Fix bug that some types of images are exported incorrectly.

pdfminer/image.py

@@ -71,19 +71,22 @@ def export_image(self, image):
         stream = image.stream
         filters = stream.get_filters()
         (width, height) = image.srcsize
+        colorspace = image.colorspace
+        if isinstance(colorspace, list):
+            colorspace = colorspace[0]
         if len(filters) == 1 and filters[0][0] in LITERALS_DCT_DECODE:
             ext = '.jpg'
         elif (image.bits == 1 or
-              image.bits == 8 and image.colorspace in (LITERAL_DEVICE_RGB, LITERAL_DEVICE_GRAY)):
+              image.bits == 8 and colorspace in (LITERAL_DEVICE_RGB, LITERAL_DEVICE_GRAY)):
             ext = '.%dx%d.bmp' % (width, height)
         else:
             ext = '.%d.%dx%d.img' % (image.bits, width, height)
         name = image.name+ext
         path = os.path.join(self.outdir, name)
         with open(path, 'wb') as fp:
             if ext == '.jpg':
-                raw_data = stream.get_rawdata()
-                if LITERAL_DEVICE_CMYK in image.colorspace:
+                raw_data = stream.get_data()
+                if colorspace is LITERAL_DEVICE_CMYK:
                     from PIL import Image
                     from PIL import ImageChops
                     ifp = BytesIO(raw_data)
@@ -101,15 +104,15 @@ def export_image(self, image):
                 for y in range(height):
                     bmp.write_line(y, data[i:i+width])
                     i += width
-            elif image.bits == 8 and image.colorspace is LITERAL_DEVICE_RGB:
+            elif image.bits == 8 and colorspace is LITERAL_DEVICE_RGB:
                 bmp = BMPWriter(fp, 24, width, height)
                 data = stream.get_data()
                 i = 0
                 width = width*3
                 for y in range(height):
                     bmp.write_line(y, data[i:i+width])
                     i += width
-            elif image.bits == 8 and image.colorspace is LITERAL_DEVICE_GRAY:
+            elif image.bits == 8 and colorspace is LITERAL_DEVICE_GRAY:
                 bmp = BMPWriter(fp, 8, width, height)
                 data = stream.get_data()
                 i = 0

pdfminer/pdfdocument.py

@@ -348,7 +348,7 @@ def compute_u(self, key):
 
     def compute_encryption_key(self, password):
         # Algorithm 3.2
-        password = (password + self.PASSWORD_PADDING)[:32]  # 1
+        password = (password.encode('ascii') + self.PASSWORD_PADDING)[:32]  # 1
         hash = md5.md5(password)  # 2
         hash.update(self.o)  # 3
         hash.update(struct.pack('<l', self.p))  # 4

pdfminer/pdftypes.py

@@ -228,7 +228,7 @@ def get_filters(self):
             params = [params]*len(filters)
         if STRICT and len(params) != len(filters):
             raise PDFException("Parameters len filter mismatch")
-        return zip(filters, params)
+        return list(zip(filters, params))
 
     def decode(self):
         assert self.data is None and self.rawdata is not None