A quick way to break the input handling of mongodb using Umlauts

umlaut_problem.rb

# encoding: UTF-8

require 'mongo'

include Mongo

mongo_client = MongoClient.new
db = mongo_client.db("test")
coll = db.collection("txt")
coll.ensure_index(
   { content: "text" },
   { default_language: "de" }
)

coll.remove({})
coll.insert({ content: "Test" })
coll.insert({ content: "test" })
coll.insert({ content: "Überlin" })

puts "contents"
puts coll.find.each { |d| puts d.inspect }

puts "search tests"
puts "-- without Umlaut"
puts "---- uppercase query"
print "number of results: "
puts coll.find("$text" => { "$search" => "Test", "$language" => "de" }).count
puts "---- lowercase query"
print "number of results: "
puts coll.find("$text" => { "$search" => "test", "$language" => "de" }).count

puts "-- with Umlaut"
puts "---- uppercase query"
print "number of results: "
puts coll.find("$text" => { "$search" => "Überlin", "$language" => "de" }).count
puts "---- lowercase query"
print "number of results: "
puts coll.find("$text" => { "$search" => "überlin", "$language" => "de" }).count

puts "-- decomposed Umlaut"
puts "---- uppercase query"
print "number of results: "
puts coll.find("$text" => { "$search" => "\u0055\u0308berlin", "$language" => "de" }).count
puts "---- lowercase query"
print "number of results: "
puts coll.find("$text" => { "$search" => "\u0075\u0308berlin", "$language" => "de" }).count