libloc: Update to version 0.9.18

libloc: Update to version 0.9.18

- Update libloc from version 0.9.17 to 0.9.18
- Update of rootfile
- Update of patch to revert installing of perl files into perl vendor directory as source
   file has changed enough.
- Changelog
    0.9.18
	* A new Lua module has been added as announced here:
	  https://www.ipfire.org/blog/ipfire-location-lua-bindings-for-fun-and-profit
	* The algorithm to detect bogons and duplicates in the tree have massively been
	  improved and should be nearing their theoretical maximum in terms of
	  performance.
	* A large number of stability and correctness fixes have been rolled out. These
	  mostly affect the code generating the database.
	* We now have a small Jenkins pipeline which will check if the library still
	  builds for a couple of major Linux distributions and various architectures.
      Python:
	* AS and Country objects are now hashable and support rich comparison operations
      Importer:
	* Exporting the database is around 200x faster due to eliminating any excessive
	  joins. Instead a new temporary table will be created and a temporary index
	  will be used to apply various updates to the data from various sources inside
	  the database. That allows us to create the export iteratively instead of
	  having one large query that runs for forever. An export that formerly took
	  around 17-20 hrs(!) will now take only ~5 mins.
	* A new source for human-friendly names for Autonomous Systems registered with
	  ARIN has been added
	* Importing feeds from AWS and Spamhaus has been split off into separate
	  database tables. This will allow us to import them separately and prioritise
	  our own rewrites over them.
	* The ARIN parser has been refactored based on csv.DictReader(), parsers for
	  the AWS and Spamhaus feeds have been rewritten, too
	* Geofeeds are now fetched concurrently with a unified downloader
	* Certain country codes will be entirely ignored. Currently this is YU for
	  former Yugoslavia and ZZ which is used to say “no country”
	* Country codes can now be corrected on the fly. This is used to change UK to
	  GB as only the latter is the valid country code for the United Kingdom.
	* Countries that are not on our list will not be imported any more.
	* Networks larger than /4 for IPv6, and /10 for IPv4 won’t be imported any
	  more. This avoids that we propagate any issues in the global routing table
	  into the database.

Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>