From a4f80082885079351d71bd3b6c3eaae1fcd5258f Mon Sep 17 00:00:00 2001
From: jvoisin
Date: Sun, 27 Oct 2013 16:07:10 +0000
Subject: Improve MAT's reliability

---
 MAT/office.py | 38 +++++++++++++++++++++-----------------
 MAT/parser.py | 19 +++++++------------
 2 files changed, 28 insertions(+), 29 deletions(-)

diff --git a/MAT/office.py b/MAT/office.py
index c44a52b..583e0f9 100644
--- a/MAT/office.py
+++ b/MAT/office.py
@@ -156,23 +156,27 @@ class PdfStripper(parser.GenericParser):
             python-cairo segfaults on unicode.
             See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=699457
         '''
-        output = tempfile.mkstemp()[1]
-        page = self.document.get_page(0)
-        # assume that every pages are the same size
-        page_width, page_height = page.get_size()
-        surface = cairo.PDFSurface(output, page_width, page_height)
-        context = cairo.Context(surface)  # context draws on the surface
-        logging.debug('PDF rendering of %s' % self.filename)
-        for pagenum in range(self.document.get_n_pages()):
-            page = self.document.get_page(pagenum)
-            context.translate(0, 0)
-            if self.pdf_quality:
-                page.render(context)  # render the page on context
-            else:
-                page.render_for_printing(context)  # render the page on context
-            context.show_page()  # draw context on surface
-        surface.finish()
-        shutil.move(output, self.output)
+        try:
+            output = tempfile.mkstemp()[1]
+            page = self.document.get_page(0)
+            # assume that every pages are the same size
+            page_width, page_height = page.get_size()
+            surface = cairo.PDFSurface(output, page_width, page_height)
+            context = cairo.Context(surface)  # context draws on the surface
+            logging.debug('PDF rendering of %s' % self.filename)
+            for pagenum in range(self.document.get_n_pages()):
+                page = self.document.get_page(pagenum)
+                context.translate(0, 0)
+                if self.pdf_quality:
+                    page.render(context)  # render the page on context
+                else:
+                    page.render_for_printing(context)  # render the page on context
+                context.show_page()  # draw context on surface
+            surface.finish()
+            shutil.move(output, self.output)
+        except:
+            logging.error('Something went wrong when cleaning %s. File not cleaned' % self.filename)
+            return False
 
         try:
             import pdfrw  # For now, poppler cannot write meta, so we must use pdfrw
diff --git a/MAT/parser.py b/MAT/parser.py
index c57eb00..c1c3f4c 100644
--- a/MAT/parser.py
+++ b/MAT/parser.py
@@ -22,8 +22,7 @@ FIELD = object()
 
 
 class GenericParser(object):
-    '''
-        Parent class of all parsers
+    ''' Parent class of all parsers
     '''
     def __init__(self, filename, parser, mime, backup, is_writable, **kwargs):
         self.filename = ''
@@ -66,8 +65,7 @@ class GenericParser(object):
         return True
 
     def remove_all(self):
-        '''
-            Remove all compromising fields
+        ''' Remove all compromising fields
         '''
         state = self._remove_all(self.editor)
         hachoir_core.field.writeIntoFile(self.editor, self.output)
@@ -75,8 +73,7 @@ class GenericParser(object):
         return state
 
     def _remove_all(self, fieldset):
-        '''
-            Recursive way to handle tree metadatas
+        ''' Recursive way to handle tree metadatas
         '''
         try:
             for field in fieldset:
@@ -90,8 +87,7 @@ class GenericParser(object):
             return False
 
     def _remove(self, fieldset, field):
-        '''
-            Delete the given field
+        ''' Delete the given field
         '''
         del fieldset[field]
 
@@ -104,8 +100,7 @@ class GenericParser(object):
         return metadata
 
     def _get_meta(self, fieldset, metadata):
-        '''
-            Recursive way to handle tree metadatas
+        ''' Recursive way to handle tree metadatas
         '''
         for field in fieldset:
             remove = self._should_remove(field)
@@ -119,7 +114,7 @@ class GenericParser(object):
 
     def _should_remove(self, key):
         '''
-            return True if the field is compromising
+            Return True if the field is compromising
             abstract method
         '''
         raise NotImplementedError
@@ -137,7 +132,7 @@ class GenericParser(object):
             but it greatly simplify new strippers implementation.
         '''
         if self.backup:
-            os.rename(self.filename, self.filename + '.bak')
+            shutil.move(self.filename, self.filename + '.bak')
         else:
             mat.secure_remove(self.filename)
         os.rename(self.output, self.filename)
-- 
cgit v1.3