Patch from MJ Ray for items without link - link might not always exist, md5sum is always present, don't fail on no link.
Make sure that we feed the parser unicode data - Check if we get a unicode string back from feedparser - If not, re-encode so that it is - Remove explicit casting to utf-8 before it goes to the parser
More utf-8 handling for images - iff we get a string object rather than a unicode one, decode it from utf-8
Stop trying to decode image tags to utf-8 - If the title/url contains a utf-8 character and we try to decode it it will fail due to not being present in the ascii set. Feedparser has already made sure that everything is utf-8 before we get it.
Change header encoding for From/To address to make sure they're utf-8 and so they are not invalidly encoded later. Bug reported by Andre Klärner with pointers to what was going wrong - many thanks!
* Add https support (thanks to Andre Klärner)
Update so that we don't get a warning in python 2.6 and above, where the md5 module has been depricated
Fix silly typo
Add item fetched date header (X-rss2maildir-rundate)
guid might not always exist, link is always present, don't fail on no guid.
Fix bug when link/guid contains characters not in ascii by encoding the keys as utf-8
Small fix to title handling code to deal with unicode better
Fix for items that actually have no content.
Fix typo in previous charref fix
Fix bug in character reference handling code
Fix for title parsing
Fix some entity handling * fixes handling of numeric entities * fixes unittest for entities.
Another blockquote fix
Fix blockquote support
Fix issue with images having the same alt value but different urls