[Len Case]

There is a bug that was introduced when adding support for handling file: urls that point to directories that prevents the proper "fetch" of non-directories. Html files will work fine, but anything with a different mime-type will either be converted to html or aborted.