
      Use a gtk.Table for GPS data layout in xgps2.
      First GPS data field goes live.
      Begin upgrading the gpslib stub to export C finctions needed by the
      Add another database entry and test log, this time for the Nokia LD-4W.
      Propagate 'interfaces' from the report CGI correctly.
      Use ATTR{}= rather than the now deprecated SYSFS{}=.
      Resolve linkage problems with extended gpslib module.
      No longer export leapsecond.py.
      Note that NTRIP seems to be fixed.
      Remove a TODO item, since NTRIP Is reported working again.
      Longitude and latitude are live in xgps2.
      First hack at supporting unit conversions in xgps2.
      Refactoring. All regression tests pass.
      en_US.UTF-8 (normal for Ubuntu) should be treated as en_US...
      Altitude, speed, and climb are all live in xgps2.
      Track is live in xgps2.
      Mode is live in xgps2.
      Error estimates are live in xgps2.
      Drawing on xgps skyview: proof of concept, not the graphics we actually 
want.
      Document our sources, set the line style properly.
      Plot both sky circles in xgps2.  Demystify some related code in xgps.
      Insert coordinate trigonometry in the required place in xgps2.
      Draw cardinal-point letters in xgps2.
      Code framework for satellite drawing in xgps2.
      Satellite icons in xgps2, no PRNs yet.
      PRNs are live in xgps2.
      Add -u option for changing display units in xgps2.
      Fix erroneous handling of satellite-used flags in test clients.
      UI polish.
      More instrumentation in gps.py.
      Put Python extension modules under "gps.".
      Replace the old Motif-based xgps with a new pygtk-based xgps (formerly 
xgps2).
      Uniform handling of unit-system defaults and -u in xgps, cgps, and lcdgps.
      Update the TODO.
      Attempt to correct the Makefile.am for the fact that gps.py has moved.
      Rename gpsclient extension module so it initializes properly.
      Add File and Units menu to xgps.
      View menu for xgps.
      Attempt to get xgps top-level window to resize dynamically when I hide 
widgets.
      Clean up after a name change.
      This might install the gps package correctly.
      Remove fulfilled items.
      Move the framework for faking a GPS so it's a submodule of the gps 
package.
      Found a kluge for making the top-level window elastic.
      Remove an obsolete caveat.
      Replace many show() method calls with one show_all()
      Remove a redundant rule.
      Fix some Python import problems.
      A cleaner way to make the window boundary elastic.
      Fix some bugs in scaled JSON output.
      Downgrade the SirF "Writing Error" message to a warning...
      Introduce the "scaled" attribute to JSON dumps...
      splint cleanup.  It caught a bug this time.
      Typo fix.
      Fix AIS dumping of altitude and speed in SAR report.
      Fully document the AIS dump format.
      Arrgh.  I confused ERR_SET with ERROR_SET in a couple of places.
      gpsfake.py now lives in the gps python package.
      Send a driver-change notification on both DEVICEID_SET *and* DEVICE_SET.
      More generous buffers to avoid generating bad JSON on AIS special strings.
      Typo fixes and polishing.
      Describe AIS messages 25 and 26, and fix some markup.
      Support for AIS type 25 and 26 messages.
      Add an AIS window (not yet updated).
      AIS window is hidden by default.
      AIS window now lives in a properly-sized ScrollWindow.
      First display of AIS data from Type 5 messages.
      Typo fixes.
      AIS storage managenent with timeouts.
      Add more message types to the AIS display.
      Various AIS corrections turned up by working on xgps's AIS display.
      More AIS tweaks.  Test client is turning up real bugs.
      Gary's last commit had a temporary debugging line in it.
      Fix sporadic client-side errors...
      Make the entry-aging logic in the AIS display actually work.
      Merge CNB information with Type 5 information.
      Include ship type information in AIS display.
      Automatically turn on AIS display when we have an AIS source.
      Make AIS display honor the selected degree format.
      Suppress a pygtk warning.
      Monitor AIS type 4 messages.
      Watch AIS message types 12 and 14.
      Mere in data from type 18 messages.
      Documentation update.
      Typo fix.
      Typo fix and some bug diagnosis.
      More bug diagnosis.
      More unfo in type 21 display.
      Arrange for setup.py to get called on installation as it should.
      Yet another attempt to beat the autotools mess into submission.
      Bump major API version.  Seems I managed to forget to do this since 2.39.
      Get rid of u_char, and some splint cleanup.
      Declare socket-valued ints as socket_t.
      Add another regression test, for the BT-451.
      Fix a mismatch between gppscap.py and the report CGI.
      David Ludlow's housekeeping and type-cleanup patch.
      >From David Ludlow: replace read(2)/write(2) on sockets with 
send(2)/recv(2).
      Type and inclusion cleanup.
      More cleanups from David Ludlow.  All regression tests pass.
      One more typedef cleanup from David Ludlow.
      Avoid a collision with typedef boolean.
      Cope better with decade and century transitions while the daemon is 
running.
      Add a warning about Gentoo systems and sys.path.
      Improved Python installation.
      Why GPSD can't do AGPS.
      Typo fix.
      Internal IPv6 support, derived from a support patch by Olivier Mehani.
      splint cleanup of IPv6 support.
      pygtk pango layout objects can indeed report their size.
      Add a FIXME comment about a known problem with IPv6.
      Improved client argument parsing to recognize IPv6 addresses...
      IPv6 support is working.
      Document support for IPV6.
      Document IPv6 support.
      Typo fix.
      Document why we use AF_INET and AF_INET6.
      Na,espace cleanup.  All regression tests pass, codebase splints clean.
      Small IPv6 cleanups. All regression tests pass.
      Corrections to GPS 76 database entry.
      Typo fix.
      Rescue the commented-out material on minimum cycle times...
      Old protocol support is removed from the daemon.
      Python xgps is ready for production, so bump version to 1.0
      Support for message type 25 in the Python decoder (untested).
      Update list of tested message types.
      Added another gpsd client.
      And yet another GPSD client.
      Detect corrupted AIS messages of type 25.  All regression tests pass.
      Typo fix.
      More message-length chexks for AIS.
      AIS message length checks throw out valid data...
      Refactor so we can control raw output better.
      We can now filter by type.
      We now have a test for AIS Type 13 decoding based on live and verified 
data
      Recover from random exceptions.
      Error-handling tweak.
      Cope with interrupts more gracefully.
      Handle and document some out-of-band AIS field values.
      Filename change and some typo tweaks.
      Added another AIS feed.
      AIS documentation tweaks.
      Improved handling and documentation of AIS type 22 messages.
      Typo fixes.
      Add AIS type 23 regression test.
      Regression test and proper field conditionalization for AIS Type 16
      Add -h option to make histograms.
      Documentation improvements.
      Correct handling of pad field.
      Checksum verification for the Python AIS decoder.
      Harden the Python AIS decoder.
      Correction and test case for two-MMSI case of message 16.
      Added packet-length check to Python decoder.
      Make the decoders use the pad fields as seems to be intended.
      Unbreak the build.
      Typo fix.
      US has stricter AIS regulations than the IMO.
      Cite the SOLAS regulations correctly.
      Add packet length checks to the C decoder
      Add an AIS Type 17 regression test.
      Python decoder shouldn't normally dump bogons.
      Fix some DTD references.
      A tiny code cleanup.
      A verified pair of AIS type 6 and type 7 messages for the regression test.
      Legend corrections.
      Clarify the status of a number of test sentences.
      More documentation of AIVDM test loads.
      More test load documentation. All regression tests pass.
      Integrated cppcheck testing, and fixed some minor bugs found by it.
      Link to Wikipedia page on WGS84.
      Documentation polishing.
      The numeric instability in earth_distance() has been fixed.
      Eliminate Python build dependency.
      Crucial typo fix.
      Pre-release splint and cppcheck cleanup. All regression tests pass.
      Version bump for release.
      Version bump.  Commits are open again.

