[Gpsd-commit-watch] [SCM] GPSD annotated tag, release-1.96, created. release-1.96

[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-1.96 has been created
        at  5a23fe154f4645d76087e834d3a50f6b604a4c0a (tag)
   tagging  d6d5b87a253e8cd375a2e5a98f578863fcc6ddb1 (commit)
  replaces  release-1.90
 tagged by  Eric S. Raymond
        on  Tue Aug 31 03:24:16 2004 -0400

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

Derrick Brashear (1):
      Update Makefile.in such that install is a .phony target, avoiding 
problems with the INSTALL file on case insensitive filesystems

Eric S. Raymond (227):
      Document some limits.
      Oops -- have to make sure the display hook actually gets called.
      First steps towards watcher mode.
      Merge -n support back in.
      Code now consistently uses timestamp objects; the old cflags are gone.
      SEEN no longer depends on the clock.
      We now compute changed bits for all fields.
      Typo fix.
      Add machinery to the makefile to build tarballs and RPMs.
      Change the application resource file names to match modern X conventions.
      Add compass points to the satellite display.
      Sanity-check to throw out some malformed BU-303 sentences.
      Associate the right stamp with the right field!.
      Added Speed to display.
      Avoid possible Year 2100 bug.
      Some filename changes,
      Added T command.
      Change structures so we can set speed explicitly.
      Restore -s option.
      Make the programs a little quieter.
      Add track field to GPS display.
      Library rename.
      Oops, get the include right.
      Split new gps.h from gpsd.h.
      Hide some things that don't seed to be visible.
      Pull the output-data structure into the library namespace.
      Hide more stuff that doesn't need to be visible.
      Typo fix.
      More things are internals now.
      errexit no longer has to fit the public namespace.
      Cleanup of date-parsing code.
      Move the initialization to a more natural place.
      Follow through on name change.
      Don't free what was never malloced.
      Better error message.
      Garbage-collect fdset slots properly.
      Change some debug levels.
      More fd garbage collection.
      Separate speed and track stamps.
      Separate speed and track stamps.
      Take satellites-in-view pout of Q, since we get it from y.
      Oops -- back out the *right* satellite field.
      Proof of concept for client library.
      Make some fieldnames a bit clearer.
      Explanatory comment.
      PRWIZCH handling conditioned out.
      Name change before redesigning 'i'.
      Verbosity reduction.
      Add libgpsd.a to production machinery.
      Build library test tool.
      Hide another #define.
      Declare the public functions.
      libgpsd is now documented.
      Satellite query command is now working.
      Some cleanup before SVN conversion.
      Cleanup before SVN conversion.
      Trailing whitespace give gmake fits.
      Ship the correct include file in the RPM!
      Improve some error messages.
      Actually return -1 on no GPS input waiting.
      gps_poll() returns an int now.
      At high verbosity levels, meter data waiting from the GPS.
      Yet more error-message cleanup.
      Important change.  Always try to reopen GPS when it's offline just
      Now, we only exit on not being able to open the GPS in monitor (-n) mode.
      New 'X' command and 'online' member in gps_data struct allows client
      Followup on X command -- online member now has its own real version stamp.
      The file renames I wanted Subversion for... :-)
      Second step of file rename somehow got skipped.
      More cleanup after PRWIZCH.  Don't build the signal-strength widget in
      Second half of file name change, and some function name changes.
      Get rid of process_exception().  All the trigger strings by which we
      This warning should save a few tears.
      Document more new features.
      Do better data validation before shipping up to the client.
      Now with prompts for interactive mode.
      Fair warning about the junk some GPSes issue.
      Raw hook is now part of the gps_data structure.  This is an essential
      Another step in raw-hook support for the client library.
      Fix a minor mess in the make macros for the X programs.
      Small code cleanup and a display tweak.
      Narrow the interface to the satellite-display code.
      Minor cleabups to narrow the interface to libgps.
      Minor cleanups to narrow the interface to libgps.
      Merged the libraries.  There is now just one libgps supporting two 
interfaces.
      Make sure the RPM exports the right things.
      Enable new commands r1, r0, w1, w0.  serial.c now uses r1 to open a
      More interface narrowing.
      Confine libgps code to as small a span as possible.
      More code cleanup.
      The gps client now goes through the daemon.  Relayerr functionality
      The gps client now goes through the daemon.  Relayer functionality
      Cleanup.
      Ready to try daemonizing this.
      Enabling raw mode in ggps does not yet work.
      Squash some excessively noisy error messages.
      Don't skip some needed initializations.
      Moved to new project site.
      Remove unused member.
      Speedometer work with the daemon now.
      Version bump -- time to do 1.91
      Fix some minor problems in the build machinery.
      Add a proper 1.91 release announcement.
      A big name change.  The gps* and gpsd* namespaces swap places, so gps_*
      Hm...looks like it will take more than two commits for the name swap.
      Contibuing the name swap.
      Continuing the name swap.
      Completing the name swap.
      Followup on the big name change -- more files have to move.
