Support blog entries with filenames in different encoding than the

Support blog entries with filenames in different encoding than the
blog's content by URL-escaping all non-standard characters in the URL.

URL-escaping done by a regexp so far. Should perhaps be converted to
some appropriate function call (doesn't CGI.pm offers URL escaping?)
before doing the next release.

Example URL generated with this patch:

http://noone.org/blog/Deutsch/Musik/Musikalisches-Podcast-St%F6ckchen.futile