Please describe the problem:
Podcast titles are not sanitized at all. A podcast with the title '<em>Example</em>' will be assigned a download folder Podcasts/<em>Example</em> (i.e. a directory called em> nested inside another called <em>Example<)

Possibly podcasts with other characters in the title could produce more dangerous results (&& rm -rf ~/*)?

Steps to reproduce:
1. create a podcast that has html tags in the title (Drupal audio module does this)
2. add podcast feed to rhythmbox
3. check podcast download directory


Actual results:
Podcast download directories are created with nested directories, and are ugly

Expected results:
The title should have special characters removed (in fact possibly anything inside <>?)

Does this happen every time?
Yes

Other information: