From 372547b31b51d871f104ae5679e98420a99a12c1 Mon Sep 17 00:00:00 2001
From: Patrick J Cherry <patrick@bytemark.co.uk>
Date: Mon, 30 Apr 2012 12:49:24 +0100
Subject: Non-UTF8 stuff is now stripped.  Added encoding tests.

---
 lib/mauve/alert.rb                      | 24 ++++++++++
 test/tc_mauve_database_peculiarities.rb | 85 +++++++++++++++++++++++++++++++++
 2 files changed, 109 insertions(+)
 create mode 100644 test/tc_mauve_database_peculiarities.rb

diff --git a/lib/mauve/alert.rb b/lib/mauve/alert.rb
index 6b08f82..a5cb885 100644
--- a/lib/mauve/alert.rb
+++ b/lib/mauve/alert.rb
@@ -75,6 +75,20 @@ module Mauve
     def size; 99; end
     
     include DataMapper::Resource
+   
+    #
+    # If a string matches this regex, it is valid UTF8.
+    #
+    UTF8_REGEXP = Regexp.new(/^(?:#{[
+         "[\x00-\x7F]",                        # ASCII
+         "[\xC2-\xDF][\x80-\xBF]",             # non-overlong 2-byte
+         "\xE0[\xA0-\xBF][\x80-\xBF]",         # excluding overlongs
+         "[\xE1-\xEC\xEE\xEF][\x80-\xBF]{2}",  # straight 3-byte
+         "\xED[\x80-\x9F][\x80-\xBF]",         # excluding surrogates
+         "\xF0[\x90-\xBF][\x80-\xBF]{2}",      # planes 1-3
+         "[\xF1-\xF3][\x80-\xBF]{3}",          # planes 4-15
+         "\xF4[\x80-\x8F][\x80-\xBF]{2}"       # plane 16
+        ].join("|")})*$/)
     
     property :id, Serial
     property :alert_id, String, :required => true, :unique_index => :alert_index, :length=>256, :lazy => false
@@ -618,6 +632,7 @@ module Mauve
       # @return [String]
       def remove_html(str, conf = Sanitize::Config::DEFAULT)
         raise ArgumentError, "Expected a string, got a #{str.class}" unless str.is_a?(String)
+        str = clean_utf8(str)
 
         if str =~ /<[^0-9 <&.-]/
           Sanitize.clean( str, conf )
@@ -631,9 +646,18 @@ module Mauve
       # @param  [String] str String to clean
       # @return [String]
       def clean_html(str)
+        str = clean_utf8(str)
         remove_html(str, Sanitize::Config::RELAXED.merge({:remove_contents => true}))
       end
 
+      def clean_utf8(str)
+        unless UTF8_REGEXP.match(str)
+          str.gsub(/[^\x00-\x7F]/,'?')
+        else
+          str
+        end
+      end
+
       # All alerts currently raised
       #
       # @return [Array]
diff --git a/test/tc_mauve_database_peculiarities.rb b/test/tc_mauve_database_peculiarities.rb
new file mode 100644
index 0000000..9eb612e
--- /dev/null
+++ b/test/tc_mauve_database_peculiarities.rb
@@ -0,0 +1,85 @@
+$:.unshift "../lib"
+
+require 'th_mauve'
+require 'mauve/datamapper'
+require 'mauve/server'
+require 'mauve/configuration'
+require 'mauve/configuration_builder'
+require 'mauve/configuration_builders'
+require 'iconv'
+
+class TcMauveDatabasePeculiarities < Mauve::UnitTest
+  include Mauve
+
+  def setup
+    super
+    setup_database
+    @temp_db = "mauve_test.#{10.times.collect{ rand(36).to_s(36) }.join}"
+  end
+
+  def teardown
+    teardown_database
+    super
+  end
+
+  def test_encoding
+    #
+    # Don't test unless the DB URL has been set.
+    #
+    return unless @db_url
+
+    config=<<EOF
+server {
+  database "#{@db_url}"
+}
+EOF
+
+    Configuration.current = ConfigurationBuilder.parse(config)
+    Server.instance.setup
+
+    x = Hash.new
+    x["en"] = "Please rush me my portable walrus polishing kit!"
+    x["fi"] = "Ole hyvä kiirehtiä minulle kannettavan mursu kiillotukseen pakki!"
+    x["jp"] = "私に私のポータブルセイウチの研磨キットを急いでください！"
+
+    %w(UTF-8 WINDOWS-1252 SHIFT-JIS).each do |enc|
+      x.each do |lang, str|
+        assert_nothing_raised("Failed to use iconv to convert to #{enc}") { str = Iconv.conv(enc+"//IGNORE", "utf8", str) }
+  
+        alert = Alert.new(
+          :alert_id  => "#{lang}:#{enc}",
+          :source    => "test",
+          :subject   => str
+        )
+
+        assert_nothing_raised("failed to insert #{enc}") { alert.save }
+      end
+    end
+  end
+end
+
+
+
+class TcMauveDatabasePostgresPeculiarities < TcMauveDatabasePeculiarities
+  def setup
+    super
+    system("createdb #{@temp_db} --encoding UTF8") || flunk("Failed to create #{@temp_db}")
+    # @pg_conn = PGconn.open(:dbname => @temp_db) 
+    @db_url = "postgres:///#{@temp_db}"
+  end
+
+  def teardown
+    # @pg_conn.finish if @pg_conn.is_a?(PGconn) and @pg_conn.status == PGconn::CONNECTION_OK
+    super
+    (system("dropdb #{@temp_db}") || puts("Failed to drop #{@temp_db}")) if @temp_db
+  end
+end
+
+class TcMauveDatabaseSqlite3Peculiarities < TcMauveDatabasePeculiarities
+  def setup
+    super
+    # @pg_conn = PGconn.open(:dbname => @temp_db) 
+    @db_url = "sqlite3::memory:"
+  end
+end
+
-- 
cgit v1.2.1