[Gpsd-commit-watch] [SCM] GPSD annotated tag, release-2.27, created. release-2.27

[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-2.27 has been created
        at  df4ded5b978f9f761b2fd6056cb3153af064f277 (tag)
   tagging  9e72a1e13589a0f0768249f23f413e316df3bf2b (commit)
  replaces  release-2.26
 tagged by  Eric S. Raymond
        on  Wed Jul 6 07:45:21 2005 -0400

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

Eric S. Raymond (93):
      Refactor, moving towards supporting multiple fake GPSes.
      Next step in gpsfake refactoring.
      gpsfake refactoring continues.
      Gather more gpsfake logic into classes.
      Daemon control is now encapsulated.
      Remove PID file on exit.
      Get a control socket for the gpsfake daemon instance.
      Factor out more predicates.
      More gpsfake cleanup.
      Add methods for editing the internal device list, but don't try to use 
them yet.
      Fixes for the hotplug/control-socket mechanism.
      Switch gpsfake to using the control socket.
      gpsfake acquires -m (monitor) option.
      Freeze before pulling the session-faking stuff into its own module.
      gpsfake.py is now a separate loadable module, so we can write more test 
loads.
      Make the spawn command that launched the test daemon visible.
      Small cleanup in gpsfake.
      Correct the arguments to gpsfake's daemon invocation.
      Start of TestSession class -- not working yet.
      Freeze before predicate unification.
      Predicate unification.
      Arrange for fake GPSes to keep their line number in the log between 
invocations.
      Guard thread starts so we don't have to track when a fake GPS is active.
      gpsfake feed function now waits for the daemon oto open the device.
      gpsfake progress prompt is now optional.
      Freeze before attempted fix to thread-launch error.
      Fix an error in the exerciser for gps.py.
      This captures O responses and dumped NMEA. but not the message log.
      Changes to support valgrind auditing.
      gpsfake now uses the TestSession object.
      TestSession now does its own signal handling.
      First version of valgrind torture test.
      Plugged a memory leak that I found with valgrind.
      Eliminate a select error.
      Second generation of torture test.
      gpsfake -v no longer does anything, so remove it. Ship valgrind-audit.
      This version makes gpsfake work, but valgrind-edit still has problems.
      This works, albeit with a long wait.
      Tilman Koschnick's fix for the DBUS code.
      Guard a reference so a -1 file descriptor won't cause havoc.
      Add progress messages to test framework.
      Fix a subtle error in channel assignment.
      Another test case is firm.
      Pre-release cleanup.
      Tilman Koschnik fxed a bug.
      Prepare for release.
      Abstract some device-allocation primitives.
      Finish abstracting the channel primitives.
      Eliminate dynamic-memory allocation from the daemon.
      Document that the low-level interface uses no dynamic allocation.
      Steps towards eliminating FIONREAD.
      Eliminate use of FIONREAD.
      Clean up after FIONREAD removal.
      Try to fix some bugs introduced by FIONREAD removal.
      Make the test main of the packet getter work again.
      Set VMIN in GPS reads.
      Stage one of joining the GPS webring.
      Pull the DGPSIP code into its own module.
      Clean up DGPSIP a bit more, some variables don't need to be per-session.
      Experimental code to hunt for a DGPSIP server.
      We can have right-hand spans now.
      Preparing for release -- make sure RPMs build OK.
      More small pre-release cleanups.
      Add some eye candy to the website.
      Add the ability to probe multiple DGPS servers...
      Improved diagnostics for DGPS server hunting.
      Get the length of an RTCM104 message right.
      This is the fix I had been waiting for Gary to check in -- I hope.
      Avoid a magic number.
      Cosmetic fix.
      Release preparation.
      Quiet down splint.
      Disable the hunt for DGPSIP servers.
      Move driver-specific members into a union in order to save some storage.
      Gather all the error-modeling into one spot.
      Try to pacify automake-1.9.5.
      We have track error computation.
      Add gpsfake code for detecting file decriptor leaks in the daemon.
      Slightly better unstrumentation of file descriptor operations.
      Better cleanup code in Python client, deinitialize file object properly.
      Hm...this has turned up a file descriptor leak.
      Attempt to banish Mark Cohen's bug.
      Introduce client timeouts.
      Fold in some advice from Wolfgang Rupprecht.
      Be aware of WAAS satellites.
      Much simpler DGPS test.
      Add status mode setting to Zodiac driver.
      Add some documentation, and a no-read timeout.
      FIONREAD can't spin anymore, because FIONREAD is gone.
      Cleanup before afddressing Peter Stranky's bug.
      Quiet splint.
      Fix Petr Stransky's bugs.
      Ready to ship 2.27.

Gary E. Miller (1):
      Modify garmin code to remove calloc().  GarminBuffer is now in 
gps_device_t.

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


hooks/post-receive
-- 
GPSD