Another patch from Alain Tauch (stripHtml enhancements) - thanks!

git-svn-id: svn://svn.berlios.de/gpodder@43 b0d088ad-0a06-0410-aad2-9ed5178a7e87

ChangeLog

@@ -1,3 +1,6 @@
+Sun, 19 Mar 2006 18:38:42 +0100 <thp@perli.net>
+	+ Applied another patch from A. Tauch (stripHTML enhancements)
+
 Sun, 19 Mar 2006 15:11:32 +0100 <thp@perli.net>
 	+ Applied two patches from A. Tauch <contrib@maisondubonheur.com>:
 	* fix for local db writer's encoding -> is now really iso-8859-1)

src/gpodder/liblocdbwriter.py

@@ -19,7 +19,7 @@ class writeLocalDB( object):
     ofile = None
     
     def __init__( self, filename, channel):
-        self.ofile = codecs.open( filename, "w", 'iso-8859-1')
+        self.ofile = codecs.open( filename, "w", 'iso-8859-1', 'replace')
 	self.ofile.write( '<?xml version="1.0" encoding="ISO-8859-1"?>'+"\n")
 	self.ofile.write( '<!-- local download database, generated by gPodder -->'+"\n")
 	self.ofile.write( '<rss version="2.0">'+"\n")

src/gpodder/libpodcasts.py

@@ -13,6 +13,7 @@
 
 import gtk
 import gobject
+import htmlentitydefs
 
 import libgpodder
 
@@ -216,5 +217,11 @@ def channelsToModel( channels):
 
 def stripHtml( html):
     # strips html from a string (fix for <description> tags containing html)
+    dict = htmlentitydefs.entitydefs
     rexp = re.compile( "<[^>]*>")
-    return rexp.sub( "", html)
+    stripstr = rexp.sub( "", html)
+    # strips html entities
+    for key in dict.keys():
+        stripstr = stripstr.replace( '&'+unicode(key,'iso-8859-1')+';', unicode(dict[key], 'iso-8859-1'))
+    return stripstr
+