Fix kiwix proxy 502ing on article links with non-ASCII titles

Sinatra decodes a route's splat before handing it to the block, so an
article link like /kiwix/content/wikipedia/Dlhé_diely arrived as raw
UTF-8, not percent-encoded. URI#path= rejects that outright
(URI::InvalidComponentError) rather than escaping it, which .proxy's
rescue turned into a silent nil -> every article link 502'd, while the
search results page linking to them worked fine (its own path,
/kiwix/search, is plain ASCII).

lib/maptool/kiwix.rb

@@ -26,9 +26,16 @@ module Maptool
     # correct without this having to touch response bodies at all. Returns
     # nil (halted as a 502 by the caller) rather than raising, same
     # not-reachable-isn't-exceptional stance as .reachable? above.
+    #
+    # path arrives already URL-*decoded* — Sinatra decodes a route's splat
+    # before handing it over — so an article title with non-ASCII
+    # characters (e.g. "/kiwix/content/wikipedia/Dlhé_diely") has to be
+    # re-escaped before URI#path= will accept it at all; it raises
+    # URI::InvalidComponentError on raw UTF-8 bytes rather than escaping
+    # them itself. Article links otherwise 502'd for exactly this reason.
     def self.proxy(base_url, path, query_string, open_timeout: 2, read_timeout: 10)
       uri = URI(base_url)
-      uri.path = path
+      uri.path = URI::RFC2396_PARSER.escape(path)
       uri.query = query_string unless query_string.to_s.empty?
       Net::HTTP.start(uri.host, uri.port, open_timeout: open_timeout, read_timeout: read_timeout) { |http| http.request(Net::HTTP::Get.new(uri)) }
     rescue StandardError