BasicIndxingQueue: Determine the mimetype without looking into the file

Baloo indexing happens in two phases - Phase 1 (Basic) and Phase 2
(File). During phase 1 only the basic stat info + mimetype + xattr are
stored. During phase 2 the file's contents are indexed.

Currently phase 1 was also looking into the contents of the file to
determine the mimetype. This was causing very high IO usage during the
inital indexing. We now determine the mimetype only using the url, and
do a proper mimetype analysis in Phase 2.

DIGEST: Reduce io usage during initial indexing