From 4e44cba5d6b65415f145ba3ded191c07a8c9c660 Mon Sep 17 00:00:00 2001
From: jvoisin
Date: Sat, 30 Jul 2011 22:15:38 +0200
Subject: Outputed opendocument do not carry any traces of possible previous
 metadata

---
 lib/office.py | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

(limited to 'lib')

diff --git a/lib/office.py b/lib/office.py
index 2320e40..ba12295 100644
--- a/lib/office.py
+++ b/lib/office.py
@@ -6,6 +6,7 @@ import os
 import logging
 import zipfile
 import re
+import fileinput
 from xml.etree import ElementTree
 
 try:
@@ -54,17 +55,23 @@ class OpenDocumentStripper(archive.GenericArchiveStripper):
             allowZip64=True)
         for item in zipin.namelist():
             name = os.path.join(self.tempdir, item)
-            if item.endswith('.xml') or item == 'mimetype':
+            if item.endswith('manifest.xml'):
+                zipin.extract(item, self.tempdir)
+                for line in fileinput.input(name, inplace=1):
+                    #remove the line which contains "meta.xml"
+                    line = line.strip()
+                    if not 'meta.xml' in line:
+                        print line
+                zipout.write(name, item)
+                mat.secure_remove(name)
+
+            elif item.endswith('.xml') or item == 'mimetype':
                 #keep .xml files, and the "manifest" file
                 if item != 'meta.xml':  # contains the metadata
                     zipin.extract(item, self.tempdir)
                     zipout.write(name, item)
                     mat.secure_remove(name)
-            elif item.endswith('manifest.xml'):
-                zipin.extract(item, self.tempdir)
-                #remove line meta.xml
-                zipout.write(name, item)
-                mat.secure_remove(name)
+
             else:
                 zipin.extract(item, self.tempdir)
                 if os.path.isfile(name):
-- 
cgit v1.3