April 8, 2004 ver 2.14
* Changed the name of Dishplayer(old) to Dishplayer_network and reversed the dev bits (as requested by Rob)

March 17, 2004 ver 2.13
* Changed the extra bit in RC5x as Jon requested in http://www.hifi-remote.com/forums/viewtopic.php?p=8577#8577

January 7, 2004 ver 2.11
* Renamed XX15 and XX16 to dBox15 and dBox16, since no one suggested a better name and
  they seem to be used only in dBox devices
* Added Teac-K for the 48 bit protocol, similar to Kaseikyo, found in a Teac device

December 31, 2003 ver 2.10
* Improved the kludged 15 and 16 bit versions of Blaupunkt, temporarily named XX15 and XX16
* Reduced (maybe eliminated) the false Async decodes that overlap Zenith decodes

December 28, 2003 ver 2.09
* Generalized the decoder for Blaupunkt to support other lengths, which
  are presently only partially supported.  This required changing an undocumented
  feature of decodeIR used only by decodeCCF, so you need a new decodeCCF to go
  with this decodeIR

December 27, 2003
* Added X10
* Added Sunfire

December 26, 2003
* Added pid_002A

December 20, 2003
* Added Panasonic2
* Added ?Mitsui (for 10 bit protocol seen only in one old Mitsui TV).

October 7, 2003
* Don't access data when length is zero (avoids crash in learned signal from TV/0135.
  That signal is bad either due to IR.EXE bug or learning bug in remote.  We don't know
  which yet).

September 1, 2003 ver 2.07
* Changed Thomson again, so EFCs are displayed for both variants of protocol 4B

August 27, 2003 ver 2.06
* Changed split between device bits and OBC bits in Thomson to match UEI
  protocol 4B

August 27, 2003 ver 2.05
* Made Denon/Sharp decoder less picky about timing

August 11, 2003 ver 2.04
* Allowed a smaller ratio between lead in and lead out in order to correctly
  decode some RCA samples.

July 30, 2003 ver 2.03
* Kludge version of Acer Keyboard

July 26, 2003 ver 2.02
* Added AirAsync (1200 Baud, dual stop bit bit version of Async, used by a version of airboard )

June 12, 2003
* Added Async

May 21, 2003
* Added Airboard

May 16, 2003
* Added Proton
* Suppress old decoder for Jerrold

Apr 25, 2003
* Made RECS80 more flexible

Apr 24, 2003
* Added Samsung20

Mar 26, 2003
* Added JVC-48

Mar 24, 2003
* Loosened the framing requirements for Sony8 and Sony15 to let the DSP
  signals decode as individual frames.

Mar 23, 2003
* Very crude improvements to the ReadMe

Mar 22, 2003
* Changed the name "Akai" to "ID-000D" since no one has ever told me about
  any connection between this protocol and Akai.
* Changed "ID-0045" to "RECS80" and changed it to msb.
* Added "Pace MSS"
* Added "Thomson"
* Added "Sony8"

Mar 11, 2003
* Added Sharp and Denon.  All the Sharp and Denon false decodes in my CCF and
  IRtxt files were eliminated and (after the change below) no correct decodes
  were lost.  No false decodes found in any of my samples.
* Made Gap framing more flexible so that it accepts varying ON values (that it
  previously rejected) as long as the largest ON is less than the smallest OFF.
  Result: Regained the Sharp and Denon decodes lost from CCF files above.
  Found some correct ID-0045 decodes in CML files.  No
  bad results in CCF files, but IRtxt files got a few extra garbage Gap
  decodes and CML files got those plus some false Mitsubishi and ViewStar
  decodes.  (Needs fine tuning).
* Added Dishplayer.  Found one false decode in a badly learned Dishplayer(old)
  signal JVC <VCR> HM-DSR100DU in master.cml (would be hard to fix this false
  decode)
* Added "Akai" (I have no idea why devices4 calls protocol 000D "Akai", but I
  found a bunch of signals in CML files that match 000D, so I report them as
  Akai).  There were a few false decodes (I think there is no way to make this
  one robust) and some brands had the same signal but 40Khz.
* Added Emerson and Sc-Atl-6 (again I have no idea why device4 calls protocol
  0078 "Sc-Atl-6", but I did too).  These signals differ only in frequency,
  protocol 0065 also differs from protocol 0078 in the way the EFC is
  processed.  My new version does my best guess of correct EFCs for each of
  these.  The Emerson differs both from KM and ccf2efc.

Mar 10, 2003
* Reduced framing requirement for Zenith based on sample:
  Runco <Video Projector> IDP-852 from anon_big
  Result: All those fixed plus 3 signals in two other devices.  No bad results.
* Added F12 (at least I think it's F12).  I used lsb8 for OBCs which gives
  pretty bogus looking values, but it's no obvious what would be better.
* Converted API to _stdcall to avoid crashing the debugger.  Dan didn't have
  _stdcall in the original DLL, but I think that was a bug.
* Added a bunch of checks for various overflows in order to get rid of a
  memory corruption bug that resisted ordinary debugging.  I found and fixed
  one bug of underflowing pDuration when preparing to decode a zero length
  signal (but that wasn't the one I was chasing).  I never found the bug I was
  looking for.  It just stopped happening.
* Added Sony

