Thanks for the excellent bashpodder script; it's just what I've been needing.  I've attached my locally changed version (as well as a diffs file) in case you'd like to include any of my changes in the main distribution.

The changes are:

* Touch podcast.log if it doesn't already exist; this eliminates grep warnings the first time you run the script

* Added ^ and $ to exactly match the URL in the main loop; the original version could, theoretically, skip a new file if it had previously downloaded another file whose URL was a superstring of the new one.  (Also note a related bug: URLs can contain grep-special characters, which could screw things up as well.  I haven't tried to fix that.)

* Quoted $datadir everywhere in case the user changes $datadir to be something that contains whitespace.

* Append dynamically created log file to permanent log file only if the dynamic file is non-empty (i.e., if something was successfully downloaded).  It was correct before, this is just a pointless optimization.

* If download directory is empty, remove it without creating a .m3u there.

The last couple of changes are useful to people like me, who subscribe only to a couple of podcasts, and those are updated only infrequently (every week or so).

Thanks again for the very nice work!