The feedparser packages has 20 years of history and is very good at parsing
RSS and Atom, so use that instead of ad-hoc regex and XML parsing.
The medium_rss and shaarli_rss parsers weren't touched because they are
probably unnecessary. (The special parse for pinboard is just needing because
of how tags work.)
Doesn't include tests because I haven't figured out how to run them in the
docker development setup.
Fixes#1171
