This is a program to replay data in NMEA-0183 format which has been logged with
NMEA-0183v4 TAG timestamps. Timestamps can be millisecond or second resolution,
but millisecond timestamps will yield playback more consistent with the original
data stream.

You can log nmea-0183 data to a file using kplex by specifying the following
additional interface on the command line:

file:direction=out,timestamp=ms,filename=<name>

Where <name> is the name of the file to log to (e.g. /tmp/foo).

Installation
------------

Just type "make" (or "gmake" where appropriate) to compile (it's only one file).
"make install" will install to /usr/local/bin, or you can just copy it wherever
you want.

Usage
-----

replay183 [-r] [-c | -n <repetitions>] [-d <delay> ] logfile ...
where logfile is one or more files containing timestamped data to replay.
if "-c" is specified replay183 will loop continuously through all logfiles
specifed on the command line.
If "-n <repetitions>" is specified, replay183 will loop <repetitions> times
through the specified logfiles.
If neither -c nor -n is specified, each logfile will be processed once.
If "-d <delay>" is specified replay183 will wait <delay> milliseconds between
each file to be processed. The default value is 1000ms (1 second).

Without the "-r" option replay183 will output recorded sentences terminated by a
single newline ("\n" or "<LF>").  With "-r" specified, the output will be
terminated by the standard NMEA-0183 sentence termination sequence "\r\n"
("<CR><LF>").

The output of replay183 can be piped into kplex's standard input (in which case
there is no need to specify "-r") or sent directly to a process or device
which accepts data in NMEA-0183 format (in which case "-r" should normally be
used).

Notes
-----
This was cobbled together without much testing so may be buggy.  Let me know
via the kplex google group.

Timestamps or any other TAG information recorded in the log file is stripped
from the output.

No checksum checking is performed on the input log file.

On a non-realtime timesharing timing is never going to be exact for playback.
Moreover the duration of a sentence's transmission at 4800 baud can be 10s of
milliseconds or more, and kplex adds the timestamp when it outputs to the log
file (at the end of transmission) whereas replay183 uses the timestamp to
start transmission.  Timing shoudl be good enough for many testing purposes
however.

replay183 times sentence transmission relative to the time it started
rather than the previous sentence.  This should result in the overall time
taken to replay a log file being close to the time it took to record.