Parsing unknown superboxes with known 1st sub box. #175

glymur/jp2box.py

@@ -164,12 +164,32 @@ class Jp2kBox(object):
             # Call the proper parser for the given box with ID "T".
             try:
                 box = _BOX_WITH_ID[box_id].parse(fptr, start, num_bytes)
-            except KeyError:
+            except KeyError as err:
                 msg = 'Unrecognized box ({0}) encountered.'.format(box_id)
                 warnings.warn(msg)
                 box = Jp2kBox(box_id, offset=start, length=num_bytes,
                               longname='Unknown box')
 
+                if fptr.tell() != start + 8:
+                    # If the file pointer has advanced, then the KeyError
+                    # ocurred during the parsing of the box.
+                    pass
+                else:
+                    # Could it be a superbox with recognizable child boxes?
+                    # Peek ahead to see.
+                    pos = fptr.tell()
+                    read_buffer = fptr.read(8)
+                    sub_length, sub_id = struct.unpack('>I4s', read_buffer)
+                    sub_id = sub_id.decode('utf-8')
+
+                    # Regardless of whether or not we recognize the box, rewind back
+                    # to properly advance to the next box.
+                    fptr.seek(pos)
+
+                    # Now process any child boxes if we actually did recognize it.
+                    if sub_id in _BOX_WITH_ID.keys():
+                        box.box = box.parse_superbox(fptr)
+
             superbox.append(box)
 
             # Position to the start of the next box.

glymur/test/test_jp2box_jpx.py

@@ -4,6 +4,7 @@ Test suite specifically targeting JPX box layout.
 """
 
 import os
+import shutil
 import struct
 import sys
 import tempfile
@@ -202,6 +203,25 @@ class TestJPX(unittest.TestCase):
         self.assertEqual(jpx.box[2].standard_flag,
                          (5, 42, 45, 2, 18, 19, 1, 8, 12, 31, 20))
 
+    @unittest.skipIf(sys.hexversion < 0x03000000, "Needs unittest in 3.x.")
+    def test_unknown_superbox(self):
+        """Verify that we can handle an unknown superbox."""
+        with tempfile.NamedTemporaryFile(suffix='.jpx') as tfile:
+            with open(self.jpxfile, 'rb') as ifile:
+                tfile.write(ifile.read())
+            
+            # Add the header for an unknwon superbox.
+            write_buffer = struct.pack('>I4s', 20, 'grp '.encode())
+            tfile.write(write_buffer)
+            write_buffer = struct.pack('>I4sI', 12, 'free'.encode(), 0)
+            tfile.write(write_buffer)
+            tfile.flush()
+
+            with self.assertWarns(UserWarning):
+                jpx = Jp2k(tfile.name)
+            self.assertEqual(jpx.box[-1].box_id, 'grp ')
+            self.assertEqual(jpx.box[-1].box[0].box_id, 'free')
+
     def test_free_box(self):
         """Verify that we can handle a free box."""
         j = Jp2k(self.jpxfile)