      More renames and explanations.
      Catch the documentation up with reality.
      Tell TripMate to ignore PRWIZCH.
      Placed HTML stuff under version control.
      Added Web pages to version control.
      Added link to hardware page.
      Documentation update.
      Added a pretty picture.
      More on the FAA mode code.
      Yet more on obscure NMEA sentences.
      Point to satellite simulation.
      Notify watchers on disconnect/reconnect events.
      Why to bother with 2.0.
      Installation production now installs gpsd as a startup-time daemon on
      Typo fixes.
      Get the initscript right.
      At -D2, command-response error messages are now copied to the client.
      Added sources listing, DTM, GBS, description of lon/lat fields, notes
      Detail corrections.
      Make sure everything shipped to clients by gpsd is NUL-terminated.
      Timeouts at the daemon level were a mistake.  This commit removes them.
      Both client should get an online/offline status report.
      More uniform ? responses.  Update the l capability string.
      Get rid of ugly macro.
      Cosmetic cleanup.
      Do a more graceful job of dealing with client disconnects.
      Get rid of some obviously bogus code.
      There is no longer any good reason for passivesock to be in the
      Clean up things so gpscli_report() is no longer required on client side.
      Ready for 1.93 release.
      Oops -- have to bump the version.
      SIGHUP resets the daeon.
      Change some verbosity levels.
      Fixed install production.
      Remove redundant tests.
      Do a better job of filtering junk in satelite data.
      Make sure old data does not get left on the display.
      Disable some data filtering that was causing problems.
      James Cameron's spelling nit.
      And yet another spelling nit.
      GGA header was out of place.
      Fill in a missing dependency.
      Better debugging of writes.
      Don't discard $ before control word matching.
      Offline check is no longer subject to false positives.
      Fixed a session-close bug that caused weird failures with the logfile
      Documentation tweaks.
      Y command now carries information about which satellites were used in
      Ready to do another point release.
      Don't ship nuls, instead check lengths.
      Clean up the web page.
      This is version 1.95.  Mainly a fix for Makefile.in.
      James Cameron's bugfixes.
      -t option is gone.
      James Cameron's patch to remove the vetigial reopen flag.
      Lose an ugly macro.
      James Cameron's flow control idea.
      We've solved James Cameron's bug.
      Fix a *nasty* array-overrun bug.  The old code was using the "used"
      Remco's nonblocking-write implementation.
      Fix a bug in the handling of watcher and raw modes.
      Fill Hamish's requist for info about defaults in the usage message.
      To make sure that N seconds of real time have elapsed, it suffices to
      Website updates.
      Important note about degree formats.
      Remove this just to soothe the reader's nerves :-)
      Do the right thing for /etc in the specfile.
      Spec file now generates a gpsd and gpsd-devel.
      More documentation updates.
      Oops. Unfrotz the default port number.
      Remove CHANGELOG from the build prerequisites, it's generated by shipper.
      Derrick pointed out that mkinstalldirs is missing.
      Acknowledge a theological dispute.
      Get rid of an unnecessary buffer copy.
      Timestamp on GPRMC might have a decimal part.  Don't throw it away.
      Last-fix updates now happen on all three of GPRMC, GPGGA, and GPGSA.
      Three sentences, GGA and GGL and RMC, contain timestamps.
      Get rid of a superfluous variable.
      Report DGPS transactions at level 2.
      gscore.c is an out-of-date copy of libgpsd_core.c.
      Timestamps to ISO8601 format -- no more US-centric date format 
assumptions.
      Document new ISO8601 format.
      Documentation fixes.
      Code cleanup.  Make some validity checks uniform.
      First cut at Python low-level interface.
      Added Python interface to daemon.
      Fail, and log the failure, when there is no GPS.
      Move some structure members used only for housekeeping out of the
      Python interfaces are actually shipped now.
      Better explanatory comments.
      We're ready to do another release.
      libgps.a interface change, everything goes in a malloced struct now.
      Cleanup, change a widget name.
      Fix bug in date copying.
      Remove unneeded argument.
      Another unnecessary argument gone.
      Try for less cryptic widget names.
      More descriptive labeling.
      Yet another step towards eliminating magic numbers.
      Remove debug member from session structure.
      Document the Python interface.
      Clean up the exerciser a little.
      Symplify the Python interface.
      Fix test maun bug.
      About to switch to autogen-based build...
      Second step towards full autogen build.
      This seems to get us to a fully automake-based build.
      We're almost ready to ship 1.9.6.
      gpsd.spec is now generated from gpsd.spec.in.
      Turn off write permissions on gpsd.spec when it's generated, and
      Backed out the use of libtool, on the grounds that any tool which
      Close these down so we can delete them.
      Get rid of things no longer needed in automake land
      Just labekl this, don't write-lock it.
      OK, we use libtool to banish an autogen.sh warning, but only because
      Make gpsd.py use a file object to talk with the GPS, that way
      Minor typo fixes.
      Prepare to ship 1.96.

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


hooks/post-receive
-- 
GPSD