I'm using rxbag (in boxturtle) to look at a bagfile that contains a topic with sensor_msgs/Image messages. I'm interested in the timestamps of these messages so I'm looking at the 'raw' view. When I do that and move the time slider, the following gets repeatedly dumped to the console: UnicodeDecodeError: 'utf8' codec can't decode byte 0xaa in position 9: unexpected code byte Traceback (most recent call last): File "/opt/ros/boxturtle/ros/tools/experimental/rxbag/src/rxbag/util/layer.py", line 259, in on_timer layer.paint_to_bitmap() File "/opt/ros/boxturtle/ros/tools/experimental/rxbag/src/rxbag/util/layer.py", line 143, in paint_to_bitmap self.paint(dc) File "/opt/ros/boxturtle/ros/tools/experimental/rxbag/src/rxbag/raw_view.py", line 87, in paint self.msg_tree.set_msg(self.msg_index, self.msg_stamp, self.msg_incoming) File "/opt/ros/boxturtle/ros/tools/experimental/rxbag/src/rxbag/raw_view.py", line 128, in set_msg self._add_msg_object(None, '', 'msg', msg, msg._type) File "/opt/ros/boxturtle/ros/tools/experimental/rxbag/src/rxbag/raw_view.py", line 240, in _add_msg_object self._add_msg_object(item, subpath, subobj_name, subobj, subobj_type) File "/opt/ros/boxturtle/ros/tools/experimental/rxbag/src/rxbag/raw_view.py", line 219, in _add_msg_object item = self.AppendItem(parent, label) File "/usr/lib/python2.6/dist-packages/wx-2.8-gtk2-unicode/wx/_controls.py", line 5469, in AppendItem return _controls_.TreeCtrl_AppendItem(*args, **kwargs) File "/usr/lib/python2.6/encodings/utf_8.py", line 16, in decode return codecs.utf_8_decode(input, errors, True) .. and the raw data window does not update, except for first line (i.e. "Fri May 28 13:21:24 2010 #871"). Note that Cturtle doesn't seem to have this problem, looking at the same bagfile. I'm guessing what's happening is that it's barfing on the data field of the message, which can contain arbitrary characters, some of which must be running afoul of the unicode converter. I'm not sure if there are any cases where you'd want to actually try to print an array of uint8's in the raw view, but I would think that std_msgs/Image is common enough that it could be special-cased to avoid this problem. Hmm, taking a look at trac.. I'm not sure but it sounds like https://code.ros.org/trac/ros/ticket/2477 may have already fixed this. Will this be backported to boxturtle? Cheers, Pat