html entities are now converted to utf-8 strings before the submission to the server

1 files changed, 7 insertions, 1 deletions

diff --git a/h-client/hlibrary.py b/h-client/hlibrary.py
index 1a86d81..c763a76 100644
--- a/h-client/hlibrary.py
+++ b/h-client/hlibrary.py
@@ -22,6 +22,7 @@ import re
 import sys
 import pycurl
 import urllib
+import htmlentitydefs
 from xml.dom import minidom
 
 
@@ -80,9 +81,14 @@ class Device(object):
 		self._post['comm_year'] = self._year;
 		self._post['pci_id'] = self._vendorId + ':' + self._productId;
 		self._post['interface'] = self._interface;
-		self._post['description'] = self._description;
+		self._post['description'] = self.htmlentitiesDecode(self._description.replace("\n","\r\n"));
 		self._post['driver'] = self._driver;
 		
+	#replace the HTML entitites with utf-8 characters
+	def htmlentitiesDecode(self,string):
+		for entity,code in htmlentitydefs.name2codepoint.iteritems():
+			string = string.replace("&"+entity+";",unichr(code))
+		return string.encode('utf-8')
 
 	#get the distro h-source allowed code from the distro codename
 	def getDistroCode(self,codenameString):