if your audio source has podcast, great news! this script will add
**BASIC** support for podcasts. It will just download the first audio it
finds in the xml.
la txdate puo' essere pescata dal corpo (attualmente la cosa e' parziale)
in base al flag replica= viene ricavata una data di fine validita per un
audio.
Sia gli audio sia gli audiogroup hanno un flag "valid"
before that, a single invalid URL in the feed would have blocked all of
it. In fact, when there is no duration, and ffprobe fails, everything
was failing. This commit make it resilient to errors.
--min-len and --max-len accept argument like 15m for 15 minutes or 2h
for two hours.
The old "number of seconds" (ie: 3600 for 1 hour) is still supported
and not deprecated.
source weights are meant to first of all pick a source, with the
defined weights. This option is unrelated to --random. In fact,
--random will switch from chronological order to random one.
--source-weights will, instead, pick only one of the sources given in
command line.
A typical example is for jingles: if you have a big number of jingles that
should run relatively rarely, and a small number of jingles that should
run often, then putting them all in the same directory isn't a good
idea. You'd better put them in two different dirs and use
--source-weights 1:2
