Read XMLTV data and write out the same data sorted in date order.
Where stop times of programmes are missing, guess them from the start
time of the next programme on the same channel. For the last programme
of a channel, no stop time can be added.
Tv_sort also performs some sanity checks such as making sure no
two programmes on the same channel overlap.
<B>--output FILEB> write to FILE rather than standard output
<B>--by-channelB> sort first by channel id, then by date within each
<B>--duplicate-errorB> If the input contains the same programme more than once,
consider this as an error. Default is to silently
ignore duplicate entries.
The time sorting is by start time, then by stop time. Without
<B>--by-channelB>, if start times and stop times are equal then two
programmes are sorted by internal channel id. With <B>--by-channelB>,
channel id is compared first and then times.
You can think of tv_sort as converting XMLTV data into a canonical
form, useful for diffing two files.