From a18af410830f9afbc963ec9e0252c29cda53a9da Mon Sep 17 00:00:00 2001
From: hukl <contact@smyck.org>
Date: Thu, 5 Mar 2009 23:06:22 +0100
Subject: get rid of the stupid entities. its the 21st century baby. unicode is
 widely available already! except for windows

---
 lib/tasks/development_init.rake | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

(limited to 'lib')

diff --git a/lib/tasks/development_init.rake b/lib/tasks/development_init.rake
index 22553fc..d3c8aff 100644
--- a/lib/tasks/development_init.rake
+++ b/lib/tasks/development_init.rake
@@ -58,4 +58,25 @@ namespace :cccms do
     
     n.publish_draft!
   end
+  
+  desc "Convert Entities to real charactes"
+  task :convert_entities  => :environment do |t|
+    Page.all.each do |page|
+      if page.body
+        puts ">> #{page.id} -- #{page.node.unique_name if page.node}"
+        tmp_body = page.body.dup
+        tmp_body.gsub!(/&auml;/, "ä")
+        tmp_body.gsub!(/&ouml;/, "ö")
+        tmp_body.gsub!(/&uuml;/, "ü")
+        tmp_body.gsub!(/&Auml;/, "ä")
+        tmp_body.gsub!(/&Ouml;/, "ö")
+        tmp_body.gsub!(/&Uuml;/, "ü")
+        tmp_body.gsub!(/&szlig;/, "ß")
+        tmp_body.gsub!(/&nbsp;/, " ")
+        tmp_body.gsub!(/&ndash;/, "–")
+        page.body = tmp_body
+        page.save
+      end
+    end
+  end
 end
\ No newline at end of file
-- 
cgit v1.3


From 3c65fcee68683f32d44ca9c3cab14d248fadea0c Mon Sep 17 00:00:00 2001
From: hukl <contact@smyck.org>
Date: Thu, 5 Mar 2009 23:32:14 +0100
Subject: more entities translated, refined xpath selector

---
 app/models/page.rb              | 2 +-
 lib/tasks/development_init.rake | 9 ++++++++-
 2 files changed, 9 insertions(+), 2 deletions(-)

(limited to 'lib')

diff --git a/app/models/page.rb b/app/models/page.rb
index 709fa4f..c2e8176 100644
--- a/app/models/page.rb
+++ b/app/models/page.rb
@@ -128,7 +128,7 @@ class Page < ActiveRecord::Base
           tmp_body = "<div>#{self.body}</div>"
           xml_string = XML::Parser.string( tmp_body )
           xml_doc = xml_string.parse
-          links = xml_doc.find("a[not(starts-with(@href, 'http://'))]")
+          links = xml_doc.find("//a[not(starts-with(@href, 'http://'))]")
           
           locales = I18n.available_locales.reject {|l| l == :root}
           
diff --git a/lib/tasks/development_init.rake b/lib/tasks/development_init.rake
index d3c8aff..789cdf6 100644
--- a/lib/tasks/development_init.rake
+++ b/lib/tasks/development_init.rake
@@ -62,7 +62,7 @@ namespace :cccms do
   desc "Convert Entities to real charactes"
   task :convert_entities  => :environment do |t|
     Page.all.each do |page|
-      if page.body
+      if page.body && page.body != ""
         puts ">> #{page.id} -- #{page.node.unique_name if page.node}"
         tmp_body = page.body.dup
         tmp_body.gsub!(/&auml;/, "ä")
@@ -74,6 +74,13 @@ namespace :cccms do
         tmp_body.gsub!(/&szlig;/, "ß")
         tmp_body.gsub!(/&nbsp;/, " ")
         tmp_body.gsub!(/&ndash;/, "–")
+        tmp_body.gsub!(/&micro;/, "µ")
+        tmp_body.gsub!(/&sup3;/, "³")
+        tmp_body.gsub!(/&eacute;/, "é")
+        tmp_body.gsub!(/&sect;/, "§")
+        tmp_body.gsub!(/&ldquo;/, "“")
+        tmp_body.gsub!(/&rdquo;/, "”")
+        tmp_body.gsub!(/&bdquo;/, "„")
         page.body = tmp_body
         page.save
       end
-- 
cgit v1.3