Gary E. Miller (36):
      Tweaks to improve Oncore GPS PPS performance.
      Add a sanity test and some logging...
      Fix typos in last patch.
      Comment conditions for the current OncoreUT+ NTP fudge.
      Fix comment typo.
      Tighten up checking for 0.5Hz 'PPS' signals to prevent falsing.
      Fix off by one bug in PPS thread.
      Note unused SiRF messages in the logs as they are received.
      Tell SiRF not to send Message ID 30 which we never decode anyway.
      More SiRF logging tweaks.  Tweaking of messages to send not working.
      2nd try at fixing a nit in PPS.
      Remove unneccessary root test.
      WHoops, make last patch compile.
      Remove another bogus root test in the ntpshm path.
      Try to be smarter about SHM permissions when starting as non-root.
      Add a good comment that gpsd CAN talk to ntpd as non-root.
      Fix typo in new IPv6 handling.  -G now works in gpsd.
      Now -G (IPv6 global listen) really does work.
      Tweak the Gentoo/PYTHONPATH warning.
      Patch to allow gpsd to reliably work with ntpd when gpsd is not run as 
root.
      Once every few days leap_seconds get set to zero on SiRF.
      Document more SiRF message types.
      More tweaks to document SiRF message IDs in a consistent manner.
      All SiRF Message IDs now prefixed with 'MID ' in comments and logging.
      One space too far.  This fixes the SiRF regressions which just broke.
      Garmin GPS18/USB does not provide PPS.
      Add the MR-350P to list of working GPS.
      Fix 2 more broken DOCTYPEs
      For some reason gpscap.py rejects my MD-350P entry.
      gpscap.py silently fails if no 'type ='.
      Tweak the www upload documentation.
      Tweaks to the Garmin GPS-18 descriptions.
      Doc tweaks on PPS from Hakan Johansson and Matthew Dunn.
      Add a bit more shm debug info in the comments
      A bit more shared memory debug info in the comments.
      Garmin USB is binary only, thus linking it to the Garmin Serial Binary

Greg Troxel (8):
      trig test - not hooked into regress
      speling
      Include tests and regress-driver in the distribution, enabling 'make 
check'.
      Fix description of bu303-nofix, and add details.  This is not about
      Add todo item with link to GeoClue...
      Rewrite comment explaining PF_INET/AF_INET (history, standards). No code 
change.
      Remove inexplicable (and unexplained) bad code.
      Use %zd to print size_t value rather than assuming size_t is long.

Michael Ledford (2):
      Fixes a problem on Mac OS X 10.6 with the shared library linking.
      Fixes up previous commit that fixed logging function linkage

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


hooks/post-receive
-- 
GPSD