From d74116dc67e47e2f361978ab8ee974c70d57faa9 Mon Sep 17 00:00:00 2001
From: Jacques Distler <distler@golem.ph.utexas.edu>
Date: Wed, 7 Mar 2007 21:06:39 -0600
Subject: [PATCH] Ensure that input is bona fide utf-8.

---
 app/controllers/file_controller.rb | 11 +++++++++--
 app/controllers/wiki_controller.rb | 18 +++++++++++-------
 lib/string_utils.rb                | 18 ++++++++++++++++++
 3 files changed, 38 insertions(+), 9 deletions(-)
 create mode 100644 lib/string_utils.rb

diff --git a/app/controllers/file_controller.rb b/app/controllers/file_controller.rb
index 865450aa..b1183e57 100644
--- a/app/controllers/file_controller.rb
+++ b/app/controllers/file_controller.rb
@@ -1,6 +1,7 @@
 # Controller responsible for serving files and pictures.
 
 require 'zip/zip'
+require 'string_utils'
 
 class FileController < ApplicationController
 
@@ -9,7 +10,7 @@ class FileController < ApplicationController
   before_filter :check_allow_uploads
 
   def file
-    @file_name = params['id']
+    @file_name = @params['id']
     if @params['file']
       # form supplied
       new_file = @web.wiki_files.create(@params['file'])
@@ -28,6 +29,7 @@ class FileController < ApplicationController
         send_data(file.content, determine_file_options_for(@file_name, :filename => @file_name))
       else
         @file = WikiFile.new(:file_name => @file_name)
+#        @file = WikiFile.new(@file_name)
         render
       end
     end
@@ -61,7 +63,8 @@ class FileController < ApplicationController
     if @web.allow_uploads?
       return true
     else
-      render :status => 403, :text => 'File uploads are blocked by the webmaster' 
+      @hide_navigation  = true
+      render(:status => 403, :text => 'File uploads are blocked by the webmaster', :layout => true)
       return false
     end
   end
@@ -77,6 +80,10 @@ class FileController < ApplicationController
       page_content = entry.get_input_stream.read
       logger.info "Processing page '#{page_name}'"
       begin
+        if !page_content.is_utf8?
+          logger.info "Page '#{page_name}' contains non-utf8 character data. Skipping."
+          next
+        end
         existing_page = @wiki.read_page(@web.address, page_name)
         if existing_page
           if existing_page.content == page_content
diff --git a/app/controllers/wiki_controller.rb b/app/controllers/wiki_controller.rb
index 1c4fedea..5117f4ed 100644
--- a/app/controllers/wiki_controller.rb
+++ b/app/controllers/wiki_controller.rb
@@ -3,6 +3,7 @@ require 'redcloth_for_tex'
 require 'parsedate'
 require 'zip/zip'
 require 'sanitize'
+require 'string_utils'
 
 class WikiController < ApplicationController
 
@@ -226,25 +227,28 @@ class WikiController < ApplicationController
   def save
     render(:status => 404, :text => 'Undefined page name') and return if @page_name.nil?
 
-    author_name = @params['author']
+    author_name = @params['author'].delete("\x01-\x08\x0B\x0C\x0E-\x1F")
     author_name = 'AnonymousCoward' if author_name =~ /^\s*$/
+    raise "Your name was not valid utf-8" if !author_name.is_utf8?
     cookies['author'] = { :value => author_name, :expires => Time.utc(2030) }
     
     begin
-      filter_spam(@params['content'])
+      the_content = @params['content'].delete("\x01-\x08\x0B\x0C\x0E-\x1F")
+      raise "Your content was not valid utf-8" if !the_content.is_utf8?
+      filter_spam(the_content)
       if @page
-        wiki.revise_page(@web_name, @page_name, @params['content'], Time.now, 
+        wiki.revise_page(@web_name, @page_name, the_content, Time.now, 
             Author.new(author_name, remote_ip), PageRenderer.new)
         @page.unlock
       else
-        wiki.write_page(@web_name, @page_name, @params['content'], Time.now, 
+        wiki.write_page(@web_name, @page_name, the_content, Time.now, 
             Author.new(author_name, remote_ip), PageRenderer.new)
       end
       redirect_to_page @page_name
     rescue => e
       flash[:error] = e
       logger.error e
-      flash[:content] = @params['content']
+      flash[:content] = the_content
       if @page
         @page.unlock
         redirect_to :action => 'edit', :web => @web_name, :id => @page_name
@@ -290,7 +294,7 @@ class WikiController < ApplicationController
 
   def s5
     if @web.markup == :markdownMML or @web.markup == :markdown
-      @s5_content = sanitize_html(Maruku.new(@page.content.delete("\r\x01-\x08\x0B\x0C\x0E-\x1F"),
+      @s5_content = sanitize_html(Maruku.new(@page.content.delete("\r"),
            {:math_enabled => true, :math_numbered => ['\\[','\\begin{equation}'], :content_only => true,
             :author => @page.author, :title => @page.plain_name}).to_s5)
     end
@@ -441,7 +445,7 @@ class WikiController < ApplicationController
       raise "Your edit was blocked by spam filtering" if content =~ pattern
     end
   end
-  
+
   def load_spam_patterns
     spam_patterns_file = "#{RAILS_ROOT}/config/spam_patterns.txt"
     if File.exists?(spam_patterns_file)
diff --git a/lib/string_utils.rb b/lib/string_utils.rb
new file mode 100644
index 00000000..2ec68f86
--- /dev/null
+++ b/lib/string_utils.rb
@@ -0,0 +1,18 @@
+# Some useful additions to the String class
+
+class String
+
+   def is_utf8?
+     self =~  /^(
+         [\x09\x0A\x0D\x20-\x7E]            # 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
+     )*$/x;
+   end
+
+end